Philippa Kaur

The overall magnitude of changes in SOC, TN, and C:N in response to chemical nitrogen fertilizers reductionThe results showed that chemical nitrogen fertilizers reduction significantly decreased SOC and TN by 2.76% and 4.19% respectively, while increased C:N by 6.11% across all database (Fig. 1). SOC mainly derives from crop residues and secretions which closely related to crops growths, and crops growths were affected by fertilization, especially nitrogen fertilization20,21. The reduction of chemical nitrogen fertilizer led to poor crop growth, which reduced the amount of crop residues return, and then decreased SOC. Similarly, TN from crops was reduced due to poor crop growth. In addition, the reduction of chemical nitrogen fertilizers directly reduced the input of soil nitrogen. The increase of C:N was the result of the decrease of TN being greater than that of SOC. The responses of C:N to chemical nitrogen fertilizers reduction enhanced the comprehension of the couple relationship between SOC and TN, which was beneficial to the evolution of the C-N coupling models. Moreover, the accuracy of C-N coupling models depends on the precise quantification of the responses of SOC and TN to nitrogen fertilization. And our results accurately quantified the difference responses of SOC and TN to different nitrogen fertilization regimes, thus optimizing the C-N coupling models.Figure 1The weighted response ratio (RR++) for the responses to chemical nitrogen fertilizers of soil organic carbon (SOC, a), total nitrogen (TN, b), and their ratios (C:N, c). Bars denote the overall mean response ratio RR++ and 95% confidence intervals (CI). The star (*) indicates significance when the 95% CI that do not go across the zero line. The vertical lines are drawn at lnRR = 0. The value represents independent sample size.Full size imageResponses of SOC, TN and C:N to chemical nitrogen fertilizers reduction magnitudeWhen grouped by chemical nitrogen fertilizers reduction magnitude, SOC showed a significant increase by 6.9% in medium magnitude, while SOC was significantly decreased by 3.10% and 7.26% in high and total magnitude respectively (Fig. 1a). Liu and Greaver22 also stated the reduction of medium nitrogen fertilizer increased the average microbial biomass from 15 to 20%, thereby increasing the SOC content. Previous studies had reported that there were strong positive correlations between soil organic matter and soil microbial biomass in both the agricultural ecosystem and natural ecosystem23,24. Numerous researchers have demonstrated the significance of nitrogen availability in soil for the plant biomass across most ecosystems25,26. Moreover, nitrogen deficient would inhibit the activity of extracellular enzymes and root activities27. Generally, soil degradation caused by continuous rising chemical nitrogen fertilizers application may inhibit the growth of crops and ultimately reduce the SOC28.TN showed no significant difference in low and medium chemical nitrogen fertilizers reduction magnitude (p  > 0.05), while TN in high magnitude and total chemical nitrogen fertilizers reduction magnitude exhibited a decrease with 3.10% and 9.37% respectively (Fig. 1b). Numerous studies described that the amount of nitrogen fertilizers used in China was higher than the demand of N for crop, which caused serious N leaching and runoff29,30. Chemical nitrogen fertilizers in low and medium magnitude would not decrease the TN of soil by reducing N leaching and runoff. However, the residual nitrogen in soil cannot meet the requirement for the sustainable growth of plant with litter or without exogenous nitrogen supplement, which resulted in the decrease of TN in high and total chemical nitrogen fertilizers magnitude. Consequently, optimal nitrogen fertilizers application rates will take into account crops yield and environment friendliness.Additionally, C:N had a significant increase with ranging from 1.82% to 8.98% under the four chemical nitrogen fertilizers reduction magnitude (Fig. 1c), suggesting the relative increase of SOC compared to TN. Previous studies have revealed that C:N had significantly influence on the soil bacterial community structures31. And there were also considerable studies indicated that chemical nitrogen fertilizers have impact on the soil bacterial communities32,33. We may speculate that the change of C:N would bring about the variations of soil bacteria communities under the chemical nitrogen fertilizers regimes.Responses of SOC, TN, and C:N to chemical nitrogen fertilizers reduction durationNegative response of SOC to short-term chemical nitrogen fertilizers reduction was observed in our study, which was consistent with the study of Gong, et al.34 that chemical nitrogen fertilizers reduction decreased SOC by reducing crop-derived carbon by one year. However, SOC was significantly increased by 1.06% and 4.65% at mid-term and long-term chemical nitrogen fertilizers reduction respectively, which was similar with the findings of Ning, et al.11 that SOC was significantly increased under more than 5 years of chemical nitrogen fertilizers reduction treatment. TN was significantly decreased by 1.96% at short-term chemical nitrogen fertilizers reduction duration, while the results converted at mid-term chemical nitrogen fertilizers reduction duration. The effect of long-term chemical nitrogen fertilizers reduction on TN was not significant (p  > 0.05). The divergent response of TN to different chemical nitrogen fertilizers duration was mainly caused by the various treatments. In terms of C:N, a greater positive response was observed at short-term chemical nitrogen fertilizers duration (9.06%) than mid-term and long-term duration (1.99%). Moreover, with the prolongation of the chemical reduction time of nitrogen, the response ratio tends to zero, suggesting that the effect of chemical fertilizers gradually decrease. This may be ascribed to the buffer capacity of soil to resist the changes from external environment, including nutrients, pollutants, and redox substances35.Responses of SOC, TN, and C:N to different chemical nitrogen fertilizers reduction patternsUnder the pattern of chemical nitrogen fertilizers reduction without organic fertilizers supplement, SOC and TN significantly decreased by 3.83% and 11.46% respectively, however, chemical nitrogen fertilizers reduction with organic fertilizers supplement significantly increased SOC and TN by 4.92% and 8.33% respectively. Moreover, C:N significantly increased under the two chemical nitrogen fertilizers patterns (p  0.05), but there was a negative effect on SOC in high and total magnitude (p  0.05). The no significant decrease at mid-term duration might result from the limited information reported in original studies of this meta-analysis36. TN showed no significant response to chemical nitrogen fertilizers without organic fertilizers supplement in the low and medium magnitude (p  > 0.05). However, TN was significantly decreased by 8.62% and 16.7% respectively in the high and total magnitude. When regarding to chemical nitrogen fertilizers reduction duration, TN was significantly reduced at all of the categories, ranging from 3.13% to 13.4% (Fig. 2c). In the pattern of chemical nitrogen fertilizers reduction with organic fertilizers supplement, chemical nitrogen fertilizers reduction at medium, high, and total magnitudes significantly increased SOC by 13.85%, 13.03%, and 5.46%respectively, however, the response of SOC in the low chemical nitrogen fertilizers magnitude was not significant. Chemical nitrogen fertilizers reduction duration significantly increased SOC by 7.01%, 1.71%, and 22.02% in the short-term, mid-term, and long-term respectively. Comparatively, TN showed a significantly increase in most chemical nitrogen fertilizers categories expect for the long-term chemical nitrogen fertilizers duration, with an increasing from 4.90% to 14.69% (Fig. 2d).Figure 2The weighted response ratio (RR++) for the responses to chemical nitrogen fertilizers of soil organic carbon (SOC, a), total nitrogen (TN, b), and their ratios (C:N, c) under the two patterns (with organic fertilizers ; without organic fertilizers). Bars denote the overall mean response ratio RR++ and 95% confidence intervals (CI). The star (*) indicates significance when the 95% CI that do not go across the zero line. The vertical lines are drawn at lnRR = 0. The values represent independent sample size.Full size imageOrganic fertilizers were mainly derived from animal manure or crops straws, which contained large amount of organic matter and nitrogen elements37,38. The application of organic fertilizers increased the input of SOC and TN directly. Moreover, organic fertilizer could promote the growth of crops by releasing phenols, vitamins, enzymes, auxins and other substances during the decomposition process, thus the SOC derived from crops would be increased37,39. In addition, organic fertilizers provide various nutrients for microbial reproduction, which increase the microbial population and organic carbon and total nitrogen content37. More importantly, the application of organic fertilizers could improve organic carbon sequestration and maintain its stability in aggregates, thereby reducing losses of SOC and TN40.C:N showed an increase under all of the chemical nitrogen fertilizers reduction with organic fertilizer supplement. The positive response of C:N to organic fertilizer supplement may be related to the higher C:N of organic fertilizer than soil. The average values of C:N of the commonly used organic fertilizers including animal manure, crop straws and biochar were 14, 60 and 30 respectively, while the C:N of soil was lower than 10 in average according to extensive literature researches41. Therefore, organic fertilizers would be a favorable alternative of chemical fertilizers for the sustainable development of agriculture.The correlation between the response of SOC, TN, and C:N and environmental variablesThe analysis of linear regression was conducted to analyze the environmental variables including mean annual temperature (MAT), mean annual precipitation (MAP), accumulated temperature above 10 °C (MATA), which may exert influence on SOC, TN and C:N. No significant correlation among the lnRR of SOC, TN, C:N and environmental variables were observed among the whole database (p  > 0.05; Fig. S1). Rule out the interference of organic fertilizers supplement, we analyzed the relationship between lnRR of SOC, TN, C:N and environmental variables as the Figures showed in Figs. 3 and 4 respectively. Under chemical nitrogen fertilizers without organic fertilizers supplement, there was a significant negative correlation between lnRR of SOC and MAT (p  More

Irigoien, X. et al. Nat. Commun. 5, 3271 (2014).Article 

Google Scholar 
Ariza, A. et al. Nat. Clim. Change https://doi.org/10.1038/s41558-022-01479-2 (2022).Article 

Google Scholar 
Klevjer, T. A. et al. Sci. Rep. 6, 19873 (2016).CAS 
Article 

Google Scholar 
Braun, C. D. et al. Annu. Rev. Mar. Sci. 14, 129–159 (2022).Article 

Google Scholar 
Heneghan, R. F. et al. Prog. Oceanogr. 198, 102659 (2021).Article 

Google Scholar 
Polovina, J. J., Dunne, J. P., Woodworth, P. A. & Howell, E. A. ICES J. Mar. Sci. 68, 986–995 (2011).Article 

Google Scholar 
Cheung, W. W. L. et al. Fish Fish. 10, 235–251 (2009).Article 

Google Scholar 
Hazen, E. L. et al. Nat. Clim. Change 3, 234–238 (2013).Article 

Google Scholar 
Powers, R. P. & Jetz, W. Nat. Clim. Change 9, 323–329 (2019).Article 

Google Scholar 
Purves, D. et al. Nature 493, 295–297 (2013).CAS 
Article 

Google Scholar 
Hobday, A. J., Spillman, C. M., Paige Eveson, J. & Hartog, J. R. Fish. Oceanogr. 25, 45–56 (2016).Article 

Google Scholar 
Pons, M. et al. Proc. Natl Acad. Sci. USA 119, e2114508119 (2022).Article 

Google Scholar  More

General description of the experimental areaThe experiment was performed for two years at the National Key Irrigation Experimental Station located on the Songnen Plain in Heping town, Qing’an County, Suihua, Heilongjiang, China, with a geographical location of 45° 63′ N and 125° 44′ E at an elevation of 450 m above sea level (Fig. 1). This region consists of plain topography and has a semiarid cold temperate continental monsoon climate, i.e., a typical cold region with a black soil distribution area. The average annual temperature is 2.5 °C, the average annual precipitation is 550 mm, the precipitation is concentrated from June to September of each year, and the average annual surface evaporation is 750 mm. The growth period of crops is 156–171 days, and there is a frost-free period of approximately 128 days year−122. The soil at the study site is albic paddy soil with a mean bulk density of 1.01 g/cm3 and a porosity of 61.8% prevails. The basic physicochemical properties of the soil were as follows: the mass ratio of organic matter was 41.8 g/kg, pH value was 6.45, total nitrogen mass ratio was 15.06 g/kg, total phosphorus mass ratio was 15.23 g/kg, total potassium mass ratio was 20.11 g/kg, mass ratio of alkaline hydrolysis nitrogen was 198.29 mg/kg, available phosphorus mass ratio was 36.22 mg/kg and available potassium mass ratio was 112.06 mg/kg.Figure 1Location of the study area. The map and inset map in this image were drawn by the authors using ArcGIS software. The software version used was ArcGIS software v.10.2, and its URL is http://www.esri.com/.Full size imageHumic acid fertilizerHumic acid fertilizer was produced by Yunnan Kunming Grey Environmental Protection Engineering Co., Ltd., China (Fig. 2). The organic matter was ≥ 61.4%, and the total nutrients (nitrogen, phosphorus and potassium) were ≥ 18.23%, of which N ≥ 3.63%, P2O5 ≥ 2.03%, and K2O ≥ 12.57%. The moisture content was ≤ 2.51%, the pH value was 5.7, the worm egg mortality rate was ≥ 95%, and the amount of faecal colibacillosis was ≤ 3%. The fertilizer contained numerous elements necessary for plants. The contents of harmful elements, including arsenic, mercury, lead, cadmium and chromium, were ≤ 2.8%, 0.01%, 7.6%, 0.1% and 4.7%, respectively; these were lower than the test standard.Figure 2Humic acid fertilizer in powder form.Full size imageExperimental design and observation methodsIrrigationIn this experiment, three irrigation practices, namely, control irrigation (C), wet irrigation (W) and flood irrigation (F), were designed (Table 1).Table 1 Different irrigation methods.Full size tableControl irrigation (C) of rice had no water layer in the rest of the growing stages, except for the shallow water layer at the regreen stage of rice, which was maintained at 0–30 mm, and the natural dryness in the yellow stage. The irrigation time and irrigation quota were determined by the root soil moisture content as the control index. The upper limit of irrigation was the saturated moisture content of the soil, the lower limit of soil moisture at each growth stage was the percentage of saturated moisture content, and the TPIME-PICO64/32 soil moisture analyser was used to determine the soil moisture content at 7:00 a.m. and 18:00 p.m., respectively. When the soil moisture content was close to or lower than the lower limit of irrigation, artificial irrigation occurred until the upper irrigation limit was reached. The soil moisture content was maintained between the upper irrigation limit and the lower irrigation limit of the corresponding fertility stage. Under the wet irrigation (W) and flood irrigation (F) conditions, it was necessary to read the depth of the water layer through bricks and a vertical ruler embedded in the field before and after 8:00 am every day to determine if irrigation was needed. If irrigation was needed, then the water metre was recorded before and after each irrigation. The difference between before and after was the amount of irrigation23.FertilizationIn our research, five fertilization methods were applied, as shown in Table 2. In this experiment, the rice cultivar “Suijing No. 18” was selected. Urea and humic acid fertilizer were applied according to the proportion of base fertilizer:tillering fertilizer:heading fertilizer (5:3:2). The amounts of phosphorus and potassium fertilizers were the same for all treatments, and P2O5 (45 kg ha−1) and K2O (80 kg ha−1) were used. Phosphorus was applied once as a basal application. Potassium fertilizer was applied twice: once as a basal fertilizer and at 8.5 leaf age (panicle primordium differentiation stage) at a 1:1 ratio22.Table 2 The fertilizer methods.Full size tableThis study was performed with a randomized complete block design with three replications. Three irrigation practices and five fertilizer methods were applied, for a total of 15 treatments as follows: CT1, CT2, CT3, CT4, CT5; WT1, WT2, WT3, WT4, WT5; FT1, FT2, FT3, FT4, and FT5 (C, W, and F represent control irrigation, wet irrigation, and flood irrigation; T represents fertilizer treatment).Measurements of the samplesA soil temperature sensor (HZTJ1-1) was buried in each experimental plot to monitor the temperature of each soil layer (5 cm, 10 cm, 15 cm, 20 cm and 25 cm depth). The transmission of photosynthetically active radiation was measured from 11:00 to 13:00 by using a SunScan Canopy Analysis System (Delta T Devices, Ltd., Cambridge, UK), and data during the crop-growing season were recorded every day24.Plant measurements were taken during the periods of tillering to ripening on days with no wind and good light. The fluorescence parameters were measured by a portable fluorescence measurement system (Li-6400XT, America). The detection light intensity was 1500 μmol m−2 s−1, and the saturated pulsed light intensity was 7200 μmolm−2 s−1. The functional leaves were dark adapted for 30 min, and then the maximum photosynthetic efficiency of PSII (Fv/Fm) was measured. Photochemical quenching (QP) and nonphotochemical quenching (NPQ) were measured with natural light. Simultaneously, the leaf chlorophyll relative content (SPAD) was monitored using SPAD 502 (Konica Minolta, Inc., Tokyo, Japan). For plant agronomic characteristics, the distance from the stem base to the stem tip was measured with a straight ruler to quantify plant height24.Statistical analysisExperimental data obtained for different parameters were analysed statistically using the analysis of variance technique as applicable to randomized complete block design. Duncan’s multiple range test was employed to assess differences between the treatment means at a 5% probability level. All statistical analyses were performed using SPSS 22.0 for Windows24.
Ethics approvalExperimental research and field studies on plants, including the collection of plant material, comply with relevant institutional, national, and international guidelines and legislation. We had appropriate permissions/licences to perform the experiment in the study area. More

Fauna, culture and chronology datasetsA geo-referenced dataset of chronometric dates covering the late MIS 3 (55–30 kyr cal bp) was compiled from the literature (dataset 1). The dataset included 363 radiocarbon, thermoluminescence, optically stimulated luminescence and uranium series dates obtained from 62 archaeological sites and seven palaeontological sites. These chronological determinations were obtained from ten palaeontological levels and 138 archaeological levels. The archaeological levels were culturally attributed to the Mousterian (n = 75), Châtelperronian (n = 6) and Aurignacian (n = 57) technocomplexes. A number of issues can potentially hamper the chronological assessment of Palaeolithic technocomplexes from radiocarbon dates, such as pretreatment protocols that do not remove sufficient contaminants or the quality of the bone collagen extracted. Moreover, discrepancies in cultural attributions or stratigraphic inconsistencies are commonly detected in Palaeolithic archaeology. Information regarding the quality of date determinations and cultural attribution or stratigraphic issues is provided in the Supplementary Information.Our dataset also included the presence of herbivore species recovered from each archaeo-palaeontological site (hereafter referred to as local faunal assemblages (LFAs)), their body masses and their chronology. The mean body mass of both sexes, for each species, was obtained from the PHYLACINE database53 and used in the macroecological modelling approach described below (see ‘Carrying capacity of herbivores’). For visual representation purposes, the herbivore species were grouped into four weight categories: small (500 kg). The chronology of the occurrence of each herbivore species was assumed to be the same as the dated archaeo-palaeontological layer where the species remains were recovered. Thus, to estimate the chronological range of each species in each region, all radiocarbon determinations were calibrated with the IntCal20 calibration curve54 and OxCAL4.2 software55. The BAMs were run to compute the upper and lower chronological boundaries at a CI of 95.4% of each LFA (see ‘Chronological assessment’ for more details). One of the purposes of the current study was to estimate the potential fluctuations in herbivore biomass during the stadial and interstadial periods of the late MIS 3. Accordingly, the time spans of the LFAs were classified into the discrete GS and GI phases provided by Rasmussen et al.51.Geographic settingsThe Iberian Peninsula locates at the southwestern edge of Europe (Fig. 1). It constitutes a large geographic area that exhibits a remarkable diversity of ecosystems, climates and landscapes. Both now and in the past, altitudinal, latitudinal and oceanic gradients affected the conformation of two biogeographical macroregions with different flora and fauna species pools: the Eurosiberian and Mediterranean regions13,46. In the north, along the Pyrenees and Cantabrian strip, the Eurosiberian region is characterized by oceanic influence and mild temperatures in the present day, whereas the Mediterranean region features drier summers and milder winters (Fig. 1). Between the Eurosiberian and Mediterranean regions, there is a transitional area termed Submediterranean or Supramediterranean. Lastly, the Mediterranean region is divided into two distinctive bioclimatic belts: (1) the Thermomediterranean region, located at lower latitudes, with high evapotranspiration rates and affected by its proximity to the coast; and (2) the Mesomediterranean region, with lower temperatures and wetter conditions (Fig. 1).Previous studies have shown that zoocoenosis and phytocenosis differed between these macroregions in the Pleistocene13,46. However, flora and fauna distributions changed during the stadial–interstadial cycles in the Iberian Peninsula, which suggests potential alterations in the boundaries of these biogeographical regions. The modelling approach used in this study to estimate the biomass of primary consumers is dependent on the reconstructed NPP and the herbivore guild structure in each biogeographical region. To test the suitability of the present-day biogeographical demarcations of the Iberian Peninsula during MIS 3, we assessed whether the temporal trends of NPP and the composition of each herbivore palaeocommunity differed between these biogeographical regions during the MUPT.Chouakria and Nagabhusan56 proposed a dissimilarity index to compare time series data by taking into consideration the proximity of values and the temporal correlation of the time series:$${rm{CORT}}(S_1,S_2) = frac{{mathop {sum}nolimits_{i = 1}^{p – 1} {left( {u_{left( {i + 1} right)} – u_i} right)} (v_{(i + 1)} – v_i)}}{{sqrt {mathop {sum}nolimits_{i = 1}^{p – 1} {(u_{(i + 1)} – u_i)^2} } sqrt {mathop {sum}nolimits_{i = 1}^{p – 1} {(v_{(i + 1)} – v)^2} } }}$$
(1)
where S1 and S2 are the time series of data, u and v represent the values of S1 and S2, respectively, and p is the length of values of each time series. CORT(S1, S2) belongs to the interval (−1,1). The value CORT(S1, S2) = 1 indicates that in any observed period (ti, ti+1), the values of the sequence S1 and those of S2 increase or decrease at the same rate, whereas CORT = −1 indicates that when S1 increases, S2 decreases or vice versa. Lastly, CORT(S1, S2) = 0 indicates that the observed trends in S1 are independent of those observed in S2. To complement this approach by considering not only the temporal correlation between each pair of time series but also the proximity between the raw values, these authors proposed an adaptive tuning function defined as follows:$$d{rm{CORT}}left( {S_1,S_2} right) = fleft({{rm{CORT}}left( {S_1,S_2} right)} right)times dleft( {S_1,S_2} right)$$
(2)
where$$fleft( x right) = frac{2}{{1 + exp left( {k,x} right)}},k ge 0$$
(3)
In this study, k was 2, meaning that the behaviour contribution was 76% and the contribution of the proximity between values was 24%57. Hence, f(x) modulates a conventional pairwise raw data distance (d(S1,S2)) according to the observed temporal correlation56. Consequently, dCORT adjusts the degree of similarity between each pair of observations according to the temporal correlation and the proximity between values. This function was used to compare the reconstructed NPP between biogeographical regions during MIS 3 in the Iberian Peninsula. However, two different biogeographical regions could have experienced similar evolutionary trends in their NPP, even though their biota composition was different. Therefore, this analysis was complemented with a JSI to assess whether the reconstructed herbivore species composition in each palaeocommunity differed among biogeographical regions during the late MIS 3. The JSI was based on presence–absence data and was calculated as follows:$${rm{JSI}} = frac{c}{{(a + b + c)}}$$
(4)
where c is the number of shared species in both regions and a and b are the numbers of species that were only present in one of the biogeographical regions. Therefore, the higher the value the more similar the palaeocommunities of both regions were.Chronological assessmentPivotal to any hypothesis of Neanderthal replacement patterns by AMHs is the chronology of that population turnover. To this end, we used three different approaches to provide greater confidence in the results: BAMs, the OLE model and SPD of archaeological assemblages. As detailed below, each of these approaches provides complementary information about the MUPT.First, we built a set of BAMs for the Mousterian, Châtelperronian and Aurignacian technocomplexes in each region during the MIS 3. As stated above, we compiled the available radiocarbon dates for Iberia between 55 and 30 kyr cal bp. However, not all dates or levels were included in the Bayesian chronology models. Radiocarbon determinations obtained from shell remains were incorporated in the dataset (dataset 1); however, the local variation of the reservoir age was unknown from 55 to 30 kyr bp. Because of uncertainties related to marine reservoir offsets, all BAMs that incorporated dates from marine shells were run twice: including and excluding these dates. All of the archaeological levels with cultural attribution issues or stratigraphic inconsistencies were excluded. The Supplementary Note provides a detailed description of the sites, levels and dates excluded and their justification. All BAMs were built for each technocomplex using the OxCAL4.2 software55 and IntCal20 calibration curve54.Bayesian chronology models were built for each archaeological and palaeontological level. Then, the dates associated with each technocomplex were grouped within a single phase to determine each culture’s regional appearance or disappearance. Our interest was not focused on the chronological duration of the Mousterian, Châtelperronian and Aurignacian cultures, but on the probability distribution function of the temporal boundaries of these cultures in each region. Thus, this chronological assessment aims to provide an updated chronological frame for Neanderthal replacement by AMHs in Iberia. For this reason, we did not differentiate between proto- and early Aurignacian cultures, since both are attributed to AMHs.In each BAM, we inserted into the same sequence the radiocarbon dates associated with a given technocomplex within a start and end boundary to bracket each culture, which allowed us to determine the probability distribution function for the beginning and end moment of each cultural phase6. The resolution of all models was set at 20 years. We used a t-type outlier model with an initial 5% probability for each determination, but when more than one radiocarbon date was obtained from the same bone remain, we used an s-type outlier model and the combine function. The thermoluminescence dating likelihoods were included in the models, together with their associated 1σ uncertainty ranges. When dates with low agreement ( More

Dataset 1: Analysis of truffle growers archivesWe selected eleven T. melanosporum orchards located across the South-West France, from Montpellier (43°44′01.4″N 3°42′13.2″E) to Jonzac (45°27′17.7″N, 0°25′26.9″W; Fig. 2). These sites were selected for (1) the quality of the records of fruitbody production and practices by truffle growers (Table S1), including the detail of inoculations since plantation (amount and frequency of added crushed sporocarps), (2) the use of truffle traps by the owners and the quality of the record from these devices, and (3) the presence of oaks (Quercus ilex, Q. pubescens and Q. suber) as the only hosts tree species. Based on the archives of truffle growers, including a systematic recording of truffle production within and outside traps, we reported at each study site the contribution of truffle traps to the annual fruitbody production of the entire truffle grounds, by using number and/or weight of collected fruitbodies within (Pin) and outside (Pout) truffle traps.Dataset 2: In situ experiment tracing the inoculation effectThree orchards located near Angoulème (45°74′35.5″N, − 0°63′78.4″W), Jonzac (45°44′09.8″N, 0°43′96.7″W), and Arles-sur-Tech (42°45′44.9″N, 2°62′89.4″W), hereafter referred to Site 1 to 3 (Fig. 2) were selected for testing both disturbance effect and inoculum effect on fruitbody production in truffle traps. These sites presented a high fruitbody production and a high Pin/Pout ratio, thus optimum conditions to test mechanisms underlying how truffle traps influence fruitbody production. Host trees were between 5 and 18 years old at the beginning of the experiment (Fig. 2). At each site, we selected three non-adjacent trees (four on Site 3) that displayed a continuous fruitbody production over the three previous years. Under each selected tree, we excavated, at two-thirds of the distance between the tree trunk and the limit of brûlé (a vegetation-poor zone that shows the extension mycelia in the soil40, eight equidistant truffle traps [20 × 20 cm large × 20 cm deep] as shown in Fig. 3a. Under each tree, two traps were filled with only a mixture of peat and vermiculite (hereafter referred as non-inoculated controls) to test for disturbance effect. The used mixture was identical to that which is currently applied in commercial orchards. In three other traps, 5 g of crushed material from a single black truffle fruitbody (including its gleba and spores) were added to the previous mixture (hereafter referred as one mating-type inoculum). In the three last traps, 5 g of crushed material from two ascocarps with gleba of opposite mating types (hereafter referred as two mating-type inoculum) were added to the previous mixture. We added the two mating-type condition to accurately test a potential contribution of the gleba (haploid and thus with a single mating type) on future production. As quoted in Introduction, maternal individuals with opposite mating types tend to exclude each other locally (spatial segregation of clusters of individuals of same mating types26. Thus, the two mating-type inoculum allows us to detect in each trap a maternal contribution by the introduced gleba, despite potential exclusion by pre-installed individuals of the locally dominant mating type in the surrounding. Moreover, it allows us to detect a paternal contribution by the introduced gleba of the mating type opposite to the locally dominant. The eight truffle traps were randomly arranged, so that two repetitions of same modality were always separated by a repetition of another modality (Fig. 3a).In March 2013, six freshly collected truffles (weighting  > 60 g) were molecularly analyzed for the mating type of their gleba as in18. On Site 1 and Site 2, the inoculum was made of fruitbodies collected at Site 1. On Site 3, fruitbodies used as inoculum originated from truffle grounds in Sarrion (Spain). In April 2013, truffles traps were installed as explained above (in all, 8 traps × 3 (or 4) trees × 3 sites) and monitored for two years by truffle growers. Harvesting was performed by trained dogs (one different dog per site) checking truffle traps and the surrounding brûlés at each visit of the orchard by truffle growers. When dogs detected truffles, a small hole was excavated to collect ascocarps without disturbing the trap further. At the end of January, 2015, all truffle traps were completely excavated, remnant truffles overlooked by dogs were systematically collected (Fig. 3b). Three soil aliquots were collected within all traps and pooled. All truffles and soil aliquots were frozen for subsequent DNA analysis.Molecular and genetic analysesDNA extractions, mating typing and genotyping were done as in18. Briefly, DNA was extracted from the gleba and from spores of each fruitbody to get access to the maternal and zygotic DNA, respectively. Simple sequence repeat (SSRs) genotyping was performed using 12 polymorphic markers and the mating-type locus as in18. Gleba extracts displaying apparent heterozygous genotypes, likely due to contamination by spore DNA were systematically discarded from further analyses. For each fruitbody, the haploid paternal genotype was then deduced by subtracting the haploid maternal genotype from the zygotic diploid genotype. This data set was used for relatedness estimations. We discarded from all further analysis the marker me11, which displayed more than 39% missing data, as well as all samples with missing data for at any locus.Multilocus genotypes comparisonsBased on the 11 remaining SSRs and the mating-type (Table S5 and Figure S2), MLGs were identified on all maternal and paternal haploid genomes using GenClone v.2.041, and the probability that MLGs represented more than once resulted from independent events of sexual reproduction was calculated (PSex41,42). On each site, clonal diversity was measured as R = (G − 1)/(N − 1) according to43, where N is the number of fruitbodies and G the number of MLGs. For testing whether the gleba of the inoculated fruitbody contributed, either paternally (H1) or maternally (H2) to the harvested fruitbodies (Fig. 1c), the inoculated maternal MLG was compared to the paternal and maternal MLG of the harvested fruitbodies.Relatedness estimationFor testing whether the spores of the inoculum, which carry many distinct haploid MLGs due to meiosis, had paternal or maternal contribution(s) to the harvested fruitbodies (H3; Fig. 1c), we used relatedness estimation.For testing whether spores of the inoculum had a paternal contribution, an individual relatedness estimate to the spore inoculum was computed for each paternal genome detected in truffle traps. Relatedness r here describes the expected frequency E[p_offpat] of each allele in a given genome, E[p_offpat] = p_pop + r * (p_inoc − p_pop), where p_pop is the allele frequency in the local population (here estimated from the glebas of other truffles collected under the focal tree), and p_inoc is the frequency of the allele in the inoculum. Thus, p_offpat takes values 0 or 1, and p_inoc takes values 0, 0.5 or 1, except when two fruitbodies were used as inoculum (two gleba mating types traps). Thus r = (p_offpat − p_pop)/(p_inoc − p_pop). An individual relatedness estimate for each genome is then obtained by summing over alleles and loci the observed values of the numerator and denominator in this expression. A population-level estimate is further obtained by summing numerators and denominators over the paternity events in each population.To test whether such estimates are compatible with the hypothesis that the paternal individuals are not from the inocula, we obtained the distribution of population-level relatedness estimates by simulating samples under this hypothesis: paternal genotypes were randomly simulated according to alleles frequencies in the local population. For each population, 10,000 samples were simulated, and p-values were estimated as the proportion of simulations with higher population-level relatedness with inocula than the observed one. Confidence intervals for these p-values were computed from the binomial distribution for 10,000 draws, and Bonferroni-corrected over the three populations.For testing whether spores of the inoculum had a maternal contribution (H4, Fig. 1c), we estimated the relatedness of the locally used spore inoculum to each maternal genome detected in truffle traps (deduced from the gleba), and we confronted it to simulated samples as previously but with one modification: if the focal fruitbody was harvested in a trap inoculated with the inoculum A1, all genomes of truffles from traps inoculated with the same inoculum (A1 or A1 + A2 + A3, see Fig. 3c.) were discarded from the estimation of p_pop.Assessment of T. melanosporum mycelium concentration in truffle trapsOn Sites 1, 2 and 3, soil samples were collected in all traps and in the surrounding brûlés at harvesting date (January, 2015). In collected soils, total DNA was extracted and quantified as in19. Briefly, after sieving and homogenizing soil collected in each trap and from out of the brûlés, aliquots (10 g) were analyzed as follows. After extraction with the kit Power Soil (MoBio Laboratories, Carlsbad, CA, USA), the extra-radical mycelium of T. melanosporum was quantified using quantitative Taqman™ PCR (qPCR) with the primers and probe described in44. Triplicate real-time PCR were performed on each sample using the same concentration of primer and the same thermocycling program as in19. Standards were prepared using fresh immature T. melanosporum ascocarp, and a standard curve was generated for each site by plotting serial tenfold dilutions against corresponding initial amount of ascocarp. Absolute quantification of mycelium biomass of T. melanosporum was expressed in mg of mycelium per g of soil.Statistical analysesStatistics were done using R version 4.0.445.Effect of truffle traps on fruitbody production—The contribution of truffle traps to the overall production of orchards was assessed by (1) data mining of truffle growers’ archives (Dataset 1) and (2) comparing the density of truffles harvested in traps (expressed in number of truffles per m2 per orchard; for each sampled tree, traps correspond to an investigated soil surface of s = 8 × 0.2 x 0.2 = 0.32 m2) with the density measured within surrounding brûlés (Dataset 1). On Dataset2, at each site, the area occupied by brûlés was evaluated by measuring in the field the surface of soil devoid of vegetation consecutively to spontaneous T. melanosporum brûlé.Fruitbody production under different conditions (i.e. non-inoculated controls versus one gleba mating type traps versus two gleba mating type traps) were compared using generalized linear mixed models with negative binominal family and log link (R, spam package46). The full model included the logarithm of the sampled area as offset to account for variations in this sampled area, interactions of trap-modality effects with site effect. Formal likelihood ratio tests are based on one-step deletions from this full model, applied to subsets of the data relevant for each hypothesis tested. Additional bootstrap tests (1000 iterations) were run to correct any bias in small sample likelihood ratio tests.Concentrations of T. melanosporum mycelium in soil—Similarly as above, the inoculum effect on mycelium concentrations was compared using generalized linear mixed models with Gamma log family.Plant materialThe use of plants in the present study complies with international, national and/or institutional guidelines. All permissions to collect T. melanosporum fruitbodies in truffle orchards were obtained. The formal identification of biological material used in the study (T. melanosporum fruitbodies) was undertaken by F. Richard and E. Taschen. Voucher specimens of all collected fruitbodies have been deposited in the Centre d’Ecologie Fonctionnelle et Evolutive herbarium in Montpellier (France).Ethical approvalAll co-authors approve the ethical statement regarding the submitted manuscript.Consent to participateAll co-authors consent to participate to the research and agree with the content of the submitted manuscript. All authors reviewed and submitted manuscript. More

Barrientos, R. et al. A review of searcher efficiency and carcass persistence in infrastructure-driven mortality assessment studies. Biol. Conserv. 222, 146–153 (2018).
Google Scholar 
Stevens, B. S., Reese, K. P. & Connelly, J. W. Survival and detectability bias of avian fence collision surveys in sagebrush steppe. J. Wildl. Manag. 75, 437–449 (2011).
Google Scholar 
Hunting, K. A Roadmap for PIER Research on Avian Collisions with Power Lines in California. (2002).Barrientos, R. et al. Wire marking results in a small but significant reduction in avian mortality at power lines: A baci designed study. PLoS ONE 7, e32569 (2012).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Costantini, D., Gustin, M., Ferrarini, A. & Dell’Omo, G. Estimates of avian collision with power lines and carcass disappearance across differing environments. Anim. Conserv. 20, 173–181 (2017).
Google Scholar 
Jenkins, A. R. et al. Estimating the impacts of power line collisions on Ludwig’s Bustards Neotis ludwigii. Bird Conserv. Int. 21, 303–310 (2011).
Google Scholar 
Shaw, J. M., Reid, T. A., Schutgens, M., Jenkins, A. R. & Ryan, P. G. High power line collision mortality of threatened bustards at a regional scale in the Karoo, South Africa. Ibis (Lond. 1859) 1859(160), 431–446 (2018).
Google Scholar 
Gómez-Catasús, J. et al. Factors affecting differential underestimates of bird collision fatalities at electric lines: a case study in the Canary Islands. Ardeola 68, 71–94 (2021).
Google Scholar 
Ponce, C., Alonso, J. C., Argandoña, G., García Fernández, A. & Carrasco, M. Carcass removal by scavengers and search accuracy affect bird mortality estimates at power lines. Anim. Conserv. 13, 603–612 (2010).
Google Scholar 
Bernardino, J. et al. Bird collisions with power lines: State of the art and priority areas for research. Biol. Conserv. 222, 1–13 (2018).
Google Scholar 
Brooks, J. W. & Sutton, L. in Veterinary Forensic Pathology (ed. Brooks, J. W.) 43–63 (2018). https://doi.org/10.1007/978-3-319-67172-7_4Brooks, J. W. Postmortem changes in animal carcasses and estimation of the postmortem interval. Vet. Pathol. 53, 929–940 (2016).CAS 
PubMed 

Google Scholar 
Ascensão, F. et al. Beware that the lack of wildlife mortality records can mask a serious impact of linear infrastructures. Glob. Ecol. Conserv. 19, e00661 (2019).
Google Scholar 
Hau, T. C., Hamzah, N. H., Lian, H. H. & Amir Hamzah, S. P. A. Decomposition process and post mortem changes: Review. Sains Malaysiana 43, 1873–1882 (2014).
Google Scholar 
Cooper, J. E. in Wildlife Forensic Investigation: Principles and Practice (eds. Cooper, J. & Cooper, M.) 237–324 (CRC Press, 2013). https://doi.org/10.1201/b14553Sutherland, A., Myburgh, J., Steyn, M. & Becker, P. J. The effect of body size on the rate of decomposition in a temperate region of South Africa. Forensic Sci. Int. 231, 257–262 (2013).CAS 
PubMed 

Google Scholar 
Valverde, I., Espín, S., María-Mojica, P. & García-Fernández, A. J. Protocol to classify the stages of carcass decomposition and estimate the time of death in small-size raptors. Eur. J. Wildl. Res. 66, 1–13 (2020).
Google Scholar 
Goff, M. L. in Current Concepts in Forensic Entomology (eds. Amendt, J., Goff, M., Campobasso, C. & Grassberger, M.) 1–24 (Springer, 2010). https://doi.org/10.1007/978-1-4020-9684-6_1Pittner, S. et al. A field study to evaluate PMI estimation methods for advanced decomposition stages. Int. J. Legal Med. 134, 1361–1373 (2020).PubMed 
PubMed Central 

Google Scholar 
Probst, C. et al. Estimating the postmortem interval of wild boar carcasses. Vet. Sci. 7, 6 (2020).PubMed Central 

Google Scholar 
Cambra-Moo, Ó., Delgado-Buscalioni, Á. & Delgado-Buscalioni, R. An approach to the study of variations in early stages of Gallus gallus decomposition. J. Taphon. 6, 21–40 (2008).
Google Scholar 
Oates, D., Coggin, J., Hartman, F. & Hoilien, G. Guide to Time of Death in Selected Wildlife Species. (Nebraska Technical Series No. 14. Lincoln, N.E., Nebraska Game and Parks Commission, 1984).Hewadikaram, K. A. & Goff, M. L. Effect of carcass size on rate of decomposition and arthropod succession patterns. Am. J. Forensic Med. Pathol. 12, 240–265 (1991).
Google Scholar 
Zhou, C. & Byard, R. W. Factors and processes causing accelerated decomposition in human cadavers—An overview. J. Forensic Leg. Med. 18, 6–9 (2011).PubMed 

Google Scholar 
Cockle, D. L. & Bell, L. S. Human decomposition and the reliability of a ‘Universal’ model for post mortem interval estimations. Forensic Sci. Int. 253(136), e1-136.e9 (2015).
Google Scholar 
Azevedo, R. R. & Krüger, R. F. The influence of temperature and humidity on abundance and richness of Calliphoridae (Diptera). Iheringia. Série Zool. 103, 145–152 (2013).
Google Scholar 
Barnes, K. M. in Wildlife Forensic Investigation: Principles and Practice (eds. Cooper, J. & Cooper, M.) 149–160 (CRC Press, 2013).Mann, R. W., Bass, W. M. & Meadows, L. Time since death and decomposition of the human body: Variables and observations in case and experimental field studies. J. Forensic Sci. 35, 103–111 (1990).CAS 
PubMed 

Google Scholar 
Gliksman, D. et al. Biotic degradation at night, abiotic degradation at day: Positive feedbacks on litter decomposition in drylands. Glob. Change Biol. 23, 1564–1574 (2017).ADS 

Google Scholar 
Araujo, P. I., Grasso, A. A., González-Arzac, A., Méndez, M. S. & Austin, A. T. Sunlight and soil biota accelerate decomposition of crop residues in the Argentine Pampas. Agric. Ecosyst. Environ. 330, 107908 (2022).
Google Scholar 
Fernández-Palacios, J. M. & Martín-Esquivel, J. L. Naturaleza de las Islas Canarias: Ecología y Conservación. (Turquesa, 2001).Kenward, M. G. & Roger, J. H. An improved approximation to the precision of fixed effects from restricted maximum likelihood. Comput. Stat. Data Anal. 53, 2583–2595 (2009).MathSciNet 
MATH 

Google Scholar 
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. http://www.r-project.org (2020).Bates, D., Mächler, M., Bolker, B. M. & Walker, S. C. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48 (2015).
Google Scholar 
Kuznetsova, A., Brockhoff, P. B. & Christensen, R. H. B. lmerTest package: Tests in linear mixed effects models. J. Stat. Softw. 82, 1–26 (2017).
Google Scholar 
Zeileis, A. & Hothorn, T. Diagnostic checking in regression relationships. R News 2, 7–10 (2002).
Google Scholar 
Halekoh, U. & Højsgaard, S. A Kenward–Roger approximation and parametric bootstrap methods for tests in linear mixed models-the R package pbkrtest. J. Stat. Softw. 59, 1–30 (2014).
Google Scholar 
Fox, J. & Weisberg, S. An {R} Companion to Applied Regression, Second Edition. (Sage, 2011).Bartoń, K. MuMIn: Multi-Model Inference. (R Package Version 1.43.6, 2019).De Rosario-Martinez, H., Fox, J. & R Core Team. Package ‘phia’ Title Post-Hoc Interaction Analysis. (R Package Version 0.2–1, 2015).Nakagawa, S. & Schielzeth, H. A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods Ecol. Evol. 4, 133–142 (2013).
Google Scholar 
Vass, A. Beyond the grave—Understanding human decomposition. Microbiol. Today 28, 190–192 (2001).
Google Scholar 
Gill-King, H. in Forensic Taphonomy: The Postmortem Fate of Human Remains (eds. Haglund, W. D. & Sorg, M. H.) 93–104 (CRC Press, 1996). https://doi.org/10.1201/9781439821923.sec2Campobasso, C. P., Di Vella, G. & Introna, F. Factors affecting decomposition and Diptera colonization. Forensic Sci. Int. 12, 18–27 (2001).
Google Scholar 
Austin, A. T., Araujo, P. I. & Leva, P. E. Interaction of position, litter type, and water pulses on decomposition of grasses from the semiarid Patagonian steppe. Ecology 90, 2642–2647 (2009).PubMed 

Google Scholar 
Brandt, L. A., Bonnet, C. & King, J. Y. Photochemically induced carbon dioxide production as a mechanism for carbon loss from plant litter in arid ecosystems. J. Geophys. Res. Biogeosci. 114, G02004 (2009).ADS 

Google Scholar 
Lee, H., Rahn, T. & Throop, H. An accounting of C-based trace gas release during abiotic plant litter degradation. Glob. Chang. Biol. 18, 1185–1195 (2012).ADS 

Google Scholar 
Zepp, R. G., Erickson, D. J., Paul, N. D. & Sulzberger, B. Interactive effects of solar UV radiation and climate change on biogeochemical cycling. Photochem. Photobiol. Sci. 6, 286–300 (2007).CAS 
PubMed 

Google Scholar 
Archer, M. S. Rainfall and temperature effects on the decomposition rate of exposed neonatal remains. Sci. Justice J. Forensic Sci. Soc. 44, 35–41 (2004).Simmons, T., Adlam, R. E. & Moffatt, C. Debugging decomposition data—Comparative taphonomic studies and the influence of insects and carcass size on decomposition rate. J. Forensic Sci. 55, 8–13 (2010).PubMed 

Google Scholar 
Spicka, A., Johnson, R., Bushing, J., Higley, L. G. & Carter, D. O. Carcass mass can influence rate of decomposition and release of ninhydrin-reactive nitrogen into gravesoil. Forensic Sci. Int. 209, 80–85 (2011).CAS 
PubMed 

Google Scholar 
Tracqui. in Encyclopaedia of Forensic Sciences (eds. Siegel, J. A., Saukko, P. J. & Max, M. H.) 1357–1363 (Academic Press, 2000).Riding, C. S. & Loss, S. R. Factors influencing experimental estimation of scavenger removal and observer detection in bird–window collision surveys. Ecol. Appl. 28, 2119–2129 (2018).PubMed 

Google Scholar  More

Changes in temperature and extreme environmental conditions can affect the dynamics of vector-borne pathogens. These include leishmaniasis, transmitted by phlebotomine sandflies, as well as mosquitoes that spread arboviruses like dengue, encephalitis, yellow fever, West Nile fever, and lymphatic filariasis19,20,21.The CCA analysis showed that maximum temperature significantly influenced the abundance of mosquito populations in the study area. In addition, the NMDS showed two different groupings that consisted of samples collected during the rainy and dry seasons. Accordingly, Refs.22,23 report that changes in temperature and relative humidity determine the abundance of mosquitoes, which can disappear entirely during the dry season. Moreover, Refs.22,24,25 note that certain species of mosquitoes increase proportionally with the regional rainfall regime. This is consistent with Ref.10, who find alternating patterns in tropical and temperate climates in some Brazilian regions.As shown by the geometric regression, there is a positive correlation between cumulative rainfall in the days before collection and the number of species found in the study period. Likewise, Ref.26 reported that under the conditions observed in the Serra do Mar State Park, climate variables directly influenced the abundance of Cq. chrysonotum and Cq. venezuelensis, favoring the occurrence of culicids during the more warm, wet, and rainy months.The current climate scenario and future projections about climate, environmental, demographic, and meteorological factors directly influence the distribution and abundance of mosquito vectors and/or diseases27,28,29,30. Environmental temperature alters mosquito population dynamics, thereby affecting the development of immature stages as well as reproduction31. While temperature has an important effect on population dynamics, rainfall and drought also affect the density and dispersal of mosquitoes in temperate and tropical regions32.To be sure, environmental changes other than climate can modify the behavior of vector insects and, subsequently, the mechanism of transmission of parasites20. Specifically, human impacts on the environment can result in drastically different disease transmission cycles in and around inhabited areas33.A previous study34 reported that changes in land use influence the mosquito communities with potential implications for the emergence of arboviruses. Another study35 noted that environmental changes negatively affect natural ecosystems with accelerated biodiversity loss. This is due to the modification and loss of natural habitat and unsustainable land use, which leads to the spread of pathogens and disease vectors.Hence, understanding the relationship between humans and the environment becomes increasingly critical, given the way in which climate changes can lead to alterations in the epidemiology of diseases such as dengue in areas considered free of the disease, as well as in endemic areas36.We found that the abundance and diversity of Mansoniini were directly influenced by the effect of the rainy season and other climatic factors. The rainfall regime has been shown to affect the development of immature forms12,37; explaining the greater frequency of these specimens in the warmer and wetter months38,39,40. According to Ref.41, stable ecosystems such as forests contain great species diversity. On the other hand, diversity tends to be reduced in biotic communities suffering from stress.Studies of insect populations in natural areas are important because they allow a direct analysis of how environmental factors influence phenomena such as the choice of breeding sites by females for oviposition, hematophagous behavior, and the distribution of species along a vegetation gradient12,26,42,43.Throughout the experimental period of the present study, we observed that Shannon light traps are an effective method for catching mosquitoes from the Mansoniini tribe. Interestingly, Ref.44 reported a species richness pattern strongly influenced by Coquillettidia fasciolata (Lynch Arribálzaga, 1891) on mosquito samples from different capture points by using CDC and Shannon light traps as sampling methods. In contrast to the results of Ref.44, where the highest population density of mosquitoes was captured with CDC traps, we observed that these traps were not effective at capturing specimens of Mansoniini in spite of being used in large numbers in the present study. Moreover, Ref.45 conducted another study on faunal diversity in an Atlantic Forest remnant of the state of Rio de Janeiro and observed the highest abundance of Cq. chrysonotum (Peryassú, 1922) and Cq. venezuelensis by using Shannon light traps, while the numbers of captures of Ma. titillans were very similar using CDC and Shannon traps.The results of this study indicate that the makeup of culicid fauna remains quite similar throughout the year, despite seasonal variations in abundance, though there was a lower variability of fauna in the dry season. Therefore, although the seasonality did not affect the temporal variation of the faunal composition in a generalized way, it was possible to detect a partial effect of the seasonality on fauna abundance.
Reference46 report that the incidence peaks of mosquitoes in the warmer and wetter months, as well as mosquito populations remaining between tolerance limits for most of the year, indicate the sensitivity of some species to the local climate.The elevated abundance and diversity of species of Mansoniini in the study area were influenced by the favorable maintenance of breeding sites, including specific water accumulations with emerging vegetation that remain present throughout the year and the well-defined rainy season in the region. In addition, the representatives of Mansoniini, which prefer breeding sites containing macrophytes, made up nearly all of the species collected7.Besides providing a greater awareness of mosquito populations’ ecological and biological aspects, research carried out in wild areas also provides information on the relationship between species diversity and the area in which they are found. Considering that wild insects may become potential vectors of diseases, research in wild areas also provides helpful information for understanding relevant epidemiological aspects. These studies facilitate the identification, monitoring, and control of mosquito populations following environmental changes caused by direct human action, which can lead to major epidemics26.We observed considerable heterogeneity among Mansoniini fauna, and the months with the highest rainfall directly influence the structure of the communities and contribute to the increase in mosquito diversity and abundance, possibly due to variations in the availability of habitat for their immature forms. More

Sightings of some common bird species increased during the UK’s 2020 lockdown.Credit: Tolga Akmen/AFP via Getty

People weren’t the only ones who changed their ways during the COVID-19 pandemic — birds did, too. Four out of five of the most commonly observed birds in the United Kingdom altered their behaviour during the nation’s first lockdown of 2020, although they did so in different ways depending on the species, according to an analysis.The study, published in Proceedings of the Royal Society B on 21 September1, is one of several that used the disruptions brought about by the pandemic — from a reduction in the number of cars on the roads to the closure of some national parks — to quantify the impact that humanity has on the natural world. Although some research has found that lockdowns had a largely positive effect on wildlife2, the latest data from the United Kingdom provide a much more nuanced picture (see Bird Behaviour).

Credit: Warrington et al/Proceedings of the Royal Society B

“People didn’t disappear during the lockdown,” says co-author Miyako Warrington, a behavioural ecologist at the University of Manitoba in Winnipeg, Canada. “We changed our behaviour, and wildlife responded.”Rare experimentIn the early months of the pandemic, social media was abuzz with reports of wild animals being seen in unusual places. These claims were partially validated when Warrington and her colleagues reported that, in 2020, many bird species in the United States and Canada were spotted moving into spaces usually occupied by people2.To see how a COVID-19 lockdown affected birds in the United Kingdom, Warrington and her colleagues tallied sightings of the 25 most common birds between March and July 2020 — during the country’s first lockdown — and compared their data set with data from previous years. In total, the study included around 870,000 observations.The team then compared this information to data showing how people split their time between home, essential shops and parks: three places people in the United Kingdom were allowed to be during the lockdown.Because people spent more time at home and in parks than before March 2020, the analysis found that 20 of the 25 bird species examined behaved differently during lockdown. Parks — which were flooded with visitors — saw an an uptick in the numbers of corvids and gulls, whereas smaller birds, such as Eurasian blue tits (Cyanistes caeruleus) and house sparrows (Passer domesticus), were spotted less frequently than in previous years. And because people spent more time at home, the number of avian species that visited domestic gardens also dropped, by around one-quarter, compared with previous years.Other species, including rock pigeons (Columba livia), didn’t react to the lockdown at all. Warrington found this surprising, because pigeons are city dwellers, so she thought they would be affected by the changes in people’s behaviour. “But they don’t give a crap about what we do,” she says.Adapting to changeThe birds that altered their habits during the lockdown were probably responding to changes in human behaviour, says Warrington. Tits and other birds whose numbers dipped might have fled when people and their pets started spending more time in parks and gardens. The reverse could be true for scavengers, such as gulls and corvids, which might have benefited from park visitors leaving behind rubbish for them to feed on.When combined with the results of other studies, the behaviour of British birds reveals the complex ways in which wildlife was affected by lockdowns and underlines the importance of reducing the disturbance of animals by people, says Raoul Manenti, a conservation zoologist at the University of Milan in Italy.For Warrington, that means acknowledging that lockdowns were not universally good for wildlife. “Our relationship with nature is complicated,” she says. By developing a better understanding of this relationship, “we know we can affect positive change as long as we do it in a thoughtful manner”. More