Philippa Kaur

AbstractSeed production, dispersal, germination, and seedling establishment are critical life phases of a tree species. Understanding these processes is crucial to recognize species composition and directional change for ecosystem restoration. This study aimed to estimate seed production, dispersal, and fate of the seed population of A. procera (Roxb.) and evaluate its seedling growth performance in relation to microclimates under natural conditions. Seed production was estimated from 15 sampled trees for three years, while seed dispersal using circular sample plots and seed traps under mother trees. The mean seed production per tree was 145,352, 43,607, and 41,490 during year 2022, 2023, and 2024 respectively, and it significantly differed between years (F = 12.09, P < 0.0001) and among individual trees (F = 4.63, P < 0.0001) while correlated positively with tree traits. Additionally, the seed density decreases with increased distance from the mother trees. A majority of the seeds (55.02% in 2022, 54.25% in 2023 and 52.92 in 2024) fell under the mother tree, while seeds disappeared due to predation and other losses reached 56.60%, 48.00%, and 49.80%, respectively. Germination rate in natural conditions were moderate (39.00%, 47.90%, and 45.40% in 2022, 2023, and 2024, respectively), and less than half (46.07%) of the germinated seedlings survived after 14 months. Further, relative seedling growth rate was strongly influenced by soil temperature, moisture and relative humidity indicating their crucial role in successful establishment. The findings provide essential insights into the population dynamics of A. procera and can inform strategies for monitoring growth and restoring degraded lands.

IntroductionAlbizia procera (Roxb.) (Family Fabaceae) is a fast growing, medium sized tree with an open canopy that can reach up to a height of 30 m1,2. It is native to India, America, Pakistan and Australia3 and found in tropical countries such as Myanmar, Thailand, Bangladesh, Malaysia, Laos, Cambodia, Vietnam4, Nepal, Indonesia, China, Andaman, Kenya, South Africa and Uganda1,5. Due to its varied adaptability and economic importance, A. procera is being planted in agroforestry systems including tea gardens, reforestation, afforestation and social forestry programmes for restoration of degraded lands6,7. Besides, the species is reported to have high medicinal properties8,9. Additionally, it is extensively used for other purposes, such as its leaves are consumed as a vegetable2 and its bark used as a fish poison5,10. These extensive uses can impact the species status and reduce its distribution in its native area.Seed dynamics and seedling establishment are critical life phases of a tree species11,12. Understanding these processes is crucial to recognize species recruitment and directional change in ecosystem restoration11. In contrast, seed production is considered a critical bottleneck of tree life cycle13 and its variation may affect regeneration and numerous biological phenomena including interaction between plants and animals, vegetation dynamics and nutrient cycle14. It may be limited by various external factors such as adverse climate, pollination failure, predation of flowers, fruits and leaves, wind speed etc15, and internal factors such as genetic condition, age and size of the tree16,17. These factors could lead to seed production reduction and significantly impact seed dynamics and seedling recruitment of a given species.Seed dispersal is another important link in the reproductive cycle of a tree from the end of the adult plant to the beginning of the new plant. The distance to which the seeds get dispersed has important ecological significance. For example, seed dispersal can avoid high density-dependent mortality in close proximity to the parent plants18, while it can facilitate maintaining species diversity and may induce spatial accumulation of seeds and seedlings in pioneer trees19. Soil seed banks, on the other hand, are influenced by seed production and dispersal and various governing factors that affect seed germination12,20. The fate of the seeds after dispersal from the parent plants can be influenced by several microclimatic and biotic factors21 and in turn primarily determine the future population structure. It is therefore of paramount importance to have a complete understanding of the population dynamics of a species from seed production to seedling recruitment. However, there are very limited studies in trees that deal with population dynamics of seeds and seedlings.A review of literature reveals that studies on seeds dynamics of A. procera is lacking or very limited. A. procera is a pioneer fast growing legume tree having hard seed coat and this property of seeds results in reduced germination in wild22,23 which may cause seedling population mosaic or heterogeneity in natural conditions24,25. The results have shown that seed pre-treatments of A. procera6,23,26, can bring seedling population homogeneity in nursery and plantations7. We hypothesized that seed production of A. procera is influenced by both intrinsic traits (e.g. tree size, and crown structure) and extrinsic environmental factors, and seedling success is modulated by microclimatic variation. Accordingly, the study aimed to quantify interannual variation in seed production and its relationship with tree characteristics, assess seed dispersal patterns and post-dispersal fate (germination, disappearance, and seed bank formation), and evaluate seedling growth and survival in relation to soil temperature, moisture, and humidity. To achieve these objectives, we conducted field experiments for A. procera species to address the following questions: (a) is there variation in seed production between years and among individual trees?, (b) how do the tree traits affect its seed production?, (c) do the seed dispersal distance influence seed germination?, and (d) which microclimatic variables affect the seedling growth the most?. The findings provide essential insights into population dynamics of the A. procera and can inform strategies for monitoring growth and restoring degraded forests.Materials and methodsStudy areaThe study was carried out at Mizoram University (MZU) campus (23°39’52”-23°48’43″N and 92°39’49”-. 92°46’39″E) (Fig. 1), located 15 km away from Aizawl, the capital city of Mizoram having elevations ranging from 300 m to 880 m asl. MZU campus encompasses roughly 980 acres of land, and harbors tropical wet evergreen forests, small biodiversity park and protected forested water catchment reserve in the north. Several streams flow through the campus27,28,29. The forest included diversified plants species, where Wapongnungsang et al.29 reported 384 plants belonging to 290 genera and 107 families and many grass species27. The area receives an average annual rainfall of approximately 1,850 mm, influenced heavily by the southwest monsoon, with most precipitation occurring between May and September. The mean annual temperature is about 21.6 °C, with summer highs reaching 30 °C and winter lows ranging between 10 and 12 °C (Meteorological data of Mizoram, Aizawl, 2023).Data collectionSeed productionSeed production of A. procera was estimated from randomly selected 15 individual trees between 2022 and 2024. However, for each tree, the number of branches (B), secondary branches (sub-branches (SB)), tertiary branches (sub-sub-branches (SSB)), and inflorescences (INF) per sub-sub-branches were recorded to calculate total seed produced per plant following formula (1). The seed were estimated during November and December of each year before seed maturation and fall16. Additionally, in each tree, the mean number of inflorescences per SSB was determined from random 5-SSB, while the average number of fruits per inflorescences was calculated from a random selection of 10 inflorescences. Conversely, the mean number of seeds per fruit were estimated from fifteen fruits (Fig. 2A) and the number of seeds per kilogram was estimated using the average weight of 25 seeds (Fig. 2B) following the formula (2).Seed dispersalTo study the seed dispersal, five individual fruiting trees were marked in the forest stand following the method suggested by Khan et al.30. The five trees selected for dispersal studies were a subset of initial 15 trees, chosen based on their health, canopy structure, accessibility, and adequate spacing (> 100 m) to avoid seed overlap. This smaller number was selected to allow intensive monitoring of seed-fall and dispersal pattern around each individual tree. Each selected individual tree base was considered as a center, of which concentric circles of 2.5 m circular increments were marked on the ground around the mother tree, extending outside the crown radius as deliberated by Sahoo and Lalfakawma17. The first circle had a radius of a minimum five meters and maximum was 25 m. In the meantime, the seeds that fell under the tree crown were not considered as dispersed seeds. However, each selected individual tree was visited at three-days interval over 8-weeks during seed-fall season (Fig. 2C). All seeds collected within each marked circle were counted separately and the seeds damaged during dispersal were excluded from the analysis.Fig. 1Study area: (a) India, (b) Mizoram state, (c) Aizawl district and Mizoram University location, and (d) sampled trees. The map was created using ArcGIS Pro and can be accessed via (https://pro.arcgis.com/en/pro-app/latest/get-started/download-arcgis-pro.htm).Full size imageFate of seed populations in the soilThe fate of seed population in the soil was assessed, where the number of seeds disappearing (fraction) during the seed fall period was studied. Five seed traps sizing (1 m x 1 m with 30 cm depth) were placed randomly on the ground under the mother tree crown at the beginning of seeds fall (Fig. 3A). The seed traps were visited every five-days until completed seed shedding following the method described by Sahoo and Lalfakawma17 and de Sá Dechoum, et al.31. Simultaneously, during each visit, all seeds in the traps were counted, and seeds damaged (due to insects and rodents) were separated from undamaged one. However, the difference between total produced seed and undamaged seeds that fell under the individual mother tree was estimated as the fraction of seeds loss during the seed fall period17.To estimate the germination of healthy seeds after fall and/or dispersal, the fate of undamaged seeds after dispersal was studied by sowing 20 seeds in plot sizing (1 m x 1 m) under the five sampled individual trees (Fig. 3B). Additionally, seed fate was assessed in relation to distance from the mother tree. At each distance e.g. (5, 15, 25, and 35 from the mother tree), plots sizing (1 m x 1 m each) were established (n = 3 replicates per distance). Twenty seeds were placed per plot (total 60 seeds per distance), and seed fate was recorded weekly for three months. The number of seeds that germinated, disappeared (e.g. translocated or consumed by dispersal agents) or rotted was recorded, and germination was defined as emergence of radicle visible through the seed coat.Soil seed bankThe study on soil seed soil bank was conducted at the end of rainy season following the method described by Souza et al.32. Four sites were selected using stratified sampling20, and from each site, five soil samples from area of (5 × 5 cm and 5 cm depth) were collected and bulked to estimate buried seed density. From each of the bulked sample, 100 g of soil was weighed and washed gently using a jet of water (Fig. 2D) following the procedure and method described by Padonou et al.20 to recover the seeds. The number of seeds found in the sample was then extrapolated to 1 m x 1 m area.Seedling growth performance in natural conditions and its relation to microclimate variablesTo evaluate seedling growth performance and its relation to microclimatic variables, 10 quadrats (1 m ×1 m) were laid randomly in the forest near A. procera species stand. All seedlings recruited from the dispersed seeds within quadrates were monitored for 14 months. The seedlings height and stem collar diameter (SCD) were measured at two-weeks interval; seedlings height was measured using 30 cm ruler, while SCD measured using digital Vernier caliper (150 mm). Monthly seedling growth increments were calculated along with relative growth. Seedling mortality and microclimatic variables such as soil temperature, soil moisture, and humidity were recorded monthly throughout the study period.In contrast, microclimate data were collected in each study plot, where soil temperature was measured using a stainless-steel dial thermometer (Model ST-9283B, 0.1 °C accuracy), soil moisture using digital soil moisture meter (Kelway HB-2), while relative humidity using portable digital hygrometer (HTC-1). One set of sensors was installed per plot (10 total). Sensors were positioned in the center of the (1 m x 1 m) seedling quadrats at 5 cm soil depth. Moreover, measurements were taken three days in a week during morning and afternoon, and averaged to monthly means for correlation with seedling growth increments.Population flux of A. procera
The life cycle of A. procera species is computed from the average of seed production, seed dispersal and fate of seeds during seed-fall and post-seed fall, and soil seed bank over three years (2022–2024), as well as seedling growth, survival, and mortality for the period of 14 months. The mean seed production from 15 trees was estimated, along with seed dispersal and seed damage. Consequently, the percentage of seed disappeared was calculated from the fate of undamaged seed at varying distances from the mother trees, which also included seed germinated and contribution to the soil seed bank. Moreover, seedling survival and mortality were monitored over 14 months to determine overall seedling survival rates.Fig. 2A. procera seed at different stages from seed development to soil bank formation: (A) Fruits pod, (B) Mature seeds, (C) Seeds in the fruit pod after fall from the mother tree, and (D) excavation of buried seeds using wet sieving technique.Full size imageFig. 3Experiments layout: (A) Laying seed trap to collect the dispersed seed, (B) Monitoring the fate of undamaged seed population (towards germination, disappearance and rotted seeds) from plots under the mother tree and at varying distance from the mother tree.Full size imageData analysisAll field data collected on seed production, dispersal, fate of seed population, and seedling growth were compiled and organized into meaningful table for analysis. Data were summarized and expressed as mean and standard deviation (SD). The mean seed production for each year was computed using formula (1), and seed germinated and disappeared was calculated using formulas (3 and 4 respectively). While the spatial distribution of seeds (under the mother tree and at varying distances) was first determined by calculating the percentage of seeds in each location, this percentage was then used to allocate the annual mean seed production across the same spatial categories, yielding the estimated number of seeds at each distance as studied by researcher33. The relationship between seed production and tree characteristics such as DBH, number of branches, number of inflorescences, crown height, crown diameter was examined. In contrast, post-hoc Fisher LSD, Tukey, and One-way ANOVA test (Sig. 0.05) were used to examine the variation of seed production between years and among individual trees. Conversely, seedlings’ growth (height and SCD) were converted to monthly basis, and growth increment was calculated. Seedling growth rates were correlated with microclimate variables such as soil moisture, temperature, and humidity. In the meantime, post-hoc Tukey (Sig. 0.05) was used to examine the variation in monthly seedling growth increments during study period. One-way ANOVA (Sig. 0.05) was used to assess the variation in rainfall, temperature, and humidity between the study years. The seed population flux was computed as accumulative mean of seed production, dispersal, germination, disappearance, soil seed bank, and survival rates. All data analysis were performed using SPSS, Version 22.0, Jamovi (Version 2.5.6)34, OriginPro 2025 and Microsoft Excel (Version 365).Formulas and equationsAll abbreviations in formulas are defined as full form in (Table 1).$${text{Total seed production}}/{text{tree }} = {text{ B }} times {text{ SB }} times {text{ SSB }} times {text{ INF }} times {text{ FI }} times {text{ SF}}$$
(1)
$$Number{text{ }}of{text{ }}seeds{text{ }}in{text{ }}kg{text{ }} = frac{{1000~}}{{Mean~seeds~weight}}~$$
(2)
$${text{Germination }}left( % right){text{ }} = frac{{Number~of~seeds~germinated~}}{{total~~seeds~sown}} times 100$$
(3)
$${text{Seeds disappeared }}left( % right){text{ }} = frac{{Number~of~missing~or~predated~seeds~}}{{total~~seeds~sown}} times 100$$
(4)
Table 1 Summary of parameters and definitions.Full size tableResultsSeed production, seed dispersal, and soil seed bankThe mean seed production of A. procera significantly differed among individual trees (F = 4.63, P < 0.0001) (Fig. 4) and between years (F = 12.09, P < 0.0001) (Fig. 5). The mean seed production per individual tree was 145,352 seed/tree in 2022, dropping sharply to 43,607 in 2023, and 41,490 in 2024 (Table 2). The variation in seed production across the years can be driven from variation in number of inflorescences (F = 12.16, P < 0.0001) and number of fruits per inflorescences (F = 8.92, P < 0.000) (Table 2). Additionally, these declines correspond with notable interannual (Table 2) and monthly (Fig. 6) variation in rainfall, mean temperature, and humidity. In contrast, the result suggests that the pronounced decline in seed output was likely driven by climatic stress during the flowering and fruit set. The mean seed weight was 0.039 g resulting estimated 25,654 seed/kg (Table 2). A strong correlation was observed between mean seed production/tree and tree traits such as DBH, number of branches, inflorescences, fruits, crown height, and crown diameter (Table 3). The seed soil bank was relatively low across the years, without showing discernable variation among sample plots and years (Table 2).The number of seeds per unit area decreased as the dispersal distance from the mother tree increased showing a negative relation (reverse J-shaped curve) between seed density and dispersal distance across the years (Fig. 7). The distance to which the seeds got dispersed and the dispersed seed density varied significantly between years (F = 13.46, P < 0.0001, Table 4). A significant proportion of the total seeds produced (55.38% in 2022, 54.25% in 2023 and 52.92 in 2024) fell directly under the mother tree (Fig. 8) and the dispersed seed density gradually decreased with increased dispersal distance and this was true for all the years (Fig. 8). Further, the maximum distance to which the seeds were found dispersed was 22.50 m from the mother trees (Fig. 8).Table 2 Inter-annual variation in seed production, soil seed bank formation, and climatic factors.Full size tableFig. 4Mean seed production of A. procera among individual trees over three years (2022 to 2024). The error bar shows the variation of seed production between years within studied tree. The different alphabetic letters show significant variation of mean seed production between individual trees (where T1 to T15 are sampled trees).Full size imageFig. 5Variation in mean seed production across study years of A. procera. Post-hoc Tukey test (Sig. 0.05).Full size imageFig. 6Interannual variation in climatic factors: (a) monthly rainfall, (b) temperature, and (c) relative humidity.Full size imageTable 3 Correlation between mean seed production and tree characteristics.Full size tableTable 4 Seed density under the mother tree (UMT) and with different distances from the mother tree.Full size tableFig. 7Post-hoc Tukey test (Sig. 0.001) shows seed density from the mother tree (m) and variation in seed dispersal across the years. The error bar (standard deviation) showing the variation in number of seeds dispersed (m) within the sampled trees, while different letter shows significant difference in number of seeds within the year and between years across various distance from the mother tree.Full size imageFig. 8The percentage of seeds fall under the mother tree (UMT) and dispersed with distance (m).Full size imageFate of seed populations in the soilDuring the seed-fall period, more than 97% of the seeds that fell under the mother tree was found undamaged and viable when observed year-wise variation (Table 5). However, during post-seed-fall, the fraction of seed disappeared was 56.6% in 2022 while it decreased to 48% in 2023, and 49.80% in 2024. Similarly, the seed germination rates of the sown seeds were quite moderate (39.00%, 47.90% and 45.40% in 2022, 2023, and 2024 respectively) and seeds those got rotten were minimal (Table 5).Seed disappearance and germination in relation to distance from the mother treeThe seed disappearance due to predation was significantly (F = 9.61, P < 0.001) higher around the mother trees and it decreased gradually as the dispersal distance increased from the mother trees, a trend commonly observed in all the studied years. Our results showed that a very high proportion (82.33%) of the sown seeds disappeared within a 5 m circle while the disappeared seeds drastically reduced (67.33%) at a distance of 35 m from the mother tree (Fig. 9).Seedlings dynamics in the forest and its relation to microclimatic variablesDuring 14-months period, the seedlings of A. procera attained an average height and SCD of 22.50 cm and 3.25 mm respectively (Fig. 10). The species showed significant change in its growth increment with time (F = 3.29, P < 0.001 for seedling height, and F = 8.36, P < 0.001 for seedling collar diameter) (Fig. 11). The soil temperature and soil moisture showed seasonal variation; the highest soil temperature was recorded in June (26.86 °C), and soil moisture in July (70.30%) (Fig. 12). Among the microclimatic variables, soil moisture showed the strongest positive correlation with height growth and collar diameter (Table 6). The other parameters (relative humidity and soil temperature) too showed positive relationship with the seedling growth increment (height and collar diameter), however, soil pH was negatively related to these seedling attributes (Table 6).Population flux of A. procera
The population flux of A. procera integrates seed production, seed dispersal, seed soil bank, and seedlings recruitment over three years (2022–2024). Based on 15-sampled trees, the mean seed production of A. procera over three years was 76,816 seeds/tree. Of these 97.63% seeds fell beneath the mother tree and dispersed at varying distances. However, during post-seed-fall, we found that 73.44%, 25.50% and 1.06% of the felled seeds got disappeared, germinated and stored in soil bank respectively. Among the germinated seeds, only 46.07% of seeds developed into seedlings that survived till the end of the experiment (Fig. 13). In the meantime, the seedling continues with vigorous growth (height and diameter) during the rainy season while dropping the leaves and stop shoot growth during dry season (Fig. 14).Table 5 Fate of seed population during seed-fall period and post seed-fall period, expressed as fractions (%) of the total seeds trapped (Mean ± SD).Full size tableFig. 9Germination and disappearance of seeds as influenced by distance from the mother tree.Full size imageFig. 10Mean seedling growth performance in the natural forest conditions over study period: (a) mean monthly seedling height (cm) and (b) mean monthly stem diameter (mm). Note: the box plot size shows the variation in seedling growth (N = 50 seedling in 10 quadrats), ANOVA test (Sig. 0.05).Full size imageFig. 11Post-hoc Tukey test (Sig. 0.05) for seedling growth increment: (a) mean seedling height and (b) stem diameter increment. The different alphabetic letters show significant variation in seedling growth increment between months during study period, while error bar (standard deviation) shows the variation in seedling growth (N = 50 seedling in 10 quadrats).Full size imageFig. 12Soil temperature and soil moisture in forest conditions during the study period.Full size imageFig. 13A. procera population flux: average of 3-years seed dynamics to seedling survival.Full size imageTable 6 Correlation between seedlings parameters and microclimate variables.Full size tableFig. 14A. procera seedling growth performance in natural forest conditions: (A) seedling during rainy season of the first year, (B) seedling response to dry winter season, and (C) seedling growth at the onset of rainy season.Full size imageDiscussionSeed productionSeed production of A. procera showed marked interannual variation, with a pronounced peak in 2022 followed by significant declines in 2023–2024. Similar variation in seed production across years have been reported in other trees species11,14,16,17,35,36,37. However, such variability is typical of tropical legumes and can be linked to fluctuations in rainfall and temperature that affect flowering and fruiting set. In contrast, the seed production variation was obviously related to the fruit loading of the species and the number of fruits bearing species in a given year. Several authors have suggested that seed production in a species will be influenced by several intrinsic and extrinsic factors during the flowering and seed setting period. For example, Iralu et al.12 reported that the annual rainfall acts as a limited factor for seed production and seedling survival while16,17,38 explained that seed production of tree species can be influenced by a variety of factors such as availability of resources, pollination failure, predation on flower, fruits, climatic condition, plant age and size. We observed a strong correlation between mean seed production and tree characteristics such as DBH, number of branches, number of inflorescences, and crown cover. On the other hand, larger trees with wider canopies tended to produce more seeds; however, this relationship must be interpreted cautiously because estimated production was driven from tree traits. Direct seed count in a validation subset are recommended to refine predictive models, and similarly used by several researchers11,13,14,16,36. In the meantime, Khan et al.30 observed that dominant trees species with large crowns, which receive a lot of light, tend to produce an optimum number of seeds. Conversely, the higher seed production occurring in warmer year, meanwhile increased temperature negatively affects seedling establishment39.Seed dispersalSeed dispersal is the movement of seed from the mother tree by help of different dispersal agents. For successful tree regeneration, it is important that the seeds should disperse to a safe location where they can germinate, survive and translate into mature plants. Thus, it determines the seeds distribution and trees during natural regeneration process which can be influenced by several factors (e.g. biological and environmental). In A. procera, seed dispersal occurred through a combination of wind dispersal and gravity. We found seed dispersal declined exponentially with distance, confirming that most seeds remain beneath or near the parent crown. A significant variation was observed in seed dispersal across the study years, this variation is mainly driven by seed production. We found the maximum seed dispersed up to 22.50 m from the mother tree, which indicated that the species dispersed its seed via explosion/gravity, while a small fraction of seeds was transported to distance places by wind. The limited seed dispersal pattern may increase density-dependent mortality and seed predation under the mother trees, a similar pattern reported for other Fabaceae species40,41. These findings suggest that the seeds with capsules tend to be exposed to secondary dispersal where wind moves it to another location. Seed dispersal, however, was highly restricted in this species, and for successful regeneration, the seeds need to be transported to far off places. The restricted seed dispersal by gravity nevertheless possesses some challenges for this species as it leads to overcrowding and competition for resources among the closely spaced individuals upon germination. To avoid for this challenge, the seeds of this species have wings/appendages which help them to be carried out and disperse to new locations. Several studies observed that seeds from same mother tree with varying seed mass are influenced by dispersal distance and differ significantly among tree species sharing the same dispersal mode41,42,43,44. Similar findings of decline the density of seed with increased dispersal distance from the mother trees have been observed for other trees species17,45. In contrast, Nathan et al.46 argued that long-dispersal distance is more common in open terrestrial landscapes and driven by migratory animals and wind. While Chen et al.35 and Kasi and Ramasubbu47 indicated that tree species that dispersed their seed by gravity are aggregated around the parent tree.Soil seed bankDespite high seed production, only a small proportion (≈ 1%) contributed to the persistent soil seed bank. Most seeds either germinated shortly after dispersal or disappeared due to predation and decay. Conversely, soil seed-bank contribution suggests that A. procera relies primarily on current-year’s recruitment rather than long-term soil storage. However, soil seed bank nevertheless plays a crucial role in regeneration of tree species, and its size is determined by seed dispersal and seed characteristics, and its ability to remain viable during the unfavourable conditions. Berihun et al.48 reported that seed bank can serve as a “memory” of past plant communities by containing seed from previous years and enhancing future plant communities. Meanwhile, the prevailing microclimate such as low temperature and moisture can have a bearing on soil seed bank by limiting germination49. The relatively low rates of seed germination under natural conditions in this species reveal that a larger fraction of the seeds remains viable-dormant in soil for a longer period, contributing to the resilience and persistent in its natural habitats. The size of the soil seed bank of a species in a given time is related to seed inputs (through seed rain) and seed outputs (through germination, predation and other losses) which are influenced by several factors including environmental and anthropogenics16,17,33,37. Though predation is reported to reduce the seed soil bank in forest floor and may affect survival and mortality12,50, the small viable seed bank can ensure the species’ survival and persistent in the nature.Seed disappearance and germination in relation to distance from the mother treeSeed disappearance in the present study was closely related to distance from the mother tree. The seeds that escaped or dispersed far off from the mother trees were found to be less predated than those which fell near the mother trees. Higher rate of predation and low survival near the mother tree were obviously due to density-dependent competition and predator preference, as also has been reported by several other workers16,17,33. Additionally, our results revealed that with increased distance from the mother tree the germination increased while seed disappearance decreased, in conformity with Souza et al.51. We observed that low seed germination compared to the fractions of the seeds that disappeared or got predated after seed fall in their natural habitats at varying distances from the mother trees. Majority of seeds that still remained in capsule could not transform into successful seedlings due to unfavourable environmental conditions (Fig. 15), which significantly reduce the seed germination and seedling recruitment in the forest. In contract, it reported that the ability of plant propagules to reach microhabitats with the adequate conditions for seed germination and establishment of sapling will have direct effect on the plant’s fitness52. The results clearly demonstrate that the seeds will have a better chance of survival if they are dispersed far away from the parent plants and get favorable germination conditions.Fig. 15Unfavorable environmental conditions limit the successful seed germination and lead to mortality.Full size imageSeedling dynamics and microclimate effectsSeedling growth and survival are often influenced by a combination of environmental factors such as soil temperature, moisture, light availability, and relative humidity. In the present study, A. procera seedlings reached an average height of 22.5 cm and SCD of 3.25 mm over study period and survival after germination was below 50% indicating the sensitivity of the seedling during early growth to various environmental stresses. This find support as the seedling growth of the species was positively correlated with soil moisture, soil temperature, and relative humidity, indicating that water availability is a key driver of recruitment success. Seasonal declines in soil moisture during the dry period sharply reduced the seedling survival, underscoring the vulnerability of young seedlings to drought. Similar findings are reported by Musa and Sahoo26, who reported that moisture availability and temperature significantly affect seedling performance of tropical species while Bebre et al.53 stated that multiple environmental factors in forests influence the seedling growth such as light, temperature, soil moisture, litter depth, intra and interspecific competition for various resources. Consequently, Greenwood et al.54 and Wieser et al.55 observed that higher soil temperature and moderate temperatures can improve physiological processes such as photosynthesis and nutrient uptake, leading to improved seedling establishment and growth. We found that seedling mortality during early growth was higher compared to their survival, a result that find support from the studies reported by Johnson et al.56. These findings further underscore the importance of the favorable microclimates during this critical stage of seedling development and can drastically affect survival rate among species57. In the meantime, the presence of canopy gaps, for instance, has been shown to provide improved conditions for seedling recruitment and survival due to increased light availability and moderated competition, and similar findings have been observed by several researchers58,59. While Awal60 and Kharuk et al.61 reported that soil temperature not only affects plant growth directly but also regulates microbial activity and nutrient cycling, which indirectly supports seedling vigor. Our study results confirm that microclimatic conditions particularly the soil temperature and moisture at the forest floor significantly influenced the seedling dynamics of A. procera.Study limitations and future directionsThis study excludes determining the viable-dormant seed fraction and association of various types of seed dormancy in the soil seed bank which would have provided better clues in understanding the role of the small-sized soil seed bank in regeneration of this fast-growing species in natural habitats. The reportedly under-estimation of dispersed seed density might have occurred due to limited seed trap coverage and secondary removal by predators. Incorporating camera monitoring or automatic traps could have enhanced better accuracy in estimation of various fractions of seeds during post seed-fall. Extending this study across climatic gradients and disturbance regimes would provide a more comprehensive understanding of A. procera regeneration ecology.ConclusionsSeed production and limited dispersal significantly constrain the natural regeneration of A. procera. Although the species produces abundant seeds, most fell beneath the mother tree canopies, where predation and low moisture reduced germination and survival. The low persistent soil seed bank ensure the regeneration of this species in its natural landscape, however, high mortality of seedling during early growth stage highlights a strong bottleneck in its seedling survival. To enhance the regeneration and restoration success, management should focus on assisted seed dispersal, moisture retention, and partial shade maintenance to improve seedling establishment.

Data availability

All data used/analyzed in this paper are available from corresponding author upon request.
ReferencesHossen, K. & Kato-Noguchi, H. Evaluation of the allelopathic activity of albizia procera (Roxb.) Benth. As a potential source of bioherbicide to control weeds. Inter J. Plant. Biol. 13, 523–534. https://doi.org/10.3390/ijpb13040042 (2022).Article 
CAS 

Google Scholar 
Kumar, S. V., Panwar, V. S., Deep, S., Verma, S. & P., & A brief review on phytopharmacological reports on albizia procera. Asian J. Pharm. Sci. 6, 144–149 (2020).
Google Scholar 
Pachuau, L., Lalhlenmawia, H. & Mazumder, B. Characteristics and composition of albizia procera (Roxb.) Benth gum. Ind. Crops Prod. 40, 90–95. https://doi.org/10.1016/j.indcrop.2012.03.003 (2012).Article 
CAS 

Google Scholar 
Shah, S. S. et al. Preparation and characterization of manganese oxide nanoparticles-coated albizia procera derived carbon for electrochemical water oxidation. J. Mater. Sci. Mater. Electron. 30, 16087–16098. https://doi.org/10.1007/s10854-019-01979-6 (2019).Article 
CAS 

Google Scholar 
Rai, E. & Ansari, S. A. A protocol for recovery of whole plants through in vitro shoots regeneration from leaflet explants of mature trees of Albizia procera (Roxb.) Benth. Indian For. 143, 487–498 (2017).
Google Scholar 
Azad, M. S., Biswas, R. K. & Matin, M. A. Seed germination of albizia procera (Roxb.) Benth. In bangladesh: a basis for seed source variation and pre-sowing treatment effect. Forestry Stud. China. 14, 124–130. https://doi.org/10.1007/s11632-012-0209-z (2012).Article 

Google Scholar 
Khurana, E. Influence of seed size on seedling growth of albizia procera under different soil water levels. Ann. Bot. 86, 1185–1192. https://doi.org/10.1006/anbo.2000.1288 (2000).Article 

Google Scholar 
Bunluepuech, K. & Tewtrakul, S. Anti-HIV-1 integrase activity of Thai medicinal plants in longevity preparations. Sonklanakarin J. Sci. Tech. 33, 693 (2011).CAS 

Google Scholar 
Khatoon, M. M., Khatun, M. H., Islam, M. E. & Parvin, M. S. Analgesic, antibacterial and central nervous system depressant activities of albizia procera leaves. Asian Pac. J. Trop. Biomed. 4, 279–284. https://doi.org/10.12980/APJTB.4.2014C348 (2014).Article 
PubMed 
PubMed Central 

Google Scholar 
Rai, E., Bouddha, S. & Ansari, S. A. Biochemical investigations during in vitro adventitious shoot regeneration in leaflet explants from nodal segments of a mature albizia procera tree. J. Res. 27, 699–705. https://doi.org/10.1007/s11676-015-0196-8 (2015).Article 
CAS 

Google Scholar 
Abiyu, A., Teketay, D., Glatzel, G. & Gratzer, G. Seed production, seed dispersal and seedling establishment of two Afromontane tree species in and around a church forest: implications for forest restoration. Ecosystem 3 https://doi.org/10.1186/s40663-016-0076-5 (2016).Iralu, V., Barbhuyan, H. S. A. & Upadhaya, K. Ecology of seed germination in threatened trees: A review. Ener Ecol. Environ. 4, 189–210. https://doi.org/10.1007/s40974-019-00121-w (2019).Article 

Google Scholar 
Mattana, E., Fenu, G. & Bacchetta, G. Seed production Andin situgermination of Lamyropsis microcephala (Asteraceae), a threatened mediterranean mountain species. Arct. Antarct. Alp. Res. 44, 343–349. https://doi.org/10.1657/1938-4246-44.3.343 (2012).Article 
ADS 

Google Scholar 
Wesołowski, T., Rowiński, P. & Maziarz, M. Interannual variation in tree seed production in a primeval temperate forest: does masting prevail? Eur. J. Res. 134, 99–112. https://doi.org/10.1007/s10342-014-0836-0 (2014).Article 

Google Scholar 
Kameswara, Rao, N., Dulloo, M. E. & Engels, J. M. A review of factors that influence the production of quality seed for long-term conservation in genebanks. Genet. Resourc Crop Evol. 64, 1061–1074. https://doi.org/10.1007/s10722-016-0425-9 (2017).Article 
CAS 

Google Scholar 
Barik, S. K., Tripathi, R. S., Pandey, H. N. & Rao, P. Tree regeneration in a subtropical humid forest: effect of cultural disturbance on seed production, dispersal and germination. J. Appl. Ecol. 33, 1551. https://doi.org/10.2307/2404793 (1996).Article 

Google Scholar 
Sahoo, U. K. & Lalfakawma, L. Population dynamics of Castanopsis tribuloides A. (DC) in an undisturbed and disturbed tropical forest stands of North-east India. J. Exp. Bio Agri Sci. 1, 454–463 (2013).
Google Scholar 
Herrero-Jáuregui, C., García-Fernández, C., Sist, P. L. J. & Casado, M. A. Recruitment dynamics of two low-density Neotropical multiple-use tree species. Plant. Ecol. 212, 1501–1512. https://doi.org/10.1007/s11258-011-9924-0 (2011).Article 

Google Scholar 
Howe, H. F. & Miriti, M. N. When seed dispersal matters. BioScience 54, 651. https://doi.org/10.1641/0006-3568(2004)054 (2004).Article 

Google Scholar 
Padonou, E. A., Akakpo, B. A., Tchigossou, B. & Djossa, B. Methods of soil seed bank estimation: a literature review proposing further work in Africa. iForest-Biogeosciences Forestry. 15, 121. https://doi.org/10.3832/ifor3850-015 (2022).Article 

Google Scholar 
Von Arx, G., Pannatier, E. G., Thimonier, A. & Rebetez, M. Microclimate in forests with varying leaf area index and soil moisture: potential implications for seedling establishment in a changing climate. J. Ecol. 101, 1201–1213. https://doi.org/10.1111/1365-2745.12121 (2013).Article 

Google Scholar 
Nongrum, A. & Kharlukhi, L. Effect of seed treatment for laboratory germination of albizia chinensis. J. Res. 24, 709–713. https://doi.org/10.1007/s11676-013-0408-z (2013).Article 
CAS 

Google Scholar 
Sahoo, U. K. Effect of pretreatments on seed germination and seedling vigour of four different species of albizzia. Seed Res. 35, 124–128 (2007).
Google Scholar 
Daibes, L. F. et al. Fire and legume germination in a tropical savanna: ecological and historical factors. Ann. Bot. 123, 1219–1229. https://doi.org/10.1093/aob/mcz028 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Koutouan-Kontchoi, M. N., Phartyal, S. S., Rosbakh, S., Kouassi, E. K. & Poschlod, P. Seed dormancy and dormancy-breaking conditions of 12 West African Woody species with high reforestation potential in the forest-savanna ecotone of Côte d’ivoire. Seed Sci. Tech. 48, 101–116. https://doi.org/10.15258/sst.2020.48.1.12 (2020).Article 

Google Scholar 
Musa, F. I. & Sahoo, U. K. Impact of pre-treatments on Albizia procera and Albizia chinensis seed germination and early growth performance in Nursery, Mizoram, India. Sci. Tech. 21, 138–148. https://doi.org/10.1080/21580103.2025.2465604 (2025).Article 

Google Scholar 
Sailo, L., Solanki, G. S., Lalhruaizela, C. & Mizoram, A. I. Z. A. W. L. Avian Diversity in Mizoram University Campus, Sci. Tech. J. 7, 54–68 https://doi.org/10.22232/stj.2019.07.01.08 (2019).Article 

Google Scholar 
Zothanpuii, J. H., Gouda, S., Parida, A. & Solanki, G. S. A study on diversity of mammalian species using camera traps and associated vegetation in Mizoram university Campus, Aizawl, Mizoram. J. Threat Taxa. 12, 17330–17339. https://doi.org/10.11609/jott.6465.12.17.17330-17339 (2020).Article 

Google Scholar 
Wapongnungsang, N., Ovung, E. Y. & Tripathi, S. Assessment of tree diversity in tropical moist deciduous forest of Mizoram University, Northeast India. J. Appl & Nat. Sci. 13, 95–100. https://doi.org/10.31018/jans.v13i1.2436 (2021).Article 

Google Scholar 
Khan, M. L., Bhuyan, P. & Tripathi, R. S. Effects of forest disturbance on fruit set, seed dispersal and predation of Rudraksh (Elaeocarpus Ganitrus Roxb.) in Northeast India. Curr. Sci. 88, 133–142 (2005).
Google Scholar 
De Sá Dechoum, M., Rejmánek, M., Castellani, T. T., Zalba, S. & Martin Limited seed dispersal May explain differences in forest colonization by the Japanese raisin tree (Hovenia dulcis thunb.), an invasive alien tree in Southern Brazil. Trop. Con Sci. 8, 610–622. https://doi.org/10.1177/194008291500800303 (2015).Article 

Google Scholar 
Souza, M. et al. Weed emergence in a soil with cover crops in an agroecological no-tillage system. Planta Daninha. 36 https://doi.org/10.1590/S0100-8358201836010065 (2018).Musa, F. I. et al. Regeneration ecology of Lagerstroemia speciosa (L.) Pers in the Eastern Himalayan region of India. J. Environ. Sci. 41, 169–181. https://doi.org/10.7747/JFES.2025.41.2.169 (2025).Article 

Google Scholar 
Fox, J., Weisberg, S. & car Companion to Applied Regression. [R package]. (2023). Retrieved from https://cran.r-project.org/package=carChen, L. et al. Seed dispersal and seedling recruitment of trees at different successional stages in a temperate forest in Northeastern China. J. Plant. Ecol. 7, 337–346. https://doi.org/10.1093/jpe/rtt024 (2014).Article 

Google Scholar 
Calama, R. & Montero, G. Cone and seed production from stone pine (Pinus Pinea L.) stands in central range (Spain). Eur. J. Res. 126, 23–35. https://doi.org/10.1007/s10342-005-0100-8 (2005).Article 

Google Scholar 
Khan, M. L., Bhuyan, P., Shankar, U. & Todaria, N. P. Seed germination and seedling fitness in Mesua Ferrea L. in relation to fruit size and seed number per fruit. Acta Oecol. 20, 599–606 (1999).Article 
ADS 

Google Scholar 
Owens, J. N. Constraints to seed production: temperate and tropical forest trees. Tree Physiol. 15, 477–484. https://doi.org/10.1093/treephys/15.7-8.477 (1995).Article 
PubMed 

Google Scholar 
Ibáñez, I., Katz, D. S. W. & Lee, B. R. The contrasting effects of short-term climate change on the early recruitment of tree species. Oecologia 184, 701–713. https://doi.org/10.1007/s00442-017-3889-1 (2017).Article 
ADS 
PubMed 

Google Scholar 
Howe, H. F. & Smallwood, J. Ecology of seed dispersal. Ann. Rev. Ecol. Syst. 13, 201–228. https://doi.org/10.1146/annurev.es.13.110182.001221 (1982).Article 

Google Scholar 
Musa, F. I., Sahoo, U. K., Mohammed, E. M. I., Ramtharmawi & Swain, S. K. Seed population dynamics and early seedling growth of albizia chinensis (Osbeck) Merr. In the forests of humid subtropics of India. Int. J. Res. 2025(1), 3199982. https://doi.org/10.1155/ijfr/3199982 (2025).Article 

Google Scholar 
Jordano, P. What is long-distance dispersal? And a taxonomy of dispersal events. J. Ecol. 105, 75–84. https://doi.org/10.1111/1365-2745.12690 (2016).Article 

Google Scholar 
Muller-Landau, H. C., Wright, S. J., Calderón, O., Condit, R. & Hubbell, S. P. Interspecific variation in primary seed dispersal in a tropical forest. J. Ecol. 96, 653–667 (2008).Article 

Google Scholar 
Vittoz, P. & Engler, R. Seed dispersal distances: a typology based on dispersal modes and plant traits. Bot. Helv. 117, 109–124. https://doi.org/10.1007/s00035-007-0797-8 (2007).Article 

Google Scholar 
Miranda, A. et al. Traits of perch trees promote seed dispersal of endemic fleshy-fruit species in degraded areas of endangered mediterranean ecosystems. J. Arid Environ. 170, 103995. https://doi.org/10.1016/j.jaridenv.2019.103995 (2019).Article 
ADS 

Google Scholar 
Nathan, R., Schurr, F. M., Spiegel, O., Steinitz, O. & Trakhtenbrot, A. Tsoar, A. Mechanisms of long-distance seed dispersal. Trends Ecol. Evol. 23 (11), 638–647 (2008).Article 
PubMed 

Google Scholar 
Kasi, K. S. & Ramasubbu, R. Phenological patterns, fruit predation, and seed dispersal in two endangered trees (Elaeocarpus spp.) of Southern Western Ghats, India. J. Asia-Pac Biodivers. 14, 275–282. https://doi.org/10.1016/j.japb.2021.02.002 (2021).Article 

Google Scholar 
Berihun, T., Bekele, T., Lulekal, E. & Asfaw, Z. Study of the soil seed bank composition in Arjo-Diga humid Afromontane forest under different land use types and its implications for the restoration of degraded lands in Western Ethiopia. Int. J. Ecol. 2024(1), 5119487. https://doi.org/10.1155/2024/5119487 (2024).Article 

Google Scholar 
Saatkamp, A., Affre, L., Dutoit, T. & Poschlod, P. Germination traits explain soil seed persistence across species: the case of mediterranean annual plants in cereal fields. Ann. Bot. 107, 415–426. https://doi.org/10.1093/aob/mcq255 (2011).Article 
PubMed 
PubMed Central 

Google Scholar 
Fenner, M. Ecology of seed banks. In: Routledge eBooks. 507–528. https://doi.org/10.1201/9780203740071-19 (2017). Souza, M. L., Silva, D. R. P. & Fantecelle, L. B. De Lemos Filho, J. P. Key factors affecting seed germination of Copaifera langsdorffii, a Neotropical tree. Acta Bot. Brasil. 29, 473–477. https://doi.org/10.1590/0102-33062015abb0084 (2015).Article 

Google Scholar 
Lomascolok, S. B. Indirect assessment of seed dispersal effectiveness for Solatium riparium (Solanaceae) based on habitat use and rate of fruitdisappearance. Ecol. Austral. 26 (1), 64–71. https://doi.org/10.25260/EA.16.26.1.0.101 (2016).Article 

Google Scholar 
Bebre, I., Riebl, H. & Annighöfer, P. Seedling growth and biomass production under different light availability levels and competition types. Forests 12 (1376). https://doi.org/10.3390/f12101376 (2021).Greenwood, S., Chen, J., Chen, C. & Jump, A. S. Temperature and sheltering determine patterns of seedling establishment in an advancing subtropical treeline. J. Veg. Sci. 26, 711–721. https://doi.org/10.1111/jvs.12269 (2015).Article 

Google Scholar 
Wieser, G., Oberhuber, W., Walder, L., Spieler, D. & Gruber, A. Photosynthetic temperature adaptation of Pinus cembra within the timberline ecotone of the central Austrian alps. Ann. Sci. 67, 201. https://doi.org/10.1051/forest/2009094 (2010).Article 
CAS 

Google Scholar 
Johnson, D. M., McCulloh, K. A. & Reinhardt, K. Tree physiology/Tree physiology. (Dordrecht) 65–87. https://doi.org/10.1007/978-94-007-1242-3_3 (2011).Brown, N. The implications of climate and gap microclimate for seedling growth conditions in a Bornean lowland rain forest. J. Trop. Ecol. 9, 153–168. https://doi.org/10.1017/s0266467400007136 (1993).Article 
CAS 

Google Scholar 
Charru, M., Seynave, I., Hervé, J. C. & Bontemps, J. D. Spatial patterns of historical growth changes in Norway Spruce across Western European mountains and the key effect of climate warming. Trees 28, 205–221. https://doi.org/10.1007/s00468-013-0943-4 (2014).Article 

Google Scholar 
Moore, C. E. et al. The effect of increasing temperature on crop photosynthesis: from enzymes to ecosystems. J. Exp. Bot. 72 (8), 2822–2844. https://doi.org/10.1093/jxb/erab090 (2021).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Awal, M. Effects of changes in soil temperature on seedling emergence and phenological development in field-grown stands of peanut (Arachis hypogaea). Environ. Exp. Bot. 47, 101–113. https://doi.org/10.1016/s0098-8472(01)00113-7 (2002).Article 

Google Scholar 
Kharuk, V. I., Ranson, K. J., Im, S. T. & Vdovin, A. S. Spatial distribution and Temporal dynamics of high-elevation forest stands in Southern Siberia. Glob Ecol. Biogeogr. 19, 822–830. https://doi.org/10.1111/j.1466-8238.2010.00555.x (2010).Article 

Google Scholar 
Download referencesAcknowledgementsThe first author (F.I.M.) thanks Indian Council for Cultural Relations (ICCR) for PhD scholarship and their valuable support and University of Blue Nile for granting study leave to carry out this research.FundingThe first author (F. I. M) grateful acknowledge the support of Indian Council for Cultural Relations (ICCR) under India-Africa Maitri Scholarship Scheme (Formerly called Africa Scholarship Scheme) through a PhD scholarship.Author informationAuthors and AffiliationsDepartment of Forestry, School of Earth Sciences and Natural Resource Management, Mizoram University, Aizawl, 796004, Mizoram, IndiaFaisal Ismail Musa, Uttam Kumar Sahoo, Uttam Thangjam & Mamta ChettriDepartment of Forestry, Faculty of Agriculture and Natural Resources, University of Blue Nile, Ad-Damzin, 26611, SudanFaisal Ismail MusaDepartment of Environmental Sciences, School of Earth Sciences and Natural Resource Management, Mizoram University, Aizawl, 796004, Mizoram, IndiaAhmed Abdallah Adam MohamedAuthorsFaisal Ismail MusaView author publicationsSearch author on:PubMed Google ScholarUttam Kumar SahooView author publicationsSearch author on:PubMed Google ScholarAhmed Abdallah Adam MohamedView author publicationsSearch author on:PubMed Google ScholarUttam ThangjamView author publicationsSearch author on:PubMed Google ScholarMamta ChettriView author publicationsSearch author on:PubMed Google ScholarContributionsFaisal Ismail Musa: Conceptualization, data curation, formal analysis, investigation, methodology, software, validation, visualization, writing—original draft; Uttam Kumar Sahoo: Conceptualization, methodology, supervision, writing—review & editing; Ahmed Abdallah Adam Mohamed: Data curation, formal analysis, validation, visualization, writing—review & editing; Uttam Thangjam: Methodology, Review & editing; Mamta Chettri writing—Review & editing. All authors read and approved the final version for publication.Corresponding authorCorrespondence to
Faisal Ismail Musa.Ethics declarations

Competing interests
The authors declare no competing interests.

Ethical approval
The seeds used in this experiment were collected from the natural forests following the standard guidelines, and this research was carried out as per the local legislation and approval from the research ethical committee of Mizoram University, India.

Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Reprints and permissionsAbout this articleCite this articleMusa, F.I., Sahoo, U.K., Mohamed, A.A.A. et al. Population dynamics of seed and seedlings of Albizia procera (Roxb.) in Mizoram, India.
Sci Rep 15, 44613 (2025). https://doi.org/10.1038/s41598-025-28651-wDownload citationReceived: 22 September 2025Accepted: 11 November 2025Published: 24 December 2025Version of record: 24 December 2025DOI: https://doi.org/10.1038/s41598-025-28651-wShare this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
Keywords
Albzia procera
Seed production and dispersalDispersal distanceGermination More

AbstractSocial familiarity within groups promotes behavioural synchrony and facilitates information transfer. Whether it shapes collective decision-making under predator threat is unknown. Here groups of six medaka (Oryzias latipes) familiarised for one month were used to test whether familiarisation promotes instantaneous collective decision-making in response to a looming stimulus (LS) mimicking a predator attack. First, we analysed behavioural transitions, defined as changes among three behavioural states: high-speed, normal and freezing-like before, during, and after LS in groups of six individuals. Individuals showing high-speed state in response to LS typically tended to shift to freezing-like state afterwards, whereas non-responders were more likely to maintain normal state. Group-level analysis revealed a bimodal distribution in the number of individuals exhibiting freezing-like state, with peaks at zero and six individuals, corresponding to ‘all non-freezing’ and ‘all-freezing’. Clustering analysis further identified three consistent group profiles: ‘freezing-dominant’, ‘non-freezing-dominant’, and ‘mixed-type’ based on behavioural tendencies across 10 trials. In contrast, in unfamiliar groups assembled immediately before testing, the ‘freezing-dominant’ profile was absent, and the distribution in the number of individuals exhibiting freezing-like state shifted to unimodal. In these groups, even at the individual level, responses more often showed a transition from high-speed state to normal state rather than freezing-like state. The results indicate that social familiarity promotes synchronous freezing-like state and consensus decisions under looming threat. Our study presents a behavioural assay for predator-evoked collective decision-making in a genetic model fish, providing a framework for future efforts to link behavioural ethology with neuroscience.

Similar content being viewed by others

Optimization, purification, and characterization of xylanase production by a newly isolated Trichoderma harzianum strain by a two-step statistical experimental design strategy

Article
Open access
22 October 2022

IoT and ML approach for ornamental fish behaviour analysis

Article
Open access
05 December 2023

Comparison of anxiety-like and social behaviour in medaka and zebrafish

Article
Open access
28 June 2022

IntroductionAnimals living in groups often exhibit synchronous behaviour in contexts such as migration, foraging, and predator avoidance. The emergence of coordinated behavioural patterns through interactions among individuals is termed collective behaviour, where group-level order and synchrony are thought to arise from local rules at the individual level, including attraction, alignment, and repulsion1. Collective behaviour includes collective decision-making, in which a group selects a single option from among multiple alternatives, and this phenomenon is observed in various contexts such as movement2,3, foraging4, and predator evasion5,6,7.In the context of predator avoidance, consensus formation during collective decision-making has been studied across various species. For instance, in sticklebacks, once the number of individuals escaping in a particular direction exceeds a threshold, the remaining group members tend to follow6. Similarly, simulations involving humans have shown that when a critical number of escape responses are observed, the group tends to adopt avoidance behaviour5. In elephants, the oldest female has been reported to be particularly sensitive to predator vocalisations and to influence the group’s decision to flee7. However, few studies have quantitatively investigated collective decision-making as an instantaneous group response under emergency conditions. In particular, how dynamic group-level responses to a rapidly approaching predator emerge remains poorly understood. While mechanisms underlying rapid individual decision-making in response to visual looming stimuli have been demonstrated8, systematic analyses of dynamic group-level responses remain scarce.Social familiarity has also been shown to affect individual recognition and interaction patterns, thereby influencing behaviour and information transmission9. At the dyadic level, social familiarisation enhances responsiveness to predators in predatory mites and reduces encounter frequency10. In sticklebacks, familiarisation reduced leadership tendencies in bold individuals, leading to more balanced coordination11. In cichlids, social familiarity promotes exploratory behaviour and reduces fear responses to novel stimuli12, while in zebrafish, the transmission of social fear is enhanced among familiar individuals13. At the group level, wild female guppies tend to associate with familiar individuals14, avoidance frequency in response to predator odour increases in fathead minnows15, and the latency to initiate avoidance of a predator model is reduced in brown trout16. In tropical damselfish, both responsiveness to fear stimuli and inter-individual information transmission are enhanced through familiarisation17. Overall, the literature indicates that social familiarity enhances alignment coordination and information transmission; however, how these factors influence collective decision-making remains unclear.Furthermore, most empirical research in this field has relied on observations in natural environments or on wild individuals6,15, yet relatively few studies have been conducted in controlled experimental settings. Although collective decision-making regarding movement direction has been demonstrated in zebrafish18, reports of such decisions in response to predators are lacking. Moreover, integrative frameworks that link decision-making mechanisms at both the individual and group levels with molecular and neural analysis, particularly in genetically tractable model organisms, are still lacking.To address these gaps, we focused on medaka Oryzias latipes, a well-established model organism in molecular genetics. Medaka are known to exhibit coordinated behaviour with conspecifics19 and to improve foraging efficiency through visual social learning20. Preliminary observations revealed that small groups of medaka responded synchronously to a human approach, either by showing ‘freezing after escape’ or by maintaining continuous movement without escape. Motivated by these findings, we aimed to develop a behavioural assay capable of quantitatively assessing instantaneous collective decision-making using a looming stimulus (LS) mimicking the sudden approach of a predator. LS has been used to elicit individual avoidance responses in mice21, zebrafish22, and fruit flies23. However, most previous studies have focused on individual-level decision-making8.Even in group contexts, the focus has remained on how individual responses are influenced by conspecifics23,24, with little attention paid to whether the group as a whole converges on a single collective choice.In this study, we optimised LS parameters and established a quantitative behavioural system capable of replicating the distinct response patterns during preliminary observations. We then tested whether medaka groups exhibit instantaneous collective decision-making in response to LS and examined the effects of social familiarity on the decision-making patterns. Our findings establish an experimental model for examining collective decisions under acute threat. They also lay the foundation for elucidating how familiarisation modulates synchrony and group-level decision-making processes in a genetically accessible model organism.MethodsEthics statementAll the methods in this study were carried out in accordance with relevant guidelines and regulations. The work in this paper was conducted using protocols specifically approved by the Animal Care and Use Committee of Tohoku University (permit number: 2022LsA-003). All efforts were made to minimise suffering following the NIH Guide for the Care and Use of Laboratory Animals. Fish and breeding conditions are described above. The study was carried out in compliance with the ARRIVE guidelines (https://arriveguidelines.org/arrive-guidelines).AnimalsMedaka (Oryzias latipes, fading strain) were obtained from Dr Tetsuro Takeuchi (Fig. 1a)25. This strain gradually loses body pigmentation in different body parts and at different timing among individuals, which initially appeared suitable for individual identification based on body colour. However, we later found that distinguishing individuals within a group of six remained difficult. During group observations, we consistently noted characteristic and reproducible collective behavioural patterns. As this strain has been maintained as a closed colony with limited genetic variation, we considered it an appropriate model for investigating the mechanisms underlying such stable group-level behavioural patterns. All individuals were hatched and bred in our laboratory. Medaka fish were maintained in groups of six individuals in plastic aquariums (22.6 cm× 14.6 cm× 14.5 cm, Sanko) or custom-made acrylic aquariums(22 cm × 14.5 cm × 14.5 cm) under controlled temperature (26 ± 1 °C) and light (14 h: 10 h light: dark) conditions. Every day, fish were fed brine shrimp between 12:00 and 13:00 and solid bait Otohime β−2 (Marubeni Nissin Feed, Tokyo, Japan) at least twice around 10:00 and 17:00 on weekdays. This study used individuals aged 2–9 months post hatch.Fig. 1Experimental design, state transition analysis, and statistical evaluation of freezing-like behaviour in medaka after looming stimulus. (a) Medaka fish (fading strain) (b) Groups of medaka were transferred, with their tanks, from the circulating water system to the apparatus after feeding. One hour later, the looming stimulus (LS) was presented five times at 30-minute intervals. The experiment was conducted over two consecutive days, giving ten LS. For analysis, three 10-second periods (before, during, and after each LS) were used, totalling 30 s per trial. (c) A state transition diagram visualises individual-level states (FS: freezing-like state, NS: normal state, and HS: high-speed state) across three intervals: before, during, and after LS. Node size represents the proportion of individuals, and numbers within nodes indicate counts. Node colours are red for HS, light blue for NS, and grey for FS. Numbers on edges indicate the transition probabilities, and edge thickness corresponds to the number of individuals. Edge colours indicate the originating state: red for transitions from HS, blue from NS, and grey from FS. (d) A bar plot showing the frequencies of transition patterns across the three intervals. Each pattern is categorised according to the sequence of states. Red bars indicate transitions to HS during LS followed by FS after LS. Blue bars indicate individuals that remained NS during and after LS. Grey bars represent all other patterns. To assess statistical significance, a binomial test with false discovery rate (FDR) correction was applied. Patterns with q < 0.001 are marked with ***. (e) The X-axis shows the number of individuals in FS after LS, and the Y-axis shows its frequency. The blue and red lines represent the observed data (17 groups) and the simulated data (17 groups × 1000 trials, seed = 1, …, 1000), respectively. A chi-square test result is shown in the graph (χ² (6) = 147, p < 0.001).Full size imageBehavioural experiments under familiar conditionsMedaka (Oryzias latipes) aged either one month or nine months were randomly selected to form groups. Ten groups of six individuals each were formed from one-month-old fish (N = 60), and seven groups were formed from nine-month-old fish (N = 42). These groups were formed without regard to sex, because the sex ratio of one-month-old fish could not be reliably determined. In contrast, the nine-month-old groups were adjusted to achieve a 1:1 sex ratio. Each group of six individuals was maintained separately for a month. In total, 17 groups of sexually mature individuals aged either two months or ten months that had undergone familiarisation, were used for experiments.Behavioural experiments under unfamiliar conditionsMedaka aged 4 or 9 months, reared in a recirculating aquaculture system, were randomly selected. From the 4-month-old fish (N = 36), six groups were formed, and from the 9-month-old fish (N = 36), another six groups, each consisting of six individuals, yielded a total of 12 groups. The sex ratio in each group was adjusted to 1:1. Group formation took place in the morning, and experiments were conducted 1 to 2 h after the afternoon feeding. Subsequent procedures were conducted in accordance with those described in the familiarisation experiments.Looming stimulus (LS)The looming stimulus (LS) was a visual stimulus mimicking an approaching predator, created using the “Zoom” animation function in Microsoft PowerPoint (Figure S1). The stimulus expanded to a width of 14.5 cm over 5.5 s, gradually darkened over 10 s, and remained black for 3 min (Figure S1). It was presented on an LCD monitor (EXLDH271DB, I-ODATA) mounted on the side of the aquarium (Figure S2). Both plastic and acrylic aquaria, also used as breeding tanks, were employed. Approximately 1–2 h after daytime feeding (12:00–13:00), each group was transferred in its breeding tank directly to the behavioural testing apparatus and the water level was adjusted to 6 cm. The LS was presented five times at 30-minute intervals for two consecutive days, beginning 1 h after transferring the aquarium from the rearing system to the testing apparatus (Fig. 1b).Behavioural recording and trackingBehaviour was recorded from above using an action camera (M80 Air, Apexcam, or HERO8, GoPro) at a resolution and frame rate of 4 K (30 fps) or 2.7 K (50 fps). Recordings lasted 5 min, beginning 2 min before the LS and ending 3 min after. For analysis, a 30-s segment was extracted for each trial: 10 s before, during, and after LS (Fig. 1b). Video files were extracted using QuickTime Player, converted to JPEG format using FFMPEG (v4.4.1), and subsequently converted to MP4 format at 5 fps. Tracking was performed using UMATracker26, and coordinate data were obtained with the UMATracker-Tracking tool, applying either the Pochi-Pochi (manual positioning) or Group Tracker GMM algorithm. Tracking errors, such as identity swaps, were manually corrected using UMATracker-TrackingCorrector.To convert pixel values to centimetres, the number of pixels along the centre of the long side of the aquarium was measured in ImageJ, which was based on the actual inner length (20.0–20.5 cm). Velocity (cm s⁻¹) was calculated from coordinate data (5 fps). A velocity matrix (6 individuals × 10 trials × 17 groups; 1020 × 150 frames) was compiled, and a moving average was applied (window size = 5) using pandas v1.4.0 to smooth short-term fluctuations.Definition of behavioural types and statesTo capture how individual fish responded to LS (looming stimulus) in terms of state transitions, we expressed behavioural responses as transition patterns across three intervals (before, during, and after LS). In total, 17 groups of six individuals each (N = 102) were tested. For each group, fish were transferred to the experimental arena and habituated for one hour, after which the looming stimulus was presented five times at 30-minute intervals over two consecutive days (Fig. 1b). This protocol yielded a total of 6 individuals × 10 trials × 17 groups of individual-level datasets for subsequent analyses. For this purpose, we first defined behavioural types for each interval based on velocity data. Using histograms and kernel density estimation (KDE) curves of velocity, we categorised behaviour into three types: ‘freezing-like behaviour’, ‘normal swimming’, and ‘high-speed swimming’. The rationale for these definitions is as follows. Some individuals exhibited freezing-like behaviour after LS. The velocity histogram for the post-LS interval showed a bimodal distribution with a trough at approximately 0.2 cm/s (Figure S3a). Therefore, frames with speeds below 0.2 cm/s were defined as ‘freezing-like behaviour’. During LS, escape-like responses characterised by high-speed swimming were observed. Such behaviours were rarely seen before the LS onset. Comparison of KDE curves for the pre-LS and LS intervals revealed minimal overlap above 6 cm/s (Figure S3b). Thus, frames with speeds of 6 cm/s or higher were defined as ‘high-speed swimming’. Frames with velocities between 0.2 cm/s and 6 cm/s were categorised as ‘normal swimming’. All histograms, KDE curves, and heat maps were generated using Python v3.8 and matplotlib v3.7.5.Based on speed-based behavioural types, we then defined the behavioural state for each 10-second interval (before LS, during LS, and after LS). An interval was categorised as a ‘freezing-like state (FS)’ if freezing-like behaviour persisted for ≥ 8 s, and as a ‘normal state (NS)’ if freezing-like behaviour lasted for < 2 s. During LS, escape behaviour occurred rapidly. Therefore, if high-speed swimming (≥ 6 cm/s) was sustained for 0.2 s (equivalent to one frame at 5 fps), we defined this as a ‘high-speed state (HS)’. This threshold reflects the minimum temporal resolution required to identify continuous motion.Statistical analysisCharacterisation of state transition patterns at the individual levelTo calculate the state transition probabilities for behavioural transitions before, during, and after LS, we constructed state transition matrices by counting the number of transitions between states and normalising each row, following established methods using Markov chain analysis27,28. To visualise the transition dynamics, we created state transition diagrams using python-graphviz v0.20.3, where each state (NS, FS, and HS) was represented by a node, with edges indicating transition probabilities.Furthermore, we used a binomial test to compare whether there were significantly more specific state transition patterns in the series of flows from before LS to after LS. In the binomial test, we set the null hypothesis that ‘the 27 behavioural patterns occur with equal probability (1/27)’ and performed a one-sided test. To control for type I errors due to multiple comparisons, we applied FDR correction to the binomial test results.Analysis of group-level freezing-like states after LSBased on the analysis of individual-level behavioural patterns, we next examined group-level freezing-like states (FS) after LS. For each trial, the number of individuals in the FS after LS was counted. To test whether synchronous FS occurred, virtual datasets were generated by randomly shuffling the states of each trial among groups (17 groups × 1,000 trials; seeds = 1, 2, …, 1,000). The proportions of individuals in each state across all trials were compared between the virtual datasets and the observational data using a chi-square test (scipy v1.10.1). The null hypothesis was defined as: “The presence or absence of the FS for each individual is independent, and synchronous FS for the entire group occur at random.”Classification of group response profilesTo classify these characteristics, we performed a principal component analysis (PCA) on 27 individual-level behavioural patterns from before to after the LS intervention. We then calculated the cumulative contribution rate and reduced the number of dimensions to the minimum required to explain > 95% of the variance. To visualise the cluster structure, we further projected the PCA-reduced data using UMAP29 (umap-learn v0.5.7). Classification was performed using spectral clustering (scikit-learn v1.2.2), and the optimal number of clusters was determined based on the silhouette coefficient. This coefficient approaches 1 when intra-cluster cohesion and inter-cluster separation are high; therefore, the number of clusters yielding the highest silhouette coefficient was selected.Comparison of state transition patterns at the individual level between clustersDifferences in the frequency of state transition patterns between clusters were evaluated using binomial tests with FDR correction, as described above.Analysis of group-level freezing-like statesTo examine differences in group-level freezing-like states (FS) across clusters and between familiar and unfamiliar groups after LS, we applied a generalized linear mixed model (GLMM)30. The dependent variable was the number of individuals exhibiting the FS (0–6) within each group. Cluster identity and the presence or absence of familiarisation were included as fixed effects. Experimental group identity, trial number, and group identity were incorporated as random effects to account for repeated measurements and inter-group variability. The model assumed a binomial distribution with a logit link function, which is appropriate for categorical or count data with hierarchical structure. Analyses were performed in Python v3.8.12. We used pyper v1.1.2 to call R v4.1.2, and the lme4 and multcomp packages for model fitting and post hoc tests. Tukey’s method was applied for multiple comparison correction.ResultsDetection of individual-level state transition characteristicsTo examine how individuals responded to the looming stimulus (LS), we analysed behavioural transitions across three intervals: before, during, and after LS. Fish behaviour was first classified into three types based on swimming velocity: freezing-like (< 0.2 cm/s), normal (0.2–6 cm/s), and high-speed (≥ 6 cm/s). Each 10-second interval was then categorized into one of three behavioural states—freezing-like, normal, or high-speed—according to duration thresholds (≥ 8 s freezing, < 2 s freezing, and ≥ 0.2 s high-speed). These definitions enabled consistent identification of state transitions across trials. Using these thresholds, each of the three temporal intervals was classified into one of the three behavioural states: ‘freezing-like state (FS)’, ‘normal state (NS)’, or ‘high-speed state (HS)’ (Figure S4).To examine how the behavioural states of individuals transitioned from before LS to during LS, and from during LS to after LS, we calculated state transition probabilities using a Markov chain and visualised them as a state transition diagram (Fig. 1c). Between the pre-LS and LS intervals, 39% of individuals transitioned from NS to HS, whereas 60% remained in NS. Among individuals that entered HS during LS, 70% transitioned to FS after LS. In contrast, individuals that remained in the NS during LS had a 76% probability of continuing in that state after LS. These findings suggest that individuals showing escape-like behaviour during LS tended to transition into FS after LS, whereas those unresponsive to LS generally maintained NS.To statistically evaluate trends in state transition patterns, we extracted behavioural state sequences across the three intervals (before, during, and after LS). Each sequence was expressed as a combination of the three defined behavioural states: HS, NS, and FS, resulting in 27 possible transition patterns. We quantified the frequency of each pattern across trials (Fig. 1d).The most frequent transition pattern was the maintenance of NS throughout the three intervals (NS→NS→NS; blue; q < 0.001). The second most frequent pattern was NS→HS→FS (red; q < 0.001), and the third was FS→HS→FS (red; q < 0.001), both involving a transition to HS during LS followed by FS: NS→HS→FS (red; q < 0.001) and FS→HS→FS (red; q < 0.001). Additional significantly overrepresented patterns were NS→HS→NS (grey; q < 0.001), NS→HS→FS (grey; q < 0.001), and FS→NS→NS (blue; q < 0.001), all of which exceeded the expected frequency under a uniform distribution (1/27).Among these six prominent transition patterns (Fig. 1d), NS→NS→NS and FS→NS→NS represent non-reactive behaviours where individuals maintained or returned to the NS during and after LS (blue). In contrast, NS→HS→FS and FS→HS→FS represent reactive responses, characterised by HS during LS followed by FS (red). These two reactive patterns accounted for 41% and 30% of all observations, respectively, and thus constitute the typical individual-level responses to LS stimulation. Notably, escape without subsequent FS (NS→HS→NS; approx. 9%) and no initial response followed by FS after LS (NS→NS→FS; approx. 8%) were also observed at appreciable frequencies.Population polarisation into synchronous freezing-like and non-freezing-like states after LSTo investigate whether medaka groups exhibited synchronous responses (either FS or non-FS) after LS, we counted the number of individuals exhibiting FS in each trial. The distribution was bimodal, with peaks at 0 and 6 individuals (Fig. 1e, blue line), suggesting that entire groups tended to respond uniformly.To determine whether this distribution could be explained by chance, we generated a virtual dataset by randomly shuffling the individual-level FS and non-FS classifications within groups of the same size (Fig. 1e, red line). This reconstructed the expected distribution under the assumption that individuals responded independently of one another. A chi-square test comparing the observed and expected distributions revealed a significant difference (χ² (6) = 147, p < 0.001). This result suggests that the strong bias towards either ‘all-freezing’ or ‘all non-freezing’ within groups after LS is unlikely to have occurred by chance alone. Instead, it indicates that individuals within a group reacted in a synchronous manner through social interaction.Group response profiles to LS classified into three typesDuring the behavioural experiments, we observed groups in which all individuals synchronously exhibited FS, as well as groups in which all individuals remained unresponsive and continued swimming. Moreover, the same groups tended to display similar response tendencies across repeated trials. Based on these preliminary observations, we hypothesised that groups exhibit consistent and characteristic behavioural tendencies, which we define as group response profiles. To evaluate this hypothesis, we classified groups according to their individual-level behavioural patterns. Specifically, behavioural data from 10 trials per group were aggregated, dimensionality reduction was performed using principal component analysis (PCA) followed by UMAP, and spectral clustering was applied to classify the groups.PCA indicated that 16 dimensions were required to exceed a cumulative variance contribution of 95%, and this was adopted as the optimal dimensionality (Figure S5). The reduced data were then projected into two dimensions using UMAP, revealing a clear distinct group-level structures (Fig. 2g). Spectral clustering, guided by the silhouette coefficient identified three as the optimal number of clusters (Figure S6). Accordingly, groups were classified into three clusters (Figure S7).Fig. 2State transition diagrams and frequencies of behavioural state transition patterns for each cluster. (a–c) State transition diagrams for each cluster were generated to represent individual-level behavioural states (FS: freezing-like state, NS: normal state, and HS: high-speed state) before, during, and after LS. All the visual elements are consistent with those in Fig. 1c. (a) State transition diagram for Cluster 0. b) State transition diagram for Cluster (1) (c) State transition diagram for Cluster (2) (d–f) Bar graphs showing the frequency of occurrence for each behavioural state transition pattern in each cluster. Details of colour coding and statistical tests are as described in Fig. 1d. (d) Distribution of transition pattern frequencies in Cluster 0. (e) Distribution of transition pattern frequencies in Cluster (1) (f) Distribution of transition pattern frequencies in Cluster (2) (g) The X-axis represents the first UMAP component and the Y-axis the second. Each point shows the group centroid, obtained by reducing the original 23 dimensions to 16 dimensions using PCA and further to two dimensions using UMAP. Colours indicate classification results based on spectral clustering. (h) The X-axis represents the number of individuals in the freezing-like state after LS, and the Y-axis represents the frequency of these counts across all trials. The lines correspond to the IDs of the three clusters. Tukey’s post hoc test based on a GLMM was performed, and the results of the cluster comparisons are shown within the graph.Full size imageTo analyse state transitions of individuals across the pre-, during-, and post-LS intervals in each cluster, we calculated state transition probabilities using a Markov chain and visualised them as state transition diagrams (Fig. 2a-c).In Cluster 0 (Fig. 2a), the probability of transitioning from NS to HS from before LS to during LS was 72.5%, while the transition probability from FS to HS was 76%. Individuals that exhibited HS during LS had an 89% probability of subsequently transitioning to the FS. In addition, even individuals that remained in NS during LS had an 80.4% probability of transitioning to FS afterwards. These results suggest that in Cluster 0, both responsive (HS) and unresponsive (NS) individuals tended to synchronise into a FS after LS. The most frequent pattern was NS→HS→FS (red; q < 0.001), and the second was FS→HS→FS (red; q < 0.001), both involving a transition to HS during LS followed by FS (Fig. 2d). These findings suggest that Cluster 0 corresponds to groups in which all individuals synchronise and exhibit FS after LS.In Cluster 1 (Fig. 2b), the probability of maintaining the NS from before LS to during LS was 86.2%, and individuals that remained unresponsive during LS continued NS after LS with a probability of 94.2%. The dominant patterns were those where individuals maintained NS throughout (NS→NS→NS and FS→NS→NS, q < 0.001; Fig. 2e, blue). These results suggest that Cluster 1 corresponds to groups in which all individuals synchronise and maintain NS after LS.In Cluster 2 (Fig. 2c, f), NS→NS→NS remained the most frequent pattern (q < 0.001). However, various other transitions were also observed, including NS→HS→NS (q < 0.001; grey), NS→HS→FS (q < 0.001; red), and NS→NS→FS (q < 0.001; grey). This diversity of transitions indicates that Cluster 2 represents a heterogeneous group response profile, reflecting a mixture of multiple individual-level response types rather than a single dominant pattern.Synchronisation of freezing-like and non-freezing-like states across the group profilesIndividual-level state transition analysis revealed that in Cluster 0, individuals frequently responded to the looming stimulus (LS) and then entered FS, whereas in Cluster 1, transitions in which individuals did not respond to LS and continued NS were predominant. We next examined whether all individuals in Cluster 0 synchronised to exhibit FS, and whether all individuals in Cluster 1 synchronised to continue NS. To this end, we counted the number of individuals in FS after LS for each group and compared these counts across clusters. A significant difference was observed in the number of individuals exhibiting FS (Fig. 2h, p < 0.001). In Cluster 0, the most frequent outcome was that all six individuals showed FS, whereas in Cluster 1, the most likely outcome was that no individual showed FS. In Cluster 2, the number of individuals exhibiting FS ranged mostly from zero to three, yielding a distribution distinct from both Clusters 0 and 1. These results indicate that in Cluster 0, individuals tended to synchronise to FS, whereas in Cluster 1 they synchronised to non-FS (continued NS). In contrast, Cluster 2 showed no clear synchronisation, with only a subset of individuals exhibiting FS after LS.Individual-level differences in behavioural transition patterns between familiar and unfamiliar groupsWe determined the behavioural patterns of individuals in the unfamiliar group (Figure S7-8), constructed a state transition diagram (Fig. 3a), and classified individual-level transition patterns into 27 categories, comparing their frequencies of occurrence (Fig. 3b). As in the familiar groups, the most frequent pattern was the non-reactive type (NS→NS→NS, blue; q < 0.001). The pattern (NS→HS→NS, grey) in which fish transitioned to HS during LS and returned to NS afterwards also appeared at a significantly high frequency (q < 0.001). In addition, the pattern in which fish transitioned to HS during LS and then entered FS after LS (NS→HS→FS, red) occurred significantly more often (q < 0.05). However, the patterns in which individuals entered FS after LS (NS→HS→FS, red; NS→NS→FS, grey), which were significantly enriched in the familiar groups, did not reach significance in the unfamiliar group. These findings suggest that individual-level transition patterns differed between the two conditions.Fig. 3State transition diagram of individual-level behaviour (Unfamiliar group). (a) For individual-level behaviours classified as FS: freezing-like state, NS: normal state, or HS: high-speed state, the state transition probabilities were shown from before to during LS and from during to after LS using a Markov chain. Details are as described in Fig. 1c. (b) Bar graphs showing transition patterns and their frequencies from before LS to during and after LS for FS, NS, and HS. Statistical significance is denoted as follows: ***q < 0.001, **q < 0.01, *q < 0.05, and no notation for q > 0.05. Details are as described in Fig. 1d.Full size imageAbsence of group-level synchronous freezing-like state in the unfamiliar groupTo test whether individuals exhibited synchronous responses after LS, we counted the number of individuals in FS per trial for each group and compared the observed data with control data generated by virtual shuffling, as in the familiar groups (Figure S9). In the observed data, the number of freezing-like individuals peaked at zero, and there was a significant difference in both the number and frequency of freezing-like individuals between the observed and virtual data (χ² (6) = 22.1, p < 0.01) (Figure S9). These results indicate that the peak at zero was not coincidental, but rather that all individuals within the unfamiliar group tended to exhibit synchronous non-FS, continuing to NS after LS.Differences in synchronous freezing-like states between familiar and unfamiliar groupsTo examine whether the occurrence of FS at the group level after LS differed depending on familiarisation, the number of individuals exhibiting FS per trial was counted for each group. The aggregated group-level data were then compared between the familiar and unfamiliar groups using GLMM (Figure S10). The results showed that, following LS, the familiar groups tended to have a higher number of individuals exhibiting FS (Figure S10, β = 2.02, p < 0.05) and displayed a bimodal distribution. In contrast, the unfamiliar group showed fewer freezing-like individuals and exhibited a unimodal distribution. This indicates that, unlike the familiar groups, the unfamiliar group lacked the peak where all six individuals exhibited FS after LS.Disappearance of the collective freezing in the unfamiliar groupTo clarify similarities and differences in collective behavioural patterns between familiar and unfamiliar groups, we integrated and analysed data from both conditions. Specifically, we performed dimensionality reduction and clustering based on 27 individual-level behavioural transition patterns. Principal component analysis (PCA) revealed that 17 dimensions were required to explain 95% of the variance, which was therefore set as the optimal number (Figure S11). The silhouette coefficient indicated that the optimal number of clusters was three (Figure S12). The clustering results were visualised using a two-dimensional UMAP embedding derived from the 17 principal components and classified into three clusters by spectral clustering (Fig. 4a). Groups in the familiar condition were distributed across all three clusters, whereas the unfamiliar groups were absent from Cluster 0, indicating a clear bias (Fig. 4b). We next verified that in Cluster 0, all individuals tended to exhibit synchronous FS, while in Cluster 1 they tended to exhibit synchronous non-FS (continued NS). In Cluster 2, synchrony was absent, with only a subset of individuals showing FS after LS. Importantly, the classification showed that the unfamiliar groups were not represented in Cluster 0, showing that the ‘freezing-dominant’ cluster was absent from their collective behavioural profiles (Fig. 4c). To evaluate whether this disappearance of the freezing-dominant response could be attributed to the immediate formation of unfamiliar groups, we compared the distributions of the number of freezing-like individuals between groups tested on the first and the following day. Although a significant difference was detected (χ² (6) = 14.2, p = 0.028), this was mainly due to a slight increase in groups with two or four freezing-like individuals, whereas the frequency of ‘all-freezing’ remained almost unchanged (Figure S13). These results suggest that handling or grouping stress immediately after formation had little effect, and that the disappearance of freezing-dominant is a robust feature of unfamiliar groups.Fig. 4Visualisation of dimensionality reduction using PCA and UMAP, and comparison of freezing-like states across clusters. (a–b) The X-axis represents the first UMAP component and the Y-axis the second. Each point corresponds to the group centroid. (a) Colours indicate cluster IDs obtained by spectral clustering. (b) Colours indicate familiarisation status: familiar groups (blue) and unfamiliar groups (orange). (c) Distribution of freezing-like states (FS) after LS exposure in the integrated dataset combining familiar and unfamiliar groups. Details are as described in Fig. 2h.Full size imageDiscussionIn this study, we established a quantitative behavioural assay to analyse collective decision-making in medaka (Oryzias latipes) in response to a looming stimulus (LS). In this system, small groups of medaka were presented with an LS that mimicked an approaching predator, and their collective behavioural choices were examined. Two dichotomous collective response patterns consistently emerged at the group level: ‘all-freezing’ and ‘all non-freezing’. Furthermore, the distribution of the number of FS individuals per trial was bimodal, with clear peaks at either zero or six individuals. These results demonstrate the presence of a dichotomous collective behavioural choice in medaka and validate as a robust tool for investigating collective decision-making under controlled laboratory conditions. Moreover, this assay will provide a platform for elucidating the genetic and neural bases of collective decision-making in vertebrates.Previous studies of collective decision-making under laboratory conditions have primarily used small fish species, such as sticklebacks and golden shiners2,4,6, often focusing on wild populations in ecological contexts. By contrast, our study employed medaka, a well-established genetic model organism, thereby enabling experimental systems in which genetic and environmental factors can be controlled. This approach enables the establishment of highly reproducible behavioural assays and allows for long-term monitoring of behavioural development, from the individual to the group level.Most previous studies of collective decision-making have focused on gradual responses to predators6,7. In contrast, our findings revealed a novel phenomenon: rapid collective responses to sudden visual threats. In coral reef fishes, escape responses of individuals can be predicted from the expansion rate of looming stimuli or the behaviour of neighbours24. However, how these individual responses converge into synchronous group-level behaviour remains unclear. Our results demonstrate that under time-constrained predatory threat, rapid collective decision-making can emerge, thereby complementing existing models of gradual escape behaviour.Our findings further suggest that a certain period of familiarisation is required for collective behavioural choices in response to the LS. In particular, in familiar groups, many individuals transitioned from HS during LS to FS afterwards, and entire groups tended to enter the FS. Such consistent behavioural synchrony was mainly observed in groups that had undergone sufficient familiarisation, suggesting that social familiarity may contribute to coordinated collective decisions. In our experiments, social familiarity increased the proportion of individuals that exhibited a FS after escape-like HS, indicating a change in behavioural regularity at the individual level. However, these individual-level changes alone cannot fully explain the emergence of dichotomous collective outcomes (all-freezing versus all non-freezing). It remains unclear whether familiarisation (1) enhanced each individual’s social sensitivity, making them more likely to be influenced by others, or (2) homogenised behavioural traits within groups. Distinguishing between these two possibilities was beyond the scope of this study. Future approaches incorporating longitudinal tracking of individually identified fish and quantitative measures of behavioural synchrony will be necessary to address this question.If explanation (1) is correct, repeated interactions during familiarisation may allow individuals to recognise and predict the behaviour of conspecifics, thereby strengthening group-level properties such as polarisation and alignment. Previous studies have reported various effects of social familiarity in fish. For example, in female guppies, 12 days of familiarisation led to preferential associations with familiar conspecifics31. Social familiarity has also been shown to promote group cohesion and alignment in guppies14 and to enhance information transfer under social threat in damselfish17. In addition, familiarisation may also induce social fear contagion. In zebrafish, individuals are known to switch from high-speed swimming to freezing when exposed to alarm cues from conspecific skin extracts13,32, suggesting that this behavioural pattern may be widespread among fishes. Moreover, zebrafish exhibit similar freezing-like responses when observing familiar conspecifics or groups displaying fear responses13,33.On the other hand, explanation (2), that familiarisation homogenises behavioural traits within groups, cannot be excluded. Previous studies have shown that bold individuals tend to maintain stable behavioural traits, whereas shy individuals are more plastic and influenced by social context. For instance, in guppies, bold individuals rely on their own information and explore independently, whereas shy individuals adjust their behaviour according to social information34. In sticklebacks, bold individuals also show stable exploratory behaviour, while shy individuals display behavioural plasticity and can change over time35. However, to our knowledge, no studies have directly demonstrated long-term homogenisation of behavioural traits caused by familiarisation in any animal species. Thus, we consider explanation (1) to be the more plausible mechanism underlying our findings.Our study therefore extends previous fish familiarity research, which has reported average increases in shoal cohesion, by showing that long-term social familiarity also structures the variability of collective decisions. Familiar groups not only became cohesive; they also differentiated into groups that reached full consensus (freezing-dominant or non-freezing-dominant) and groups that failed to do so (mixed-type), revealing that familiarity regulates the probability—rather than the inevitability—of consensus formation under threat.In summary, our results indicate that social familiarity promotes the dichotomisation of collective behavioural choices in medaka, and that factors such as social familiarity or changes in social sensitivity may contribute to this process. Although further investigation will be required to directly verify these mechanisms, our study provides a foundation for exploring how social experience shapes collective decision-making under time-constrained predatory threats in vertebrates.

Data availability

All data generated or analysed during this study are available from the corresponding author on reasonable request.
ReferencesKrause, J. & Ruxton, G. D. in Living in Groups. (eds Krause, J. & Ruxton, G. D.) (Oxford University Press, 2002). https://doi.org/10.1093/oso/9780198508175.002.0001Couzin, I. et al. Uninformed individuals promote democratic consensus in animal groups. Science 334, 1578–1580 (2011).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Strandburg-Peshkin, A., Farine, D. R., Couzin, I. D. & Crofoot, M. C. Shared decision-making drives collective movement in wild baboons. Science 348, 1358–1361 (2015).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Ward, A. J., Krause, J. & Sumpter, D. J. Quorum decision-making in foraging fish shoals. PloS One. 7, e32411 (2012).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Clément, R. J. G., Wolf, M., Snijders, L., Krause, J. & Kurvers, R. H. J. M. Information transmission via movement behaviour improves decision accuracy in human groups. Anim. Behav. 105, 85–93 (2015).Article 

Google Scholar 
Ward, A. J. W., Sumpter, D. J. T., Couzin, I. D., Hart, P. J. B. & Krause, J. Quorum decision-making facilitates information transfer in fish shoals. Proc. Natl. Acad. Sci. 105, 6948–6953 (2008).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
McComb, K. et al. Leadership in elephants: the adaptive value of age. Proc. R Soc. B Biol. Sci. 278, 3270–3276 (2011).Article 

Google Scholar 
Shang, C. et al. Divergent midbrain circuits orchestrate escape and freezing responses to looming stimuli in mice. Nat. Commun. 9, 1232 (2018).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Ward, A. J. W. & Hart, P. J. B. The effects of kin and familiarity on interactions between fish. Fish. Fish. 4, 348–358 (2003).Article 

Google Scholar 
Strodl, M. A. & Schausberger, P. Social familiarity reduces reaction times and enhances survival of group-living predatory mites under the risk of predation. PLOS ONE. 7, e43590 (2012).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Riley, R. J., Kwon, Y. M., Manica, A. & Savage, J. L. Familiarity dampens the effect of boldness on coordination in three-spined sticklebacks. Behaviour 162, 191–206 (2025).Article 

Google Scholar 
Galhardo, L., Vitorino, A. & Oliveira, R. F. Social familiarity modulates personality trait in a cichlid fish. Biol. Lett. 8, 936–938 (2012).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Fernandes Silva, P., Garcia de Leaniz, C. & Luchiari, A. C. Fear contagion in zebrafish: a behaviour affected by familiarity. Anim. Behav. 153, 95–103 (2019).Article 

Google Scholar 
Davis, S., Lukeman, R., Schaerf, T. M. & Ward, A. J. W. Familiarity affects collective motion in shoals of guppies (Poecilia reticulata). R Soc. Open. Sci. 4, 170312 (2017).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Chivers, D., Brown, G. & Smith, J. Familiarity and shoal cohesion in Fathead minnows (Pimephales promelas): implications for antipredator behaviour. Can. J. Zool. 73, 955–960 (1995).Article 
ADS 

Google Scholar 
Griffiths, S., Brockmark, S., Höjesjö, J. & Johnsson, J. Coping with divided attention: the advantage of familiarity. Proc. Biol. Sci. 271, 695–699 (2004).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Nadler, L. E., McCormick, M. I., Johansen, J. L. & Domenici, P. Social familiarity improves fast-start escape performance in schooling fish. Commun. Biol. 4, 897 (2021).Article 
PubMed 
PubMed Central 

Google Scholar 
Kadak, K. & Miller, N. Follow the straggler: zebrafish use a simple heuristic for collective decision-making. Proc. R. Soc. B Biol. Sci. 287:20202690 (2020).Imada, H. et al. Coordinated and cohesive movement of two small conspecific fish induced by eliciting a simultaneous optomotor response. PLoS One. 5, e11248 (2010).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Ochiai, T., Suehiro, Y., Nishinari, K., Kubo, T. & Takeuchi, H. A new data-mining method to search for behavioral properties that induce alignment and their involvement in social learning in Medaka fish (Oryzias latipes). PLoS One. 8, e71685 (2013).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Yilmaz, M. & Meister, M. Rapid innate defensive responses of mice to looming visual stimuli. Curr. Biol. 23, 2011–2015 (2013).Article 
CAS 
PubMed 

Google Scholar 
Temizer, I., Donovan, J. C., Baier, H. & Semmelhack, J. L. A visual pathway for looming-evoked escape in larval zebrafish. Curr. Biol. 25, 1823–1834 (2015).Article 
CAS 
PubMed 

Google Scholar 
Ferreira, C. H. & Moita, M. A. Behavioral and neuronal underpinnings of safety in numbers in fruit flies. Nat. Commun. 11, 4182 (2020).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Hein, A. M., Gil, M. A., Twomey, C. R., Couzin, I. D. & Levin, S. A. Conserved behavioral circuits govern high-speed decision-making in wild fish shoals. Proc. Natl. Acad. Sci. 115, 12224–12228 (2018).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Takeuchi, T. & Manabe, E. Genetical study on the new mutant of the fading medaka, Oryzias latipes. Res. Bull. Shujitsu Women’s Coll. Shujitsu Jr Coll. 14, 1–18 (1984).
Google Scholar 
Yamanaka, O. & Takeuchi, R. UMATracker: an intuitive image-based tracking platform. J. Exp. Biol. 221, jeb182469 (2018).Article 
PubMed 

Google Scholar 
Tuqan, M. & Porfiri, M. Mathematical modeling of zebrafish social behavior in response to acute caffeine administration. Front. Appl. Math. Stat. 7, 751351. https://doi.org/10.3389/fams.2021.751351 (2021).Article 
PubMed 
PubMed Central 

Google Scholar 
Syme, J., Kiszka, J. J. & Parra, G. J. Behavioural variation facilitates coexistence and explains the functions of mixed-species groups of sympatric delphinids. Anim. Behav. 210, 395–408 (2024).Article 

Google Scholar 
McInnes, L., Healy, J. & Melville, J. U. M. A. P. Uniform Manifold Approximation and Projection for dimension reduction. arXiv:1802.03426v3 (2020).Bolker, B. M. et al. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24, 127–135 (2009).Article 
PubMed 

Google Scholar 
Griffiths, S. W. & Magurran, A. E. Familiarity in schooling fish: how long does it take to acquire? Anim. Behav. 53, 945–949 (1997).Article 

Google Scholar 
Masuda, M. et al. Identification of olfactory alarm substances in zebrafish. Curr. Biol. 34, 1377–1389e7 (2024).Article 
CAS 
PubMed 

Google Scholar 
Akinrinade, I. et al. Evolutionarily conserved role of Oxytocin in social fear contagion in zebrafish. Science 379, 1232–1237 (2023).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Trompf, L. & Brown, C. Personality affects learning and trade-offs between private and social information in guppies, Poecilia reticulata. Anim. Behav. 88, 99–106 (2013).Article 

Google Scholar 
Jolles, J., Briggs, H., Araya, Y. & Boogert, N. Personality, plasticity and predictability in sticklebacks: bold fish are less plastic and more predictable than shy fish. Anim. Behav. 154, 193–202 (2019).Article 

Google Scholar 
Download referencesAcknowledgementsWe thank Dr. Tetsuro Takeuchi for sharing the fading strain. We thank Drs. Masahiro Daimon and Masayuki Koganezawa for their advice on the development of the behavioural assay. We thank Drs. Ken-Ichiro Tsutsui, Hiromu Tanimoto, Jamie M. Kass and Towako Hiraki-Kajiyama for comments on the manuscript.FundingThis work was supported by the National Institute for Basic Biology Priority Collaborative Research Project 10–104 (to H.T.), 19–347 (to H.T.), and 21–335 (to H.T.); a grant for Joint Research (#01111904) by the National Institutes of Natural Sciences (to H.T.); Japan Society for the Promotion of Science (JSPS) KAKENHI Grants 21H04773 (to H.T.), 20H04925 (to H.T.), 18H02479 (to H.T.), 22H05483 (to H.T.), 23K27205 (to H.T.), 24H01216 (to H.T.) and 24K21957 (to H.T.). Takeda Science Foundation (to H.T.), and the natural science grant of the Mitsubishi Foundation (to H.T.); Japan Science and Technology Agency (JST) SPRING, Grant Number JPMJSP2114(to R.N.).Author informationAuthors and AffiliationsMolecular Ethology Laboratory, Graduate School of Life Science, Tohoku University, Sendai, 980-8577, JapanRyohei Nakahata & Hideaki TakeuchiDepartment of Cardiac Regeneration Biology, National Cerebral and Cardiovascular Centre, Osaka, 564-8565, JapanRyohei NakahataAuthorsRyohei NakahataView author publicationsSearch author on:PubMed Google ScholarHideaki TakeuchiView author publicationsSearch author on:PubMed Google ScholarContributionsR.N. and H.T. designed experiments. R.N. conducted experiments, wrote code and analysed data. R.N. and H.T. co-wrote and edited the paper and supervised the project.Corresponding authorsCorrespondence to
Ryohei Nakahata or Hideaki Takeuchi.Ethics declarations

Competing interests
The authors declare no competing interests.

Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissionsAbout this articleCite this articleNakahata, R., Takeuchi, H. Social familiarity shapes collective decision-making in response to looming stimuli in Medaka fish.
Sci Rep 15, 43650 (2025). https://doi.org/10.1038/s41598-025-30656-4Download citationReceived: 14 September 2025Accepted: 26 November 2025Published: 23 December 2025Version of record: 23 December 2025DOI: https://doi.org/10.1038/s41598-025-30656-4Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative More

AbstractPeri-urban agroforestry can provide affordable, fresh, and nutritious food and a departure from conventional forms of cropping. Indigenous foods are well-adapted to local conditions, and may hold cultural and economic value for peri-urban residents. Social, ecological, and economic variables influence the feasibility of indigenous agroforestry in peri-urban areas. This study uses participatory mapping and geographic information systems (GIS) to assess these variables and to map suitable spaces and species for peri-urban indigenous agroforestry at three peri-urban sites in Durban, South Africa. We find that: land tenure, livelihood opportunities, and indigenous food perceptions factor into socioeconomic preferences; topography and soil quality influence ecological feasibility; access to water and roads influences perceived economic viability. Although GIS techniques can identify land suitability, participatory mapping adds local fine-scale context to enhance decision-making. Based on the social-ecological conditions at the three sites, we suggest specific configurations of locally adapted foods and farm designs for peri-urban agroforestry. Our study demonstrates how agroforestry is more feasible in places where basic living conditions are fulfilled, and how co-design can improve recognition of local needs, accessibility to services, and balancing urban green equity.

Similar content being viewed by others

The impact of multiple agricultural land uses in sustaining earthworm communities in agroecosystems – A global meta-analysis

Article
Open access
04 December 2024

Analysis of food system drivers of deforestation highlights foreign direct investments and urbanization as threats to tropical forests

Article
Open access
16 July 2024

Impacts of commodity prices and governance on the expansion of tropical agricultural frontiers

Article
Open access
22 April 2024

IntroductionUrban and peri-urban food production can improve the resilience of food systems in multiple ways1. From a logistic viewpoint, it can reduce the risk of supply chain failure and subsequent food and nutritional insecurity. Ecologically, it can reduce impacts from large-scale farming and transportation2, and socioeconomically, it can provide urban residents accessible, affordable, and nutritious alternatives to mass-produced and processed foods3. However, availability of land, labour, and materials are the main determinants of urban and peri-urban food production, in the global North and South4. In the face of densification and development, green space is a critical yet contested component of the urban landscape5,6. Cities worldwide have various legislations and allocations for food production in urban and peri-urban areas in communal gardens, private farms, and food forests7. These allocations help city planning to balance local economies, development interests, and urban environments, often in collaboration with local residents8. City-level adaptations are often crucial for successful implementation of national and regional food policies9,10.Urban and peri-urban agriculture is an emerging and potent response to provisioning fresh and nutritious food, closing nutrient loops, creating circular economies, and reducing carbon footprints11. It can take various forms, from intensive indoor vertical farms, to communal agroecological (including agroforestry) spaces, with several intermediate configurations of social, ecological, and technological variables12. In this article, we focus on urban agroforestry as the proposed intervention to improve food and nutritional security among urban and peri-urban dwellers. Urban agroforestry systems, defined as urban landscapes combining crops and trees are increasingly recognised as productive landscapes with greater allied cultural and ecological benefits than conventional agriculture13. The cultural and recreational values associated with urban agroforestry systems can facilitate more equitable and widespread uptake of the food and nutritional produce yielded by these landscapes (ibid).The feasibility of urban and peri-urban agroforestry could vary across different urban contexts. For example, in densely populated or historically established sections of cities, it could involve planting fruit trees and/or food crops along verges of transportation and utility lines that provide substantial nutrient yields14,15,16. In some cases, urban and peri-urban brownfields (previously developed land) may be reclaimed by municipalities or citizen collectives to grow food17,18,19. Urban parks and gardens established primarily for recreation may also be a significant and legitimate source of food and nutrition20,21,22. Structural constraints to urban and peri-urban agroforestry include the availability of contiguous land and arable soil4,11 and also resident and developer preferences for gentrified forms of nature, neighbourhoods, and greenspaces23. While biophysical and infrastructural variables can help determine suitability for urban agroforestry, social structures and perceptions are crucial to its long-term sustainability24.Indigenous crops and trees are important components of agroecological systems. They are often resilient to local ecological stresses and shocks25, as well as human disturbance and extraction26,27. On farms, indigenous crops and trees provide pollination services, alternative income, and nutrition for farmers28. In urban and peri-urban areas, they can also provide habitat connectivity to wildlife29,30, including pollinators important to rural and urban food production31. This makes them ideal candidates for fragmented landscapes of high-intensity human use, such as urban and peri-urban areas, where large-scale farming is impractical. Foods from indigenous crops and trees are rich in high-quality micronutrients32,33, which are generally deficient in urban diets due to constrained accessibility and affordability. Recent research on indigenous crops has focussed on nutritional yields and land suitability for annual crops such as grains and tubers34,35. Although the potential of indigenous food-bearing tree species has been recognised36,37,38, the research and application of these in agroforestry is still nascent114,115. Therefore, in this study, we also attempt to identify the feasibility of planting indigenous crops and trees, comprising indigenous agroforestry, in urban and peri-urban areas, identifying synergies and constraints as applicable.Participatory mapping was employed to document the cultural, economic, and social values of peri-urban agroforestry, seeking to establish its potential in enhancing livelihoods and promoting environmental sustainability. This involved mapping the variables favouring indigenous food production at three study sites. Thus, a suite of social science methods were used to elicit spatial and temporal data, trends, and preferences in landscapes and land uses39. The participatory mapping was conducted to promote democratic, inclusive, and locally appropriate decision-making when combined with GIS modelling techniques40. This is especially important in urban and peri-urban areas, where land use and land cover are fast-changing, and can often leave under-resourced communities impoverished41. People’s values for landscape features and uses can play an important role in successful landscape governance42, including the implementation and observance of regulations43. Peri-urban areas in the Global South differ from many in the Global North, in that the regulations and infrastructure in the former are not as organised and developed as urban areas44. This situation underscores the need for participatory mapping and GIS modelling to enable better planning and service provisioning in peri-urban areas in the Global South45. Our study follows a mixed methods approach giving equal importance to communities and experts in the mapping process46.In this study, we seek to design locally appropriate indigenous agroforestry systems for urban and peri-urban areas with the aim of improving their food and nutritional security47. We combine data on social perceptions, spatial modelling, and indigenous agroforestry species to generate these designs, which are intended to inform local communities and municipal departments on feasible agroforestry and food security initiatives. Our study demonstrates how government policies and programmes, e.g.55,56,115 can be operationalised at local scale, e.g.9,93,96,116 by combining participatory research and different forms of knowledge. This study is part of the Durban Research Action Partnership between the local metropolitan eThekwini Municipality and University of KwaZulu-Natal, which aims to generate knowledge and learning to address the gap between scientific research, policy development and management within a local government113.The local contextAs in the case of many developing nations, households in South Africa experience the triple burden of malnutrition, which includes undernutrition (stunting and wasting), micronutrient deficiencies (often termed hidden hunger), and overnutrition (overweight and obesity)48. Urbanising and westernising lifestyles influence the preference for cheap, convenient, ultra-processed and packaged food over traditional, nutritious, and fresh, diverse farm-based food49. Post-apartheid market liberalisation has facilitated the penetration of cheap and calorie-dense low-nutrient foods into local markets for consumers and incentivised the export of high-quality foods such as fruit and vegetables to foreign markets for producers50. Smallholder farmers who cannot export or sell to mainstream domestic markets often struggle with a lack of infrastructure and institutional support to improve yields and sales51.In the broader socioeconomic sense, unemployment and economic inequality result in income poverty and food poverty, and limited opportunities for people experiencing poverty to engage in either primary production or secondary activities to secure an income52. Particularly in cities, legacy spatial planning also constrains access to greenfields (previously undeveloped land) and greenspace, which can often be a source of food or materials to support the household economy37,53 As a result of apartheid-era policies, green infrastructure and public access to it is well-developed in affluent neighbourhoods, but is severely lacking in poorer (often racially differentiated) neighbourhoods54. Allocating and enriching urban and peri-urban spaces for food production are a priority on the National Development Plan for South Africa55, and could also contribute to national-level cross-cutting initiatives like the Integrated Food Security and Nutrition Programme and the Natural Resources Management Programme56. In this study three communities in the peri-urban areas of the eThekwini Metropolitan Municipality (that houses Durban, hereafter eThekwini) were consulted to identify spaces where food production can be undertaken, to enhance food and nutritional security.MethodsConceptual framingThe study aimed to identify the most compatible configurations of peri-urban food production given the social-ecological conditions at each site. The study used an overarching landscape ecology approach57 to define the landscape configuration, land use, land cover change, and landscape management guidelines. The study combined participatory mapping and GIS suitability analyses to identify suitable areas for peri-urban food production. The research objective was achieved by answering three questions (Fig. 1) in both the participatory mapping and suitability analysis approaches. The methodology characterised local social-ecological factors and existing and potential land use for food production at each site (Fig. 1). These were analysed to produce socioeconomic and land use guidelines suggesting configurations of peri-urban food production suitable to each site. Landscape configuration was determined using spatial datasets (Table 1). The land use and cover change were determined from social data, to account for socio-polictical nuance about these changes. The management guidelines emerging from the analysis of these datasets were combined to propose locally appropriate landscape management guidelines that included biophysical suitability maps, social aspirations and needs. The conceptual framing and resulting management guidelines resonate with recent calls for agriculture to incorporate indigenous species as ‘trade-ons’, through appropriate design for ‘land maxing’20,114. The management guidelines resulting from this study are applicable examples of mainstreaming biodiversity and indigenous knowledge into productive and adaptive peri-urban agroforestry115.Fig. 1Conceptual framing of the study, research questions, data collected, and analyses.Full size imageTable 1 Influence percentage for each factor used to produce the final suitability maps.Full size tableStudy areaThe eThekwini municipality is host to a population of 3.9 million people, in its urban centre of Durban, as well as several peri-urban areas58. Due to legacy planning and diverse tenure systems, the city centre and suburbs have designated greenspace, whereas the peri-urban areas are more informal and sporadic in structure. About 44% of the land in eThekwini falls under the Ingonyama Trust, governed by traditional chiefs, and is not subject to the same planning requirements as municipal land37. The Durban Metropolitan Open Space System (DMOSS) was instituted in the 1990s to plan and govern land use across formal, informal, protected, and indigenous greenspace in urban and peri-urban areas59,60,61. Under this system, land use is restricted in areas of ecological importance, ecological restoration offsets are required where feasible, and urban greening and agroforestry are promoted in collaboration with municipal departments and NGOs62. The municipality routinely undertakes reforestation and restoration across these open spaces, with the dual intention of improving biodiversity and supporting local bio-economy livelihoods63. There is also a strong emphasis on removing and controlling invasive alien species and planting endemic and indigenous species in greenspaces37. Given this background, our research questions consider the diversity of land tenure and indigenous species, especially trees, that are of interest at each study site (Fig. 2). The three study sites were suggested by eThekwini Municipality as areas of interest for rolling out peri-urban agroforestry interventions, as part of their municipal agroecology programme9,116.Fig. 2Location of study sites in the local context and South Africa (inset). Some sites are fragmented because of dual land tenure—traditional and municipal.Full size image

Maphephetheni: situated on the mountainous area surrounding the Inanda reservoir on the Umgeni River, built in the 1980s, contiguous with the suburb of Inanda. The peri-urban settlement consists of savanna and grassland vegetation. Areas degraded by invasive species (e.g. Acacia spp.) and fire are being replanted with useful food and forage species (e.g. Canthium spp., Ficus spp., Searsia spp.) to encourage sustainable land use64. The municipality’s erstwhile Environmental Planning and Climate Protection Department (EPCPD), now the Biodiversity Management Department, engages members within these communities to nurture saplings for restoration, increase awareness and stewardship and prevent cyclical degradation and restoration. The area has a population density of 344 persons per km2 as per the 2011 census.

Ntshongweni: situated along the ridge of the Shongweni dam on the Mlazi River, built in the 1920s, with vegetation consisting of riparian forest, grassland, and some wetland. Rail and road connections to the urban centres of Durban and Pietermaritzburg have attracted investment in transport and logistics centres in the vicinity, and more recently, in commercial retail and residential development. The EPCPD also runs invasive alien control and reforestation programmes at Ntshongweni. The area has a population density of 399 persons per km2 as per the 2011 census.

Osindisweni: situated along the ridge of the Hazelmere dam on the Mdloti River, built in the 1970s, with vegetation consisting of riparian forest and grassland. It is adjacent to Buffelsdraai, a site historically degraded by intensive sugarcane farming, and currently serving (since 2008) as a suburban landfill ring-fenced by indigenous forest fragments65. These fragments are gradually expanded and connected by ongoing planting, and although most of the forest is protected, the periphery and a small section of the site have been earmarked to grow indigenous food-bearing tree species for surrounding communities. The EPCPD does not yet run restoration programmes at Osindisweni. The area has a population density of 439 persons per km2 as per the 2011 census.

Participatory mapping workshopsOne participatory mapping workshop was conducted in each community between September 2021 and March 2022. Local chiefs and councillors were approached for their consent to engage with the community, and for assistance in recruiting community members to participate in the workshops. We acknowledge that this recruitment strategy may have resulted in a representational bias, but assert that we communicated to each chief and councillor the need to engage with all sections of the community including youth, elders, employed, unemployed, and women. The aim of the research was introduced at the beginning, and informed consent was obtained from all participants to record their responses and take photographs for research purposes only. The study was ethically reviewed and approved by the Humanities and Social Sciences Research Ethics Committee of the University of KwaZulu-Natal in June 2021 (Protocol Reference Number HSS/1971/017D). All methods were performed in accordance with the Economic and Social Research Council guidelines on ethical scientific research.The outline map (Online Appendix Fig. A1) of the community with key features, namely, rivers, roads, schools, and hospitals, was presented to the participants. They were asked: (i) What are the various greenspaces in the community, and what are their tenure and access terms? (ii) What are the resources and uses associated with each of these greenspaces? (iii) What are the positive and negative characteristics of these greenspaces? (iv) What changes have these greenspaces undergone in the past 10 years? (v) What changes, if any, would the community like to see in these greenspaces? (vi) What species of food, especially indigenous trees, grow or are grown in the community, and where? (vii) What food species would the community want growing in their greenspaces, and where? We used the most open and commonly accepted definition of greenspace, implying undeveloped land that harbours some form (cultivated or wild) of vegetation, and is used for one or more of the purposes of: agriculture and food cultivation, cultural and recreational activities, foraging, fishing, and grazing66.A native isiZulu speaker interpreted the questions and responses, and all responses were recorded on the map during the discussion. Names of places and indigenous plants were recorded in isiZulu. “Tree” spaces were recorded as a separate category overlapping with other types of spaces. They included home gardens, sports fields, and open spaces, as they may be fragmented yet productive in their food and non-food yields (e.g. fibre, fuel, medicine, wood). Food production was recorded as a use only when explicitly mentioned by participants (e.g. home gardens or open spaces where food was gathered from plants but not grown as crops were not deemed used for food production). The data collected were analysed qualitatively using reflexive grounded theory, e.g.67 for emergent themes and descriptions in MS Word. Quotes from participants were anonymised using the monikers ‘Respondent n’, and presented to illustrate examples, claims, and arguments.Species selectionSocial-ecological attributes of land use and cover change at each site were derived from the data shared by respondents. Landscape design configurations were suggested in response to these attributes, with functions such as biophysical tolerance and cultural importance. The landscape management guidelines recommended biome-appropriate indigenous food species from36,68, and69. The underlying philosophy of this species allocation was to increase land productivity for local food and nutritional security, regenerating natural capital for human and environmental health, and enabling local communities to generate human, social, physical and financial capital through the planting of useful indigenous species114.Suitability analysesSeveral factors influence land suitability for urban agricultural farming. Biophysical, socio-economic, and technical aspects are some of the primary factors. The principal purpose of land suitability for urban crop farming is to predict the potential and limitation of land for crop production70. Generally, determining suitable areas for crop farming in urban areas revolves around making the most sustainable use of land resources while avoiding depleting other resources71. Crop farming land suitability analysis requires an efficient decision support system to analyse and interpret the related ecological, environmental and spatial information. GIS and participatory GIS are combined with multicriteria decision analysis (MCDA) methods to deliver a better spatial decision72.This study determined suitable areas for peri-urban food production in three stages. First, the factors affecting the agricultural uses were set up as criterion maps. Secondly, all the factors were scored in the suitability range based on expert opinion and the results from the participatory mapping workshops. Finally, GIS spatial analysis modelling techniques were used to generate suitability maps for the three sites.The study adopted six factors, as suggested by73, to set up criterion maps. The factors include land cover, agricultural land capability, dominant soils, slope, proximity to water sources and proximity to the main road. The weighted overlay in ArcGIS Pro was used to generate the final suitability maps based on the percentage of influence for each geographic factor. Here, the influence of each factor (weights) was arbitrarily chosen based on the results of the interviews and experts’ knowledge. Thus, each layer contributes to the influence based on the type of agricultural land use (Table 1). In this study, the final suitability maps were reclassified into five classes with suitability scales ranging from highly suitable to not suitable.Results and discussionCharacterising greenspace attributes, perceptions, preferences and potential for peri-urban agroforestry through participatory mappingThe workshops lasted about 90 min at each site, and involved between 11 and 29 participants. Participants included but were not limited to, representatives of the ward, workers with different departments of the municipality such as community services, education, environment, and health and sanitation, local smallholder farmers, part-time employed and unemployed youth and elders, private sector employees, and representatives of local NGOs, churches, and cooperatives. The participation in the workshops was variable and limited due to the ongoing Covid-19 restrictions at the time. Nevertheless, we believe the depth and diversity of the discussions are representative of the sites.Current land use of greenspaces included provisioning and recreation, although the former was reported as significant only at Maphephetheni and Ntshongweni (Table 2). Greenspaces are an important avenue for urban and peri-urban foraging at all three sites, providing residents with resources and recreational opportunities76,77. Food production was intentionally undertaken in communal and public greenspaces at the aforementioned sites, but not at Osindisweni. Tenure over greenspaces also varied across the sites, and areas under the traditional authority, i.e. the local chief, were used for recreation and grazing, but not to grow food. Land tenure is an important driver of land use and stewardship, and traditional authority tenure can deter land-based livelihoods such as food production and agroforestry. For example, lack of accountability and definition in spatial allocation in communal areas can result in violation of land use agreements78, reducing certainty of long-term land use, and subsequent investment of labour and capital in food production79. It may also result in rent appropriation by powerful stakeholders at the expense of the community80 and undemocratic development on land intended for food production, especially in urban and peri-urban areas81. This may partly explain why at Osindisweni, where most greenspaces are under communal tenure, participants expressed low interest in food production and agroforestry.Table 2 Existing land use: types of greenspaces, their tenure, and uses at the three study sites. Species information is listed in Online Appendix Table A1.Full size tableThe productivity of greenspaces varied across sites, with Ntshongweni residents earning and saving money from the sale of food produced in home and public greenspaces (Table 3). Participants at Ntshongweni expressed an interest in diversifying their food production by including indigenous crop and animal species.Table 3 Perceived land use potential and food production feasibility in greenspaces at the three study sites.Full size table“The municipality [representative] tells us that there is a market for indigenous crops and chickens. We would like to learn about how to farm these so that we can sell not just within our communities, but also to the urban market.”—Respondent 1.Natural greenspaces were “far away” for residents of Maphephetheni and Osindisweni.“[That place] is far away, so we visit only on some weekends, maybe once or twice a year. When we go there, it is with family and friends. We can take our time and be one with nature.”—Respondent 2.These observations make a case for the development of more accessible parks and gardens for residents closer to residential areas. Planning for such should consider local perceptions of safety and environmental quality to minimise unintended consequences such as dereliction or gentrification82. Across all three sites, lack of plant material, stable water supply, livestock predation, and know-how were reported as hindrances to food production in community food and school gardens.“There are times in the summer when we don’t have water [on tap] for some ten, twenty days. This is when the plants also need water, and we also need [drinking] water. That’s why our [community food] gardens are not successful. The crops die.”—Respondent 3.“What we need to know is how to grow crops and trees properly. Both common and indigenous ones. We need to learn how to water them care for them, how to harvest them at the right time.”—Respondent 4.Participants made different site-specific recommendations to improve food productivity. For example, in Maphephetheni, home gardens were considered more effective than public gardens, as protecting them from water shortages, flooding, and livestock and human predation was easier. On the other hand, Osindisweni respondents prioritised shops and soup kitchens as means to improve food security, as they believed their land to be no longer viable for food production due to pollution associated with the landfill.“We live close to the city. We do not need to grow our own food. What we need is more shops to buy our food from. We need schools and soup kitchens to support our people with meals for food security. This is the support we need from the government.”—Respondent 5.“The soil here is very degraded. There is so much dumping, so many fires. People suffer from respiratory problems because of this environment. Crops and trees will never grow here. If the municipality wants to help us, they should collect our garbage more regularly.”—Respondent 6.Suggesting locally occurring, useful indigenous species suited to respective site attributes for peri-urban agroforestryBased on the pros, cons, and potential identified in the previous stages, we characterise seven site attributes and six response functions that can be served by greening for urban food production, in addition to improving food and nutritional security (Table 4). We suggest using thorny plants as fencing structures to prevent livestock predation while simultaneously maintaining biomass for humans and non-humans in the form of fruits and fodder. Given the use of greenspaces for non-food and non-timber products and the need for invasive alien replacement, indigenous trees with multiple uses can be planted in various greenspaces. Some of these species already grow in greenspaces across these sites (Table A1) but were not referred to as serving the proposed functions. None of the herbs or crops were specifically mentioned by participants during the elicitation at the workshops.Table 4 Design configurations based on synthesised site attributes, desired response functions, reviewed literature on indigenous food species, and participatory mapping locations, for Maphephetheni (M), Ntshongweni (N), and Osindisweni (O). (Y = Yes, N = No, indicating species suitability at site).Full size tableWe acknowledge that cultivation of some of the trees and crops suggested in Table 4 may require significant investment in technical training and infrastructure. For example, Carissa, Dovyalis, and Harpephyllum are dioecious species, requiring careful selection and planting of sufficient male and female plants in close proximity to ensure fruiting. Cultivation of crops, especially grains, may require knowledge of seed accessions, and access to postharvest facilities for processing and storage110,111,112. The local-scale matching of trees and crops to sites undertaken in this study is validated by species distribution based on biophysical parameters35,37. Findings from our research present the first step towards operationalising national policy on indigenous knowledge and agriculture115, and further directions for developing the required physical and social infrastructure by the local municipality.A number of these indigenous trees serve as a significant conduit to the intergenerational transfer of ecological knowledge and a connection to nature69, which in turn forms an important part of biocultural diversity and landscape stewardship83. Fast-growing herbs that require little input can be grown in marginal areas where the terrain poses difficulties, or where land tenure induces uncertainty. Crops that can resist waterlogging, enrich soil, and improve local productivity are also suggested where appropriate. Similarly, choices of crops exist for areas that are more prone to drought or heatwaves, or for marginal soils or shaded or windy areas68. Spatiotemporal intercropping of these with conventional crops can help remediate soil84,85. Surplus production of indigenous crops and trees can feed into short and high value supply chains to urban centres, e.g.38,86.Determining biophysical land suitability for peri-urban agroforestry using geospatial analysesFigure 3 shows the maps produced using expert-derived weights and value functions in each area. According to experts` knowledge, a higher weight was suggested for land cover than for agricultural land capability, dominant soils, slope, proximity to water sources and proximity to the main road. It should be noted that bare land plays a major role in delineating suitable urban areas for food production. Based on the results, a final weight of 0.35 was assigned to land cover. The final suitability maps for each area were divided into five agriculture suitability quality classes defined at discrete levels, allowing for comparisons between the three maps. The classes include suitable, moderately suitable, marginally suitable and not suitable areas for peri-urban agriculture. A simple visual comparison of the suitability patterns revealed by the three maps shows that Osindisweni has the greatest proportion of highly suitable and suitable areas for peri-urban agriculture. The Osindisweni area has suitable areas such as land cover, agricultural suitability, and open spaces, which favour the area’s suitability for agriculture. Also, this area has a good road and river network.Fig. 3Peri-urban agriculture suitability maps for (a) Osindisweni, (b) Ntshongweni and (c) Maphephetheni.Full size imageFor further analysis, the highly suitable, suitable and moderately Suitable areas were combined and overlaid with PGIS-identified suitable areas. Areas identified in the participatory mapping workshops tended to overlap with the high- to moderately-suitable classes of land identified in the GIS (biophysical) model (Fig. 4). This shows an agreement between the two methods used in this study to identify areas suitable for peri-urban food production at the three sites. Using both approaches strengthens the estimates of suitable areas by identifying the areas for which both approaches identify while minimising the number of wrongly identified areas. These maps will significantly value future land use and land cover change analysis for urban crop production.Fig. 4The suitable areas for peri-urban agriculture after overlaying both the PGIS and GIS layers in Osindisweni, Maphephetheni and Ntshongweni districts.Full size imageTable 5 shows that 75% of the study area in Maphephetheni, 4.53% in Ntshongweni and 0.21% in Osindisweni is permanently unsuitable for peri-urban crop production. These areas have unsuitable land cover and steep slopes, far from the road and river network. With a 10.74% suitability rate, the Osindisweni area has the highest potential for peri-urban food production, followed by Maphephetheni (1.2%) and Ntshongweni (0.84%). Generally, a small portion of the total area in all three study areas is suitable for urban crop production. For successful and effective peri-urban food production, growing crops with high production over a small piece of land, such as onions, herbs, garlic and leaf vegetables, is advisable.Table 5 The distribution of land suitability for each site from the PGIS land suitability analysis model.Full size tableStatistics comparing the number of cells assigned to each suitability class for the three maps are presented in Table 5 and Fig. 3. The difference between the areas of the site was up to 50 km2, with Ntshongweni being the smallest (167.61 km sq.), followed by Maphephetheni (183.83 km sq.) and Osindisweni (217.95 km sq.). The GIS suitability analysis indicated that Osindisweni has the largest absolute area of suitable land and the largest ratio of suitable to unsuitable land, with over 99% of its area being suitable for greening for food (Table 5). Conversely, Maphephetheni had the smallest suitable land area, accounting for about 25% of its total area. Ntshongweni also had a high ratio of 96% of its land suitable for greening for food.Osindisweni’s proximity to erstwhile sugarcane fields65 corroborates the finding that it has a greater proportion of suitable to marginally suitable agricultural land. However, more recent social-political developments, such as rapid urbanisation and the expansion of landfills have resulted in food production being perceived as untenable in the area. Indeed, expanding industrial and urban activities can accelerate a shift from land-based livelihoods and a decline in soil and water quality87,88. Our findings reiterate the importance of triangulating land use planning across large to fine scales through participatory methods. Alongside soil depth and nutrients, the slope is an important landscape determinant of land suitability for conventional food production89. Notwithstanding, results from our participatory and synthesis process offer options to reinforce local food and nutritional security through innovative design elements12. Despite the proximity to water sources, last-mile connectivity to arable land emerged as a significant limitation for agroforestry at the three sites. Plans to promote food production should consider strategies to manage nutrient flows in soil and water90,91. Excess runoff of agricultural enrichment materials may threaten water quality and safety. This is especially important given the immediate dependence of peri-urban and urban dwellers on surface and groundwater92. Agroecological strategies, including organic and circular inputs, are likely to alleviate environmental and food safety issues93. We acknowledge that our model considers arable land suitability in general, but that this may vary depending upon crop and tree species. Future research on multi-species indigenous agroforestry could be more species-specific, e.g.94.LimitationsOur study has two main limitations, namely the sample frame, and analytical depth. Our strategy to enlist participatory mapping workshop participants relied mainly on the local ward councillors and traditional chiefs, as this is standard practice to demonstrate respect of local authorities and build trust with local communities in the area. Participatory mapping work was carried out during a period when pandemic lockdown restrictions were in effect to varying degrees. Under these conditions, the participation in workshops was variable across the three sites, and it is possible that certain sections of communities that are less socially empowered may have been under-represented in the sample. Secondly, the scope of this study was to combine basic social, spatial, and species data to suggest locally appropriate peri-urban agroforestry design. Therefore, detailed evaluations of equitable access, agroclimatic variability, agronomic feasibility, etc. were not possible at this stage. We suggest that future research can delve into these specificities at site and regional scale. We posit that findings from this study provide a valuable baseline for research and implementation.Policy, practice, and research implicationsThis study highlights the role of participatory co-design in developing urban agriculture configurations. It combines a social-ecological systems lens95 with a landscape ecology approach57 to derive locally appropriate designs using locally adapted species. The communities expressed their aspirations for local food and nutrition security, which took on different forms. The participatory mapping outcomes demonstrate how local social-ecological and political situations influence preferences and feasibility of urban food production. Where food production is ongoing, diversification is welcomed, but where basic living conditions such as water supply and environmental quality are compromised, food production becomes secondary to expectations of urban living standards. This reiterates the need for participatory planning in the development of urban agriculture as a sustainable and citizen-driven enterprise in South Africa96. Our study demonstrates an interdisciplinary and participatory design approach to designing urban green infrastructure for ecosystem services97. Co-design has the benefits of recognising local needs, making services more accessible, balancing environmental regulatory frameworks with land use guidelines, and improving local resilience for urban green equity98,99. Further work in these communities has included the planting of a community agroforestry trial100 and an agroecology demonstration hub101 using indigenous species. These sites will serve as living learning laboratories for indigenous urban and peri-urban agroforestry. Communities will be involved in research related to assessing biodiversity and ecological implications such as species richness and plant biomass102,103, and also with development of market linkages for urban and peri-urban agroforestry, e.g.24,114.ConclusionThis study finds that while GIS tools can generate detailed information on land use suitability, the participatory process allows for the democratic exchange of knowledge, particularly in fast-changing socioeconomic landscapes like peri-urban areas. Specifically, the site where people have secure land tenure, service delivery, and existing involvement in agroforestry was found to be more suited to diversified multifunctional agroforestry configurations. Conversely, the site with uncertain land tenure and service delivery could support fewer configurations despite having the most are of suitable land according to GIS modelling. The site with most area of non-suitable land could also be matched with a high number of agroforestry configurations. This integrated approach can aid the development of site-specific solutions, forming dynamic governance co-produced by communities and institutions8,83. The participatory research component also helps to build an adaptive, responsive community of practice around each site by engaging with relevant stakeholders in non-political terms104,105,106. Through innovative design considerations, our findings aim to enable synergistic improvements in food and nutritional security, agroecology, and multifunctional urban green infrastructure107,108. The outputs contribute to achieving the sustainable development goals (SDGs) 2 (reducing hunger), 3 (promoting health and wellbeing), 9 (infrastructure innovation), 11 (sustainable cities and communities), 12 (responsible consumption and production), 13 (climate action), 15 (life on land), and 17 (partnerships) (SDG 2015)109. Policymakers and planners can draw from this partnership template using participatory research to feed into programmatic implementation at local scale8,9,10,113,114,115,116. In turn, participatory action research can be a conduit to intervention scaling, and to generating social-ecological evidence through living laboratories. This form of exchange is especially applicable to developing the field of indigenous local ecological knowledge28,35,38,86,94,115.

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].
ReferencesAugstburger, H., Käser, F. & Rist, S. Assessing food systems and their impact on common pool resources and resilience. Land 8(4), 71 (2019).Article 

Google Scholar 
Sarkodie, S. A., Owusu, P. A. & Leirvik, T. Global effect of urban sprawl, industrialization, trade and economic development on carbon dioxide emissions. Environ. Res. Lett. 15(3), 034049 (2020).Article 
ADS 

Google Scholar 
Bergius, M. & Buseth, J. T. Towards a green modernization development discourse: the new green revolution in Africa. J. Polit. Ecol. 26(1), 57–83 (2019).
Google Scholar 
Follmann, A., Willkomm, M. & Dannenberg, P. As the city grows, what do farmers do? A systematic review of urban and peri-urban agriculture under rapid urban growth across the Global South. Landsc. Urban Plan. 215, 104186 (2021).Article 

Google Scholar 
Haaland, C. & van den Bosch, C. K. Challenges and strategies for urban green-space planning in cities undergoing densification: A review. Urban For. Urban Green. 14(4), 760–771 (2015).Article 

Google Scholar 
Kabisch, N., Qureshi, S. & Haase, D. Human–environment interactions in urban green spaces—A systematic review of contemporary issues and prospects for future research. Environ. Impact Assess. Rev. 50, 25–34 (2015).Article 

Google Scholar 
Hajzeri, A. & Kwadwo, V. O. Investigating integration of edible plants in urban open spaces: Evaluation of policy challenges and successes of implementation. Land Use Policy 84, 43–48 (2019).Article 

Google Scholar 
Buijs, A. et al. Mosaic governance for urban green infrastructure: Upscaling active citizenship from a local government perspective. Urban For. Urban Green. 40, 53–62 (2019).Article 

Google Scholar 
Greenberg, S., Drimie, S., Losch, B. & May, J. From local initiatives to coalitions for an effective agroecology strategy: Lessons from South Africa. Sustainability 15(21), 15521 (2023).Article 
ADS 

Google Scholar 
Sardeshpande, M., Vanak, A., Ashwini, A., Casiker, C. V., Hariya, H., Mukherjee, R., Chatterjee, S., Lepcha, P. Y., Biswas, D., Devy, M. S., Hiremath, A. J., Jamwal, P., Khaling, S., & Setty, S. (In Press). Harnessing synergies across networks to drive actionable change for sustainable and healthy food systems in India. Journal of Agriculture and Food Research.Opitz, I., Berges, R., Piorr, A. & Krikser, T. Contributing to food security in urban areas: differences between urban agriculture and peri-urban agriculture in the Global North. Agric. Hum. Values 33(2), 341–358 (2016).Article 

Google Scholar 
Armanda, D. T., Guinée, J. B. & Tukker, A. The second green revolution: Innovative urban agriculture’s contribution to food security and sustainability—A review. Glob. Food Sec. 22, 13–24 (2019).Article 

Google Scholar 
Taylor, J. R. & Lovell, S. T. Designing multifunctional urban agroforestry with people in mind. Urban Agric. Regional Food Syst. 6(1), e20016 (2021).Article 

Google Scholar 
Botzat, A., Fischer, L. K. & Kowarik, I. Unexploited opportunities in understanding liveable and biodiverse cities. A review on urban biodiversity perception and valuation. Glob. Environ. Chang. 39, 220–233 (2016).Article 

Google Scholar 
Larondelle, N. & Strohbach, M. W. A murmur in the trees to note: Urban legacy effects on fruit trees in Berlin, Germany. Urban For. Urban Green. 17, 11–15 (2016).Article 

Google Scholar 
Säumel, I., Weber, F. & Kowarik, I. Toward livable and healthy urban streets: Roadside vegetation provides ecosystem services where people live and move. Environ. Sci. Policy 62, 24–33 (2016).Article 

Google Scholar 
Bonthoux, S., Brun, M., Di Pietro, F., Greulich, S. & Bouché-Pillon, S. How can wastelands promote biodiversity in cities? A review. Landsc. Urban Plan. 132, 79–88 (2014).Article 

Google Scholar 
Rupprecht, C. D., Byrne, J. A., Garden, J. G. & Hero, J. M. Informal urban green space: A trilingual systematic review of its role for biodiversity and trends in the literature. Urban For. Urban Green. 14(4), 883–908 (2015).Article 

Google Scholar 
Sardeshpande, M., Rupprecht, C. & Russo, A. Edible urban commons for resilient neighbourhoods in light of the pandemic. Cities 109, 103031 (2021).Article 

Google Scholar 
Hurley, P. T. & Emery, M. R. Locating provisioning ecosystem services in urban forests: Forageable woody species in New York City, USA. Landsc. Urban Plan. 70, 266–275 (2018).Article 

Google Scholar 
Colinas, J., Bush, P. & Manaugh, K. The socio-environmental impacts of public urban fruit trees: A Montreal case-study. Urban For. Urban Green. 45, 126132 (2019).Article 

Google Scholar 
Bunge, A., Diemont, S. A., Bunge, J. A. & Harris, S. Urban foraging for food security and sovereignty: quantifying edible forest yield in Syracuse, New York using four common fruit-and nut-producing street tree species. J. Urban Ecol. 5(1), juy028 (2019).Article 

Google Scholar 
Zhu, J., He, B. J., Tang, W. & Thompson, S. Community blemish or new dawn for the public realm? Governance challenges for self-claimed gardens in urban China. Cities 102, 102750 (2020).Article 

Google Scholar 
Wiek, A. & Albrecht, S. Almost there: On the importance of a comprehensive entrepreneurial ecosystem for developing sustainable urban food forest enterprises. Urban Agric. Reg. Food Syst. 7(1), e20025 (2022).Article 

Google Scholar 
Mabhaudhi, T., O’Reilly, P., Walker, S. & Mwale, S. Opportunities for underutilised crops in southern Africa’s post–2015 development agenda. Sustainability 8(4), 302 (2016).Article 
ADS 

Google Scholar 
Gaoue, O. G., Jiang, J., Ding, W., Agusto, F. B. & Lenhart, S. Optimal harvesting strategies for timber and non-timber forest products in tropical ecosystems. Thyroid Res. 9(3), 287–297 (2016).
Google Scholar 
Lankoandé, B., Ouédraogo, A., Boussim, J. I. & Lykke, A. M. Natural stands diversity and population structure of Lophira lanceolata Tiegh. ex Keay, a local oil tree species in Burkina Faso, West Africa. Agroforestry Syst. 91(1), 85–96 (2017).Article 

Google Scholar 
Leakey, R. R. A re-boot of tropical agriculture benefits food production, rural economies, health, social justice and the environment. Nature Food 1(5), 260–265 (2020).Article 

Google Scholar 
Champness, B. S., Palmer, G. C. & Fitzsimons, J. A. Bringing the city to the country: relationships between streetscape vegetation type and bird assemblages in a major regional centre. J. Urban Ecol. 5(1), juz018 (2019).Article 

Google Scholar 
Zietsman, M. Y., Montaldo, N. H. & Devoto, M. Plant–frugivore interactions in an urban nature reserve and its nearby gardens. J. Urban Ecol. 5(1), juz021 (2019).Article 

Google Scholar 
Bennett, A. B. & Lovell, S. Landscape and local site variables differentially influence pollinators and pollination services in urban agricultural sites. PLoS ONE 14(2), e0212034 (2019).Article 
PubMed 
PubMed Central 
CAS 

Google Scholar 
Broegaard, R. B. et al. Wild food collection and nutrition under commercial agriculture expansion in agriculture-forest landscapes. Forest Policy Econ. 84, 92–101 (2017).Article 

Google Scholar 
Bvenura, C. & Sivakumar, D. The role of wild fruits and vegetables in delivering a balanced and healthy diet. Food Res. Int. 99, 15–30 (2017).Article 
PubMed 

Google Scholar 
Mugiyo, H. et al. Multi-criteria suitability analysis for neglected and underutilised crop species in South Africa. PLoS ONE 16(1), e0244734 (2021).Article 
PubMed 
PubMed Central 
CAS 

Google Scholar 
Fredenberg, E., Karl, K., Passarelli, S., Porciello, J., Rattehalli, V., Auguston, A., Chimwaza, G., Grande, F., Holmes, B., Kozlowski, N., Laborde, D., MacCarthy, D., Masikati, P., McMullin, S., Mendez Leal, E., Valdivia, R., Van Deynze, A., van Zonneveld, M. Vision for adapted crops and soils (VACS) research in action: Opportunity crops for Africa (2024). https://doi.org/10.7916/3hd1-8t86Sardeshpande, M. & Shackleton, C. Wild edible fruits: A systematic review of an under-researched multifunctional NTFP. Forests 10(6), 467 (2019).Article 

Google Scholar 
Sardeshpande, M. & Shackleton, C. Fruits of the veld: Ecological and socioeconomic patterns of natural resource use across South Africa. Hum. Ecol. 48, 665–677. https://doi.org/10.1007/s10745-020-00185-x (2020).Article 

Google Scholar 
Nkosi, N. N., Mostert, T. H. C., Dzikiti, S. & Ntuli, N. R. Prioritization of indigenous fruit tree species with domestication and commercialization potential in KwaZulu-Natal, South Africa. Genet. Resour. Crop Evol. 67, 1567–1575 (2020).Article 

Google Scholar 
Rich, K. M., Rich, M. & Dizyee, K. Participatory systems approach for urban and peri-urban agriculture planning: The role of system dynamics and spatial group model building. Agric. Syst. 160, 110–123. https://doi.org/10.1016/j.agsy.2016.09.022 (2018).Article 

Google Scholar 
Bilgilioglu, S. et al. A GIS-based multi-criteria decision-making method for the selection of potential municipal solid waste disposal sites in Mersin, Turkey. Environ Sci Pollut Res 29, 5313–5329. https://doi.org/10.1007/s11356-021-15859-2 (2022).Article 

Google Scholar 
Schubert, H., Rauchecker, M., Caballero Calvo, A. & Schütt, B. Land use changes and their perception in the Hinterland of Barranquilla, Colombian Caribbean. Sustainability 11(23), 6729. https://doi.org/10.3390/su11236729 (2019).Article 
ADS 

Google Scholar 
Kirby, M. G., Scott, A. J. & Walsh, C. L. Translating policy to place: exploring cultural ecosystem services in areas of Green Belt through participatory mapping. Ecosyst. People 19(1), 2276752. https://doi.org/10.1080/26395916.2023.2276752 (2023).Article 

Google Scholar 
Depietri, Y. & Orenstein, D. E. Managing fire risk at the wildland-urban interface requires reconciliation of tradeoffs between regulating and cultural ecosystem services. Ecosyst. Serv. 44, 101108 (2020).Article 

Google Scholar 
Carrilho, J. & Trindade, J. Sustainability in Peri-Urban Informal Settlements: A Review. Sustainability 14(13), 7591. https://doi.org/10.3390/su14137591 (2022).Article 
ADS 

Google Scholar 
Sahana, M., Ravetz, J., Patel, P. P., Dadashpoor, H. & Follmann, A. Where is the peri-urban? A systematic review of peri-urban research and approaches for its identification and demarcation worldwide. Remote Sens 15(5), 1316. https://doi.org/10.3390/rs15051316 (2023).Article 
ADS 

Google Scholar 
Johnson, M. S. et al. The benefits of Q + PPGIS for coupled human-natural systems research: A systematic review. Ambio 51, 1819–1836. https://doi.org/10.1007/s13280-022-01709-z (2022).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Sardeshpande, M., Bangira, T., Matongera, T. N., Azong Cho, M., & Mabhaudhi, T. (2023). Appraising peri-urban food production in Durban, South Africa, with Participatory Geographic Information Systems (PGIS), 15 Nov 2023, PREPRINT (Version 1). Accessed on 08 Dec 2023, at Research Square.Department of Health (DoH). (2013). Roadmap for nutrition in South Africa 2013–2017. http://www.adsa.org.za/Portals/14/Documents/DOH/Nutrition%20Road%20Map%20 2013–2017.pdf (Accessed on 30/07/2017).Van der Hoeven, M. et al. Indigenous and traditional plants: South African parents’ knowledge, perceptions and uses and their children’s sensory acceptance. J. Ethnobiol. Ethnomed. 9, 1–12 (2013).
Google Scholar 
Porkka, M., Kummu, M., Siebert, S. & Varis, O. From food insufficiency towards trade dependency: a historical analysis of global food availability. PLoS ONE 8, e82714 (2013).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Bizikova, L. et al. A scoping review of the contributions of farmers’ organizations to smallholder agriculture. Nature Food 1(10), 620–630 (2020).Article 
PubMed 

Google Scholar 
StatsSA. (2017). Poverty trends in South Africa: An examination of absolute poverty between 2006 and 2015. Report 031006, Statistics South Africa. Pretoria, South Africa, 138pp.Sardeshpande, M. & Shackleton, C. Urban foraging: Land management policy, perspectives, and potential. PLoS ONE 15(4), e0230693 (2020).Article 
PubMed 
PubMed Central 
CAS 

Google Scholar 
Venter, Z. S., Shackleton, C. M., Van Staden, F., Selomane, O. & Masterson, V. A. Green Apartheid: Urban green infrastructure remains unequally distributed across income and race geographies in South Africa. Landsc. Urban Plan. 203, 103889 (2020).Article 

Google Scholar 
NDP. (2013). National Development Plan 2030. Our Future – make it work. Republic of South Africa. ISBN: 978–0–621–41180–5. 498pp. https://www.gov.za/documents/national-development-plan-2030-our-future-make-it-work accessed 08/04/2020NDA. (2021). National Department of Agriculture, Republic of South Africa. Programmes webpage. National Department of Agriculture, Land Reform, and Rural Development, Republic of South Africa. Available online at https://www.nda.agric.za/Programmes accessed on 17/08/2021.Hersperger, A. M., Grădinaru, S. R., Pierri Daunt, A. B., Imhof, C. S., & Fan, P. (2021). Landscape ecological concepts in planning: review of recent developments. Landscape Ecology, 1–17.eThekwini website. (2023). About us (eThekwini Municipality). Available online at https://www.durban.gov.za/pages/government/about-ethekwini accessed on July 13 2023Roberts, D. C. The design of an urban open-space network for the city of Durban (South Africa). Environ. Conserv. 21(1), 11–17 (1994).Article 

Google Scholar 
Davids, R., Rouget, M., Boon, R. & Roberts, D. Identifying ecosystem service hotspots for environmental management in Durban South Africa. Bothalia-African Biodiver. Conserv. 46(2), 1–18 (2016).Article 

Google Scholar 
Davids, R. et al. Civic ecology uplifts low-income communities, improves ecosystem services and well-being, and strengthens social cohesion. Sustainability 13(3), 1300 (2021).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Bisaga, I., Parikh, P. & Loggia, C. Challenges and opportunities for sustainable urban farming in South African low-income settlements: A case study in Durban. Sustainability 11(20), 5660 (2019).Article 
ADS 

Google Scholar 
Moyo, H. et al. Adaptive management in restoration initiatives: Lessons learned from some of South Africa’s projects. S. Afr. J. Bot. 139, 352–361 (2021).Article 

Google Scholar 
eThekwini Municipality. (2021). iNanda Mountain Community Reforestation Project Available online: http://www.durban.gov.za/City_Services/development_planning_management/environmental_planning_climate_protection/Projects/Pages/Inanda-Mountain-Community-Reforestation.aspx (accessed on August 13, 2021).Douwes, E. et al. The buffelsdraai landfill site community reforestation project. Unasylva 67, 12–19 (2016).
Google Scholar 
Taylor, L. & Hochuli, D. F. Defining greenspace: Multiple uses across multiple disciplines. Landsc. Urban Plan. 158, 25–38 (2017).Article 

Google Scholar 
Braun, V. & Clarke, V. Reflecting on reflexive thematic analysis. Qualitative Res. Sport Exercise Health 11(4), 589–597 (2019).Article 

Google Scholar 
Modi, A. T., & Mabhaudhi, T. M. (2018). Underutilised crop species in South Africa. Report to the water research commission of South Africa as part of WRC Project K5/2603//4 on developing a research agenda for promoting underutilised, indigenous and traditional crops.Sardeshpande, M. & Shackleton, C. Fruits of the city: The nature, nurture and future of urban foraging. People Nat. 5, 213–227. https://doi.org/10.1002/pan3.10428 (2023).Article 

Google Scholar 
El Baroudy, A. A. Mapping and evaluating land suitability using a GIS-based model. CATENA 140, 96–104 (2016).Article 

Google Scholar 
Kapoor, N., Jain, M. & Bansal, V. K. A methodological approach for weighting factors in land suitability assessment: a tool for facilitating spatial planning. J. Mt. Sci. 17(3), 724–739 (2020).Article 

Google Scholar 
Kazemi, H. & Akinci, H. A land use suitability model for rainfed farming by multi-criteria decision-making analysis (MCDA) and geographic information system (GIS). Ecol. Eng. 116, 1–6 (2018).Article 

Google Scholar 
Taherdoost, H. & Madanchian, M. Multi-criteria decision making (MCDM) methods and concepts. Encyclopedia 3(1), 77–87 (2023).Article 

Google Scholar 
Mazahreh, S., Bsoul, M. & Hamoor, D. A. GIS approach for assessment of land suitability for different land use alternatives in semi arid environment in Jordan: Case study (Al Gadeer Alabyad-Mafraq). Inform. Process. Agric. 6(1), 91–108 (2019).
Google Scholar 
Rstudio Team (2020). Rstudio: Integrated Development for R. Rstudio, PBC, Boston, MA URL http://www.rstudio.com/.Shackleton, C. M. et al. How important is green infrastructure in small and medium-sized towns? Lessons from South Africa. Landsc. Urban Plan. 180, 273–281 (2017).Article 

Google Scholar 
Shackleton, C. M., Hurley, P. T., Dahlberg, A. C., Emery, M. R. & Nagendra, H. Urban foraging: A ubiquitous human practice overlooked by urban planners, policy, and research. Sustainability 9(10), 1884 (2017).Article 
ADS 

Google Scholar 
Bennett, J., Ainslie, A. & Davis, J. Contested institutions? Traditional leaders and land access and control in communal areas of Eastern Cape Province, South Africa. Land Use Policy 32, 27–38 (2013).Article 

Google Scholar 
Shackleton, C. M. et al. Deactivation of field cultivation in communal areas of South Africa: Patterns, drivers and socio-economic and ecological consequences. Land Use Policy 82, 686–699 (2019).Article 

Google Scholar 
Olofsson, M. Expanding commodity frontiers and the emergence of customary land markets: A case study of tree-crop farming in Venda, South Africa. Land Use Policy 101, 105203 (2021).Article 

Google Scholar 
Lidzhegu, Z. & Kabanda, T. Declining land for subsistence and small-scale farming in South Africa: A case study of Thulamela local municipality. Land Use Policy 119, 106170 (2022).Article 

Google Scholar 
Combrinck, Z., Cilliers, E. J., Lategan, L. & Cilliers, S. Revisiting the proximity principle with stakeholder input: Investigating property values and distance to urban green space in 26otchefstroom. Land 9(7), 235 (2020).Article 

Google Scholar 
Elands, B. H. M. et al. Biocultural diversity: A novel concept to assess human-nature interrelations, nature conservation and stewardship in cities. Urban For. Urban Green. 40, 29–34 (2019).Article 

Google Scholar 
Tiroesele, B., Obopile, M. & Karabo, O. Insect diversity and population dynamics of natural enemies under sorghum–legume intercrops. Trans. Royal Soc. South Africa 74(3), 258–267 (2019).Article 

Google Scholar 
Malviya, M. K. et al. Sugarcane-legume intercropping can enrich the soil microbiome and plant growth. Front. Sustain. Food Syst. 5, 606595 (2021).Article 

Google Scholar 
Pfukwa, T. M. et al. Southern African indigenous fruits and their byproducts: Prospects as food antioxidants. J. Funct. Foods 75, 104220 (2020).Article 
CAS 

Google Scholar 
Ahmad, W. et al. Impact of land use/land cover changes on water quality and human health in district Peshawar Pakistan. Sci. Rep. 11(1), 16526 (2021).Article 
ADS 
PubMed 
PubMed Central 
CAS 

Google Scholar 
Ozsahin, E. & Ozdes, M. Agricultural land suitability assessment for agricultural productivity based on GIS modeling and multi-criteria decision analysis: the case of Tekirdağ province. Environ. Monit. Assess. 194(1), 41 (2022).Article 

Google Scholar 
Seyedmohammadi, J. & Navidi, M. N. Applying fuzzy inference system and analytic network process based on GIS to determine land suitability potential for agricultural. Environ. Monit. Assess. 194(10), 712 (2022).Article 
PubMed 
CAS 

Google Scholar 
Chang, N. B., Imen, S. & Vannah, B. Remote sensing for monitoring surface water quality status and ecosystem state in relation to the nutrient cycle: a 40-year perspective. Crit. Rev. Environ. Sci. Technol. 45(2), 101–166 (2015).Article 
CAS 

Google Scholar 
Zhang, Q. R. et al. Spatial distribution and quantitative source identification of nutrients and beneficial elements in the soil of a typical suburban area, Beijing. Environ. Monit. Assess. 195(1), 1–14 (2023).Article 
CAS 

Google Scholar 
Calijuri, M. L., Castro, J. D. S., Costa, L. S., Assemany, P. P. & Alves, J. E. M. Impact of land use/land cover changes on water quality and hydrological behavior of an agricultural subwatershed. Environ. Earth Sci. 74, 5373–5382 (2015).Article 
ADS 

Google Scholar 
Greenberg, S., & Drimie, S. (2021). The state of the debate on agroecology in South Africa A scan of actors, discourses and policies. Report to CIRAD, available online at https://foodsecurity.ac.za/wp-content/uploads/2022/11/TAFS-project_The-state-of-the-debate-on-agroecology-in-South-Africa-A-scan-of-actors-discourses-and-policies.pdf Accessed on 19/06/2023Wotlolan, D. L., Lowry, J. H., Wales, N. A. & Glencross, K. Land suitability evaluation for multiple crop agroforestry planning using GIS and multi-criteria decision analysis: A case study in Fiji. Agrofor. Syst. 95(8), 1519–1532 (2021).Article 

Google Scholar 
Andersson, E. et al. What are the traits of a social-ecological system: Towards a framework in support of urban sustainability. NPJ Urban Sustain. 1(1), 14 (2021).Article 

Google Scholar 
Cilliers, E. J., Lategan, L., Cilliers, S. S. & Stander, K. Reflecting on the potential and limitations of urban agriculture as an urban greening tool in South Africa. Front. Sustain. Cities 2, 43 (2020).Article 

Google Scholar 
Kamjou, E., Scott, M. & Lennon, M. A bottom-up perspective on green infrastructure in informal settlements: Understanding nature’s benefits through lived experiences. Urban For. Urban Green. 94, 128231 (2024).Article 

Google Scholar 
Nesbitt, L., Meitner, M. J., Girling, C. & Sheppard, S. R. Urban green equity on the ground: Practice-based models of urban green equity in three multicultural cities. Urban For. Urban Green. 44, 126433 (2019).Article 

Google Scholar 
Vargas-Hernández, J. G. & Zdunek-Wielgołaska, J. Urban green infrastructure as a tool for controlling the resilience of urban sprawl. Environ. Dev. Sustain. 23(2), 1335–1354 (2021).Article 

Google Scholar 
Burgdorf, R., Sardeshpande, M. & Cooper, G. Indigenous tree forestry: Economic potential or potential fantasy?. Timber Indus. Presents Mag. 4(2), 14–20 (2023).
Google Scholar 
Green Trust (2024). Woza Nami: Trees combating hunger in South Africa. Available online at: https://www.greentrust.org.za/2024/09/03/woza-nami-phase-ii-trees-combating-hunger-in-south-africa/ Accessed on 27/11/2024Rockwell, C. A., Crow, A., Guimarães, É. R., Recinos, E. & La Belle, D. Species richness, stem density, and canopy in food forests: contributions to ecosystem services in an urban environment. Urban Plan. 7(2), 139–154 (2022).Article 

Google Scholar 
Balima, L. H. et al. Higher diversity, denser stands and greater biomass in peri-urban forests than in adjacent agroforestry systems in Western Burkina Faso: implications for urban sustainability. Environ. Monit. Assess. 195(9), 1077 (2023).Article 
PubMed 

Google Scholar 
Fischer, J. et al. Advancing sustainability through mainstreaming a social–ecological systems perspective. Curr. Opinion Environ. Sustain. 14, 144–149 (2015).Article 

Google Scholar 
Wyborn, C. Co-productive governance: a relational framework for adaptive governance. Glob. Environ. Chang. 30, 56–67 (2015).Article 

Google Scholar 
Goodwin, G. The problem and promise of coproduction: Politics, history, and autonomy. World Dev. 122, 501–513 (2019).Article 

Google Scholar 
Isaac, M. E., Sinclair, F., Laroche, G., Olivier, A. & Thapa, A. The ties that bind: how trees can enhance agroecological transitions. Agrofor. Syst. 98, 2369–2383 (2024).Article 

Google Scholar 
Shackleton, C. M. et al. Embedding opportunities for poverty alleviation in urban green infrastructure design and management using South Africa as a case example. Cities 155, 105442 (2024).Article 

Google Scholar 
Hák, T., Janoušková, S. & Moldan, B. Sustainable Development Goals: A need for relevant indicators. Ecol. Ind. 60, 565–573 (2016).Article 

Google Scholar 
Alemneh, S. T., Emire, S. A., Hitzmann, B. & Zettel, V. Comparative study of chemical composition, pasting, thermal and functional properties of teff (Eragrostis tef) flours grown in Ethiopia and South Africa. Int. J. Food Prop. 25(1), 144–158 (2022).Article 
CAS 

Google Scholar 
Bultosa, G. Teff: Overview. In 2015 (eds Wrigley, C. W. et al.) 209–220 (Academic Press, 2016).
Google Scholar 
Heuzé V., Thiollet H., Tran G., Lebas F., 2017. Tef (Eragrostis tef) hay. Feedipedia, a programme by INRAE, CIRAD, AFZ and FAO. https://www.feedipedia.org/node/22768 Last updated on March 6, 2017, 14:38 Accessed on 15 October 2025.Cockburn, J., Rouget, M., Slotow, R., Roberts, D., Boon, R., Douwes, E., & Willows-Munro, S. (2016). How to build science-action partnerships for local land-use planning and management: lessons from Durban, South Africa. Ecology and Society, 21(1).Leakey, R. R. & Harding, P. E. ‘Land maxing’: Regenerative, remunerative, productive and transformative agriculture to harness the six capitals of sustainable development. Sustainability 17(13), 5876 (2025).Article 
ADS 

Google Scholar 
Republic of South Africa. (2019). Act No. 6 of 2019: Protection, Promotion, Development and Management of Indigenous Knowledge Act, 2019. Government Gazette 650: 42647.Greenberg, S., Drimie, S., Losch, B., & Jila, N. (2022). Agroecological initiatives in eThekwini Metropolitan Municipality, KwaZulu-Natal. TAFS phase 2. Final site report. Report to CIRAD, available online at https://agritrop.cirad.fr/605098/1/TAFS%20eThekwini%20research%20report_July%202022%20final%20.pdf Accessed on 22 October 2025Download referencesAcknowledgementsMS is grateful to all workshop participants at the three sites and the local chiefs and ward councillors for their time and efforts. MS thanks Kuhlekonke Mathenjwa for isiZulu and English interpretation at the workshops, and Nathi Ngcobo, Linda Mhlotshwa, and their associates for organising permissions and logistics. Thanks also to Errol Douwes, Mbali Goge, and Nomzamo Mncube for their support through the Durban Research Action Partnership. This research was funded in part by the Wellcome Trust [Grant number: 205200/Z/16/Z]. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. Fieldwork was funded in part by the WoodRIGHTS Project, which the University of KwaZulu-Natal Strategic Flagships funds.FundingThis research was funded in part by the Wellcome Trust [Grant number: 205200/Z/16/Z]. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. Fieldwork was funded in part by the WoodRIGHTS Project, which the University of KwaZulu-Natal Strategic Flagships funds.Author informationAuthors and AffiliationsCentre for Transformative Agricultural and Food Systems, University of KwaZulu-Natal, Scottsville, 3209, South AfricaMallika Sardeshpande, Tsitsi Bangira, Matilda Azong Cho, Trylee Nyasha Matongera & Tafadzwanashe MabhaudhiDepartment of Geography, School of Agriculture and Science, University of KwaZulu-Natal, Scottsville, 3209, South AfricaMatilda Azong Cho & Trylee Nyasha MatongeraCentre on Climate Change and Planetary Health, London School of Hygiene and Tropical Medicine, London, UKTafadzwanashe MabhaudhiAuthorsMallika SardeshpandeView author publicationsSearch author on:PubMed Google ScholarTsitsi BangiraView author publicationsSearch author on:PubMed Google ScholarMatilda Azong ChoView author publicationsSearch author on:PubMed Google ScholarTrylee Nyasha MatongeraView author publicationsSearch author on:PubMed Google ScholarTafadzwanashe MabhaudhiView author publicationsSearch author on:PubMed Google ScholarContributionsAll authors have read and consented to publication of this manuscript. Conceptualisation, Investigation, Data Curation, Writing—original: Mallika Sardeshpande; Methodology, Formal Analyses, Visualisation, Writing—review and editing: Mallika Sardeshpande, Tsitsi Bangira, Trylee Matongera, Matilda Azong Cho; Funding Acquisition, Supervision: Tafadzwanashe Mabhaudhi.Corresponding authorCorrespondence to
Mallika Sardeshpande.Ethics declarations

Competing interests
The authors declare no competing interests.

Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissionsAbout this articleCite this articleSardeshpande, M., Bangira, T., Azong Cho, M. et al. Exploring the social-ecological potential for indigenous agroforestry in peri-urban areas: a participatory mapping approach.
Sci Rep 15, 44344 (2025). https://doi.org/10.1038/s41598-025-27864-3Download citationReceived: 25 June 2025Accepted: 06 November 2025Published: 23 December 2025Version of record: 23 December 2025DOI: https://doi.org/10.1038/s41598-025-27864-3Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
KeywordsAgroecologyFarm designIndigenous foodsPeri-urban agricultureSuitability analysis More

AbstractEnvironmental heterogeneity plays a crucial role in shaping the distribution and composition of natural vegetation, including understory ferns. This study investigated the influence of environmental variation on understory fern communities within a one-hectare permanent plot in a low-altitude mountain forest in central Taiwan. Twenty-two environmental factors, including topographic, soil, and biotic variables, were recorded. Multiple regression, cluster (two-way indicator species analysis, TWINSPAN), and ordination (detrended correspondence analysis, DCA, and canonical correspondence analysis, CCA) analyses were conducted. A total of 51 fern species (including Lycophytes) belonging to 20 families and 30 genera were recorded. Among these, 43 were terrestrial, and eight were epiphytic; however, only terrestrial species were analyzed because of the limited representation of epiphytes. Multiple regression analyses revealed that environmental variables significantly affected fern richness, abundance, and community composition. Specifically, stream distance and the importance value (IV) of the saplings significantly influenced fern richness; herb/vine IV affected abundance; and the carbon-to-nitrogen ratio (C/N), manganese concentration (Mn), and herb/vine IV impacted the first axis of the DCA. Furthermore, the elevation, curvature, slope, and topographic wetness index (TWI) significantly influenced the second axis of the DCA. In all the models, topographic variables—particularly stream distance—were one of the most influential drivers. TWINSPAN categorized the ferns into four distinct groups (Diplazium donianum var. donianum [DIPLDO], D. donianum var. aphanoneuron [DIPLAP], Blechnopsis orientalis [BLECOR], and Angiopteris lygodiifolia [ANGILY]), and CCA revealed that environmental factors structured the community compositions in line with the TWINSPAN grouping. The DIPLDO and DIPLAP groups were associated with ridges and upper slope habitats characterized by higher elevations and drier conditions. In contrast, the BLECOR and ANGILY groups were associated with lower elevations, stream proximity, steeper slopes, and higher humidity. This study highlights the role of topographic and soil C/N heterogeneity in structuring fern communities in fine-scale plots for future ecological monitoring in subtropical forest ecosystems.

Similar content being viewed by others

Species diversity and spatial pattern of heritage trees in Taiyuan

Article
Open access
21 May 2025

Eco-systematic assessment of the spring herbaceous vegetation under edaphic and topographic effects

Article
Open access
23 July 2024

Changes in taxonomic and functional diversity of plants in a chronosequence of Eucalyptus grandis plantations

Article
Open access
24 May 2021

IntroductionEnvironmental gradients strongly influence the diversity and spatial distribution of plant communities. At regional scales, climate variables such as temperature and precipitation are dominant drivers1,2,3. In contrast, at local scales, fine-scale topography, soil characteristics, and biological interactions play significant roles4,5,6,7,8,9. Among biotic factors, canopy structure and density are widely recognized for their influence on understory plant communities10,11. Ferns, the second largest group of vascular plants, are a dominant component of understory vegetation in tropical and subtropical forests3,12 and are particularly sensitive to environmental heterogeneity.Topography—such as elevation, slope, aspect, and stream proximity—shapes microhabitats by altering light, temperature, and moisture regimes13,14. Topographic variation at the local scale is known to affect both fern diversity10 and abundance15. Stream proximity, in particular, is a strong predictor of fern assemblages16.Soil characteristics are also closely tied to fern distribution. Soil moisture (or humidity) is critical for the growth and development of ferns5,17. Variables such as nutrient content (e.g., N, P, K, Ca, and Mg), pH, and organic matter significantly influence fern performance5,16,18. The carbon-to-nitrogen (C/N) ratio, in particular, serves as a proxy for soil fertility and has been linked to fern richness in several tropical studies19,20. However, soil properties are partially influenced by topographic variations21,22, which in turn may affect the distribution of ferns.Understory ferns depend on canopy-mediated light availability for growth and reproduction23,24. Canopy openness not only alters photosynthetically active radiation but also modulates temperature and humidity in the understory25. Research in Southeast Asia and Taiwan suggests that canopy openness is an important factor influencing fern richness and cover10,26. Furthermore, ferns interact with other plant groups. Dense fern layers can suppress tree seedling recruitment27,28, whereas the diversity of co-occurring understory taxa appears to influence fern richness in varying ways29,30. These biotic interactions may result in mutual inhibition or facilitation depending on local environmental conditions.Ferns reproduce via spores that are readily dispersed by wind. Despite the fact that many ferns may produce spores capable of travelling long distances, chances of establishing new populations are low31. Allopatric differentiation may be associated with gametophytes that are highly sensitive to microclimatic and edaphic parameters32,33. In addition, environmental factors at the mesoscale—such as soil moisture, humidity, temperature, wind speed, rainfall, vegetation type, and canopy openness—significantly influence fern distribution by affecting sporophytes’ water requirements, temperature tolerance, and photosynthetic capacity5,26,34,35,36. Therefore, the diversity of forest microenvironments varies across regions and is reflected in corresponding differences in fern diversity and composition. We surveyed the relationship between environmental heterogeneity and fern diversity within a one-hectare plot embedded in a broader 25-hectare permanent plot in the Lienhuachih region, central Taiwan. We addressed the following questions: (1) Which environmental factors most strongly influence fern richness, abundance, and composition? (2) How do ferns cluster into ecological groups on the basis of these factors?Materials and methodsStudy areaThe study site is located in the Lienhuachih Experimental Forest (23°55’N, 120°52’E), which is located in a low-altitude mountainous area of central Taiwan (Fig. 1). A one-hectare permanent plot was established within a natural forest and represents the northwestern section of a broader 25-hectare forest dynamics plot initiated in 2007. The plot encompasses both mid-slope and riparian habitats, with an elevation range spanning from 755 m to 814 m (Table S1). On the basis of earlier tree surveys (DBH ≥ 1 cm), two forest types were identified: one dominated by Diospyros morrisiana and Cryptocarya chinensis and the other by Machilus japonica var. kusanoi and Helicia formosana37. The region experiences a mean annual temperature of 21.2℃ and receives approximately 2,178 mm of precipitation annually, with rainfall concentrated from March to September and a dry season from October to November26. The soil properties of the 25-hectare plot are partially influenced by topography21.Fig. 1Map of Taiwan (left) and a topographic map (right) of the one-hectare natural forest plot in Lienhuachih within the Houloun stream catchment, a mountainous region in central Taiwan.Full size imageSampling designThe one-hectare plot (100 m × 100 m) was divided into 100 subplots (10 m × 10 m) used as survey units. Vegetation data were collected per subplot, within which all ferns, herbs, vines, and tree saplings (< 1 cm in DBH, > 30 cm in height) were recorded. Fern abundance was assessed by counting individual clumps (treating each clump as one individual) and by estimating percent cover. Epiphytic ferns and vines were measured by the horizontal projection of their canopy cover. Tree data (DBH ≥ 1 cm) were also recorded. Ferns were categorized as either terrestrial or epiphytic, with the former defined as those growing on soil or rocks and the latter as those occurring on tree trunks. Taxonomy follows Volume 6 of the Flora of Taiwan38 and the classification by Kuo et al. (2019)39. Surveys were conducted from July 13 to 15, 2023.Environmental variablesThe topographic variables included elevation, plan curvature, slope, aspect, topographic wetness index (TWI), and distance to the stream. These data were collected from the center of each subplot. Soil variables (data were collected by Chang et al. (2013)40 (2023)41 ) included pH, the carbon-to-nitrogen ratio (C/N), nitrogen, phosphorus, potassium, calcium, magnesium, manganese, zinc, iron, and copper. Soil data were collected at a 20 m × 20 m resolution, with each 10 m × 10 m subplot assigned the values of its nearest soil sample point. Soil moisture was measured at a depth of 5 cm using a RiXEN M-700 S meter between February 26 and March 3, 2024, and the data were averaged from three diagonal points per subplot. Canopy openness was measured from March 15 to 22, 2024, using spherical crown densiometers at a height of 1.3 m at the center of each 10 m × 10 m subplot.Statistical analysisSince the soil properties of the 25-hectare plot are influenced by topography, a two-factor Pearson correlation coefficient analysis was conducted to examine the relationships between topography and soil properties in this study plot. Forward stepwise multiple regression (FSMR) with Poisson and linear models was employed to examine the effects of environmental factors on fern diversity. The abundance data for each species were quantified using the importance value (IV), which was calculated as the sum of its relative density (individuals per species/total individuals, unit: %) and relative cover (cover per species/total cover, unit: %). Twenty-two environmental variables—spanning topography (elevation, plan curvature, slope, aspect, TWI, and stream distance), soil properties (pH, C/N, N, P, K, Ca, Mg, Mn, Zn, Fe, Cu, and soil moisture), and biotic factors (canopy openness, tree density, sapling IV, and herb/vine IV)—were included in the analyses (Supplementary Table S1). The dependent variables included fern richness (species count), abundance (IV), and community composition (first two DCA axes). FSMR was performed for model selection using the “MuMIn” package in R. Significant predictors were selected (p < 0.05, chi-square test [Poisson] for fern richness and F test for fern abundance and composition [linear]), and then collinearity was assessed using the variance inflation factor (VIF). Variables with a VIF > 5 were iteratively excluded. Finally, Poisson regression was used to analyze fern richness and its selected predictor variables, whereas linear regression was applied to abundance, composition, and their respective selected predictors.Community classification was performed using two-way indicator species analysis (TWINSPAN) (dissimilarity metric = total inertia)42. Detrended correspondence analysis (DCA)43 was used to ordinate species and subplots. Canonical correspondence analysis (CCA)44 related species distributions to environmental gradients. The raw data utilized in the aforementioned analysis were derived from the species-subplot matrix, with the data comprising the previously described importance value (IV). All analyses were performed in R v4.3.1.ResultsA total of 51 fern species representing 20 families and 30 genera were recorded within the one-hectare plot. Of these, 43 species were terrestrial, and eight were epiphytic. Diplazium dilatatum was the most abundant species, with 1,011 individuals observed in 98 subplots, followed by Pleocnemia winitii, with 567 individuals in 90 subplots. These two species accounted for 55.8% of the total abundance of terrestrial ferns (Table 1). In contrast, ten terrestrial species were found in only one subplot (Supplementary Table S2), representing 23.3% of the total terrestrial fern richness.Table 1 Number of subplots, relative density, and relative coverage for terrestrial species of fern in the one-hectare plot of the low-altitude natural forest of central Taiwan. The species did not occur in fewer than 2 subplots.Full size tableIn addition to ferns, 37 herb, 43 vine and 76 sapling species were recorded, resulting in a total of 207 understory species (including 8 epiphytic fern species). Owing to their limited abundance and patchy distribution, epiphytic ferns were excluded from further analyses but are documented in Supplementary Table S2. Among the environmental variables, elevation was most significantly correlated with soil properties (11), followed by slope (nine) and stream distance (seven) (Supplementary Table S3).The regression models (Table 2) revealed that among the six selected variables, fern richness was significantly influenced by stream distance (negatively) and sapling abundance (positively). Fern abundance was most strongly associated with herb/vine IVs. With respect to fern composition, DCA1 was associated with stream distance, the C/N ratio, manganese, and herb/vine IV—with only stream distance being not significant. The best model of DCA2 included eight variables, among which elevation, curvature, slope, and TWI were significant.Table 2 Environmental factor models in a one-hectare natural forest plot in Lienhuachih used Poisson regression for fern richness and linear regression for fern abundance and composition (two DCA axes). “VIF” indicates the variance inflation factor test. *: p < 0.05; **: p < 0.01; and ***: p < 0.001.Full size tableTWINSPAN classified the fern community into four groups (Fig. 2a; Supplementary Figure S1): the Diplazium donianum var. donianum group (DIPLDO; n = 52), the D. donianum var. aphanoneuron group (DIPLAP; n = 21), the Blechnopsis orientalis group (BLECOR; n = 8), and the Angiopteris lygodiifolia group (ANGILY; n = 19). The mean fern richness was lowest in DIPLDO (4.2 ± 1.9) and highest in ANGILY (6.4 ± 2.5) (Supplementary Figure S2). Fern abundance (log-transformed) was positively correlated with richness (r = 0.46, p < 0.001), a pattern that was consistent across groups (Fig. 3). However, significant correlations were observed only in the DIPLDO and ANGILY groups, whereas the other two groups showed no significant correlation.Fig. 2TWINSPAN and CCA from a one-hectare natural forest plot in the Lienhuachih area of central Taiwan. (a) TWINSPAN identified four fern groups: DIPLDO (□), DIPLAP (△), BLECOR (○), and ANGILY (●). (b) The figure of the first two axes from the CCA; the words beside the lines represent environmental and biological factors, and the direction indicates the trend in which the value increases. (c) The same analysis as in b, with the letters representing the fern species (see Table 1).Full size imageFig. 3Relationships between fern abundance (log-transformed) and richness (r = 0.46, p < 0.001). The Pearson correlation in the four fern groups was r = 0.48 (p < 0.001, DIPLDO), 0.26 (p = 0.264, DIPLAP), 0.50 (p = 0.205, BLECOR), and 0.65 (p = 0.002, ANGILY).Full size imageThe cumulative explained variance of the first three CCA axes was 9.7%, 16.5%, and 21.2%, respectively. On the first axis of the CCA, herb/vine IV had the highest absolute score (0.80), followed by C/N (0.54), elevation (–0.50), and stream distance (–0.46); on the second axis of the CCA, stream distance (–0.61) had the highest absolute score, followed by Ca (–0.58), elevation (–0.50), and slope (0.43). In addition, herb/vine IV, stream distance, Ca and C/N were the most important determinants of one of the first two CCA axes (Table 3). As shown in Fig. 2b, the BLECOR group is more distinct, whereas the ANGILY group somewhat overlaps with the other groups, and the DIPLDO and DIPLAP groups exhibit greater overlap. Our results showed that these fern groups have adapted to different environments (Fig. 4).Table 3 Scores of the first two CCA axes with the environmental factors in the 1 ha plot of Lienhuachih in the low-altitude natural forest of central Taiwan. * shows the significance test (p < 0.05) for Pearson correlation between these factors and the CCA axes.Full size tableFig. 4Variation in elevation (a), stream distance (b), slope (c), and C/N (d) among different fern communities. Different letters denote statistically significant differences among the different types of fern vegetation (p < 0.05).Full size imageThe species ordination (Fig. 2c) revealed dominant ferns (Diplazium dilatatum and Pleocnemia winitii) near the plot center, whereas the species of named TWINSPAN groups aligned with the environmental characteristics of their respective groups. For example, DIPLDO’s D. donianum var. donianum was located in a topographic and edaphic space that is indicative of drier, upland sites.DiscussionTopographic effectsTopography has long been recognized as a key determinant of forest vegetation patterns8,45,46. In this study, elevation and stream distance emerged as primary predictors of fern richness and composition, despite the modest elevation range (~ 59 m) within the plot. These gradients reflect moisture availability: ridges with higher elevations and well-drained soils tend to be drier, whereas lower streamside zones retain more moisture. The strong correlation between elevation and stream distance (r = 0.40, p < 0.001) reinforces this interpretation. Our findings align with those of previous studies5,10,16 in montane forests where even fine-scale topographic variation influences fern diversity. For the other plant taxa in the 25-ha plot (of which our 1-ha plot was a part), topography was the most important factor affecting the changes in the plant community and species composition37.Soil effectsSoil properties such as nutrient concentrations and organic matter content often covary with topography because of erosion, leaching, and deposition21,47. While some studies have indicated a positive correlation between soil fertility and fern richness49,50, others have shown that lower fertility results in more fern species15,20. In our plot, C/N was significantly associated with fern composition, particularly along the first two CCA axes. Stream-adjacent soils, which are rich in organic matter and nitrogen, presented elevated carbon-to-nitrogen (C/N) ratios because of the greater accumulation of organic matter than that associated with decomposition in moist areas. Although fern richness was not directly correlated with C/N, its indirect effects via topographic mediation were evident. Calcium and manganese were also included in the regression models, although their contributions were relatively modest.Water serves as a critical determinant of both fern richness and distribution4,5,24. Water availability, inferred through the topographic wetness index (TWI), slope, and stream distance, likely exerts a dominant control on fern distributions. While soil moisture in the 25-ha plot (6 transects) was an important factor for the understory plants, including fern species30, it did not emerge as a significant predictor in the models in this study plot. The strong influence of hydrologically relevant topographic variables suggests their overriding importance in determining local fern composition.Biotic influencesLight availability is a well-documented driver of fern performance and affects morphology, abundance, and richness10,24. Although canopy openness was not retained in the final regression models, its significant correlation with tree density (r = − 0.22, p < 0.05) implies indirect effects. Denser tree canopies may reduce understory light, thus constraining fern growth.Interestingly, this study revealed positive associations between fern richness (or abundance) and both sapling and herb/vine cover, contrary to previous findings that emphasized competitive suppression28,30,48. The factors contributing to this outcome are likely multifaceted. The underlying mechanism may involve moisture availability, which is influenced by proximity to the stream. Although herb and vine cover are strongly correlated with fern abundance and composition, the primary determinant appears to be the distance from streams, as areas closer to streams generally exhibit higher moisture levels. Tuomisto et al. (2002) similarly reported that fern and Melastomataceae diversity co-occurred with tree richness in fertile tropical soils29.Fern community grouping and habitat differentiationTWINSPAN and CCA revealed clear compositional differentiation among the four fern groups, corresponding to distinct habitat types. The DIPLDO and DIPLAP groups were associated with ridges and upper slope habitats characterized by higher elevations and drier conditions. In contrast, the BLECOR and ANGILY groups were associated with lower elevations, stream proximity, steeper slopes, and higher humidity. These habitat preferences support the role of environmental filtering in fern assembly and align with prior vegetation classifications within the same forest37.In terms of the correlation between fern richness and abundance, compared with the DIPLDO group, the ANGILY group exhibited communities with greater evenness. These findings suggest that the environment inhabited by the ANGILY group is more conducive to the survival of a diverse range of ferns.ConclusionThis study highlights the significant role of environmental heterogeneity in shaping fern diversity and community composition in a low-altitude subtropical forest. Among the examined factors, topographic variables—particularly stream distance—exerted one of the most influential drivers of fern richness, abundance, and species assemblage. Soil properties, especially C/N, further mediated these relationships and reflected microhabitat variation. The classification of ferns into four ecological groups on the basis of environmental gradients highlights the structuring effect of habitat differentiation. The DIPLDO and DIPLAP groups occupied ridges and upper slope habitats characterized by higher elevations and drier conditions, whereas the BLECOR and ANGILY groups were associated with lower elevations, greater proximity to streams, steeper slopes, and higher humidity. This study highlights the role of topographic and soil-related heterogeneity in structuring fern communities in fine-scale plots. Long-term monitoring incorporating both abiotic and biotic variables is essential for understanding how fern communities respond to environmental change and for informing conservation strategies in subtropical forest ecosystems.

Data availability

The datasets utilized and/or analyzed during the current study are available from the first author upon reasonable request.
ReferencesAntonelli, A. et al. Geological and Climatic influences on mountain biodiversity. Nat. Geosci. 11, 718–725 (2018).Article 
ADS 
CAS 

Google Scholar 
Lin, H. Y. et al. Climate-based approach for modeling the distribution of montane forest vegetation in Taiwan. Appl. Veg. Sci. 23, 239–253 (2020).Article 

Google Scholar 
Suissa, J. S., Sundue, M. A. & Testo, W. L. Mountains, climate and niche heterogeneity explain global patterns of fern diversity. J. Biogeogr. 48, 1296–1308 (2021).Article 

Google Scholar 
Richard, M., Bernhardt, T. & Bell, G. Environmental heterogeneity and the Spatial structure of fern species diversity in one hectare of old-growth forest. Ecography 23, 231–245 (2000).Article 
ADS 

Google Scholar 
Karst, J., Gilbert, B. & Lechowicz, M. J. Fern community assembly: the roles of chance and the environment at local and intermediate scales. Ecology 86, 2473–2486 (2005).Article 

Google Scholar 
Bohlman, A. S. et al. Importance of soils, topography and geographic distance in structuring central Amazonian tree communities. J. Veg. Sci. 19, 863–874 (2008).Article 

Google Scholar 
Da Silva, K. E. et al. Tree species community Spatial structure in a Terra firme Amazon forest, Brazil. Bosque (Valdivia). 35, 347–355 (2014).Article 

Google Scholar 
Zhou, T. et al. Biodiversity of Jinggangshan mountain: the importance of topography and geographical location in supporting higher biodiversity. PLoS ONE. 10, e0120208 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Zuleta, D. et al. Importance of topography for tree species habitat distributions in a Terra firme forest in the Colombian Amazon. Plant. Soil. 450, 133–149 (2020).Article 
CAS 

Google Scholar 
Jones, M. M., Cicuzza, D., Straaten, O., Veldkamp, E. & Kessler, M. Determinants of fern and angiosperm herb community structure in lower montane rainforest in Indonesia. J. Veg. Sci. 25, 1216–1224 (2014).Article 

Google Scholar 
Mendes, C. N., Diniz, E. S., Terra, M. C. N. S., Jeannot, K. K. & Fontes, M. A. L. Light conditions imposed by canopy: allometric strategies of an understorey palm (Geonoma schottianamart) in Atlantic forest. J. Trop. Sci. 31, 332–342 (2019).
Google Scholar 
Weigand, A. et al. Global fern and lycophyte richness explained: how regional and local factors shape plot richness. J. Biogeogr. 47, 59–71 (2020).Article 

Google Scholar 
Chao, K. J., Chao, W. C., Chen, K. M. & Hsieh, C. F. Vegetation dynamics of a lowland rainforest at the Northern border of the paleotropics at Nanjenshan, Southern Taiwan. Taiwan. J. Sci. 25, 29–40 (2010).
Google Scholar 
Moeslund, J. E., Arge, L., Bøcher, P. K., Dalgaard, T. & Svenning, J. C. Topography as a driver of local terrestrial vascular plant diversity patterns. Nord J. Bot. 31, 129–144 (2013).Article 

Google Scholar 
Cicuzza, D. & Mammides, C. Soil, topography and forest structure shape the abundance, richness and composition of fern species in the fragmented tropical landscape of Xishuangbanna, Yunnan, China. Forests 13, 1453 (2022).Article 

Google Scholar 
Bonari, G. et al. Fine-scale fern ecological responses inform on riparian forest habitat conservation status. Biodivers. Conserv. 31, 2141–2161 (2022).Article 

Google Scholar 
Poulsen, A. D., Tuomisto, H. & Balslev, H. Edaphic and floristic variation within a 1-ha plot of lowland Amazonian rain forest. Biotropica 38, 468–478 (2006).Article 

Google Scholar 
Viana, J. L., Turner, B. L. & Dalling, J. W. Compositional variation in understorey fern and palm communities along a soil fertility and rainfall gradient in a lower montane tropical forest. J. Veg. Sci. 32, 12947 (2021).Article 

Google Scholar 
Lwanga, J. S., Balmford, A. & Badaza, R. Assessing fern diversity: relative species richness and its environmental correlates in Uganda. Biodivers. Conserv. 7, 1387–1398 (1998).Article 

Google Scholar 
Zuquim, G. et al. Predicting environmental gradients with fern species composition in Brazilian Amazonia. J. Veg. Sci. 25, 1195–1207 (2014).Article 

Google Scholar 
Chang, L. W. et al. Lienhuachih Subtropical Evergreen Broadleaf Forest Dynamics Plot Tree Species Characteristics and Distribution Patterns 334–335 (Taiwan Forestry Research Institute, 2012).Wei, J. B., Xiao, D. N., Zeng, H. & &Fu, Y. K. Spatial variability of soil properties in relation to land use and topography in a typical small watershed of the black soil region, Northeastern China. Environ. Geol. 53, 1663–1672 (2008).Article 
ADS 
CAS 

Google Scholar 
Messier, C., Parent, S. & Bergeron, Y. Effects of overstory and understory vegetation on the understory light environment in mixed boreal forests. J. Veg. Sci. 9, 511–520 (1998).Article 

Google Scholar 
Hill, J. D. & Silander, J. A. Distribution and dynamics of two ferns: Dennstaedtia punctilobula (Dennstaedtiaceae) and Thelypteris noveboracensis (Thelypteridaceae) in a Northeast mixed hardwoods–hemlock forest. Amer. J. Bot. 88, 894–902 (2001).Article 
CAS 

Google Scholar 
Angelini, A., Corona, P., Chianucci, F. & Portoghesi, L. Structural attributes of stand overstory and light under the canopy. Ann. Silvic Res. 39, 23–31 (2015).
Google Scholar 
Lee, P. H., Huang, Y. M., Chiou, W. L., Tseng, Y. H. & Tzeng, H. Y. The correlation between fern diversity and environmental factors-case study in a Chinese fir plantation in Lienhuachih, central Taiwan. Taiwan J. For. Sci. 39, 271–289 (2024).
Google Scholar 
George, L. O. & Bazzaz, F. A. The fern understory as an ecological filter: growth and survival of canopy-tree seedlings. Ecology 80, 846–856 (1999).Article 

Google Scholar 
Royo, A. A. & Carson, W. P. On the formation of dense understory layers in forests worldwide: consequences and implications for forest dynamics, biodiversity, and succession. Can. J. Res. 36, 1345–1362 (2006).Article 
ADS 

Google Scholar 
Tuomisto, H. et al. Distribution and diversity of pteridophytes and Melastomataceae along edaphic gradients in Yasuní National Park, Ecuadorian Amazonia. Biotropica 34, 516–533 (2002).Article 

Google Scholar 
Chang, L. W., Huang, J. L., Luo, S. F. & Lee, P. H. Understory plant composition and its relations with environmental factors of the Lienhuachih forest dynamics plot at a subtropical evergreen broadleaf forest in central Taiwan. Taiwan J. For. Sci. 30, 245–257 (2015).
Google Scholar 
Wolf, P. G., Schneider, H. & Ranker, T. A. Geographic distributions of homosporous ferns: does dispersal obscure evidence of vicariance? J. Biogeogr. 28, 263–270 (2001).Article 

Google Scholar 
Page, C. N. The diversity of ferns: an ecological perspective. In: The Experimental Biology of Ferns (ed. Dyer, A. F.) 10–56 (1979).Leake, J. R. The biology of myco-heterotrophic (saprophytic) plants. New. Phytol. 127, 171–216 (1994).Article 
PubMed 

Google Scholar 
Sanginés-Franco, C., Luna-Vega, I., Ayala, O. A. & Contreras-Medina, R. Distributional patterns and biogeographic analysis of ferns in the Sierra madre Oriental, Mexico. Amer Fern J. 101, 81–104 (2011).Article 

Google Scholar 
Chiu, T. Y. et al. Ecophysiological characteristics of three Cyathea species in Northeastern Taiwan. Taiwan J. For. Sci. 30, 147–155 (2015).
Google Scholar 
Silva, V. L., Mehltreter, K. & Schmitt, J. L. Ferns as potential ecological indicators of edge effects in two types of Mexican forests. Ecol. Indic. 93, 669–676 (2018).Article 

Google Scholar 
Chang, L. W. et al. Changes of plant communities classification and species composition along the micro-topography at the Lienhuachih forest dynamics plot in the central Taiwan. Taiwania 57, 359–371 (2012).CAS 

Google Scholar 
Boufford, D. E. et al. A Checklist of the Vascular Plants of Taiwan. in Flora of Taiwan, 2nd edition, Vol. 6 (ed. Editorial Committee of the Flora of Taiwan, S. E.) 15–139 (2003).Kuo, L. Y. et al. Updating Taiwanese pteridophyte checklist: A new phylogenetic classification. Taiwania 64, 367–395 (2019).
Google Scholar 
Chang, L. W., Zelený, D., Li, C. F., Chiu, S. T. & Hsieh, C. F. Better environmental data May reverse conclusions about niche- and dispersal-based processes in community assembly. Ecology 94, 2145–2151 (2013).Article 
PubMed 

Google Scholar 
Chang, L. W., Chiu, S. T. & Hsieh, C. F. Contrasting Spatial distribution of pioneer versus non-pioneer saplings in a Taiwanese forest: A multiple scale approach. Ecol. Res. 38, 604–616 (2023).Article 

Google Scholar 
Rolecěk, J., Tichy, L., Zeleny, D. & Chytry, M. Modified TWINSPAN classification in which the hierarchy respects cluster heterogeneity. J. Veg. Sci. 20, 596–602 (2009).Article 

Google Scholar 
Hill, M. O. & Gauch, H. G. Detrended correspondence analysis: an improved ordination technique. Vegetatio 42, 47–58 (1980).Article 

Google Scholar 
Ter Braak, C. J. F. Canonical correspondence analysis: A new eigenvector technique for multivariate direct gradient analysis. Ecology 67, 1167–1179 (1986).Article 

Google Scholar 
Itoh, A. et al. Importance of topography and soil texture in the Spatial distribution of two sympatric dipterocarp trees in a Bornean rainforest. Ecol. Res. 18, 307–320 (2003).Article 

Google Scholar 
Jucker, T. et al. Topography shapes the structure, composition and function of tropical forest landscapes. Ecol. Lett. 21, 989–1000 (2018).Article 
PubMed 
PubMed Central 

Google Scholar 
Wilcke, W. et al. Soil properties and tree growth along an altitudinal transect in Ecuadorian tropical montane forest. J. Plant. Nutr. Soil. Sci. 171, 220–230 (2008).Article 
CAS 

Google Scholar 
Liang, M. et al. Research progress on the ecology of dense fern understory. J. Trop. Subtrop Bot. 30, 219–300 (2022).ADS 

Google Scholar 
Ruokolainen, K., Tuomisto, H., Macía, M. J., Higgins, M. A. & Yli-Halla, M. Are floristic and edaphic patterns in Amazonian rain forests congruent for trees, pteridophytes and melastomataceae? J. Trop. Ecol. 23, 13–25 (2007).Article 

Google Scholar 
Jones, M. M., Ruokolainen, K., Martinez, N. C. L. & Tuomisto, H. Differences in topographic and soil habitat specialization between trees and two understorey plant groups in a Costa Rican lowland rain forest. J. Trop. Ecol. 32, 482–497 (2016).Article 

Google Scholar 
Download referencesAcknowledgementsWe thank the students of the Department of Forestry at National Chung Hsing University for their assistance in the wild investigation.Author informationAuthors and AffiliationsDepartment of Forestry, National Chung Hsing University, No. 145, Xingda Rd., South Dist, Taichung, 402202, TaiwanPei-Hsuan Lee, Yen-Hsueh Tseng & Hsy-Yu TzengTaiwan Forestry Research Institute, No. 53, Nanhai Rd, Taipei, 100051, TaiwanPei-Hsuan Lee, Yao-Moan Huang, Li-Wan Chang, Jian-Hong Yang, Wen-Liang Chiou & Yen-Hsueh TsengAuthorsPei-Hsuan LeeView author publicationsSearch author on:PubMed Google ScholarYao-Moan HuangView author publicationsSearch author on:PubMed Google ScholarLi-Wan ChangView author publicationsSearch author on:PubMed Google ScholarJian-Hong YangView author publicationsSearch author on:PubMed Google ScholarWen-Liang ChiouView author publicationsSearch author on:PubMed Google ScholarYen-Hsueh TsengView author publicationsSearch author on:PubMed Google ScholarHsy-Yu TzengView author publicationsSearch author on:PubMed Google ScholarContributionsConception or design of the work: Pei-Hsuan Lee, Hsy-Yu Tzeng. Data collection: Pei-Hsuan Lee, Li-Wan Chang, Jian-Hong Yang. Data analysis and interpretation: Pei-Hsuan Lee. Drafting the article: Pei-Hsuan Lee. Critical revision of the article: Wen-Liang Chiou, Yao-Moan Huang, Hsy-Yu Tzeng, Yen-Hsueh Tseng. All the authors reviewed the manuscript.Corresponding authorCorrespondence to
Hsy-Yu Tzeng.Ethics declarations

Competing interests
The authors declare no competing interests.

Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Reprints and permissionsAbout this articleCite this articleLee, PH., Huang, YM., Chang, LW. et al. Environmental heterogeneity and its influence on fern diversity in a low-altitude mountain forest in central Taiwan.
Sci Rep 15, 44383 (2025). https://doi.org/10.1038/s41598-025-28048-9Download citationReceived: 29 May 2025Accepted: 07 November 2025Published: 23 December 2025Version of record: 23 December 2025DOI: https://doi.org/10.1038/s41598-025-28048-9Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
KeywordsFern communityTopographySoil propertiesBiotic factors More

AbstractThe seal louse (Echinophthirius horridus) is a remarkable example of evolutionary adaptation, thriving as an obligate ectoparasite on deep-diving marine mammals under extreme environmental conditions, including high hydrostatic pressure, extreme drag force, salinity, and fluctuating temperatures. To investigate the anatomical and functional specializations enabling this lifestyle, we compared the leg morphology and musculature of E. horridus with its terrestrial relative, the human head louse (Pediculus humanus capitis), using synchrotron-based 3D microtomography and confocal laser scanning microscopy. Our findings reveal that the seal louse has developed a highly compact and robust leg structure with a fused tibiotarsus, an additional set of leg muscles, and a shortened claw tendon—an unprecedented adaptation among insects. These features allow for greater force transmission and reduced metabolic cost during sustained attachment. Behavioral assays further show that E. horridus can only move effectively on hair-like substrates, underscoring its complete reliance on host fur. These findings suggest a highly specialized muscular control system enabling strong, reliable, and reversible attachment in a challenging aquatic environment.

Similar content being viewed by others

Attachment performance of the ectoparasitic seal louse Echinophthirius horridus

Article
Open access
05 January 2024

The ectoparasitic seal louse, Echinophthirius horridus, relies on a sealed tracheal system and spiracle closing apparatus for underwater respiration

Article
Open access
03 June 2025

The deeper the rounder: body shape variation in lice parasitizing diving hosts

Article
Open access
09 September 2024

Data availability

All data is provided in the Supplementary Material of the manuscript. Synchrotron data and histological sectioning series90 can be provided upon request or online under: [https://doi.org/10.6084/m9.figshare.28596953.v2] (https:/doi.org/https://doi.org/10.6084/m9.figshare.28596953.v2).
ReferencesGorb, S.N. (2001) Attachment Devices of Insect Cuticle. 1st edn, Springer Netherlands. 1st edn. Dordrecht, Netherlands: Springer, Dordrecht. https://doi.org/10.1007/0-306-47515-4.Beutel, R. G. & Gorb, S. N. Ultrastructure of attachment specializations of hexapods (Arthropoda): Evolutionary patterns inferred from a revised ordinal phylogeny. J. Zool. Syst. Evol. Res. 39(4), 177–207. https://doi.org/10.1046/j.1439-0469.2001.00155.x (2001).
Google Scholar 
Gorb, S. N. Biological attachment devices: exploring nature’s diversity for biomimetics. Philos. Trans. R. Soc. A: Math., Phys. Eng. Sci. 366(1870), 1557–1574. https://doi.org/10.1098/rsta.2007.2172 (2008).
Google Scholar 
Gorb, S. N. & Beutel, R. Evolution of locomotory attachment pads of hexapods. Naturwissenschaften 88(12), 530–534. https://doi.org/10.1007/s00114-001-0274-y (2001).
Google Scholar 
Bajerlein, D. et al. To attach or not to attach? The effect of carrier surface morphology and topography on attachment of phoretic. Naturwissenschaften 103(7–8), 61. https://doi.org/10.1007/s00114-016-1385-9 (2016).
Google Scholar 
Federle, W. & Labonte, D. Dynamic biological adhesion: Mechanisms for controlling attachment during locomotion. Philos. Trans. R. Soc. B: Biol. Sci. 374(1784), 20190199. https://doi.org/10.1098/rstb.2019.0199 (2019).
Google Scholar 
Li, D., Huson, M. G. & Graham, L. D. Proteinaceous adhesive secretions from insects, and in particular the egg attachment glue of Opodiphthera sp. moths. Arch. Insect Biochem. Physiol.: Publ. Collab. Entomol. Soc. Am. 69(2), 85–105. https://doi.org/10.1002/arch.20267 (2008).
Google Scholar 
Autumn, K. et al. Adhesive force of a single gecko foot-hair. Nature 405(6787), 681–685. https://doi.org/10.1038/35015073 (2000).
Google Scholar 
Autumn, K. et al. Evidence for van der Waals adhesion in gecko setae. Proc. Natl. Acad. Sci. 99(19), 12252–12256. https://doi.org/10.1073/pnas.192252799 (2002).
Google Scholar 
Autumn, K. & Hansen, W. Ultrahydrophobicity indicates a non-adhesive default state in gecko setae. J. Comp. Physiol. A. 192, 1205–1212. https://doi.org/10.1007/s00359-006-0149-y (2006).
Google Scholar 
Autumn, K. & Peattie, A. M. Mechanisms of adhesion in geckos. Integr. Comp. Biol. 42(6), 1081–1090. https://doi.org/10.1093/icb/42.6.1081 (2002).
Google Scholar 
Badge, I. et al. The role of surface chemistry in adhesion and wetting of gecko toe pads. Sci. Rep. 4(1), 6643. https://doi.org/10.1038/srep06643 (2014).
Google Scholar 
Maderson, P. F. A. Keratinized epidermal derivatives as an aid to climbing in gekkonid lizards. Nature 203(4946), 780–781. https://doi.org/10.1038/203780a0 (1964).
Google Scholar 
Mitchell, C. T. et al. The effect of substrate wettability and modulus on gecko and gecko-inspired synthetic adhesion in variable temperature and humidity. Sci. Rep. 10(1), 19748. https://doi.org/10.1038/s41598-020-76484-6 (2020).
Google Scholar 
Ruibal, R. & Ernst, V. The structure of the digital setae of lizards. J. Morphol. 117(3), 271–293. https://doi.org/10.1002/jmor.1051170302 (1965).
Google Scholar 
Williams, E. E. & Peterson, J. A. Convergent and alternative designs in the digital adhesive pads of scincid lizards. Science 215(4539), 1509–1511. https://doi.org/10.1126/science.215.4539.1509 (1982).
Google Scholar 
Higham, T. E. & Russell, A. P. Geckos running with dynamic adhesion: Towards integration of ecology, energetics and biomechanics. J. Exp. Biol. 228, 247980. https://doi.org/10.1242/jeb.247980 (2025).
Google Scholar 
Russell, A. P. The morphological basis of weight-bearing in the scansors of the tokay gecko (Reptilia: Sauria). Can. J. Zool. 64(4), 948–955. https://doi.org/10.1139/z86-144 (1986).
Google Scholar 
Chen, Y. et al. Underwater attachment using hairs: The functioning of spatula and sucker setae from male diving beetles. J. R. Soc. Interface 11(97), 20140273. https://doi.org/10.1098/rsif.2014.0273 (2014).
Google Scholar 
Kampowski, T. et al. Exploring the attachment of the Mediterranean medicinal leech (Hirudo verbana) to porous substrates. J. R. Soc. Interface 17(168), 20200300. https://doi.org/10.1098/rsif.2020.0300 (2020).
Google Scholar 
Kier, W. M. & Smith, A. M. The structure and adhesive mechanism of octopus suckers. Integr. Comp. Biol. 42(6), 1146–1153. https://doi.org/10.1093/icb/42.6.1146 (2002).
Google Scholar 
Smith, A. M. Negative pressure generated by octopus suckers: a study of the tensile strength of water in nature. J. Exp. Biol. 157(1), 257–271. https://doi.org/10.1242/jeb.157.1.257 (1991).
Google Scholar 
Smith, A. M. Cephalopod sucker design and the physical limits to negative pressure. J. Exp. Biol. 199(4), 949–958. https://doi.org/10.1242/jeb.199.4.949 (1996).
Google Scholar 
Busshardt, P. & Gorb, S. N. Walking on smooth and rough ground: Activity and timing of the claw retractor muscle in the beetle Pachnoda marginata peregrina (Coleoptera, Scarabaeidae). J. Exp. Biol. 216(2), 319–328. https://doi.org/10.1242/jeb.075614 (2013).
Google Scholar 
Dunlop, J. A. Movements of scopulate claw tufts at the tarsus tip of a tarantula spider. Netherlands J. Zool. 45(3–4), 513–520 (1994).
Google Scholar 
Federle, W. et al. Biomechanics of the movable pretarsal adhesive organ in ants and bees. Proc. Natl. Acad. Sci. 98(11), 6215–6220. https://doi.org/10.1073/pnas.111139298 (2001).
Google Scholar 
Frantsevich, L. & Gorb, S. Structure and mechanics of the tarsal chain in the hornet, Vespa crabro (Hymenoptera: Vespidae): implications on the attachment mechanism. Arthropod Struct. Dev. 33(1), 77–89. https://doi.org/10.1016/j.asd.2003.10.003 (2004).
Google Scholar 
Frazier, S. F. et al. Elasticity and movements of the cockroach tarsus in walking. J. Comp. Physiol. – A Sensory, Neural, Behav. Physiol. 185(2), 157–172. https://doi.org/10.1007/s003590050374 (1999).
Google Scholar 
Heming, B. S. Functional morphology of the thysanopteran pretarsus. Can. J. Zool. 49(1), 91–108. https://doi.org/10.1139/z71-014 (1971).
Google Scholar 
Niederegger, S. & Gorb, S. N. Tarsal movements in flies during leg attachment and detachment on a smooth substrate. J. Insect Physiol. 49(6), 611–620. https://doi.org/10.1016/S0022-1910(03)00048-9 (2003).
Google Scholar 
Gorb, S. N. et al. The insect unguitractor plate in action: Force transmission and the micro CT visualizations of inner structures. J. Insect Physiol. 117, 103908. https://doi.org/10.1016/j.jinsphys.2019.103908 (2019).
Google Scholar 
Gorb, S. N., Gorb, E. V. & Kastner, V. Scale effects on the attachment pads and friction forces in syrphid flies (Diptera, Syrphidae). J. Exp. Biol. 204(8), 1421–1431. https://doi.org/10.1242/jeb.204.8.1421 (2001).
Google Scholar 
Büscher, T. H. et al. The exceptional attachment ability of the ectoparasitic bee louse Braula coeca (Diptera, Braulidae) on the honeybee. Physiol. Entomol. 47(2), 83–95. https://doi.org/10.1111/PHEN.12378 (2022).
Google Scholar 
Petersen, D. S. et al. Holding tight to feathers – structural specializations and attachment properties of the avian ectoparasite Crataerina pallida (Diptera, Hippoboscidae). J. Exp. Biol. 221(13), jeb179242. https://doi.org/10.1242/jeb.179242 (2018).
Google Scholar 
Preuss, A. et al. Attachment performance of the ectoparasitic seal louse Echinophthirius horridus. Commun. Biol. 7(1), 36. https://doi.org/10.1038/s42003-023-05722-0 (2024).
Google Scholar 
Bush, A. O. et al. Parasitism: The diversity and ecology of animal parasites (Cambridge University Press, Cambridge, 2001).
Google Scholar 
Kim, K. C. Coevolution of parasitic arthropods and mammals (Wiley-Interscience, 1985).
Google Scholar 
Anderson, R. C. ‘Host-parasite relations and evolution of the Metastrongyloidea (Nematoda)’, Memoires du Museum National d’Histoire Naturelle. Serie A. Zoologie 123, 129–132. https://doi.org/10.5281/zenodo.16007555 (1982).
Google Scholar 
Raga, J. A. et al. Parasites. In Encyclopedia of Marine Mammals (eds Perrin, W. F. et al.) 821–830 (Academic Press, London, UK, 2009). https://doi.org/10.1016/B978-0-12-373553-9.00193-0.
Google Scholar 
Rybczynski, N., Dawson, M. R. & Tedford, R. H. A semi-aquatic Arctic mammalian carnivore from the Miocene epoch and origin of Pinnipedia. Nature 458(7241), 1021–1024. https://doi.org/10.1038/nature07985 (2009).
Google Scholar 
Durden, L. A. & Musser, G. G. The sucking lice (Insecta, Anoplura) of the world – a taxonomic checklist with records of mammalian hosts and geographical distributions. Bull. Am. Mus. Nat. Hist. 218, 1–90 (1994).
Google Scholar 
Grzimek, B. Grzimek’s encyclopedia of mammals (McGraw-Hill Publishing Company, 1990).
Google Scholar 
Leonardi, M. S. & Palma, R. L. Review of the systematics, biology and ecology of lice from pinnipeds and river otters (Insecta: Phthiraptera: Anoplura: Echinophthiriidae). Zootaxa 3630(3), 445–466. https://doi.org/10.11646/zootaxa.3630.3.3 (2013).
Google Scholar 
Eguchi, T. & Harvey, J. ‘Diving behavior of the Pacific harbor seal (Phoca vitulina richardii) in Monterey Bay California. Mar. Mamm. Sci. 21, 283–295. https://doi.org/10.1111/j.1748-7692.2005.tb01228.x (2006).
Google Scholar 
Frost, K. J., Simpkins, M. A. & Lowry, L. F. Diving behavior of subadult and adult harbor seals in Prince William Sound, Alaska. Mar. Mamm. Sci. 17(4), 813–834. https://doi.org/10.1111/j.1748-7692.2001.tb01300.x (2001).
Google Scholar 
Gjertz, I., Lydersen, C. & Wiig, Ø. Distribution and diving of harbour seals (Phoca vitulina) in Svalbard. Polar Biol. 24(3), 209–214. https://doi.org/10.1007/s003000000197 (2001).
Google Scholar 
Hastings, K. K. et al. Regional differences in diving behavior of harbor seals in the Gulf of Alaska. Can. J. Zool. 82(11), 1755–1773. https://doi.org/10.1139/z04-145 (2004).
Google Scholar 
Kolb, P. M. A harbor seal Phoca vitulina richardsi, taken from a sablefish trap. California Fish Game 68, 123–124 (1982).
Google Scholar 
Rosing-Asvid, A. et al. Deep diving harbor seals (Phoca vitulina) in South Greenland: movements, diving, haul-out and breeding activities described by telemetry. Polar Biol. 43(4), 359–368. https://doi.org/10.1007/s00300-020-02639-w (2020).
Google Scholar 
Dehnhardt, G., Mauck, B. & Hyvärinen, H. Ambient temperature does not affect the tactile sensitivity of mystacial vibrissae in harbour seals. J. Exp. Biol. 201(22), 3023–3029. https://doi.org/10.1242/jeb.201.22.3023 (1998).
Google Scholar 
Mauck, B. et al. Thermal windows on the trunk of hauled-out seals: Hot spots for thermoregulatory evaporation?. J. Exp. Biol. 206(10), 1727–1738. https://doi.org/10.1242/jeb.00348 (2003).
Google Scholar 
Hansen, S. & Lavigne, D. M. Ontogeny of the thermal limits in the harbor seal (Phoca vitulina). Physiol. Zool. 70(1), 85–92. https://doi.org/10.1086/639549 (1997).
Google Scholar 
Watts, P. Thermal constraints on hauling out by harbor seals (Phoca vitulina). Can. J. Zool. 70, 553–560. https://doi.org/10.1139/z92-083 (2011).
Google Scholar 
Leonardi, M. S. et al. Under pressure: the extraordinary survival of seal lice in the deep sea. J. Exp. Biol. 223(17), jeb226811. https://doi.org/10.1242/jeb.226811 (2020).
Google Scholar 
Williams, T. M. & Kooyman, G. L. Swimming performance and hydrodynamic characteristics of harbor seals Phoca vitulina. Physiol. Zool. 58(5), 576–589. https://doi.org/10.1086/physzool.58.5.30158584 (1985).
Google Scholar 
Preuss, A., Gorb, S.N., et al. (2025) ‘Role of the Setae in an Ectoparasitic Seal Louse in Reducing Surface Drag: Numerical Modeling Approach’, Advanced Theory and Simulations, p. e00429. https://doi.org/10.1002/adts.202500429.Herzog, I. et al. Heartworm and seal louse: Trends in prevalence, characterisation of impact and transmission pathways in a unique parasite assembly on seals in the North and Baltic Sea. Int. J. Parasitol.: Parasites Wildlife 23, 100898. https://doi.org/10.1016/j.ijppaw.2023.100898 (2024).
Google Scholar 
Herzog, I., Siebert, U. & Lehnert, K. High prevalence and low intensity of Echinophthirius horridus infection in seals revealed by high effort sampling. Sci. Rep. 14(1), 14258. https://doi.org/10.1038/s41598-024-64890-z (2024).
Google Scholar 
Siebert, U. et al. Pathological findings in harbour seals (Phoca vitulina): 1996–2005. J. Comp. Pathol. 137(1), 47–58. https://doi.org/10.1016/j.jcpa.2007.04.018 (2007).
Google Scholar 
Michels, J. & Gorb, S. N. Detailed three-dimensional visualization of resilin in the exoskeleton of arthropods using confocal laser scanning microscopy. J. Microsc. 245(1), 1–16. https://doi.org/10.1111/j.1365-2818.2011.03523.x (2012).
Google Scholar 
Andersen, S. O. Biochemistry of insect cuticle. Annu. Rev. Entomol. 24(1), 29–59. https://doi.org/10.1146/annurev.en.24.010179.000333 (1979).
Google Scholar 
Büsse, S. & Gorb, S. N. Material composition of the mouthpart cuticle in a damselfly larva (Insecta: Odonata) and its biomechanical significance. R. Soc. Open Sci. 5(6), 172117. https://doi.org/10.1098/rsos.172117 (2018).
Google Scholar 
Josten, B., Gorb, S. N. & Büsse, S. The mouthparts of the adult dragonfly Anax imperator (Insecta: Odonata), functional morphology and feeding kinematics. J. Morphol. 283(9), 1163–1181. https://doi.org/10.1002/jmor.21497 (2022).
Google Scholar 
Vincent, J. F. V. Arthropod cuticle: a natural composite shell system. Compos. A Appl. Sci. Manuf. 33(10), 1311–1315. https://doi.org/10.1016/S1359-835X(02)00167-7 (2002).
Google Scholar 
Cecilia, A. et al. The IMAGE beamline at the KIT light source. J. Synchrotron Radiat. 32(4), 1036–1051. https://doi.org/10.1107/s1600577525003777 (2025).
Google Scholar 
Douissard, P. A. et al. A versatile indirect detector design for hard X-ray microimaging. J. Instrum. 7(9), P09016. https://doi.org/10.1088/1748-0221/7/09/P09016 (2012).
Google Scholar 
Vogelgesang, M. et al. Real-time image-content-based beamline control for smart 4D X-ray imaging. J. Synchrotron Radiat. 23(5), 1254–1263. https://doi.org/10.1107/S1600577516010195 (2016).
Google Scholar 
Vogelgesang, M. et al. (2012) ‘UFO: A Scalable GPU-based Image Processing Framework for On-line Monitoring’, in Proceedings of HPCC-ICESS., pp. 824–829. https://doi.org/10.1109/HPCC.2012.116.Faragó, T. et al. Tofu: A fast, versatile and user-friendly image processing toolkit for computed tomography. J. Synchrotron Radiat. 29, 916–927. https://doi.org/10.1107/S160057752200282X (2022).
Google Scholar 
Gray, P. T. A., Mill, P. J. & Dodd, J. M. ‘The musculature of the prothoracic legs and its innervation in Hierodula membranacea (Mantidea)’. Philos. Trans. R. Soc. London B, Biol. Sci. 309(1140), 479–503. https://doi.org/10.1098/rstb.1985.0094 (1997).
Google Scholar 
Preuss, A. et al. The ectoparasitic seal louse, Echinophthirius horridus, relies on a sealed tracheal system and spiracle closing apparatus for underwater respiration. Commun. Biol. 8(1), 1–14. https://doi.org/10.1038/s42003-025-08285-4 (2025).
Google Scholar 
Leonardi, M. S. et al. The deeper the rounder: body shape variation in lice parasitizing diving hosts. Sci. Rep. 14(1), 1–10. https://doi.org/10.1038/s41598-024-71541-w (2024).
Google Scholar 
Appel, E. et al. Ultrastructure of dragonfly wing veins: Composite structure of fibrous material supplemented by resilin. J. Anat. 227(4), 561–582. https://doi.org/10.1111/joa.12362 (2015).
Google Scholar 
Vaughan, J. A. & Azad, A. F. Patterns of erythrocyte digestion by bloodsucking insects: constraints on vector competence. J. Med. Entomol. 30(1), 214–216. https://doi.org/10.1093/jmedent/30.1.214 (1993).
Google Scholar 
Waniek, P. J. The digestive system of human lice: Current advances and potential applications. Physiol. Entomol. 34(3), 203–210. https://doi.org/10.1111/j.1365-3032.2009.00681.x (2009).
Google Scholar 
Leonardi, M. S. et al. Host-parasite coevolution leads to underwater respiratory adaptations in extreme diving insects, seal lice (Lepidophthirus macrorhini). Communications Biology 8(1), 1–11. https://doi.org/10.1038/s42003-025-08306-2 (2025).
Google Scholar 
Soler Cruz, M. D. & Martín Mateo, M. P. Scanning electron microscopy of legs of two species of sucking lice (Anoplura: Phthiraptera). Micron 40(3), 401–408. https://doi.org/10.1016/j.micron.2008.10.001 (2009).
Google Scholar 
Wolfram, L. J. Human hair: A unique physicochemical composite. J. Am. Acad. Dermatol. 48(6), S106–S114. https://doi.org/10.1067/mjd.2003.276 (2003).
Google Scholar 
Aibekova, L. et al. The skeletomuscular system of the mesosoma of Formica rufa Workers (Hymenoptera: Formicidae). Insect Systematics and Diversity 6(2), 1–26. https://doi.org/10.1093/isd/ixac002 (2022).
Google Scholar 
Woodworth, C. W. The leg tendons of insects. Am. Nat. 42(499), 452–456. https://doi.org/10.1086/278953 (1908).
Google Scholar 
Lee, S. S. M. & Piazza, S. J. Built for speed: Musculoskeletal structure and sprinting ability. J. Exp. Biol. 212(22), 3700–3707. https://doi.org/10.1242/jeb.031096 (2009).
Google Scholar 
Baxter, J. R. et al. Ankle joint mechanics and foot proportions differ between human sprinters and non-sprinters. Proc. R. Soc. B: Biol. Sci. 279(1735), 2018–2024. https://doi.org/10.1098/rspb.2011.2358 (2012).
Google Scholar 
Fletcher, J. R. & MacIntosh, B. R. Achilles tendon strain energy in distance running: consider the muscle energy cost. J. Appl. Physiol. 118(2), 193–199. https://doi.org/10.1152/japplphysiol.00732.2014 (2015).
Google Scholar 
MacIntosh, B.R. and Holash, R.J. (2000) ‘Power output and force-velocity properties of muscle’, In: B.M. Nigg, B.R. MacIntosh, and J. Mester (Eds) Biomechanics and biology of movement. Human Kinetics. Champaign, Illinois (US): Human Kinetics, pp. 193–210.Kim, K. C. Ecology and morphological adaptations of the sucking lice on the nothern fur seal. J. Cons. Int. Explor. Mer. 169, 504–515 (1975).
Google Scholar 
Koštál, V. Eco-physiological phases of insect diapause. J. Insect Physiol. 52(2), 113–127. https://doi.org/10.1016/j.jinsphys.2005.09.008 (2006).
Google Scholar 
Leonardi, M. S. & Lazzari, C. R. Uncovering deep mysteries: The underwater life of an amphibious louse. J. Insect Physiol. 71, 164–169. https://doi.org/10.1016/j.jinsphys.2014.10.016 (2014).
Google Scholar 
Gorb, S. N. Design of insect unguitractor apparatus. J. Morphol. 230(2), 219–230. https://doi.org/10.1002/(SICI)1097-4687(199611)230:2%3c219::AID-JMOR8%3e3.0.CO;2-B (1996).
Google Scholar 
Hörger, V., Labisch, S. & Dirks, J.-H. Biomimetic tag attachment inspired by the seal louse. Bioinspiration Biomimetics 20(6), 066015. https://doi.org/10.1088/1748-3190/adfbb8 (2025).
Google Scholar 
Preuss, A., Schwaha, T., et al. (2025a) ‘Nano-CT data and histological sections of Pediculus humanus capitis and Echinophthirius horridus’, figshare. https://doi.org/10.6084/m9.figshare.28596953.v2.Download referencesAcknowledgementsWe express our profound gratitude to Dr. Thies Büscher, Dr. Helen Gorges, Fabian Bäumler, Julian Thomas, Simon Züger, and Benedikt Josten for their invaluable guidance and assistance throughout this study. We also extend our appreciation to Esther Appel and Dr. Alexander Kovalev for their technical support during the experiments. We would also like to thank the TiHo Hannover, especially Dr. Insa Herzog and Dr. Kristina Lehnert, for supplying us with seal louse samples. Additionally, we acknowledge Marta Tischer for generously supplying head louse samples preserved in ethanol. We are indebted to Dr. Angelica Cecilia for her assistance during beamtime and to Dr. Tomáš Faragó for his work on tomographic reconstructions. We express our gratitude to the KIT Light Source for providing instruments at their beamlines and to the Institute for Beam Physics and Technology (IBPT) for operating the storage ring, the Karlsruhe Research Accelerator (KARA). Furthermore, we acknowledge the funding provided to S.N.G. by the grant GO 995 46-1 from the German Science Foundation (DFG) within the Special Priority Program (SPP 2332) “Physics of Parasitism.” The funders had no involvement in the study design, data collection and analysis, publication decisions, or manuscript preparation.FundingOpen Access funding enabled and organized by Projekt DEAL. Funding for this study was granted to S.N.G by the German Science Foundation (DFG) under grant number GO 995 46–1, as part of the Special Priority Program (SPP 2332) known as “Physics of Parasitism”.Author informationAuthors and AffiliationsDepartment of Functional Morphology and Biomechanics, Zoological Institute, Kiel University, Am Botanischen Garten 1–9, 24118, Kiel, GermanyAnika Preuss, Lina Ornowski & Stanislav N. GorbInstitute for Photon Science and Synchrotron Radiation (IPS), Karlsruhe Institute of Technology (KIT), Hermann-Von-Helmholtz-Platz 1, 76344, Eggenstein-Leopoldshafen, GermanyThomas van de Kamp, Elias Hamann & Marcus ZuberLaboratory for Applications of Synchrotron Radiation (LAS), Karlsruhe Institute of Technology (KIT), Kaiserstr.12, 76131, Karlsruhe, GermanyThomas van de KampAuthorsAnika PreussView author publicationsSearch author on:PubMed Google ScholarThomas van de KampView author publicationsSearch author on:PubMed Google ScholarElias HamannView author publicationsSearch author on:PubMed Google ScholarMarcus ZuberView author publicationsSearch author on:PubMed Google ScholarLina OrnowskiView author publicationsSearch author on:PubMed Google ScholarStanislav N. GorbView author publicationsSearch author on:PubMed Google ScholarContributionsA.P., S.N.G.—Conceptualization; A.P., T.v.d.K., S.N.G.—Data curation; A.P., T.v.d.K., S.N.G.—Formal analysis; S.N.G.—Funding acquisition; A.P., T.v.d.K., L.O.—Investigation; A.P., T.v.d.K., E.H., M.Z., S.N.G.—Methodology; S.N.G.—Project administration; A.P.—Software; S.N.G.—Supervision; A.P., S.N.G.—Validation; A.P., L.O.—Visualization; A.P.—Writing—original draft; T.v.d.K., E.H., M.Z., L.O., S.N.G.—Writing – review & editing.Corresponding authorCorrespondence to
Anika Preuss.Ethics declarations

Competing interests
The authors declare no competing interests.

Ethical approval
Ethical review and approval were not required for this study, as all host animals were either found dead, died naturally, or were euthanized on welfare grounds, with none being killed specifically for this research. The authors were not involved in the euthanasia of the hosts, which was carried out by certified seal rangers for reasons unrelated to this study. All regulations regarding animal use were strictly followed.

Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Supplementary Material 2Supplementary Material 3Supplementary Material 4Supplementary Material 5Supplementary Material 6Supplementary Material 7Supplementary Material 8Supplementary Material 9Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissionsAbout this articleCite this articlePreuss, A., van de Kamp, T., Hamann, E. et al. Functional morphology of the leg musculature in the marine seal louse: adaptations for high-performance attachment to diving hosts.
Sci Rep (2025). https://doi.org/10.1038/s41598-025-32804-2Download citationReceived: 16 September 2025Accepted: 12 December 2025Published: 23 December 2025DOI: https://doi.org/10.1038/s41598-025-32804-2Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
KeywordsParasitismSeal louseHuman head louseMarine mammalsBiomechanicsExtremitiesSkeleton-muscle organizationSupplementary Material 4Supplementary Material 5Supplementary Material 6 More

AbstractHeavy-metal (HM) contamination undermines soil functions and food safety, while risk appraisals often rely on chemical indices that can be unstable in the presence of extremes and only indirectly reflect biological integrity. We present an integrative framework that couples standardized contamination metrics with soil microbiome profiling to deliver stable, interpretable classifications and actionable bioindicators. Twelve peri-urban soils from Southern Italy were analysed for potentially toxic elements, including Arsenic (As), Cadmium (Cd), Chromium (Cr), Copper (Cu), Nickel (Ni), Lead (Pb), and Zinc (Zn) and profiled by shotgun metagenomics. We introduce a Standardized Ecological Risk index (SPERI) that preserves the ranking conveyed by conventional composites yet reduces outlier leverage. SPERI strongly agreed with Improved Potential Ecological Risk Index (IPERI) while stabilizing variance (R² = 0.896) and improved between-site comparability. Along the contamination gradient, community structure shifted consistently: families such as Pseudomonadaceae, Xanthomonadaceae and Rhodospirillaceae increased with risk, whereas Geodermatophilaceae and Nocardiaceae declined. Simple decision-tree models trained on family-level relative abundances reliably separated SPERI classes and repeatedly selected Zn- and Cd-enriched sites as primary split drivers, aligning microbial signals with chemical risk. By combining open, reproducible analytics with jointly chemical- and microbiome-informed endpoints, this workflow improves the interpretability and transferability of ecological risk assessment and supports targeted remediation and monitoring in contaminated agro-ecosystems.

Similar content being viewed by others

Integrating ryegrass-maize intercropping systems for mitigating Cd-Pb contamination: enhancing ecosystem services and habitat sustainability

Article
Open access
24 November 2025

Heavy metal pollution migration and its ecological impact on microbial communities in the karst region of Guangxi

Article
Open access
28 April 2025

The spatial distribution of potentially toxic elements in the mountain forest topsoils (the Silesian Beskids, southern Poland)

Article
Open access
03 January 2024

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
ReferencesGatta, G. et al. Irrigation with treated municipal wastewater on artichoke crop: assessment of soil and yield heavy metal content and human risk. Water 10, 255 (2018).
Google Scholar 
Nyiramigisha, P., Sajidan & Komariah, & Harmful impacts of heavy metal contamination in the soil and crops grown around dumpsites. Rev. Agric. Sci. 9, 271–282 (2021).
Google Scholar 
Sodhi, K. K., Mishra, L. C., Singh, C. K. & Kumar, M. Perspective on the heavy metal pollution and recent remediation strategies. Curr. Res. Microb. Sci. 3, 100166 (2022).
Google Scholar 
Rashid, A. et al. Heavy metal contamination in agricultural soil: environmental pollutants affecting crop health. Agronomy 13, 1521 (2023).
Google Scholar 
Hama Aziz, K. H. et al. Heavy metal pollution in the aquatic environment: efficient and low-cost removal approaches to eliminate their toxicity: a review. RSC Adv. 13, 17595–17610 (2023).
Google Scholar 
Mavakala, B. K. et al. Evaluation of heavy metal content and potential ecological risks in soil samples from wild solid waste dumpsites in developing country under tropical conditions. Environ. Chall. 7, 100461 (2022).
Google Scholar 
Tiabou, A. F., Tanyi, T. A. M., Yiika, L. P. & Ayuk, M. M. A. Spatial distribution, ecological and ecotoxicity evaluation of heavy metals in agricultural soils along Lala-Manjo Highway, Cameroon volcanic line. Discov Soil 1, (2024).Xu, X. et al. Risk hotspots and influencing factors identification of heavy metal(loid)s in agricultural soils using Spatial bivariate analysis and random forest. Sci. Total Environ. 954, 176359 (2024).
Google Scholar 
Marvelli, E. www.gat.report – Foggia, 618 thousand tons of illegal waste. (2025). https://www.gat.report/28593/foggia-618-thousand-tons-of-illegal-waste/Brevik, E. C. et al. Soil and human health: current status and future needs. Air Soil. Water Res 13, (2020).Khatun, J., Intekhab, A. & Dhak, D. Effect of uncontrolled fertilization and heavy metal toxicity associated with arsenic (As), lead (Pb) and cadmium (Cd), and possible remediation. Toxicology 477, 153274 (2022).
Google Scholar 
Munir, N. et al. Heavy metal contamination of natural foods is a serious health issue: A review. Sustainability 14, 161 (2021).
Google Scholar 
Patel, N. et al. Contamination and health impact of heavy metals. in Environmental Chemistry for a Sustainable World 259–280 (Springer International Publishing, Cham, doi:https://doi.org/10.1007/978-3-030-52421-0_10. (2021).
Google Scholar 
Grados, L. et al. How advanced are we on the consequences of oral exposure to food contaminants on the occurrence of chronic Non communicable diseases? Chemosphere 303, 135260 (2022).
Google Scholar 
Shea, D. & Thorsen, W. Ecological risk assessment. in Progress in Molecular Biology and Translational Science vol. 112 323–348 (Elsevier, (2012).Rattner, B. A. et al. Wildlife ecological risk assessment in the 21st century: promising technologies to assess toxicological effects. Integr. Environ. Assess. Manag. 20, 725–748 (2024).
Google Scholar 
Bhaduri, D. et al. A review on effective soil health bio-indicators for ecosystem restoration and sustainability. Front Microbiol 13, (2022).Naylor, D., McClure, R. & Jansson, J. Trends in microbial community composition and function by soil depth. Microorganisms 10, 540 (2022).
Google Scholar 
Campillo-Cora, C., Rodríguez-Seijo, A., Pérez-Rodríguez, P. & Fernández-Calviño, D. Santás-Miguel, V. Effect of heavy metal pollution on soil microorganisms: influence of soil physicochemical properties. A systematic review. Eur. J. Soil. Biol. 124, 103706 (2025).
Google Scholar 
Wang, Z. et al. Microbial diversity and community assembly in heavy metal-contaminated soils: insights from selenium-impacted mining areas. Front Microbiol 16, (2025).Ma, F. et al. Development of microbial indicators in ecological systems. Int. J. Environ. Res. Public. Health. 19, 13888 (2022).
Google Scholar 
Visca, A. et al. Microbial bioindicators for monitoring the impact of emerging contaminants on soil health in the European framework. Sustainability 17, 1093 (2025).
Google Scholar 
Wilhelm, R. C., Amsili, J. P., Kurtz, K. S. M., Van Es, H. M. & Buckley, D. H. Correction: Ecological insights into soil health according to the genomic traits and environment-wide associations of bacteria in agricultural soils. ISME Commun 3, (2023).Zheng, M. et al. Microbial taxonomic diversity and functional genes mirror soil ecosystem multifunctionality in nonferrous metal mining areas. J. Hazard. Mater. 492, 138149 (2025).
Google Scholar 
Ferreira, S. L. C. et al. Use of pollution indices and ecological risk in the assessment of contamination from chemical elements in soils and sediments – Practical aspects. Trends Environ. Anal. Chem. 35, e00169 (2022).
Google Scholar 
Li, Y. et al. Application of machine learning in soil heavy metals pollution assessment in the southeastern Tibetan plateau. Sci. Rep. 15, 13579 (2025).
Google Scholar 
Cicchella, D. et al. Spatial distribution of chemical elements in agricultural and grazing land soil of Italy. J. Geochem. Explor. 154. GEMAS, 129–142 (2015).
Google Scholar 
Official Gazette of the Italian Republic. Legislative decree n. 152, 3rd April 2006 (2006).Yang, C. et al. Ecological risk assessment and identification of the distinct microbial groups in heavy metal-polluted river sediments. Environ. Geochem. Health. 45, 1311–1329 (2023).
Google Scholar 
De’ath, G. & Fabricius, K. E. Classification and regression trees: a powerful yet simple technique for ecological data analysis. Ecology 81, 3178–3192 (2000).
Google Scholar 
Sun, C., Wu, P., Wang, G. & Kong, X. Heavy metal pollution decreases the stability of microbial co-occurrence networks in the rhizosphere of native plants. Front Environ. Sci 10, (2022).Zhao, Z. et al. Response of soil microbial community composition and function to prolonged heavy metal exposure. Environ Monit. Assess 197, (2025).Cao, Y. et al. Improved calculations of heavy metal toxicity coefficients for evaluating potential ecological risk in sediments based on seven major Chinese water systems. Toxics 11, 650 (2023).
Google Scholar 
Wang, F., Wang, F., Yang, H., Yu, J. & Ni, R. Ecological risk assessment based on soil adsorption capacity for heavy metals in Taihu basin, China. Environ. Pollut. 316, 120608 (2023).
Google Scholar 
Zerizghi, T., Guo, Q., Tian, L., Wei, R. & Zhao, C. An integrated approach to quantify ecological and human health risks of soil heavy metal contamination around coal mining area. Sci. Total Environ. 814, 152653 (2022).
Google Scholar 
ST-11299-2024-INIT. Proposal for a Directive of the European Parliament and of the Council on Soil Monitoring and Resilience (Soil Monitoring Law) – General Approach. (T_11).Hakanson, L. An ecological risk index for aquatic pollution control.a sedimentological approach. Water Res. 14, 975–1001 (1980).
Google Scholar 
Li, J. et al. Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils. Sci. Rep. 15, 17451 (2025).
Google Scholar 
Zhang, J. et al. Assessment of soil heavy metal pollution in provinces of China based on different soil types: from normalization to soil quality criteria and ecological risk assessment. J. Hazard. Mater. 441, 129891 (2023).
Google Scholar 
Bao, Z. et al. Tardiphaga Alba sp. nov., a heavy-metal-tolerant bacterium isolated from garden soil. Int J. Syst. Evol. Microbiol 74, (2024).Gao, T. et al. The variations of bacterial community structures in tailing soils suffering from heavy metal contaminations. Water Air Soil. Pollut 232, (2021).Salmi, A. & Boulila, F. Heavy metals multi-tolerant Bradyrhizobium isolated from mercury mining region in Algeria. J. Environ. Manage. 289, 112547 (2021).
Google Scholar 
Qi, R. et al. Heavy metal(loid)s shape the soil bacterial community and functional genes of desert grassland in a gold mining area in the semi-arid region. Environ. Res. 214, 113749 (2022).
Google Scholar 
Li, D., Chen, J., Zhang, X., Shi, W. & Li, J. Structural and functional characteristics of soil microbial communities in response to different ecological risk levels of heavy metals. Front Microbiol 13, (2022).Lu, L. et al. Long-term metal pollution shifts microbial functional profiles of nitrification and denitrification in agricultural soils. Sci. Total Environ. 830, 154732 (2022).
Google Scholar 
Pu, C. et al. Intercropping grapevine with solanum nigrum enhances their cadmium tolerance through changing rhizosphere soil microbial diversity. Front Microbiol 16, (2025).Xu, N. et al. Machine learning predicts ecological risks of nanoparticles to soil microbial communities. Environ. Pollut. 307, 119528 (2022).
Google Scholar 
Wijaya, J., Byeon, H., Jung, W., Park, J. & Oh, S. Machine learning modeling using Microbiome data reveal microbial indicator for oil-contaminated groundwater. J. Water Process. Eng. 53, 103610 (2023).
Google Scholar 
Cipullo, S., Snapir, B., Prpich, G., Campo, P. & Coulon, F. Prediction of bioavailability and toxicity of complex chemical mixtures through machine learning models. Chemosphere 215, 388–395 (2019).
Google Scholar 
Song, J. et al. Effects of Cd, Cu, Zn and their combined action on microbial biomass and bacterial community structure. Environ. Pollut. 243, 510–518 (2018).
Google Scholar 
Yang, W. et al. Dissolve organic matter of mature chicken compost contributes to the protection of microorganisms from the stress of heavy metals. J. Environ. Chem. Eng. 12, 113590 (2024).
Google Scholar 
Zeng, K. et al. Bacterial community regulation of soil organic matter molecular structure in heavy metal-rich Mangrove sediments. J. Hazard. Mater. 465, 133086 (2024).
Google Scholar 
Zhang, H. et al. Effects of dissolved organic matter on distribution characteristics of heavy metals and their interactions with microorganisms in soil under long-term exogenous effects. Sci. Total Environ. 947, 174565 (2024).
Google Scholar 
Diba, H. et al. Isolation and characterization of halophilic bacteria with the ability of heavy metal bioremediation and nanoparticle synthesis from Khara salt lake in Iran. Arch. Microbiol. 203, 3893–3903 (2021).
Google Scholar 
Orji, O. U. et al. Halotolerant and metalotolerant bacteria strains with heavy metals biorestoration possibilities isolated from Uburu salt Lake, southeastern. Nigeria Heliyon. 7, e07512 (2021).
Google Scholar 
Official Gazette of the Italian Republic. Ministerial decree 13th September 1999 (1999).Olsen, S. R., Cole, C. V., Watanabe, F. S. & Dean, L. A. Estimation of available phosphorus in soils by extraction with sodium bicarbonate. Circ N939, 1–19 (1954).
Google Scholar 
Walkley, A. A critical examination of a rapid method for determining organic carbon in soils—effect of variations in digestion conditions and of inorganic soil constituents. Soil. Sci. 63, 251 (1947).
Google Scholar 
Faé, G. S., Montes, F., Bazilevskaya, E., Añó, R. M. & Kemanian, A. R. Making soil particle size analysis by laser diffraction compatible with standard soil texture determination methods. Soil. Sci. Soc. Am. J. 83, 1244–1252 (2019).
Google Scholar 
Link, D. D., Walter, P. J. & Kingston, H. M. Development and validation of the new EPA Microwave-Assisted leach method 3051A. Environ. Sci. Technol. 32, 3628–3632 (1998).
Google Scholar 
Grattacaso, M. et al. Health and ecological risk from potentially toxic elements in soils and extra Virgin Olive oils from Valdichiana Senese, Tuscany (Italy). J. Plant. Nutr. Soil. Sci. 188, 241–250 (2025).
Google Scholar 
Hammer, Ø., Harper, D. & Ryan, P. P. A. S. T. Paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4, 1–9 (2001).
Google Scholar 
Download referencesFundingNo funding received for this research.Author informationAuthors and AffiliationsDepartment of Agriculture, Food Natural Resources and Engineering, University of Foggia, Foggia, ItalyAlessandro De Santis, Antonio Bevilacqua, Maria Rosaria Corbo, Barbara Speranza, Matteo Francavilla, Giuseppe Gatta & Milena SinigagliaDepartment of Agricultural and Forestry scieNcEs (DAFNE), University of Tuscia, Viterbo, ItalyFederica CarucciAuthorsAlessandro De SantisView author publicationsSearch author on:PubMed Google ScholarAntonio BevilacquaView author publicationsSearch author on:PubMed Google ScholarMaria Rosaria CorboView author publicationsSearch author on:PubMed Google ScholarBarbara SperanzaView author publicationsSearch author on:PubMed Google ScholarMatteo FrancavillaView author publicationsSearch author on:PubMed Google ScholarGiuseppe GattaView author publicationsSearch author on:PubMed Google ScholarFederica CarucciView author publicationsSearch author on:PubMed Google ScholarMilena SinigagliaView author publicationsSearch author on:PubMed Google ScholarContributionsConceptualization, A.B., M.S., M.R.C. and A.D.S.; methodology, A.B., M.S. and M.R.C.; investigation, A.D.S., B.S., G.G., F.C., and M.F.; data curation, A.B. and A.D.S.; software, A.D.S., and A.B.; writing original draft preparation, A.D.S. and A.B.; writing—review and editing, all authors; supervision, A.B. All authors have read and agreed to the published version of the manuscript.Corresponding authorCorrespondence to
Antonio Bevilacqua.Ethics declarations

Competing interests
The authors declare no competing interests.

Permission
This research is an exploratory action focusing on soil samples collected solely for methodological purposes. Therefore, according to national and local guidelines and regulations a preliminary approval is not required.

Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Reprints and permissionsAbout this articleCite this articleDe Santis, A., Bevilacqua, A., Corbo, M.R. et al. A statistical approach to model soil microbiota versus heavy metals: a case study on soil samples from Foggia, Southern Italy.
Sci Rep (2025). https://doi.org/10.1038/s41598-025-32485-xDownload citationReceived: 03 September 2025Accepted: 10 December 2025Published: 23 December 2025DOI: https://doi.org/10.1038/s41598-025-32485-xShare this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
KeywordsClassification and regression treesShotgun metagenomicsStandardized ecological risk indexCorrelationEcological risk assessmentMicrobial communities More

AbstractDenitrification, a major source of gaseous nitrogen emissions from agricultural soils, is influenced by management. Practices promoting belowground diversity are suggested to support sustainable agriculture, but how they modulate nitrogen losses via denitrification remains inconclusive. Here we sampled 106 cereal fields spanning a 3000 km North-South gradient across Europe and compiled 56 associated climatic, soil, microbial and management variables. We show that increased denitrification potential was associated with higher proportion of time with crop cover over the last ten years and was best predicted by microbial biomass and microbial functional guilds involved in nitrogen cycling, in particular denitrification. We also demonstrate that several diversification practices affect the variation in denitrification potential predictors, suggesting a trade-off between agricultural diversification and nitrogen losses via denitrification. However, increased crop diversity in rotations improved yield-scaled denitrification, highlighting the potential of this practice to minimize nitrogen losses while contributing to sustainable food production.

Similar content being viewed by others

Increasing crop rotational diversity can enhance cereal yields

Article
Open access
23 March 2023

Crop cover is more important than rotational diversity for soil multifunctionality and cereal yields in European cropping systems

Article

13 January 2021

Aggregation of activity data on crop management can induce large uncertainties in estimates of regional nitrogen budgets

Article
Open access
29 April 2024

Data availability

Data and OTU tables used in this study as well as source data for the figures are available at Zenodo (https://doi.org/10.5281/zenodo.14760398).
Code availability

The R code used in this study is available at Zenodo (https://doi.org/10.5281/zenodo.14760398).
ReferencesSutton, M. A. et al. Too much of a good thing. Nature 472, 159–161 (2011).
Google Scholar 
Ludemann, C. I. et al. A global FAOSTAT reference database of cropland nutrient budgets and nutrient use efficiency (1961–2020): nitrogen, phosphorus and potassium. Earth Syst. Sci. Data 16, 525–541 (2024).
Google Scholar 
Richardson, K. et al. Earth beyond six of nine planetary boundaries. Sci. Adv. 9, eadh2458 (2023).
Google Scholar 
Scheer, C. et al. Estimating global terrestrial denitrification from measured N2O:(N2O + N2) product ratios. Curr. Opin. Environ. Sustain. 47, 72–80 (2020).
Google Scholar 
Tian, H. et al. A comprehensive quantification of global nitrous oxide sources and sinks. Nature 586, 248–256 (2020).
Google Scholar 
Bowles, T. M. et al. Addressing agricultural nitrogen losses in a changing climate. Nat. Sustain. 1, 399–408 (2018).
Google Scholar 
Sinha, E. et al. Eutrophication will increase during the 21st century as a result of precipitation changes. Science 357, 405–408 (2017).
Google Scholar 
Springmann, M. et al. Options for keeping the food system within environmental limits. Nature 562, 519–525 (2018).
Google Scholar 
You, L. et al. Optimized agricultural management reduces global cropland nitrogen losses to air and water. Nat. Food 5, 995–1004 (2024).
Google Scholar 
Garland, G. et al. Crop cover is more important than rotational diversity for soil multifunctionality and cereal yields in European cropping systems. Nat. Food 2, 28–37 (2021).
Google Scholar 
Kim, N. et al. Do cover crops benefit soil microbiome? A meta-analysis of current research. Soil Biol. Biochem. 142, 107701 (2020).
Google Scholar 
Tamburini, G. et al. Agricultural diversification promotes multiple ecosystem services without compromising yield. Sci. Adv. 6, eaba1715 (2020).
Google Scholar 
Smith, M. E. et al. Increasing crop rotational diversity can enhance cereal yields. Commun. Earth Environ. 4, 89 (2023).
Google Scholar 
Kremen, C. et al. Ecosystem services in biologically diversified versus conventional farming systems: benefits, externalities, and trade-offs. Ecol. Soc. 17, 40 (2012).
Google Scholar 
Domeignoz-Horta, L. A. et al. Peaks of in situ N2O emissions are influenced by N2O-producing and reducing microbial communities across arable soils. Glob. Change Biol. 24, 360–370 (2017).
Google Scholar 
Wang, J. et al. No-till increases soil denitrification via its positive effects on the activity and abundance of the denitrifying community. Soil Biol. Biochem. 142, 107706 (2020).
Google Scholar 
Pan, B. et al. A global synthesis of soil denitrification: driving factors and mitigation strategies. Agric. Ecosyst. Environ. 327, 107850 (2022).
Google Scholar 
Philippot, L. et al. Ecology of denitrifying prokaryotes in agricultural soil. Adv. Agron. 96, 249–305 (2007).
Google Scholar 
Fageria, N. K. et al. Role of cover crops in improving soil and row crop productivity. Commun. Soil Sci. Plant Anal. 36, 2733–2757 (2005).
Google Scholar 
Duan, Y. F. et al. Catch crop residues stimulate N2O emissions during spring, without affecting the genetic potential for nitrite and N2O reduction. Front. Microbiol. 9, 2629 (2018).
Google Scholar 
Olesen, J. E. et al. Challenges of accounting nitrous oxide emissions from agricultural crop residues. Glob. Change Biol. 29, 6846–6855 (2023).
Google Scholar 
Kravchenko, A. N. et al. Hotspots of soil N2O emission enhanced through water absorption by plant residue. Nat. Geosci. 10, 496–500 (2017).
Google Scholar 
Højberg, O. et al. Denitrification in soil aggregates analyzed with microsensors for nitrous oxide and oxygen. Soil Sci. Soc. Am. J. 58, 1691–1698 (1994).
Google Scholar 
MacLaren, C. et al. Long-term evidence for ecological intensification as a pathway to sustainable agriculture. Nat. Sustain. 5, 770–779 (2022).
Google Scholar 
Costa, A. et al. Crop rotational diversity can mitigate climate-induced grain yield losses. Glob. Change Biol. 30, e17298 (2024).
Google Scholar 
Renard, D. et al. National food production stabilized by crop diversity. Nature 571, 257–260 (2019).
Google Scholar 
Fageria, N. K. et al. Enhancing nitrogen use efficiency in crop plants. Adv. Agron. 88, 97–185 (2005).
Google Scholar 
Yi, B. et al. Diversified cropping systems with limited carbon accrual but increased nitrogen supply. Nat. Sustain. 8, 152–161 (2025).
Google Scholar 
Gelfand, I. et al. Long-term nitrous oxide fluxes in annual and perennial agricultural and unmanaged ecosystems in the upper Midwest USA. Glob. Change Biol. 22, 3594–3607 (2016).
Google Scholar 
Putz, M. et al. Relative abundance of denitrifying and DNRA bacteria and their activity determine nitrogen retention or loss in agricultural soil. Soil Biol. Biochem. 123, 97–104 (2018).
Google Scholar 
Thompson, K. A. et al. Soil microbial communities as potential regulators of in situ N2O fluxes in annual and perennial cropping systems. Soil Biol. Biochem. 103, 262–273 (2016).
Google Scholar 
Dai, Z. et al. Long-term nitrogen fertilization decreases bacterial diversity and favors the growth of Actinobacteria and Proteobacteria in agro-ecosystems across the globe. Glob. Change Biol. 24, 3452–3461 (2018).
Google Scholar 
Jones, C. M. et al. Reactive nitrogen restructures and weakens microbial controls of soil N2O emissions. Commun. Biol. 5, 273 (2022).
Google Scholar 
Isobe, K. et al. Phylogenetic conservation of bacterial responses to soil nitrogen addition across continents. Nat. Commun. 10, 2499 (2019).
Google Scholar 
Thompson, R. L. et al. Acceleration of global N2O emissions seen from two decades of atmospheric inversion. Nat. Clim. Change 9, 993–998 (2019).
Google Scholar 
Charles, A. et al. Global nitrous oxide emission factors from agricultural soils after addition of organic amendments: a meta-analysis. Agric. Ecosyst. Environ. 236, 88–98 (2017).
Google Scholar 
Graf, D. R. H. et al. Intergenomic comparisons highlight modularity of the denitrification pathway and underpin the importance of community structure for N2O emissions. PLoS ONE 9, e114118 (2014).
Google Scholar 
Saghaï, A. et al. Diversity and ecology of NrfA-dependent ammonifying microorganisms. Trends Microbiol. 32, 602–613 (2024).
Google Scholar 
Saghaï, A. et al. Phyloecology of nitrate ammonifiers and their importance relative to denitrifiers in global terrestrial biomes. Nat. Commun. 14, 8249 (2023).
Google Scholar 
Morrissey, E. M. et al. Phylogenetic organization of bacterial activity. ISME J. 10, 2336–2340 (2016).
Google Scholar 
Jones, C. M. et al. Phylogenetic analysis of nitrite, nitric oxide, and nitrous oxide respiratory enzymes reveal a complex evolutionary history for denitrification. Mol. Biol. Evol. 25, 1955–1966 (2008).
Google Scholar 
Shoun, H. et al. Fungal denitrification and nitric oxide reductase cytochrome P450nor. Philos. Trans. R. Soc. B Biol. Sci. 367, 1186–1194 (2012).
Google Scholar 
Bösch, Y. et al. Distribution and environmental drivers of fungal denitrifiers in global soils. Microbiol. Spectr. 11, e00061–23 (2023).
Google Scholar 
FAL et al. Referenzmethoden Der Eidg. Landwirtschaftlichen Forschungsanstalten. 1. Bodenuntersuchung Zur Düngeberatung (Zürich- Reckenholz, 1996).Gregorich, E. G. et al. Soil Sampling and Methods of Analysis (CRC Press, 2007).Hijmans, R. J. Raster: Geographic Data Analysis and Modeling https://doi.org/10.32614/CRAN.package.raster (2024).R Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing https://www.R-project.org/ (Vienna, Austria 2024).Jin, Y. et al. V.PhyloMaker: an R package that can generate very large phylogenies for vascular plants. Ecography 42, 1353–1359 (2019).
Google Scholar 
Oksanen, J. et al. Vegan: community ecology package. https://doi.org/10.32614/CRAN.package.vegan (2018).Magurran, A. et al. Explaining the excess of rare species in natural species abundance distributions. Nature 422, 714–716 (2003).
Google Scholar 
Hubbell, S. The Unified Neutral Theory of Biodiversity and Biogeography (Princeton University Press, Princeton, NJ, USA, 2001).Krebs, C. J. Ecological Methodology (Addison-Wesley Educational Publishers, Inc: New York, NY, USA, 1999).Pell, M. et al. Potential denitrification activity assay in soil – With or without chloramphenicol? Soil Biol. Biochem. 28, 393–398 (1996).
Google Scholar 
Philippot, L. et al. Importance of denitrifiers lacking the genes encoding the nitrous oxide reductase for N2O emissions from soil. Glob. Change Biol. 17, 1497–1504 (2011).
Google Scholar 
van Groenigen, J. W. et al. Towards an agronomic assessment of N2O emissions: a case study for arable crops. Eur. J. Soil Sci. 61, 903–913 (2010).
Google Scholar 
Wood, S. N. Generalized Additive Models: An Introduction with R (Chapman and Hall/CRC, 2017).Marra, G. et al. Practical variable selection for generalized additive models. Comput. Stat. Data Anal. 55, 2372–2387 (2011).
Google Scholar 
Wood, S. N. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. R. Stat. Soc. B. 73, 3–36 (2011).
Google Scholar 
Breheny, P. et al. Visualization of regression models using visreg. R. J. 9, 56–71 (2017).
Google Scholar 
Genuer, R. et al. VSURF: an R package for variable selection using random forests. R. J. 7, 19–33 (2015).
Google Scholar 
Liaw, A. et al. Classification and regression by randomForest. R. N. 2, 18–22 (2002).
Google Scholar 
Molnar, C. et al. iml: an R package for interpretable machine learning. J. Open Source Softw. 3, 786 (2018).
Google Scholar 
Apley, D. W. et al. Visualizing the effects of predictor variables in black box supervised learning models. J. R. Stat. Soc. Ser. B Stat. Methodol. 82, 1059–1086 (2020).
Google Scholar 
Download referencesAcknowledgementsThe Digging Deeper project was funded through the 2015–2016 BiodivERsA call, with national funding from the Swiss National Science Foundation (grant 31BD30-172466 to M.G.A.v.d.H), the Deutsche Forschungsgemeinschaft (grant 317895346 to M.C.R.), the Swedish Research Council Formas (grant 2016-0194 to S.H. and 2018-02321 to R.B.), the Spanish Ministerio de Economía y Competitividad (grant PCIN-2016-028 to F.T.M.) and the Agence Nationale de la Recherche (grant ANR-16-EBI3-0004-01 to L.P.). We thank Claudia von Brömssen (Swedish University of Agricultural Sciences) for advice on the generalized additive models.FundingOpen access funding provided by Swedish University of Agricultural Sciences.Author informationAuthor notesThese authors contributed equally: Aurélien Saghaï, Monique E. Smith.Authors and AffiliationsDepartment of Forest Mycology and Plant Pathology, Swedish University of Agricultural Sciences, Uppsala, SwedenAurélien Saghaï & Sara HallinDepartment of Ecology, Swedish University of Agricultural Sciences, Uppsala, SwedenMonique E. Smith, Giulia Vico & Riccardo BommarcoAgroscope, Plant-Soil Interactions Group, Zurich, SwitzerlandSamiran Banerjee, Anna Edlinger, Gina Garland, Marcel G. A. van der Heijden & Chantal HerzogDepartment of Microbiological Sciences, North Dakota State University, Fargo, ND, USASamiran BanerjeeDepartment of Plant and Microbial Biology, University of Zurich, Zurich, SwitzerlandAnna Edlinger, Pablo García-Palacios, Marcel G. A. van der Heijden & Chantal HerzogWageningen Environmental Research, Wageningen University & Research, Wageningen, The NetherlandsAnna EdlingerInstituto de Ciencias Agrarias, Consejo Superior de Investigaciones Científicas, Madrid, SpainPablo García-PalaciosSoil Quality and Use Group, Agroscope, Zurich, SwitzerlandGina GarlandDepartment of Environmental System Sciences, Soil Resources Group, ETH Zurich, Zurich, SwitzerlandGina GarlandEnvironmental Sciences and Engineering, Biological and Environmental Science and Engineering Division, King Abdullah University of Science and Technology, Thuwal, Kingdom of Saudi ArabiaFernando T. MaestreDepartamento de Biología y Geología, Física y Química Inorgánica, Escuela Superior de Ciencias Experimentales y Tecnología, Universidad Rey Juan Carlos, Móstoles, SpainDavid S. PescadorInstitut Agro Dijon, Agroecologie, INRAE, Université de Bourgogne, Dijon, FranceLaurent Philippot & Sana RomdhaneInstitute of Biology, Freie Universität Berlin, Berlin, GermanyMatthias C. RilligBerlin-Brandenburg Institute of Advanced Biodiversity Research, Berlin, GermanyMatthias C. RilligAuthorsAurélien SaghaïView author publicationsSearch author on:PubMed Google ScholarMonique E. SmithView author publicationsSearch author on:PubMed Google ScholarGiulia VicoView author publicationsSearch author on:PubMed Google ScholarSamiran BanerjeeView author publicationsSearch author on:PubMed Google ScholarAnna EdlingerView author publicationsSearch author on:PubMed Google ScholarPablo García-PalaciosView author publicationsSearch author on:PubMed Google ScholarGina GarlandView author publicationsSearch author on:PubMed Google ScholarMarcel G. A. van der HeijdenView author publicationsSearch author on:PubMed Google ScholarChantal HerzogView author publicationsSearch author on:PubMed Google ScholarFernando T. MaestreView author publicationsSearch author on:PubMed Google ScholarDavid S. PescadorView author publicationsSearch author on:PubMed Google ScholarLaurent PhilippotView author publicationsSearch author on:PubMed Google ScholarMatthias C. RilligView author publicationsSearch author on:PubMed Google ScholarSana RomdhaneView author publicationsSearch author on:PubMed Google ScholarRiccardo BommarcoView author publicationsSearch author on:PubMed Google ScholarSara HallinView author publicationsSearch author on:PubMed Google ScholarContributionsS.H., M.G.A.v.d.H., F.T.M., L.P., and M.C.R. initiated the study, planned the field work, and contributed materials. A.S., S.B., F.D., A.E., P.G-P., G.G., C.H., D.S.P., and S.R. contributed to data collection. A.S. and M.E.S. performed the analyses, and A.S., M.E.S., G.V., R.B., and S.H. interpreted the results. A.S., M.E.S., and S.H. drafted the manuscript. All authors commented on and approved the final manuscript.Corresponding authorCorrespondence to
Sara Hallin.Ethics declarations

Competing interests
The authors declare no competing interests.

Peer review

Peer review information
Communications Earth and Environment thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Jinfeng Chang and Mengjie Wang. [A peer review file is available].

Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationTransparent Peer Review fileSupplementary_infonr-reporting-summaryRights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissionsAbout this articleCite this articleSaghaï, A., Smith, M.E., Vico, G. et al. Diverse crop rotations offset yield-scaled nitrogen losses via denitrification.
Commun Earth Environ (2025). https://doi.org/10.1038/s43247-025-03116-0Download citationReceived: 26 July 2025Accepted: 08 December 2025Published: 23 December 2025DOI: https://doi.org/10.1038/s43247-025-03116-0Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative More

AbstractIn response to global climate change and China’s “dual carbon” goals, forest carbon sinks, as a key nature-based solution, have gained importance in balancing human-induced carbon emissions and ecological restoration. This study examines the effectiveness of forest carbon sinks across 31 Chinese provinces from 2003 to 2018, using the forest stock expansion method to quantify the validity of carbon sinks. We explore the spatiotemporal evolution and regional disparities of forest carbon sink validity and identify the influence of research and development intensity, industrial structure upgrading, urbanisation level, government intervention degree, and economic development level factors. A spatial Durbin model is employed to assess both direct and indirect effects of natural and policy factors on the carbon sink’s effectiveness in both local and neighbouring provinces. Our findings reveal that forest carbon sink effectiveness follows a pattern of “higher in the west, faster in the east, and catching up in the central region”. The results indicate that increased research and development investment and optimised industrial structure positively influence carbon sink growth, whereas excessive government intervention hampers development. Urbanisation and economic development were found to have no significant direct effect. The spatial analysis shows that research and development intensity and industrial optimisation yield positive spillover effects on neighbouring provinces’ carbon sink growth, whereas government intervention and urbanisation yield negative, non-significant spillover effects. These findings suggest the need for strengthened regional innovation policies, improved forestry governance, and optimised forestry services to support the high-quality development of the forestry sector.

Similar content being viewed by others

Forest vegetation increased across China’s carbon offset projects and positively impacted neighboring areas

Article
Open access
18 December 2024

Assessing the potential for carbon storage enhancement in forests of Xinjiang Uygur autonomous region, China

Article
Open access
14 August 2025

Evidence and attribution of the enhanced land carbon sink

Article

25 July 2023

Data availability

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
ReferencesVitousek, P. M. Beyond global warming: Ecology and global change. Ecology 75, 1861–1876 (1994).
Google Scholar 
Feng, B. Forest Carbon sink resource asset evaluation with case study of fujian province in China. Math. Probl. Eng. 2022, 7391473 (2022).
Google Scholar 
Chang, W.-Y., Li, Z., Lu, K. & Chang, S. J. Optimal eco-compensation for forest-based carbon sequestration programs: A case study of larch carbon sink plantations in Gansu, Northwest China. Forests 13, 268 (2022).
Google Scholar 
Breuer, G. Can forest policy contribute to solving the CO2 problem?. Environ. Int. 2, 449–451 (1979).
Google Scholar 
Wong, C. S. Atmospheric Input of carbon dioxide from burning wood. Science 200, 197–200 (1978).
Google Scholar 
Woodwell, G. M. et al. The biota and the world carbon budget: The terrestrial biomass appears to be a net source of carbon dioxide for the atmosphere. Science 199, 141–146 (1978).
Google Scholar 
Lin, J. Y. Rural reforms and agricultural growth in China. Am. Econ. Rev. 34–51 (1992).Xu, M. Can the ecological protection red line policy promote food security?—based on the empirical analysis of land protection in China. Front. Environ. Sci. 13, 1654217 (2025).
Google Scholar 
Shi, L. et al. Achieving sustainable green agriculture: analyze the enabling role of data elements in agricultural carbon reduction. Front. Earth Sci. 13, 1618999 (2025).
Google Scholar 
Yao, Y., Piao, S. & Wang, T. Future biomass carbon sequestration capacity of Chinese forests. Sci. Bull. 63, 1108–1117 (2018).
Google Scholar 
Xiao, J. et al. Assessing net ecosystem carbon exchange of U.S. terrestrial ecosystems by integrating eddy covariance flux measurements and satellite observations. Agric. Forest Meteorol. 151, 60–69 (2011).
Google Scholar 
Chen, K. et al. Forest Carbon Sink Evaluation – An Important Contribution for Carbon Neutrality. IOP Conf. Ser. Earth Environ. Sci. 811, 012009 (2021).
Google Scholar 
Wang, J., Shi, K. & Hu, M. Measurement of forest carbon sink efficiency and its influencing factors empirical evidence from China. Forests 13, 1909 (2022).
Google Scholar 
Bonilla-Bedoya, S., Mora, A., Vaca, A., Estrella, A. & Herrera, M. Á. Modelling the relationship between urban expansion processes and urban forest characteristics: An application to the Metropolitan District of Quito. Comput. Environ. Urban Syst. 79, 101420 (2020).
Google Scholar 
Liu, S. et al. Forest fragmentation and landscape connectivity change associated with road network extension and city expansion: A case study in the Lancang River Valley. Ecol. Ind. 36, 160–168 (2014).
Google Scholar 
Kurz, W. A., Stinson, G., Rampley, G. J., Dymond, C. C. & Neilson, E. T. Risk of natural disturbances makes future contribution of Canada’s forests to the global carbon cycle highly uncertain. Proc. Natl. Acad. Sci. USA 105, 1551–1555 (2008).
Google Scholar 
Seidl, R. et al. Invasive alien pests threaten the carbon stored in Europe’s forests. Nat. Commun. 9, 1626 (2018).
Google Scholar 
Cho, S.-H., Lee, J., Roberts, R., Yu, E. T. & Armsworth, P. R. Impact of market conditions on the effectiveness of payments for forest-based carbon sequestration. Forest Policy Econ. 92, 33–42 (2018).
Google Scholar 
Sharma, B. P., Cho, S.-H. & Yu, T. E. Designing cost-efficient payments for forest-based carbon sequestration: An auction-based modeling approach. Forest Policy Econ. 104, 182–194 (2019).
Google Scholar 
Xu, M. et al. The influence mechanism of environmental regulations on food security: the mediating effect of technological innovation. Environ. Sci. Eur. 37, 83 (2025).
Google Scholar 
Xu, M. & Lu, Z. Achieving green agricultural development: Analyzing the impact of agricultural non-point source pollution on food security and the regulation effect of environmental regulation. PLoS One 20, 0324899 (2025).
Google Scholar 
He, D., Deng, X., Gao, Y. & Wang, X. How does digitalization affect carbon emissions in animal husbandry? A new evidence from China. Resour. Conserv. Recycl. 214, 108040 (2025).
Google Scholar 
Lu, Z., Xu, M., Shi, L. & Lei, T. The criticality of environmental sustainability in agriculture: The decarbonization role of green finance in China. J. Environ. Manag. 394, 127470 (2025).
Google Scholar 
Wang, Y. et al. Persistent and enhanced carbon sequestration capacity of alpine grasslands on Earth’s third pole. Sci. Adv. 9, 6875 (2023).
Google Scholar 
Xu, M. et al. Achieving agricultural sustainability: analyzing the impact of digital financial inclusion on agricultural green total factor productivity. Front. Sustain. Food Syst. 8, 1515207 (2025).
Google Scholar 
Lu, Z., Gou, D., Wu, Q. & Feng, H. Does the rural digital economy promote shared prosperity among farmers? Evidence from China. Front. Sustain. Food Syst. 9, 1649753 (2025).
Google Scholar 
Tan, Y. et al. Environmental regulation, spatial spillovers and high quality economic development – The moderating effect of green technology innovation. Pol. J. Environ. Stud. 34, 6763–6779 (2025).
Google Scholar 
Lu, Z., Gou, D., Yang, L., Wu, Z. & Feng, H. The impact of environmental tax reform on industrial green development: evidence from China. Front. Environ. Sci. 13, 1593549 (2025).
Google Scholar 
Tan, Z., Chen, K. & Liu, P. Possibilities and challenges of China’s forestry biomass resource utilization. Renew. Sustain. Energy Rev. 41, 368–378 (2015).
Google Scholar 
Zhang, X. et al. A large but transient carbon sink from urbanization and rural depopulation in China. Nat. Sustain. 5, 321–328 (2022).
Google Scholar 
Kassinis, G. & Vafeas, N. Stakeholder pressures and environmental performance. AMJ 49, 145–159 (2006).
Google Scholar 
Sun, R., He, D., Yan, J. & Tao, L. Mechanism analysis of applying blockchain technology to forestry carbon sink projects based on the differential game model. Sustainability 13, 11697 (2021).
Google Scholar 
Khanna, N., Fridley, D. & Hong, L. China’s pilot low-carbon city initiative: A comparative assessment of national goals and local plans. Sustain. Cities Soc. 12, 110–121 (2014).
Google Scholar 
Van Den Bergh, J. C. J. M. & Grazi, F. Ecological footprint policy? Land use as an environmental indicator. J. Ind. Ecol. 18, 10–19 (2014).
Google Scholar 
Zhai, H. et al. Understanding spatio-temporal patterns of land use/land cover change under urbanization in Wuhan, China, 2000–2019. Remote Sens. 13, 3331 (2021).
Google Scholar 
Que, Z., Zhang, Y. & Cheng, X. Judicial innovation in enhancing forest carbon sinks: Evidence from China. Forests 16, 369 (2025).
Google Scholar 
Beckerman, W. Economic growth and the environment: Whose growth? whose environment?. World Dev. 20, 481–496 (1992).
Google Scholar 
Li, M. et al. Assessment of forest geospatial patterns over the three giant forest areas of China. J. Forestry Res. 19, 25–31 (2008).
Google Scholar 
Li, Y., Wang, X., Wang, Y., Peng, J. & Chen, S. How does market-oriented environmental regulation affect carbon emission performance? A quasinatural experiment based on the pilot policy of energy-use rights trading. Int. J. Energy Res. 2024, 6771192 (2024).
Google Scholar 
Download referencesFunding1. This research was funded by the Sichuan Police Law Enforcement Research Centre (grant number: JCZFQN202402). 3. This research was funded by the Research Centre for Social Governance Innovation (grant number: SHZLQN2404).Author informationAuthors and AffiliationsSchool of National Security, Southwest University of Political Science and Law, Chongqing, ChinaLidong Shi, Ming Xu & Jiahui ZhaoCollege of Business Administration, Chongqing Technology and Business University, Chongqing, ChinaYuntao TanLegal Affairs Department, Chongqing People’s Hospital, Chongqing, ChinaYucen WuAuthorsLidong ShiView author publicationsSearch author on:PubMed Google ScholarMing XuView author publicationsSearch author on:PubMed Google ScholarYuntao TanView author publicationsSearch author on:PubMed Google ScholarYucen WuView author publicationsSearch author on:PubMed Google ScholarJiahui ZhaoView author publicationsSearch author on:PubMed Google ScholarContributionsConceptualisation, L.S. and J.Z.; methodology, M.X.; software, Y.T.; validation, M.X. and Y.W.; formal analysis, L.S.; investigation, J.Z.; resources, Y.T.; data curation, L.S. and Y.T.; writing—original draft preparation, Y.W., M.X., J.Z.; writing—review and editing, Y.T.; visualisation, L.S.; supervision, J.Z.; project administration, J.Z.; funding acquisition, L.S. All authors reviewed the manuscript.Corresponding authorCorrespondence to
Jiahui Zhao.Ethics declarations

Competing interests
The authors declare no competing interests.

Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Reprints and permissionsAbout this articleCite this articleShi, L., Xu, M., Tan, Y. et al. Responding to climate change: assessing the current situation and influencing factors of forest carbon sinks in China.
Sci Rep (2025). https://doi.org/10.1038/s41598-025-33255-5Download citationReceived: 05 September 2025Accepted: 17 December 2025Published: 23 December 2025DOI: https://doi.org/10.1038/s41598-025-33255-5Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
KeywordsForest carbon sink validitySpatiotemporal heterogeneitySpatial Durbin modelSpatial spillover effect More