Ecology

The shell of the endangered hawksbill sea turtle (pictured) is prized for trinkets and jewellery.Credit: Reinhard Dirscherl/SPL

Poaching is less of a threat to the survival of sea turtles than it once was, a new analysis suggests1. Illegal sea-turtle catch has dropped sharply since 2000, with most of the current exploitation occurring in areas where turtle populations are relatively healthy.This study is the first worldwide estimate of the number of adult sea turtles moved on the black market. According to the analysis, more than one million sea turtles were illegally harvested between 1990 and 2020. But the researchers also found that the illegal catch from 2010 to 2020 was nearly 30% lower than that in the previous decade.“The silver lining is that, despite the seemingly large illegal take, exploitation is not having a negative impact on sea-turtle populations on a global scale. This is really good news,” says co-author Jesse Senko, a marine conservation scientist at Arizona State University in Tempe. The research was published 7 September in Global Change Biology.Turtles for trinketsFor millennia, humans have used both adult sea turtles and their eggs as a food source and for cultural practices. In the past 200 years, however, many sea turtle populations declined steeply as hunting rose to meet a growing demand for turtle-based goods. In Europe, North America and Asia, sea-turtle shells were used to make combs, jewelry and furniture inlays. Turtles were also hunted for meat and for use in traditional medicine.The rise in turtle hunting meant that, by 2014, an estimated 42,000 sea turtles were legally harvested every year, and an unknown number of sea turtles were sold on the black market. Today, six of the seven sea-turtle species found around the globe are endangered owing to a deadly combination of habitat destruction, poaching and accidental entanglement in fishing gear.To pin down how many sea turtles were illegally harvested, Senko and his colleagues surveyed sea-turtle specialists and sifted through 150 documents, including reports from non-governmental organizations, papers in peer-reviewed journals and news articles.

Source: Ref. 1

By combining this information, the researchers made a conservative estimate that around 1.1 million sea turtles were illegally caught between 1990 and 2020. Nearly 90% of these turtles were funneled into China and Japan, largely from a handful of middle- and low-income countries (see ‘Long-distance turtle transport’). Of the species that could be identified, the most frequently exploited were the endangered green turtles (Chelonia mydas), hunted for meat, and the critically endangered hawksbill turtles (Eretmochelys imbricata), prized for their beautiful shells.However, the data also showed that the number of illegally caught turtles decreased from around 61,000 each year between the start of 2000 and the end of 2009 to around 44,000 in the past decade (see ‘More sea turtles swim free’). And, although there were exceptions, most sea turtles were taken from relatively robust populations that were both large and genetically diverse.

Source: Ref. 1

Although sea turtles seem to be doing well globally, this doesn’t mean that threats to regional populations can be ignored, says Emily Miller, an ecologist at the Monterey Bay Aquarium Research Institute in California. The study pins down where — and for whom — sea turtles are being exploited, which could help conservationists to target communities for advocacy, she says.Overall, the numbers signal that conservation efforts could be working, says Senko. “Contrary to popular belief, most sea-turtle populations worldwide are doing quite well,” he says. “The number of turtles being exploited is a shocker, but the ocean is big, and there are a lot of turtles out there.” More

WoSIS and permafrost-affected soil profilesThe World Soil Information Service (WoSIS) collates and manages the largest database of explicit soil profile observations across the globe29. In this study, we used the quality-assessed and standardised snapshot of 2019 (ISRIC Data Hub). We further screened the snapshot, and excluded soil profiles with obvious errors (e.g., negative depth values of mineral soil, the value of the depth for the deeper layer is smaller than that of the upper layer). Finally, there is a total of 110,695 profiles with records of SOC content (SOCc, g C kg–1 soil) in the fine earth fraction < 2 mm. The soil layer depths are inconsistent between soil profiles. We harmonised SOCc to three standard depths (i.e., 0–0.3, 0.3–1 and 1–2 m) using mass-preserving splines61,62, which makes it possible to directly compare among soil profiles. We also calculated SOC stock (SOCs, kg C m–2) in each standard depth as:$${{{{{{rm{SOC}}}}}}}_{{{{{{rm{s}}}}}}}=frac{{{{{{{rm{SOC}}}}}}}_{{{{{{rm{c}}}}}}}}{100}cdot Dcdot {{{{{rm{BD}}}}}}cdot left(1-frac{G}{100}right),$$ (1) where D is the soil depth (i.e., 0.3, 0.7, or 1 m in this study), BD is the bulk density of the fine earth fraction 2 mm) of soil. Amongst the 110,695 soil profiles, unfortunately, only 18,590 profiles have measurements of both BD and G. To utilise and take advantage of all SOCc measurements, we used generalised boosted regression modelling (GBM) to perform imputation (i.e., filling missing data). As such, SOCs can be estimated. To do so, for BD and G in each standard soil depth, GBM was developed based on all measurements of that property (e.g., BD) in the 110,695 profiles with other 32 soil properties recorded in the WoSIS database. The detailed approach for missing data imputation has been described in ref. 41.Together with the WoSIS soil profiles, a total of 2,703 soil profiles with data of SOCs from permafrost-affected regions were obtained from ref. 30. The original data used in ref. 30 have been obtained, and we used the data of SOCs in the 0–0.3, 0.3–1, and 1–2 m soil layers in this study. These permafrost-affected profiles compensate for the scarce soil profiles in high latitudinal regions in the WoSIS database. Overall, the soil profiles cover 13 major biome groups although the profile numbers vary among biome types (Supplementary Fig. 1). The profiles also cover various climate conditions across the globe with mean annual temperature (MAT) ranging from –20.0 to 30.7 °C and mean annual precipitation (MAP) ranging from 0 to 6,674 mm.Environmental covariatesMAT and MAP for each soil profile were obtained from the WorldClim version 2 (ref. 63). The WorldClim version 2 calculates biologically meaningful variables using monthly temperature and precipitation during the period 1970–2000. We obtained global spatial layers of MAT and MAP at the resolution of 30 arcsecond (i.e., 0.0083° which is equivalent to ~1 km at the equator). Soil profiles in the same 0.0083° grid (i.e., ~1 km2) share the same MAT and MAP. Besides MAT and MAP, other climatic variables for each soil profile were also obtained from the WorldClim version 2. The WWF (World Wildlife Fund) map of terrestrial ecoregions of the world (https://www.worldwildlife.org/publications/terrestrial-ecoregions-of-the-world) was used to extract the biome type at each soil profile. The MODIS land cover map64 at the same resolution of NPP databases was used to identify that if the land is cultivated (i.e., land cover type of croplands and cropland/natural vegetation mosaic) at the location of each soil profile.Space-for-time substitution: grouping soil profilesWe used a hybrid approach of space-for-time substitution and meta-analysis to estimate the response of SOC to warming. Traditionally, space-for-time substitution involves determining regression relationships across gradients at one time31. The regression was then used to predict future status under conditions when one or more of the covariates has changed31. However, the approach was compromised when the effects of other driving variables such as soil type and landform were not minimised. Regarding SOC dynamics, they would show non-linear relationships19 with temperature modulated by a series of other environmental covariates (e.g., precipitation, vegetation type).Based on the idea of space-for-time approach31, first, we sorted all soil profiles by MAT at the soil-profile locations and designated them into MAT classes with an increment of 1 °C (Fig. 1). Then, we derived pairs of soil profiles, with each pair including a “ambient” and “warm” class (i.e., control vs treatment in meta-analysis language) distinguished by MAT (Fig. 1). The ambient class includes soil profiles with MAT ranging from i to i + 1 degree Celsius, where i is the lowest temperature in the class. If 1 °C warming is of interest, for example, the warm class will be identified as the class with MAT ranging from i + 1 to i + 2 degree Celsius (i.e., one degree higher than that of the ambient class; Fig. 1). To control the effects of precipitation, soil type and topography, soil profiles in both ambient and warm classes were further grouped; and each group must have the same following characteristics: (1) Landform. A global landform spatial layer was obtained from Global Landform classification - ESDAC - European Commission (europa.eu), and global terrestrial lands were divided into three general landform types: plains including lowlands, plateaus, and mountains including hills. (2) Soil type. The 12 USDA soil orders were used to distinguish soil types. A global spatial layer of soil orders was obtained from The Twelve Orders of Soil Taxonomy | NRCS Soils (usda.gov). We also independently tested the sensitivity of the results to different soil classification systems by including FAO and WRB soil groups (Soil classification | FAO SOILS PORTAL|Food and Agriculture Organization of the United Nations). (3) Mean annual precipitation (MAP). MAP cannot be exactly the same between the ambient and warm groups. In practice, we considered that soils meet this criterion if the absolute difference of MAP between ambient and warm soils is less than 50 mm. We also tested the sensitivity of the results to this absolute MAP difference using another value of 25 mm, and found that this difference has negligible effect (Supplementary Fig. 11). (4) Precipitation seasonality. Precipitation seasonality indicates the temporal distribution of precipitation. In this study, we focused on warming alone, and global warming would also have less effect on this seasonal distribution of precipitation. The seasonal distribution pattern of precipitation was classified into three categories: summer-dominated precipitation, winter-dominated precipitation and uniform precipitation. Precipitation concentration index (PCI) was calculated in R precintcon package to distinguish the three patterns65: $${{{{{rm{PCI}}}}}}=frac{mathop{sum }nolimits_{{{{{{rm{i}}}}}}=1}^{12}{p}_{{{{{{rm{i}}}}}}}^{2}}{{left(mathop{sum }nolimits_{{{{{{rm{i}}}}}}=1}^{12}{p}_{{{{{{rm{i}}}}}}}right)}^{2}}cdot 100,$$ (2) where pi is the precipitation in month i in a particular year. In this study, we used the monthly precipitation from 1970 to 2000 obtained from WorldClim version 2 (ref. 63) to calculate the average (overline{{{{{{rm{PCI}}}}}}}) at the location of each profile. If (overline{{{{{{rm{PCI}}}}}}})  8.3 and total precipitation from April to September (from October to March in the Southern Hemisphere) is larger than that from October to March (from April to September in the Southern Hemisphere), precipitation mainly occurs in summer (i.e., summer precipitation); otherwise, it is winter precipitation.By applying these selection criteria to all soil profiles, we obtained pairs (i.e., an “ambient” group vs a “warm” group) of soil profiles mainly distinguished by MAT (i.e., warming). Amongst pairs, they would be different in landform, soil type, MAP and precipitation seasonality, which enables us to address their effects on the response of SOC to warming. We are interested in five warming levels including 1, 2, 3, 4, and 5 °C.Meta-analysis: estimation of the response of SOC to warmingMeta-analysis techniques were used to estimate the percentage response of SOC to warming by comparing SOC content and stock in groups in the warm group to that in the ambient group. The log response ratio of soil C (lnRR) to warming for each pair (i.e., an ambient group vs a warm group) of soil profiles was calculated as:$${{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}={{{{{rm{ln}}}}}}left(frac{bar{{{{{{{rm{SOC}}}}}}}^{*}}}{overline{{{{{{rm{SOC}}}}}}}}right),$$ (3) where (overline{{{{{{rm{SOC}}}}}}}) and (bar{{{{{{{rm{SOC}}}}}}}^{*}}) are the mean SOC (either content or stock) in groups from ambient and warm class, respectively. In order to provide a robust estimate of global mean response ratio, the individual lnRR values were weighted by the inverse of the sum of within- (v) and between-group (τ2) variances. As such, the global mean response ratio ((overline{{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}})) could be estimated as:$$overline{{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}}=frac{{sum }_{{{{{{rm{i}}}}}}}left({{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}}_{{{{{{rm{i}}}}}}}times {w}_{{{{{{rm{i}}}}}}}right)}{{sum }_{{{{{{rm{i}}}}}}}{w}_{{{{{{rm{i}}}}}}}},$$ (4) where ({w}_{{{{{{rm{i}}}}}}}=frac{1}{{v}_{{{{{{rm{i}}}}}}}+{tau }^{2}}) is the weight for the ith lnRR. In addition, we estimated and compared the mean response ratios under different soil orders, landforms, and precipitation concentration patterns. These mean response rates were calculated in weighted, mixed-effects models using the rma.mv function in R package metafor. To assist interpretation, the results of (overline{{{{{{rm{ln}}}}}}{{{{{rm{RR}}}}}}}) were back-transformed and reported as percentage change under warming, i.e., (({{{{{{rm{e}}}}}}}^{{{{{{rm{RR}}}}}}}-1)times)100. These back-transformed values were also used for subsequent data analyses.An implicit assumption underlying the space-for-time substitution approach is that important events or processes which substantially change the succession direction of studied system (e.g., volcano disruption in one class but not in another class, cultivation in one class but not in another class) are independent of space and time (which includes the past and future)66. We conducted two sensitivity assessment to test this assumption. First, we repeated all above assessment by excluding soil profiles from croplands since preferential choice of land clearing for cultivation should be common. Second, we repeated all assessment by including only groups having at least 20 soil profiles. This allows the assessed pairs to cover a higher diversity of land history and future land cover/use, diluting the effect of a typical event at a specific soil profile on the estimates.Comparison with SOC turnover modelsWe compared our estimation with predictions by SOC models. A simple one-pool SOC model can be written as:$$frac{{{{{{rm{d}}}}}}C}{{{{{{rm{d}}}}}}t}=I-kcdot C,$$ (5) where I is the amount of carbon input, k is the decay rate of SOC, and C is the stock of SOC. At steady state, (C=I/k). A Q10 function can be applied to estimate k under warming (kw):$${k}_{{{{{{rm{w}}}}}}}=kcdot {{exp }}left(0.1cdot triangle Tcdot {{log }}left({Q}_{10}right)right),$$ (6) where (triangle T) is the warming level. Thus, when soil reaches a new steady state under warming, SOC stock (Cw) can be estimated as:$${C}_{{{{{{rm{w}}}}}}}=frac{{I}_{{{{{{rm{w}}}}}}}}{kcdot {{exp }}left(0.1cdot triangle Tcdot {{log }}left({Q}_{10}right)right)},$$ (7) where Iw is the carbon input amount under warming condition. Finally, the response of SOC to warming (R) can be calculated as:$$R=frac{{C}_{{{{{{rm{w}}}}}}}-C}{C}=frac{{I}_{{{{{{rm{w}}}}}}}}{I}cdot {{exp }}left(-0.1cdot triangle Tcdot {{log }}left({Q}_{10}right)right)-1.$$ (8) Using Eq. (8), we calculated R under a series of ensembles of (frac{{I}_{{{{{{rm{w}}}}}}}}{I}), (triangle T), and ({Q}_{10}), and compared R with that estimated using our space-for-time substitution approach.Comparison with field warming experimentsA number of meta-analyses based on data from field warming experiments had been performed to assess the response of SOC to warming7,26,46,47,48,49,50, which enable us to conduct comparisons with the estimates using our hybrid approach combining space-for-time substitution and meta-analysis techniques. A total of five meta-analysis papers have been found by searching the Web of Science. We retrieved the response ratios from the identified papers, and compared them to our estimations. Here, it should be noted that most field warming experiments focused on SOC changes (stock or content) in the top 0.2 m soil layer. We compared them with our estimation of the response of SOC stock in the top 0.3 m soil.Besides the published results of meta-analysis, we also conducted an independent meta-analysis using data from field warming experiments. The meta-analysis dataset was mainly from published papers on meta-analysis from 2013 to 2020 (see Supplementary Data 1). It should be noted that the field warming experiments manipulate temperature using different approaches such as open/closed-top chamber, infrared radiators and heating cables. For the comparison, we did not explicitly distinguish these approaches. The experimental duration ranged from 0.42 to 25 years with a mean of 4.7 years, and the warming magnitude ranged from 0.1 to 7°C with a mean of 1.92 °C. To ease comparison, field warming levels were classified into 0–1, 1–2, 2–3, 3–4, 4–5, and >5 °C. The same meta-analysis to that assessing soil profile data was used to predict the response ratio of SOC to the above six warming levels. In addition, we divided the data into four ecosystems (i.e., tundra, forest, shrublands and grasslands) and estimated the response ratio in each ecosystem. These estimates based on field warming experiments were compared with those estimated using our space-for-time approach.Variable importance and global mappingWe included 15 environmental predictors to derive a meta-forest model, a machine learning-based random forest model adapted for meta-analysis, to map the response of SOC stock/content to warming across the globe at the resolution of 0.0083°. The 15 environmental predictors reflect generally four broad groups of environmental conditions: baseline SOC conditions represented by current standing SOC stock or content, soil order and soil depth; current baseline climatic conditions represented by MAT, MAP, aridity index, precipitation seasonality represented by PCI, the fraction of precipitation in summer, the difference of temperature between ambient and warm groups, the difference of precipitation between ambient and warm groups; topography represented by elevation and landform; and vegetation represented by NPP and biome type.The metaforest function in the metafor package was used to derive the model. To fit the model, a fivefold cross-validation was conducted. That is, 80% of the derived response ratios was used to train the model, and the remaining 20% to validate the model. The best model hypeparameters were targeted by running the model under a series of parameter combinations, and the model performance was assessed by the rooted mean squared error (RMSE) and determination coefficient (R2). The meta-forest model allows the estimation of the relative influence of each individual variable in predicting the response, i.e. the relative contribution of variables in the model. The relative influence is calculated based on the times a variable selected for splitting when growing a tree, weighted by squared model improvement due to that splitting, and then averaged over all fitted trees which are determined by the algorithm when adding more trees cannot reduce prediction residuals. As such, the larger the relative influence of a variable, the stronger the effect of the variable on the response variable.Combining with spatial layers of predictors, the meta-forest model for SOC stock was used to predict the response of SOC to warming across the globe at the resolution of 1 km (most data layers are already at the 1 km resolution as abovementioned, for those layers that are not at the target resolution, they were resampled to the 1 km resolution). In the meta-forest model, current standing SOC stock is the most important predictor (Fig. 4). We use three global maps of SOC stocks including WISE51 (WISE Soil Property Databases | ISRIC), HWSD52 (Harmonized World Soil Database (HWSD v 1.21) – HWSD – IIASA) and SoilGrids53 (SoilGrids250m 2.0) to obtain current standing SOC stocks. These three global maps represent the major mapping products of SOC stock at the global level, and had been widely used for large scale modelling. The derived meta-forest model was applied across the globe to estimate the response ratio of SOC stock in each 1 km pixel. To do so, the same procedure to group the observed soil profiles (Fig. 1) was applied to group global land pixels (section Space-for-time substitution: grouping soil profiles). The only difference is that global mapping uses all pixels instead of the 113,013 soil profiles. In each 1 km pixel, prediction uncertainty was also quantified using estimates of randomly drawn 500 trees of the fitted meta-forest model to calcuate standard deviation of the predictions. More

The European Union’s new (2023–2027) Common Agricultural Policy (CAP) aims to reverse current environmental degradation and biodiversity declines in European farmland1 through the achievement of three green objectives: contribute to climate change mitigation, support efficient natural resource management, and reverse biodiversity loss2,3. Following the outbreak of war in Ukraine, the European Commission proposed a series of short and medium-term relaxations to CAP’s environmental commitments to offset expected shortages in grain imports and enhance food security4.Here, we argue that policy changes to allow cultivation of fallow land will disproportionately impact biodiversity and support further intensification of livestock production. Thus, ultimately, these changes in policy may sacrifice long term biodiversity and agricultural sustainability in Europe, in favour of modest increases in current agricultural production and alleged improvements of food security.A catalyst for reversing green policiesRussia and Ukraine are world-leading producers and exporters of cereal and fodder production (notably, oleo-proteaginous crops)5. The Ukraine war and international sanctions on Russia are threatening the import of these products to the EU. Ukrainian winter cereal, maize and sunflower production is expected to decrease by 20–30%, at least during the 2022–2023 season, and similar reductions in Russian exports are also expected5. Therefore agro-industry lobbies and farmers’ organisations in Brussels, some political parties in the European parliament and some countries’ administrations perceive a need to increase agricultural production6 and, as a means to offset expected shortages, are pressing to relax or remove CAP’s environmental commitments. Mechanisms supporting these commitments include enhanced conditionality (compulsory for all farmers receiving subsidies), voluntary measures of Rural Development Programmes (i.e. agri-environment-climate-measures) and Greening measures (crop diversification, maintenance of permanent grasslands and promotion of Ecological Focus Areas). A call made to mobilise all relevant international groups during the informal meeting held on 2 March 2022 by Member States’ agriculture and food ministers, with the exception of Denmark, Germany and Italy, may reflect such pressure6. Indeed, the European Commission has finally proposed a series of “short- and medium-term actions to enhance global food security and to support farmers and consumers in the EU”4. In regard to land-use, actions refer to the cultivation of fallows, which are protected by green payments for keeping land in good agricultural and environmental conditions and, adequately managed (both long-term and annual), support high levels of biodiversity and ecosystem services7 (Fig. 1). More precisely, the European Commission proposes that “To enlarge the EU’s production capacity, the Commission has today adopted an implementing act to exceptionally and temporarily allow Member States to derogate from certain greening obligations. In particular, they may allow for production of any crops for food and feed on fallow land that is part of Ecological Focus Areas in 2022, while maintaining the full level of the greening payment”4. This measure was recently extended for 2023.Fig. 1: Arable field left fallow and allowed to develop a grassy vegetation cover.Under non-intensive management, fallow areas become a genuine semi-natural habitat, key for the conservation of farmland biodiversity. Credit: Jordi Bas, taken in the cereal steppes of the Lleida plain (Catalonia, Spain).Full size imageConsidering food sovereigntyHowever, the FAO does not draw the same conclusions about the possible world impacts of the conflict and recommends finding alternative suppliers, instead suggesting using existing food stocks, diversifying domestic crops and reducing fertiliser dependence and food waste as mechanisms to help guarantee Europe’s food supplies and sovereignty5. Even the European Commission, while acknowledging the vulnerability of European farmers to animal feed import shortages and increased costs, clearly stated that food supply is not at risk in the EU4. Indeed, EU-based production supplies 79% of the feed proteins consumed in European livestock farming, 90% of feed cereals and 93% of other products such as Dried Distillers’ Grains and Solubles or beet pulp8. In 2020, the EU was completely self-sufficient with respect to dairy products, pork, beef, veal, poultry, and cereals. It remained the largest global exporter of agri-food products, in spite of the COVID-19 pandemic8.Counterproductive policiesAny increase in production from cultivating fallow land will therefore likely be used to feed intensively reared livestock and sustain cattle feed exports. Supporting the increasing trend of feed exports and industrial intensive livestock farming does not align with the EU’s Green Deal due to the negative impacts on air, soil and water quality8,9,10. In addition, cultivating fallow land to support intensive livestock-based agriculture will undermine the EU’s Farm-to-Fork strategy and CAP’s ‘Food and Health’ objective of reducing meat consumption to favour a more sustainable and healthier diet among European consumers2,11. Encouraging the growth of intensive livestock farming through enabling cultivation of fallow lands will increase environmental damage, biodiversity loss and public health risks. Thus, the recent relaxations of the new CAP compromise several of its fundamental objectives, along with those of other elements of the Green Deal, such as the EU’s Nature Restoration Law2,9,12.The duration of the war in Ukraine and its effects on provision of raw materials to Europe is hard to foresee. We acknowledge the uncertainties and input costs faced by farmers but calls for further agricultural intensification may be largely unjustified at this stage. Specifically, cultivating semi-natural habitats like long-term or unploughed annual fallows will have serious environmental costs, including an increase in pesticide and fertiliser application, since fallows often occupy less productive land13. Even a moderate increase in food production at the expense of the semi-natural habitats remaining in farmland landscapes (field margins, grasslands, and fallow land), which support most of Europe’s farmland biodiversity and its associated ecosystem services14, will seriously damage farmland biodiversity and sustainability in European agricultural landscapes3,15. For example, a comprehensive study carried out on 169 farms across 10 European countries showed that semi-natural habitats, including fallows, occupied 23% of the land but hosted 49% of vascular plant, earthworm, spider, and wild bee species; a 10% decrease of these habitats if reclaimed for food production would cause exponential decreases in biodiversity, but only moderate linear increases in production15. Furthermore, the loss of semi-natural habitats in arable systems, fallows among them, would negatively affect arthropod functional diversity and the ecosystem services it supports, which may affect agriculture production14.Sustainable alternativesThere are alternatives to cultivating semi-natural habitats that may (and need to) be assessed to achieve a more strategic European agricultural policy able to meet food demands while maintaining the sustainability principles and improvements of the food-production chain sought by the Farm-to-Fork strategy. Proposals include agro-ecological approaches to increase production through the enhancement of ecosystem services such as pollination and biological control16,17,18, adjusting the amount of cultivated surface in relation to landscape structure and composition19, or relocating crops that are more in demand to areas where production is optimal without increasing the total cultivated area20.After decades of costly implementation and reforms of agricultural and conservation policies1, the EU is at risk of engaging in a hasty and misguided strategy on food production jeopardising the green transition13. As an alternative to such a ‘business as usual’ reaction, the EU has now the opportunity to consolidate the mentioned environmental and social objectives of the new CAP2,3. A more sustainable agriculture, resilient to food supply crises (present and future), should be based on ecological functionality of farmland, which ultimately depends on the conservation of its biodiversity16, along with measures to counter climate change. Responses to this and other challenges on the new CAP should be assessed with a long-term perspective and based on robust scientific evidence before undermining its environmental ambitions3,13. More

Study areaSoutheastern Chongqing, China (107° 14′–109° 19′ E, 28° 9′–30° 32′ N), has an area of about 19,800 km2 (Fig. 1). The study area has a subtropical monsoon climate. And the area has four distinct seasons, with an annual average temperature of 16.2 °C and abundant rainfall, with an average annual rainfall of 1209 mm. This region is located in the central part of the Wuling mountains, which is characterized by medium and low mountainous landforms, with an average altitude of greater than 1000 m. The water system (the Wujiang River system) in the study area is well developed, with a large drainage area and rich groundwater resources. The soil is dominated by yellow soil and limestone soil, and the sensitivity to soil erosion is high. The district exhibits the typical ecological fragility of karst areas, with barren soil, fragmented surfaces, a single community, and a low ecological carrying capacity. The area includes six counties: Qianjiang district, Shizhu Tujia Autonomous county, Xiushan Tujia and Miao Autonomous county, Youyang Tujia and Miao Autonomous county, Wulong district, and Pengshui Miao and Tujia Autonomous county. The coverage rates of the carbonatite layers in these counties are 42.11, 67.77, 25.70, 34.80, 59.70 and 88.46%, respectively38, and the average coverage of the carbonatite layers is 53.09%, making this a representative area of karst rocky desertification.Data and image pre-processingIn the study, the remote sensing data were obtained from the United States Geological Survey (USGS, https://earthexplorer.usgs.gov/), including landsat-5 thematic mapper (TM) images acquired in 2001, 2006 and 2011 and Landsat-8 operational land imager (OLI) images obtained in 2016 and 2021 (Table 1). The spatial resolution is 30 m. In order to ensure the comparability of spectral characteristics, the data collection was conducted from May to September when the vegetation grew better. In order to meet the usage requirements, the cloud cover of each image used is below 10%. For the images with poor quality, the adjacent years were selected for replacement. The difference in ecological quality between adjacent years in the same region was not particularly large. In order to represent the actual situation of the ecological environment quality in the target year as much as possible, we tried to minimize the replaced part in each target year. A total of 20 images were collected in this study. The images downloaded were all L1T products, which had undergone systematic radiometric correction and geometric correction, so precise geometric correction was no longer performed. Before the subsequent processing, all 20 images were preprocessed by radiometric calibration, atmospheric correction, image mosaicking and cropping. Then these images were calculated to obtain NDVI, WET, NDBSI, LST and RI. And based on the preprocessed Landsat images, support vector machine classification was performed to obtain the land use (LU) status.Table 1 Information of images used in this study.Full size tableThe topographical data included the elevation (EV) and slope (SP) data. Among them, the elevation data was provided by the official website of the United States Geological Survey (USGS, https://earthexplorer.usgs.gov/). And the slope data was calculated from the elevation data. The meteorological data, including the monthly average temperature (MT), monthly mean precipitation (PR), monthly even relative humidity (RH), and monthly total sunshine hours (SH) from May to September of the target year, were got from the China Meteorological Data Network (http://data.cma.cn/). In addition, socioeconomic data, including the population density (PD) and gross domestic product (GDP), were obtained from the statistical yearbooks of each district and county in the study area. The nighttime light (NTL) data were obtained from the National Oceanic and Atmospheric Administration (NOAA, https://www.noaa.gov/). The above data and LU were used as the influencing factors of ecological quality to analyze the reasons for the change of local ecological environment quality. The statistical data and monitoring data of each evaluation index used to construct the EI come from the statistical yearbooks, water resources bulletin and soil and water conservation bulletin of each district and county.MethodologyStudy frameworkA framework was developed for evaluating the ecological quality in southeastern Chongqing from 2001 to 2021 in the study. And the framework included three parts: data preparation, construction of the MRSEI, and the analysis of the ecological status in the region. Figure 2 presents the detailed information about the framework. The operations of band calculation, normalization and PCA were all carried out using the ENVI 5.3 software (https://www.harrisgeospatial.com).Figure 2The study framework.Full size imageIndicators used in MRSEIThe greenness, humidity, heat, dryness, and degree of rocky desertification were used to construct the MRSEI. The NDVI39 was chosen to characterize the greenness. The humidity component acquired from the tasseled cap transformation (WET)40 was selected to represent the humidity. The LST41 was used to represent the heat, the normalized difference build-up soil index (NDBSI)42 was used to characterize the dryness. The RI was applied to characterize the degree of rocky desertification.The NDVI is an important indicator for monitoring the physical and chemical properties of vegetation, and it can be employed to calculate the vegetation coverage, leaf area index, and so on19. In addition, it eliminates some radiation errors and has a stronger response to surface vegetation. It has been widely used in vegetation remote sensing monitoring. The equation for calculating the NDVI is as follows39:$$ {text{NDVI}} = {{(uprho }}_{{{text{NIR}}}} – {uprho }_{{{text{Red}}}} {)}/{{(uprho }}_{{{text{NIR}}}} {{ + uprho }}_{{{text{Red}}}} ), $$
(1)
where ({uprho }_{{{text{NIR}}}}) is the reflectance of the near-infrared band and ({uprho }_{{{text{Red}}}}) refers to the reflectance of the red band corresponding to each image.The WET can effectively reflect the humidity conditions of the surface vegetation, water, and soil, and can reveal the changes in the ecological environment, such as soil degradation. Therefore, it is commonly used in ecological environment monitoring43. The WET can be expressed as40,43:$$ {text{WET}}_{{{text{TM}}}} { = 0}{{.3102uprho }}_{{{text{Red}}}} { + 0}{{.2021uprho }}_{{{text{Green}}}} { + 0}{{.0315uprho }}_{{{text{Blue}}}} { + 0}{{.1594uprho }}_{{{text{NIR}}}} – {0}{{.6806uprho }}_{{{text{SWIR1}}}} – {0}{{.6109uprho }}_{{{text{SWIR2}}}} , $$
(2)
$$ {text{WET}}_{{{text{OLI}}}} { = 0}{{.3283uprho }}_{{{text{Red}}}} { + 0}{{.1972uprho }}_{{{text{Green}}}} { + 0}{{.1511uprho }}_{{{text{Blue}}}} { + 0}{{.3407uprho }}_{{{text{NIR}}}} – {0}{{.7117uprho }}_{{{text{SWIR1}}}} – {0}{{.4559uprho }}_{{{text{SWIR2}}}} , $$
(3)
where ({uprho }_{{text{i}}} ,) is the reflectance of band i.The NDBSI is expressed as the average of two indicators, the bare soil index (SI)44 and the index-based built-up index (IBI)45. It can be applied to characterize the dryness. The calculation formulas are44,45:$$ {text{IBI }} = {text{ }}left[ {2uprho _{{{text{SWIR1}}}} /left( {uprho _{{{text{SWIR1}}}} + {text{ }}uprho _{{{text{NIR}}}} } right) – uprho _{{{text{NIR}}}} /(uprho _{{{text{NIR}}}} + {text{ }}uprho _{{{text{Red}}}} } right) – uprho _{{{text{Green}}}} /(uprho _{{{text{Green}}}} + {text{ }}uprho _{{{text{SWIR1}}}} )]/[2uprho _{{{text{SWIR1}}}} /left( {uprho _{{{text{SWIR1}}}} + {text{ }}uprho _{{{text{NIR}}}} } right) + {text{ }}uprho _{{{text{NIR}}}} /(uprho _{{{text{NIR}}}} + {text{ }}uprho _{{{text{Red}}}} ) + {text{ }}uprho _{{{text{Green}}}} /(uprho _{{{text{Green}}}} + {text{ }}uprho _{{{text{SWIR1}}}} )], $$
(4)
$$ {text{SI = }}left[ {{uprho }_{{{text{SWIR1}}}} {{ + uprho }}_{{{text{red}}}} – left( {{uprho }_{{{text{Blue}}}} {{ + uprho }}_{{{text{NIR}}}} } right)} right]/left[ {{uprho }_{{{text{SWIR1}}}} {{ + uprho }}_{{{text{red}}}} { + }left( {{uprho }_{{{text{Blue}}}} {{ + uprho }}_{{{text{NIR}}}} } right)} right], $$
(5)
$$ {text{NDBSI = (IBI + SI)/2,}} $$
(6)
where ({uprho }_{{text{i}}} ,) is the reflectance of band i.The LST is closely related to natural processes and human phenomena such as crop yield, vegetation growth and distribution, surface water cycle, etc. It can well reflect the state of the surface ecological environment. The atmospheric correction method is used to invert the LST here46,47, it can be expressed as:$$ {text{L = gain}} times {text{DN + bias,}} $$
(7)
$$ {text{T = K}}_{{2}} /{text{ln}}left( {frac{{{text{K}}_{{1}} }}{{text{L}}}{ + 1}} right){,} $$
(8)
$$ {text{LST = T}}/left[ {{1 + }left( {frac{{{lambda T}}}{{upalpha }}} right){{lnvarepsilon }}} right]{,} $$
(9)
where L is the radiation value in the thermal infrared band, DN is the gray value, gain and bias is the gain value and offset value of the L-band, which was got from the image header file. And T is the temperature value at the sensor; K1 and K2 are calibration parameters respectively (for TM, K1 = 607.76 W/(m2 sr μm), K2 = 1260.56 K; for TIRS, K1 = 774.89 W/(m2 sr μm), K2 = 1321.08 K); λ is the central wavelength of thermal infrared band; α = 1.438 × 10−2 m K. ε is the surface emissivity and the value is estimated by the vegetation index mixture model48,49. It is calculated as follows:$$ {text{VFC = }}frac{{{text{NDVI}} – {text{NDVI}}_{{{text{Soil}}}} }}{{{text{NDVI}}_{{{text{Veg}}}} – {text{NDVI}}_{{{text{Soil}}}} }}, $$
(10)
$$ {text{d}}_{{upvarepsilon }} { = }left( {{1} – {upvarepsilon }_{{text{s}}} } right){{ times (1}} – {text{VFC) }}times text{F} times upvarepsilon _{{text{v}}} , $$
(11)
$$ {{upvarepsilon = upvarepsilon }}_{{text{v}}} times {text{ VFC}} + varepsilon _{{text{s}}} {{ times }}left( {{1} – {text{FVC}}} right){text{ + d}}_{{upvarepsilon }} , $$
(12)
where VFC is the vegetation fractional cover, ({text{NDVI}}_{{{text{Veg}}}}) is the NDVI of the pixel covered by full vegetation and the pixels with NDVI  > 0.72 are regarded as pure vegetation pixels; ({text{NDVI}}_{{{text{Soil}}}}) is the NDVI of the bare pixel and the pixels with NDVI  More

Essential oils yield and compositionAmong the 14 seed sample populations collected, the content of EOs among populations ranged from 3.16 to 5% (v/w). The lowest and highest EO content was determined in Ghayen (P2) and Fars (P8) populations, respectively (Table 1). Similarly, the percentage of EO in ajwain samples has been reported from Pakistan 3.5–5.2%31, India 2–4%4,32, and Iran 2–6%5,33,34,35. EO yield may vary in plants depending on species, quality (chemotype of the plant), condition (fresh or dry), the layout of plant material (e.g., leaf/stem ratio), harvest time, and also extraction method15,16,36. The EO yield is an important quality factor to bring medicinal plants to the pharmaceutical, and food industries. Seed EO constituents of the 14 ajwain populations and chromatograms are shown in Table 1 and Fig. S1. In this study, eleven constituents were identified in all 14 populations, and thymol was the major constituent ranging from 59.92 to 96.4 percent (Fig. S2). Other major constituents were p-cymene (0.55–21.15%), γ-terpinene (0.23–17.78%), and carvacrol (0.41–2.77%) among populations studied. The highest content of thymol (96.4%) and its structural isomer carvacrol (2.77%) were found in the Ghayen population (P2). Additionally, the lowest thymol content was detected in the Isfahan population (P13) (59.92%). The highest (17.78%) and lowest (0.23%) γ-terpinene content was found in the Isfahan (P13) and Ghayen (P2) populations, respectively. The Birjand population (P3) displayed the highest p-cymene content (21.15%) and (P2) showed the lowest content (0.55%).Table 1 The essential oil composition of the fourteen Trachyspermum ammi populations.Full size tableThe GC–MS spectra obtained from the Hamedan population (P7) are represented in the graphical diagram in Fig. 1. According to our results, the Ghayen population (P2) has the highest levels of thymol and carvacrol and lowest levels of p-cymene and γ-terpinene. So, a higher rate of precursors (γ-terpinene and p-cymene) to final products (thymol/carvacrol) can be converted in isolated EO35. According to the biosynthetic pathway, γ-terpinene precursor converts to thymol and carvacrol during the developmental stages37.Figure 1Represent of graphical design of the present research.Full size imageIn this context, EO compositions of ajwain have been reported from various geographical areas. According to the chemical composition of ajwain oils, major constituents of thymol, γ-terpinene, and p-cymene11,12,33,35 carvone, limonene, and dillapiole13 and carvacrol and p-cymene14 have been documented. Up to now, the high-thymol content populations from Iran were between 34 to 55%33 48.8 to 61.435, and 65.411. However, no chemotype of the plant EO has been reported with a very high percentage of thymol ( > 90%). Thymol and carvacrol percentages of seed EO of 14 populations are shown in Fig. 2. As can be seen in this figure, populations P2 and P8 have the highest thymol content (more than 90% of EO). The presence of a high percentage of thymol in the P8 and P2 can be industrially valuable. Chemotypes are named based on the main constituents in EO within single botanical species38. Normally ajwain oils on the market are those rich in thymol and/or carvacrol with strong antibacterial properties and high antioxidant potential. High purity thymol is interested in the market and will not have the subsequent purification costs. Therefore, chemotypes P2 and P8 with a high percentage of thymol 96.4. 90.57% can be significant respectively.Figure 2Thymol + carvacrol (%) in EO in studied populations. Chemotype determined with hierarchical cluster analysis (HCA).Full size imageEstimation of phyto-constituents of extractSignificant differences were obtained among the population for total phenolic (TPC), total flavonoid (TFC), and total coumarin contents (TCC) (P ≤ 0.01) (Table 2). Natural phenolic compounds are including simple phenolics, phenolic acids, flavonoids, coumarins, tannins, stilbenes, curcuminoids, lignans, quinones, and others39. Phenolic compounds and flavonoids are major bioactive components in medicinal plants and thus can comprise an essential part of the human diet40. The present study assessed the total phenolic, flavonoid, and coumarin contents of ajwain populations, and the results are presented in Fig. 3A–C. Up to now, no studies have reported total phenol, flavonoid, and coumarin contents of Iranian ajwain populations.
Table 2 Analysis of variance for nine phytochemical traits in fourteen populations of Trachyspermum ammi.Full size tableFigure 3Phyto-constituents analysis of seed samples of 14 studied populations of Trachyspermum ammi (A); total phenolic content (TPC) as mg Gallic acid/g DW equivalent. (B) Total flavonoid content (TFC) quantified based on mg Quercetin/g DW. (C) Total coumarin (mg Coumarin E/g DW).Full size imageTotal phenol content (TPC)The total phenolic content in the evaluated extracts varied from 26.91 (P13) in the Isfahan population to 43.20 (P2) mg GAE/g DW in the Ghayen population, Results demonstrated that TPC in the populations varied as the following the order P2  > P10  > P8  > P1  > P11  > P14  > P6, P9  > P3, P5  > P4  > P7  > P12  > P13 (Fig. 3A). In the few evaluable sources, the total phenolic content of ajwain seeds extracted with CHCl3: MeOH (1: 2) solvent was 69 mg/g DW41. In the present study, the highest phenol content (43.2 mg GAE/g DW) was recorded in the P2 population. The difference in TPC with the available report may be due to genetic diversity and differences in extraction methods. According to the presence of apolar thymol in the seed structure, a combination of polar and non-polar solvents to extract compounds may optimize the extraction performance. Various environmental conditions in different places influence the content and metabolic profile of phenolic compounds in plant populations. It seems that high temperature and high UV radiation levels, and differences in genotypes are the reasons why the Isfahan population has a high content of TPC15,16.Total flavonoid content (TFC)Analysis of variance showed a significant difference in TFC content at levels P ≤ 0.01. The total flavonoid contents ranged from 4.45 (P7) in the Hamedan population to 8.03 (P8) mg QE/g DW in the Fars population. P6 and P10 with 7.38 mg QE/g DW were also among the high content TFC populations (Fig. 3B). It seems that the reason for the lack of total flavonoids in Hamedan is due genetic differences and the low temperature of this region compared to other regions. Also, the reason for the high level of flavonoids in the Fars population may be due to genetic differences and high temperatures during the growing period. It has been reported that seeds and spurts of ajwain contain 0.58 and 1.15 mg/ g FW of TFC respectively42. Also, TFC of methanolic extract of Anethum graveolens L. (dill) seeds from the Apiaceae family have been reported to be 5.07 (mg QE /g)43. Flavonoid accumulation with many protective roles may be influenced by the combination of genetics (i.e., adaptation to local conditions) and environmental effects (i.e., phenotypic plasticity)44,45. Flavonoid accumulation rates among geographically different ajwain populations concerning climate can be correlated positively with temperature and UV-B radiation and negatively with precipitation (Chalker-Scott, 1999; Koski and Ashman, 2015).Total coumarin content (TCC)The TCC content of the T. ammi populations examined ranges from 0.079 (P12) to 0.26 (P1) mg coumarin equivalent to dry weight. The highest coumarin content was obtained from the methanolic extract of Kalat (P1) (0.260 mg CE/g DW) and the lowest value of coumarin was recorded for the population of Ardabil (Fig. 3C). Seed coumarin levels in populations can result from genetic and environmental differences. It seems that coumarin accumulation is decreased due to the coolness condition in Ardabil city during the seed maturation stage. Ajwain is a coumarin-rich source of coumarins (umbelliferone, scopoletin, xanthotoxin, bergapten) mostly found in its sprouts46. However, no literature source was found to report the amount of total coumarin in ajwain seeds. These compounds have valuable medicinal properties, including edema reduction and possible anticancer activity47 Furthermore, they are widely used as a flavoring in foods and pastries. Human exposure to coumarin from the diet has been calculated to be around 0.02 mg/kg/day and its maximum daily intake was estimated to be 0.07 mg/kg BW/day48.Free radical scavenging effects and antioxidant activity of essential oils and extractsThe antioxidant activities of EOs and extracts were assessed using the DPPH, FRAP free-radical scavenging, and total antioxidant capacity (TAC) assays (Fig. 4A–C).Figure 4Antioxidant activities of methanolic extracts and essential oils obtained from Trachyspermum ammi seed populations and seven antioxidant standards (A); Antioxidant activity (DPPH) IC50 (µg/ml) (B); antioxidant activity (FRAP) quantified by µmol Fe+2/g DW (C); total antioxidant capacity (TAC) quantified by mg Ascorbic acid equivalent (AAE).Full size imageIn the DPPH assay, the samples were capable to decrease the DPPH free radical to evaluate their in vitro antioxidant activity. Analysis of variance on DPPH IC50 showed a significant difference in antioxidant activity of EOs and extracts among populations (P  BHT  > RU. Also, this value ranged from 8.3 to 16.6 among EO samples with the highest value in P2. TCA values in extracts were recorded in the range of 1.83–4.59 with the highest value obtained in P11. Other detailed information is shown in Fig. 4C.Antibacterial activityThe antibacterial activity of ajwain EOs was evaluated against two antibiotic resistance bacteria and their ability was compared with Cefixime as a standard. In the present study, we tried to use both gram-positive bacteria and gram-negative bacteria as samples. Staphylococcus aureus is a gram-positive pathogenic and antibiotic-resistant bacteria. It is also one of the most common causes of nosocomial infections. Also, Escherichia coli is available and inexpensive, and easily cultured in the laboratory. It is one of the most common causes of urinary tract infections. Gram-negative bacteria are also resistant to antibiotics and are an important species in the field of microbiology. One of the main problems in the field of microbiology is the resistance of microbes to antibiotics and so introducing new antibiotics is necessary53. The reasons for using Cefixime in the present study are due to its widely used, great therapeutic power, and effectiveness against a wide range of microbes.In this study, EOs exhibited bacteriostatic activities against S. aureus (0.06–64 µg/mL) and E. coli (1–64 µg/mL) (Table 3). High thymol content EO (P2) showed high antibacterial activity (MIC = 0.06 µg/mL) against S. aureus. Also, the EO from the Isfahan population (P13) showed the lowest antibacterial activity with the highest MIC value (64 µg/mL). In the present study, the mean MIC was not significantly different on gram-negative and positive bacteria, and populations with high thymol had a high antibacterial ability, indicating the antibacterial effects of thymol. Some researchers have evaluated the antimicrobial activity of ajwain oil14,54,55. Thymol and carvacrol were found to be more effective in killing bacteria3,4,5,6,7,9. The antibacterial properties of natural products, such as essential oils and their components, are widely explored by both industrial and academic fields56. The antibacterial activity of the EOs is dependent on the composition and concentration, type, and dose of the target microorganism57. The high antibacterial potential of cumin essential oil compared to Ferula essential oil has already been identified due to the high ratio of phenolic monoterpene compounds to other monoterpenes58. It seems that the antibacterial effects of C. copticum are also mainly due to the presence of phenolic monoterpenes such as thymol, carvacrol, p-cymene, and γ-terpinene. Therefore, ajwain EO can be used as a natural agent with antibacterial properties in the food industry and the treatment of infectious diseases, especially antibiotic-resistant strains.Table 3 Minimal Inhibitory Concentrations (MIC) essential oil Iranian 14 populations of Trachyspermum ammi against Escherichia coli and Staphylococcus aureus.Full size tableHierarchical cluster analysis (HCA) of essential oil constituentsHCA was performed by using the 11 identified compounds and 14 populations (Fig. 5A). All used populations were divided into two clusters; Cluster I included P4, P6, P7, P10, P11, P12, P13, and P14 and cluster II consist of P1, P2, P5, P8, and P9 samples. In cluster I the major constituents were thymol (59.92–72.86), p-cymene (15.66–21.15), and γ-terpinene (10.22–17.78). In the second cluster thymol (80.09–96.4) and carvacrol (0.5–2.77) were the major constituents. Cluster analysis can classify studied populations into several groups, according to the chemical composition by ‘magnifying’ their similarities59. Forasmuch as, plant sources from environmentally different origins led to the emergence of new chemotypes to baring domestication and cultivation to obtain uniform chemical plants along with appropriate agricultural features60.Figure 5(A) Heat-map diagram of two-way hierarchical cluster analysis (HCA) of fourteen Trachyspermum ammi populations based on 11 essential oil constituents quantified by GC and GC–MS. Blue color with a great positive share and red color with a great negative share affects cluster formation. (B) Principal component analysis (PCA) based on EO constituents. (C) PCA is based on all studied traits. (D) PCA is based on all studied traits according to populations.Full size imagePrincipal component analysis (PCA)Principal component analysis (PCA) is one of the multivariate statistical techniques used to explain differentiation between populations and to obtain more information on the variables that mainly influence the population’s similarities and differences61. The PCA was performed to identify the most significant variables in the data set (Fig. 5B). The same data set (14 population × 11 components) was used in this section. The PCA showed two components with explain 83.3% of the total variance. The first principal component (PC1) had the most portion of variance (74.5%) which was given by compounds such as γ-Terpinene, α-pinene, α-Thujene, p-cymene, and limonene. The second component (PC2), explaining 8.8% of the total variance, consisted of compounds thymol, carvacrol, and 1, 8-cineol (Fig. 6). The results of PCA agreed with those of the cluster analysis the populations similarly were divided into two distinct groups including high thymol/carvacrol and high thymol/p-cymene/γ-terpinene groups (Fig. 5B). Heat map analyses were drowned to determine how constituents effect on clustering. Based on heat map analysis samples were well-classified.Figure 6Correlation between 24 traits on the studied Trachyspermum ammi populations: TPC: Total phenolic content, TFC: Total flavonoid content, TCC: Total coumarin, EO: Essential Oil yield, TSW: One thousand seed weight (g), MIC: minimum inhibitory concentration, Ec: E. coli, MIC: minimum inhibitory concentration, Sa: S. aureus, DPPH Ext.: DPPH assay Extract is expressed as IC50 index, DPPH EO: DPPH assay EO is expressed as IC50 index, FRAP Ext.: FRAP assay Extract, FRAP EO: FRAP assay Essential oil, TAC Ext: The total antioxidant capacity Extract, TAC EO: The total antioxidant capacity Essential oil.Full size imageAlso, in the analysis of the principal factors (PCA) between all the evaluated traits in the populations, the first principal factor (PC1) showed 53.8% and the second principal factor (PC2) 14.7% of the variance. This analysis determined the principal component, correlation of traits, and their relationship with populations. Accordingly, traits with positive arrows show a positive correlation and two traits with non-directional arrows show a negative correlation. Accordingly, thymol and carvacrol have a high correlation with antioxidant properties and this property is correlated with populations of chemotype 1 (P1, P2, P5, P8, P9). Other relationships and details correlations are shown in Fig. 5C, D.CorrelationSimple correlation estimated the relationship between variables. Simple correlations between 24 studied traits in the present study are shown in Fig. 6. Thymol as the major constituent of EOs showed a high positive correlation with TPC (0.71), carvacrol (0.64), FRAP EO (0.85), and FRAP ext. (0.66). Thymol also had a significant negative correlation with Mic EO (-0.74), Mic Sa (-0.69), α-Thujene (-0.84), α-Pinene (-0.77), β-Pinene (-0.75), β-Myrcene (-0.9), α-Terpinene (-0.85), p-Cymene (-0.98), Limonene (-0.89), γ-Terpinene (-0.97). TPC had a positive correlation with TFC, thymol, carvacrol, FRAP Ext., TAC Ext., and a significant negative correlation with DPPH Ext. The antioxidant methods in extracts DPPH50 vs FRAP (-0.8), DPPH50 vs TAC (-0.67) and FRAP vs TAC (0.59) were highly correlated. Similarly, in estimating the antioxidant activity of essential oil DPPH50 vs FRAP (-0.79), DPPH50 vs TAC (-0.48), and FRAP vs TAC Ext (0.55) were highly correlated. Also, the high correlation of all antioxidant methods with thymol can explain its positive effect on the antioxidant activity of the extracts and EOs. The correlations found between each of the traits can be very important in breeding programs. More

Andela, N. et al. A human-driven decline in global burned area. Science 356, 1356 (2017).CAS 
Article 

Google Scholar 
Sloan, S., Locatelli, B., Wooster, M. J. & Gaveau, D. L. A. Fire activity in Borneo driven by industrial land conversion and drought during El Niño periods, 1982–2010. Glob. Environ. Change 47, 95–109 (2017).Article 

Google Scholar 
Kelley, D. I. et al. How contemporary bioclimatic and human controls change global fire regimes. Nat. Clim. Change 9, 690–96 (2019).Article 

Google Scholar 
Jolly, W. M. et al. Climate-induced variations in global wildfire danger from 1979 to 2013. Nat. Commun. 6, 7537 (2015).Ward, D. S., Shevliakova, E., Malyshev, S. & Rabin, S. Trends and variability of global fire emissions due to historical anthropogenic activities. Glob. Biogeochem. Cycles 32, 122–42 (2018).CAS 
Article 

Google Scholar 
Earl, N. & Simmonds, I. Spatial and temporal variability and trends in 2001–2016 global fire activity. J. Geophys. Res. Atmos. 123, 2524–36 (2018).Article 

Google Scholar 
Giglio, L., Randerson, J. T. & van der Werf, G. R. Analysis of daily, monthly, and annual burned area using the fourth-generation Global Fire Emissions Database (GFED4). J. Geophys. Res. Biogeosci. 118, 317–28 (2013).Article 

Google Scholar 
Doerr, S. H. & Santín, C. Global trends in wildfire and its impacts: perceptions versus realities in a changing world. Philos. Trans. R. Soc. B Biol. Sci. 371, 20150345 (2016).Article 

Google Scholar 
van Lierop, P., Lindquist, E., Sathyapala, S. & Franceschini, G. Global forest area disturbance from fire, insect pests, diseases and severe weather events. Forest Ecol. Manag. 352, 78–88 (2015).Article 

Google Scholar 
Zheng, B. et al. Increasing forest fire emissions despite the decline in global burned area. Sci. Adv. 7, eabh2646 (2021).Article 

Google Scholar 
Andela, N. & van der Werf, G. R. Recent trends in African fires driven by cropland expansion and El Niño to La Niña transition. Nat. Clim. Change 4, 791–95 (2014).Article 

Google Scholar 
Van der Werf, G. R. et al. Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires (1997–2009). Atmos. Chem. Phys. 10, 11707–35 (2010).Article 
CAS 

Google Scholar 
Balch, J. K. et al. Negative fire feedback in a transitional forest of southeastern Amazonia. Glob. Change Biol. 14, 2276–87 (2008).Article 

Google Scholar 
Cochrane, M. A. & Laurance, W. F. Synergisms among fire, land use, and climate change in the Amazon. Ambio 37, 522–27 (2008).Article 

Google Scholar 
Gaveau, D. L. A. et al. Major atmospheric emissions from peat fires in Southeast Asia during non-drought years: evidence from the 2013 Sumatran fires. Sci. Rep. 4, 6112 (2014).Vadrevu, K. P. et al. Trends in vegetation fires in South and Southeast Asian countries. Sci. Rep. 9, 7422 (2019).Article 
CAS 

Google Scholar 
Sloan, S., Tacconi, L. & Cattau, M. E. Fire prevention in managed landscapes: recent successes and challenges in Indonesia. Mitig. Adapt. Strateg. Glob. Change 26, Article 32 (2021).Article 

Google Scholar 
Gaveau, D. L. A., Descales, A., Salim, M. A., Shields, D. & Sloan, S. Refined burned-area mapping protocol using Sentinel-2 data increases estimate of 2019 Indonesian burning. Earth Syst. Sci. Data, https://doi.org/10.5194/essd-2021-113, (2021).Randerson, J. T., Chen, Y., van der Werf, G. R., Rogers, B. M. & Morton, D. C. Global burned area and biomass burning emissions from small fires. J. Geophys. Res. Biogeosci. 117, G04012 (2012).Article 
CAS 

Google Scholar 
Field, R. D., van der Werf, G. R. & Shen, S. S. P. Human amplification of drought-induced biomass burning in Indonesia since 1960. Nat. Geosci. 2, 185–88 (2009).CAS 
Article 

Google Scholar 
Huijnen, V. et al. Fire carbon emissions over maritime Southeast Asia in 2015 largest since 1997. Sci. Rep. 6, 26886 (2016).CAS 
Article 

Google Scholar 
Tacconi, L. Preventing fires and haze in Southeast Asia. Nat. Clim. Change 6, 640–43 (2016).Article 

Google Scholar 
Koplitz, S. N. et al. Public health impacts of the severe haze in Equatorial Asia in September–October 2015: demonstration of a new framework for informing fire management strategies to reduce downwind smoke exposure. Environ. Res. Lett. 11, 094023 (2016).Article 

Google Scholar 
Kiely, L. et al. Air quality and health impacts of vegetation and peat fires in Equatorial Asia during 2004–2015. Environ. Res. Lett.15, 094054 (2020).Article 

Google Scholar 
Crippa, P. et al. Population exposure to hazardous air quality due to the 2015 fires in Equatorial Asia. Sci. Rep. 6, 37074 (2016).CAS 
Article 

Google Scholar 
Glauber, A. J. & Gunawan, I. The Cost of Fire: An Economic Analysis of Indonesia’s 2015 Fire Crisis. (The World Bank, Washington, D.C., (2016).Tan, Z. D., Carrasco, L. R. & Taylor, D. Spatial correlates of forest and land fires in Indonesia. Int. J. Wildland Fire 29, 1088–99 (2020).Article 

Google Scholar 
Marlier, M. E. et al. Fire emissions and regional air quality impacts from fires in oil palm, timber, and logging concessions in Indonesia. Environ. Res. Lett. 10, 085005 (2015).Article 
CAS 

Google Scholar 
Vetrita, Y. & Cochrane, M. A. Fire frequency and related land-use and land-cover changes in Indonesia’s peatlands. Remote Sens. 12, 5 (2020).Nikonovas, T., Spessa, A., Doerr, S. H., Clay, G. D. & Mezbahuddin, S. Near-complete loss of fire-resistant primary tropical forest cover in Sumatra and Kalimantan. Commun. Earth Environ. 1, 65 (2020).Article 

Google Scholar 
Field, R. Biomass burning in Indonesia: Signs of Progress in 2019?, http://www.columbia.edu/~rf2426/index_files/20200128.Field.GSFC.NoOz.pdf, January, NASA Goddard Space Flight Center, (2019).Watts, J. et al. Incentivising compliance: evaluating the effectiveness of targeted village incentives for reducing forest and peat fires. Forest Policy Econ. 108, 101956 (2019).Wijedasa, L. et al. Carbon emissions from peat forests will continue to increase despite emission-reduction schemes. Glob. Change Biol. 24, 4598–613 (2018).Article 

Google Scholar 
Sloan, S., Meyfroidt, P., Rudel, T. K. & Bongers, F. & Chazdon Robin, L. The forest transformation: Planted tree cover and regional dynamics of tree gains and losses. Glob. Environ. Change 59, 101988 (2019).Article 

Google Scholar 
Albar, I., Jaya, I. N. S., Saharjo, B. H., Kuncahyo, B. & Vadrevu, K. P. Spatio-temporal analysis of land and forest fires in Indonesia using MODIS active fire dataset, in Land-Atmospheric Research Applications in South and Southeast Asia (eds K P Vadrevu et al.), p. 105-27 (Springer International Publishing, 2018).Miettinen, J., Shi, C. & Liew, S. C. Fire distribution in Peninsular Malaysia, Sumatra and Borneo in 2015 with special emphasis on peatland fires. Environ. Manage. 60, 747–57 (2017).Article 

Google Scholar 
Fanin, T. & van der Werf, G. R. Precipitation–fire linkages in Indonesia (1997–2015). Biogeosciences 14, 3995–4008 (2017).Article 

Google Scholar 
Wiggins, E. B. et al. Smoke radiocarbon measurements from Indonesian fires provide evidence for burning of millennia-aged peat. Proc. Natl. Acad. Sci. USA 115, 12419 (2018).CAS 
Article 

Google Scholar 
Page, S. E. et al. The amount of carbon released from peat and forest fires in Indonesia during 1997. Nature 420, 61–65 (2002).CAS 
Article 

Google Scholar 
Lohberger, S., Stängel, M., Atwood, E. C. & Siegert, F. Spatial evaluation of Indonesia’s 2015 fire-affected area and estimated carbon emissions using Sentinel-1. Glob. Change Biol. 24, 644–54 (2018).Article 

Google Scholar 
van der Werf, G. R. et al. Global fire emissions estimates during 1997–2016. Earth Syst. Sci. Data 9, 697–720 (2017).Article 

Google Scholar 
Field, R. D. et al. Indonesian fire activity and smoke pollution in 2015 show persistent nonlinear sensitivity to El Niño-induced drought. Proc. Natl Acad. Sci. USA 113, 9204–09 (2016).CAS 
Article 

Google Scholar 
Austin, K. G. et al. Shifting patterns of oil palm driven deforestation in Indonesia and implications for zero-deforestation commitments. Land Use Policy 69, 41–48 (2017).Article 

Google Scholar 
Pan, X., Chin, M., Ichoku, C. & Field, R. Connecting Indonesian fires and drought with the type of El Niño and phase of the Indian Ocean Dipole during 1979–2016. J. Geophys. Res. Atmos. 123, (2018).van der Werf, G. R. et al. Climate regulation of fire emissions and deforestation in Equatorial Asia. Proc. Natl Acad. Sci. USA 105, 20350–55 (2008).Article 

Google Scholar 
Wooster, M. J., Roberts, G., Perry, G. L. W. & Kaufman, Y. J. Retrieval of biomass combustion rates and totals from fire radiative power observations: FRP derivation and calibration relationships between biomass consumption and fire radiative energy release. J. Geophys. Rese. Atmos. 110, (2005).Spessa, A. et al. Seasonal forecasting of fires over Kalimantan, Indonesia. Nat. Hazards Earth Syst. Sci. 15, 429–42 (2015).Article 

Google Scholar 
Siegert, F., Ruecker, G., Hinrichs, A. & Hoffmann, A. A. Increased damage from fires in logged forests during droughts caused by El Niño. Nature 414, 437–40 (2001).CAS 
Article 

Google Scholar 
Fernandes, K. et al. Heightened fire probability in Indonesia in non-drought conditions: the effect of increasing temperatures. Environ. Res. Lett. 12, 054002 (2017).Article 

Google Scholar 
Herawati, H. & Santoso, H. Tropical forest susceptibility to and risk of fire under changing climate: a review of fire nature, policy and institutions in Indonesia. Forest Policy Econ. 13, 227–33 (2011).Article 

Google Scholar 
Nepstad, D. et al. Slowing Amazon deforestation through public policy and interventions in beef and soy supply chains. Science 344, 1118–23 (2014).CAS 
Article 

Google Scholar 
Dennis, R. A Review of Fire Projects In Indonesia, 1982-1998. (CIFOR, Bogor, Indonesia, 1999).de Groot, W. J., Field, R. D., Brady, M. A., Roswintiarti, O. & Mohamad, M. Development of the Indonesian and Malaysian fire danger rating systems. Mitig. Adapt. Strateg. Glob. Change 12, 165 (2006).Article 

Google Scholar 
Clough, Y. et al. Land-use choices follow profitability at the expense of ecological functions in Indonesian smallholder landscapes. Nat. Commun. 7, 13137 (2016).CAS 
Article 

Google Scholar 
Bissonnette, J.-F. & De Koninck, R. The return of the plantation? Historical and contemporary trends in the relation between plantations and smallholdings in Southeast Asia. J. Peasant Stud. 44, 918–38 (2017).Article 

Google Scholar 
Gaveau, D. L. A. et al. Slowing deforestation in Indonesia follows declining oil palm expansion and lower oil prices. PLOS ONE 17, e0266178 (2022).Svatoňová, T., Herák, D. & Kabutey, A. Financial profitability and sensitivity analysis of palm oil plantation in Indonesia. Acta Univ. Agric. Silvic. Mendelianae Brunensis 63, 1365–73 (2015).Article 

Google Scholar 
Gaveau, D. L. A. et al. Rapid conversions and avoided deforestation: examining four decades of industrial plantation expansion in Borneo. Scientific Reports 6, (2016).Simamora, A. P. Govt says no to converting peatland into plantations, The Jakarta Post. August (2010).Satriastanti, F. E. Jokowi bans new oil palm and mining concessions, Mongabay.com April (2016).Sloan, S., Edwards, D. P. & Laurance, W. F. Does Indonesia’s REDD+ moratorium on new concessions spare imminently-threatened forests? Conserv. Lett. 5, 222–31 (2012).Article 

Google Scholar 
Busch, J. et al. Reductions in emissions from deforestation from Indonesia’s moratorium on new oil palm, timber, and logging concessions. Proc. Natl Acad Sci USA 112, 1328–33 (2015).CAS 
Article 

Google Scholar 
Forsyth, T. Public concerns about transboundary haze: a comparison of Indonesia, Singapore, and Malaysia. Glob. Environ. Change 25, 76–86 (2014).Article 

Google Scholar 
Carbon Conservation. Fire Free Village Program – Review 2017. (Carbon Conservation, Singapore, (2017).Gaveau, D. L. A. et al. Overlapping land claims limit the use of satellites to monitor no-deforestation committments and no-burning compliance. Conserv. Lett. 10, 257–64 (2017).Article 

Google Scholar 
EarthData. MODIS Collection 6 Active-Fire Detections standard scientific data (MCD14ML), NASA EarthData, https://earthdata.nasa.gov/firms (2019).Giglio, L., Schroeder, W. & Justice, C. O. The Collection 6 MODIS active fire detection algorithm and fire products. Remote Sens. Environ. 178, 31–41 (2016).Article 

Google Scholar 
Sloan, S., Cattau, M.E. Discrete Fire Events, their Severity, and their Ignitions, as Derived from MODIS MCD 14ML Active-Fire Detection Data for Indonesia, 2002-2019. Sean Sloan and Megan E. Cattau, Datadryad.org. (2022).Cattau, M. E. et al. Sources of anthropogenic fire ignitions on the peat-swamp landscape in Kalimantan, Indonesia. Glob. Environ. Change 39, 205–19 (2016).Article 

Google Scholar 
Wooster, M. J., Perry, G. L. W. & Zoumas, A. Fire, drought and El Niño relationships on Borneo during the pre-MODIS era (1980–2000). Biogeosciences 9, 317–40 (2012).Article 

Google Scholar 
Tansey, K., Beston, J., Hoscilo, A., Page, S. E. & Paredes Hernández, C. U. Relationship between MODIS fire hot spot count and burned area in a degraded tropical peat swamp forest in Central Kalimantan, Indonesia. J. Geophys. Res. 113, (2008).Oom, D., Silva, P. C., Bistinas, I. & Pereira, J. M. C. Highlighting biome-specific sensitivity of fire size distributions to time-gap parameter using a new algorithm for fire event individuation. Remote Sens. 8, 663 (2016).Schroeder, W. et al. Validation of GOES and MODIS active fire detection products using ASTER and ETM plus data. Remote Sens. Environ. 112, 2711–26 (2008).Article 

Google Scholar 
Hantson, S., Padilla, M., Corti, D. & Chuvieco, E. Strengths and weaknesses of MODIS hotspots to characterize global fire occurrence. Remote Sens. Environ. 131, 152–59 (2013).Article 

Google Scholar 
Tanpipat, V., Honda, K. & Nuchaiya, P. MODIS hotspot validation over Thailand. Remote Sens. 1, 1043–54 (2009).Article 

Google Scholar 
Liew, S. C., Shen, C., Low, J., Lim, A. & Kwoh, L. K. The 24th Asian Conference on Remote Sensing and 2003 International Symposium on Remote Sensing (ACRS2003). p. 671-73 (Asian Association on Remote Sensing), November 3–7.Fornacca, D., Ren, G. & Xiao, W. Performance of three MODIS fire products (MCD45A1, MCD64A1, MCD14ML), and ESA Fire_CCI in a mountainous area of northwest Yunnan, China, characterized by frequent small fires. Remote Sens. 9, 1131 (2017).Article 

Google Scholar 
Schroeder, W., Oliva, P., Giglio, L. & Csiszar, I. A. The New VIIRS 375m active fire detection data product: algorithm description and initial assessment. Remote Sens. Environ. 143, 85–96 (2014).Article 

Google Scholar 
Giglio, L., Boschetti, L., Roy, D. P., Humber, M. L. & Justice, C. O. The Collection 6 MODIS burned area mapping algorithm and product. Remote Sens. Environ. 217, 72–85 (2018).Article 

Google Scholar 
Roy, D. P., Boschetti, L., Justice, C. O. & Ju, J. The Collection 5 MODIS burned area product — Global evaluation by comparison with the MODIS active fire product. Remote Sens. Environ. 112, 3690–707 (2008).Article 

Google Scholar 
Miettinen, J., Langner, A. & Siegert, F. Burnt area estimation for the year 2005 in Borneo using multi-resolution satellite imagery. Int. J. Wildland Fire 16, 45–53 (2007).Luo, R., Hui, D., Miao, N., Liang, C. & Wells, N. Global relationship of fire occurrence and fire intensity: a test of intermediate fire occurrence-intensity hypothesis. J. Geophys. Res. Biogeosci. 122, 1123–36 (2017).Article 

Google Scholar 
Andela, N. et al. The Global Fire Atlas of individual fire size, duration, speed, and direction. Earth Syst. Sci. Data 11, 529–52 (2019).Article 

Google Scholar 
Andela, N., Morton, D. C., Giglio, L. & Randerson, J. T. Global Fire Atlas with Characteristics of Individual Fires, 2003-2016, ORNL Distributed Active Archive Center, https://doi.org/10.3334/ORNLDAAC/1642, https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=1642 (2019).Field, R. D. & Shen, S. S. P. Predictability of carbon emissions from biomass burning in Indonesia from 1997 to 2006. J. Geophys. Res. Biogeosci. 113, G04024 (2008).Article 
CAS 

Google Scholar 
Fuller, D. O. & Murphy, K. The ENSO-fire dynamic in insular Southeast Asia. Clim. Change 74, 435–55 (2006).Article 

Google Scholar 
Field, R. D. et al. Development of a global fire weather database. Nat. Hazards Earth Syst. Sci. 15, 1407–23 (2015).Article 

Google Scholar 
Huffman, G. J. GPM IMERG Final Precipitation gridded data, L3 1 month 0.1 degree x 0.1 degree, version 06B. NASA Precipitation Processing System, Goddard Earth Sciences Data and Information Services Center (GES DISC). https://storm-pps.gsfc.nasa.gov/storm/; https://pmm.nasa.gov/data-access/downloads/gpm (2019).Huffman, G. J. et al. The TRMM Multisatellite Precipitation Analysis (TMPA): quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeorol. 8, 38–55 (2007).Article 

Google Scholar 
Funk, C. et al. The climate hazards infrared precipitation with stations—a new environmental record for monitoring extremes. Sci. Data 2, 150066 (2015).Article 

Google Scholar 
Hsu, J., Huang, W.-R., Liu, P.-Y. & Li, X. Validation of CHIRPS precipitation estimates over taiwan at multiple timescales. Remote Sens. 13, 254 (2021).Rozante, J. R., Vila, D. A., Barboza Chiquetto, J., Fernandes, A. D. A. & Souza Alvim, D. Evaluation of TRMM/GPM blended daily products over Brazil. Remote Sens. 10, 882 (2018).Prakash, S., Mitra, A. K., Pai, D. S. & AghaKouchak, A. From TRMM to GPM: how well can heavy rainfall be detected from space? Adv. Water Resour. 88, 1–7 (2016).Article 

Google Scholar 
Ma, Q. et al. Performance evaluation and correction of precipitation data using the 20-year IMERG and TMPA precipitation products in diverse subregions of China. Atmos. Res. 249, 105304 (2021).Article 

Google Scholar 
Nwachukwu, P. N., Satge, F., Yacoubi, S. E., Pinel, S. & Bonnet, M.-P. From TRMM to GPM: how reliable are satellite-based precipitation data across Nigeria? Remote Sens. 12, 3964 (2020).Popovych, V. F. & Dunaieva, I. A. Assessment of the GPM IMERG and CHIRPS precipitation estimations for the steppe part of the Crimea. Meteorol. Hydrol. Water Manage 9, (2021).Navarro, A. et al. Assessment of IMERG precipitation estimates over Europe. Remote Sens. 11, 2470 (2019).Dezfuli, A. K. et al. Validation of IMERG precipitation in Africa. J. Hydrometeorol. 18, 2817–25 (2017).Article 

Google Scholar 
Efron, B. & Tibshirani, R. J. An Introduction to the Bootstrap. (Chapman and Hall, Boca Raton, FL, USA, 1993).Pérez-Hoyos, A., Rembold, F., Kerdiles, H. & Gallego, J. Comparison of global land cover datasets for cropland monitoring. Remote Sens. 9, 1118 (2017).ESA. Annual land-cover product, 1992 to 2019/present, based on MERIS 300-m and ancillary SPOT, AVHRR, Sentinel-3 and PROB-V satellite data. European Space Agency (ESA) European Centre for Medium-Range Weather Forecasts (ECMFW) Copernicus Climate Change Service (C3S) Climate Change Initiative (CCI), https://cds.climate.copernicus.eu/cdsapp#!/dataset/satellite-land-cover?tab=overview; http://maps.elie.ucl.ac.be/CCI/viewer/download.php; http://www.esa-landcover-cci.org/ (2020).Defourny, P. Product User Guide and Specification: ICDR Land Cover 2016 to 2019 (Version 2.1.1 of ESA Coperninus Climate Change Intitiative Annual 300-m Land-Cover Classifications). (Universitie Catholique du Lovain, Louvain, Belgium, (2020).Vetrita, Y. & Cochrane, M. A. Annual Burned Area from Landsat, Mawas, Central Kalimantan, Indonesia, 1997-2015, ORNL Distributed Active Archive Center, https://doi.org/10.3334/ORNLDAAC/1708, https://daac.ornl.gov/CMS/guides/Annual_Burned_Area_Maps.html; https://daac.ornl.gov/cgi-bin/dataset_lister.pl?p=33 (2019). More

Transforming our World: The 2030 Agenda for Sustainable Development (UN, 2015).May, R. M. Will a large complex system be stable? Nature 238, 413–414 (1972).CAS 
PubMed 
Article 

Google Scholar 
Pimm, S. L. The complexity and stability of ecosystems. Nature 307, 321–326 (1984).Article 

Google Scholar 
McCann, K. S. The diversity–stability debate. Nature 405, 228–233 (2000).CAS 
PubMed 
Article 

Google Scholar 
Ives, A. R. & Carpenter, S. R. Stability and diversity of ecosystems. Science 317, 58–62 (2007).CAS 
PubMed 
Article 

Google Scholar 
Donohue, I. et al. Navigating the complexity of ecological stability. Ecol. Lett. 19, 1172–1185 (2016).PubMed 
Article 

Google Scholar 
Yachi, S. & Loreau, M. Biodiversity and ecosystem productivity in a fluctuating environment: the insurance hypothesis. Proc. Natl Acad. Sci. USA 96, 1463–1468 (1999).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Isbell, F. et al. Biodiversity increases the resistance of ecosystem productivity to climate extremes. Nature 526, 574–577 (2015).CAS 
PubMed 
Article 

Google Scholar 
Tilman, D., Reich, P. B. & Knops, J. M. Biodiversity and ecosystem stability in a decade-long grassland experiment. Nature 441, 629–632 (2006).CAS 
PubMed 
Article 

Google Scholar 
Loreau, M. et al. Biodiversity as insurance: from concept to measurement and application. Biol. Rev. 96, 2333–2354 (2021).PubMed 
Article 

Google Scholar 
Thibaut, L. M. & Connolly, S. R. Understanding diversity–stability relationships: towards a unified model of portfolio effects. Ecol. Lett. 16, 140–150 (2013).PubMed 
Article 

Google Scholar 
Xu, Q. et al. Consistently positive effect of species diversity on ecosystem, but not population, temporal stability. Ecol. Lett. 24, 2256–2266 (2021).PubMed 
Article 

Google Scholar 
Hector, A. et al. General stabilizing effects of plant diversity on grassland productivity through population asynchrony and overyielding. Ecology 91, 2213–2220 (2010).CAS 
PubMed 
Article 

Google Scholar 
Hautier, Y. et al. Anthropogenic environmental changes affect ecosystem stability via biodiversity. Science 348, 336–340 (2015).CAS 
PubMed 
Article 

Google Scholar 
Cardinale, B. J. et al. Biodiversity loss and its impact on humanity. Nature 486, 59–67 (2012).CAS 
PubMed 
Article 

Google Scholar 
Tilman, D., Isbell, F. & Cowles, J. M. Biodiversity and ecosystem functioning. Annu. Rev. Ecol. Evol. Syst. 45, 471–493 (2014).Article 

Google Scholar 
Isbell, F. et al. Linking the influence and dependence of people on biodiversity across scales. Nature 546, 65–72 (2017).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Gonzalez, A. et al. Scaling-up biodiversity–ecosystem functioning research. Ecol. Lett. 23, 757–776 (2020).PubMed 
PubMed Central 
Article 

Google Scholar 
Hooper, D. U. et al. Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecol. Monogr. 75, 3–35 (2005).Article 

Google Scholar 
Wang, S. & Loreau, M. Ecosystem stability in space: alpha, beta and gamma variability. Ecol. Lett. 17, 891–901 (2014).PubMed 
Article 

Google Scholar 
Wang, S. & Loreau, M. Biodiversity and ecosystem stability across scales in metacommunities. Ecol. Lett. 19, 510–518 (2016).PubMed 
PubMed Central 
Article 

Google Scholar 
Wang, S. et al. Biotic homogenization destabilizes ecosystem functioning by decreasing spatial asynchrony. Ecology 102, e03332 (2021).Zhang, Y., He, N., Loreau, M., Pan, Q. & Han, X. Scale dependence of the diversity–stability relationship in a temperate grassland. J. Ecol. 106, 1227–1285 (2018).PubMed 
Article 

Google Scholar 
Wang, S., Lamy, T., Hallett, L. M. & Loreau, M. Stability and synchrony across ecological hierarchies in heterogeneous metacommunities: linking theory to data. Ecography 42, 1200–1211 (2019).Article 

Google Scholar 
Hautier, Y. et al. General destabilizing effects of eutrophication on grassland productivity at multiple spatial scales. Nat. Commun. 11, 5375 (2020).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Liang, M., Liang, C., Hautier, Y., Wilcox, K. R. & Wang, S. Grazing-induced biodiversity loss impairs grassland ecosystem stability at multiple scales. Ecol. Lett. 24, 2054–2064 (2021).PubMed 
Article 

Google Scholar 
Qiao, X. et al. Spatial asynchrony matters more than alpha stability in stabilizing ecosystem productivity in a large temperate forest region. Glob. Ecol. Biogeogr. 31, 1133–1146 (2022).Article 

Google Scholar 
Catano, C. P., Fristoe, T. S., LaManna, J. A. & Myers, J. A. Local species diversity, beta-diversity and climate influence the regional stability of bird biomass across North America. Proc. R. Soc. B 287, 20192520 (2020).PubMed 
PubMed Central 
Article 

Google Scholar 
Patrick, C. J. et al. Multi‐scale biodiversity drives temporal variability in macrosystems. Front. Ecol. Environ. 19, 47–56 (2021).Article 

Google Scholar 
Wilcox, K. R. et al. Asynchrony among local communities stabilises ecosystem function of metacommunities. Ecol. Lett. 20, 1534–1545 (2017).PubMed 
PubMed Central 
Article 

Google Scholar 
Zhang, Y. et al. Nitrogen addition does not reduce the role of spatial asynchrony in stabilising grassland communities. Ecol. Lett. 22, 563–571 (2019).PubMed 
PubMed Central 
Article 

Google Scholar 
Garcia, R. A., Cabeza, M., Rahbek, C. & Araujo, M. B. Multiple dimensions of climate change and their implications for biodiversity. Science 344, 1247579 (2014).PubMed 
Article 
CAS 

Google Scholar 
García-Palacios, P., Gross, N., Gaitán, J. & Maestre, F. T. Climate mediates the biodiversity–ecosystem stability relationship globally. Proc. Natl Acad. Sci. USA 115, 8400–8405 (2018).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
Hillebrand, H. On the generality of the latitudinal diversity gradient. Am. Nat. 163, 192–211 (2004).PubMed 
Article 

Google Scholar 
Qian, H. & Ricklefs, R. E. A latitudinal gradient in large-scale beta diversity for vascular plants in North America. Ecol. Lett. 10, 737–744 (2007).PubMed 
Article 

Google Scholar 
Kraft, N. J. B. et al. Disentangling the drivers of β diversity along latitudinal and elevational gradients. Science 333, 1755–1758 (2011).CAS 
PubMed 
Article 

Google Scholar 
Ma, Z. et al. Climate warming reduces the temporal stability of plant community biomass production. Nat. Commun. 8, 15378 (2017).PubMed 
PubMed Central 
Article 

Google Scholar 
Song, J. et al. A meta-analysis of 1,119 manipulative experiments on terrestrial carbon-cycling responses to global change. Nat. Ecol. Evol. 3, 1309–1320 (2019).PubMed 
Article 

Google Scholar 
Valencia, E. et al. Synchrony matters more than species richness in plant community stability at a global scale. Proc. Natl Acad. Sci. USA 117, 24345–24351 (2020).CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Gilbert, B. et al. Climate and local environment structure asynchrony and the stability of primary production in grasslands. Glob. Ecol. Biogeogr. 29, 1177–1188 (2020).Article 

Google Scholar 
Hallett, L. M. et al. Biotic mechanisms of community stability shift along a precipitation gradient. Ecology 95, 1693–1700 (2014).PubMed 
Article 

Google Scholar 
Hong, P. et al. Biodiversity promotes ecosystem functioning despite environmental change. Ecol. Lett. 25, 555–569 (2022).PubMed 
Article 

Google Scholar 
. Plant presence and percent cover, RELEASE-2021. NEON (National Ecological Observatory Network) https://doi.org/10.48443/abge-r811 (2021).Barnett, D. T. et al. The plant diversity sampling design for The National Ecological Observatory. Netw. Ecosphere 10, e02603 (2019).
Google Scholar 
Lasky, J. R., Uriarte, M. & Muscarella, R. Synchrony, compensatory dynamics, and the functional trait basis of phenological diversity in a tropical dry forest tree community: effects of rainfall seasonality. Environ. Res. Lett. 11, 115003 (2016).Article 

Google Scholar 
Inchausti, P. & Halley, J. Investigating long-term ecological variability using the global population dynamics database. Science 293, 655–657 (2001).CAS 
PubMed 
Article 

Google Scholar 
Luo, M. et al. The effects of dispersal on spatial synchrony in metapopulations differ by timescale. Oikos 130, 1762–1772 (2021).Article 

Google Scholar 
Pimm, S. L. & Redfearn, A. The variability of population densities. Nature 334, 613–614 (1988).Article 

Google Scholar 
Craven, D. et al. Multiple facets of biodiversity drive the diversity–stability relationship. Nat. Ecol. Evol. 2, 1579–1587 (2018).PubMed 
Article 

Google Scholar 
Peet, R. K., Wentworth, T. R. & White, P. S. A flexible, multipurpose method for recording vegetation composition and structure. Castanea 63, 262–274 (1998).
Google Scholar 
Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G. & Jarvis, A. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25, 1965–1978 (2005).Article 

Google Scholar 
Loreau, M. & de Mazancourt, C. Species synchrony and its drivers: neutral and nonneutral community dynamics in fluctuating environments. Am. Nat. 172, E48–E66 (2008).PubMed 
Article 

Google Scholar 
Lefcheck, J. S. piecewiseSEM: piecewise structural equation modelling inr for ecology, evolution, and systematics. Methods Ecol. Evol. 7, 573–579 (2016).Article 

Google Scholar 
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & R Core Team. nlme: Linear and nonlinear mixed effects models. R package v.3.1–152 https://CRAN.R-project.org/package=nlme (2021).R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2019). More

Beer, C. et al. Terrestrial gross carbon dioxide uptake: global distribution and covariation with climate. Science 329, 834 (2010).Article 

Google Scholar 
Anderegg, W. R. L. et al. Pervasive drought legacies in forest ecosystems and their implications for carbon cycle models. Science 349, 528–532 (2015).Article 

Google Scholar 
Ciais, P. et al. Europe-wide reduction in primary productivity caused by the heat and drought in 2003. Nature 437, 529–533 (2005).Article 

Google Scholar 
Peñuelas, J. et al. Shifting from a fertilization-dominated to a warming-dominated period. Nat. Ecol. Evol. 1, 1438–1445 (2017).Article 

Google Scholar 
Morin, X. et al. Temporal stability in forest productivity increases with tree diversity due to asynchrony in species dynamics. Ecol. Lett. 17, 1526–1535 (2014).Article 

Google Scholar 
Isbell, F. et al. Biodiversity increases the resistance of ecosystem productivity to climate extremes. Nature 526, 574 (2015).Article 

Google Scholar 
De Boeck, H. J. et al. Patterns and drivers of biodiversity–stability relationships under climate extremes. J. Ecol. 106, 890–902 (2018).Article 

Google Scholar 
Grossiord, C. Having the right neighbors: how tree species diversity modulates drought impacts on forests. N. Phytol. 228, 42–49 (2020).Article 

Google Scholar 
O’Brien, M. J. et al. Resistance of tropical seedlings to drought is mediated by neighbourhood diversity. Nat. Ecol. Evol. 1, 1643–1648 (2017).Article 

Google Scholar 
Gazol, A. & Camarero, J. J. Functional diversity enhances silver fir growth resilience to an extreme drought. J. Ecol. 104, 1063–1075 (2016).Article 

Google Scholar 
Pretzsch, H., Schütze, G. & Uhl, E. Resistance of European tree species to drought stress in mixed versus pure forests: evidence of stress release by inter-specific facilitation. Plant Biol. 15, 483–495 (2013).Article 

Google Scholar 
Grossiord, C. et al. Tree diversity does not always improve resistance of forest ecosystems to drought. P. Natl Acad. Sci. USA 111, 14812–14815 (2014).Article 

Google Scholar 
Grossiord, C. et al. Does drought influence the relationship between biodiversity and ecosystem functioning in boreal forests. Ecosystems 17, 394–404 (2014).Article 

Google Scholar 
Loreau, M., Mouquet, N. & Gonzalez, A. Biodiversity as spatial insurance in heterogeneous landscapes. P. Natl Acad. Sci. USA 100, 12765 (2003).Article 

Google Scholar 
Lloret, F. et al. Woody plant richness and NDVI response to drought events in Catalonian (northeastern Spain) forests. Ecology 88, 2270–2279 (2007).Article 

Google Scholar 
He, Q. & Bertness, M. D. Extreme stresses, niches, and positive species interactions along stress gradients. Ecology 95, 1437–1443 (2014).Article 

Google Scholar 
Hafner, B. D. et al. Hydraulic redistribution under moderate drought among English oak, European beech and Norway spruce determined by deuterium isotope labeling in a split-root experiment. Tree Physiol. 37, 950–960 (2017).Article 

Google Scholar 
Forrester, D. I. & Bauhus, J. A review of processes behind diversity–productivity relationships in forests. Curr. For. Rep. 2, 45–61 (2016).Article 

Google Scholar 
Vitali, V., Forrester, D. I. & Bauhus, J. Know your neighbours: drought response of Norway spruce, silver fir and Douglas fir in mixed forests depends on species identity and diversity of tree neighbourhoods. Ecosystems 21, 1215–1229 (2018).Article 

Google Scholar 
The State of the World’s Forests 2020: Forests, Biodiversity and People (FAO and UNEP, 2020).Zhang, J., Fu, B., Stafford-Smith, M., Wang, S. & Zhao, W. Improve forest restoration initiatives to meet Sustainable Development Goal 15. Nat. Ecol. Evol. https://doi.org/10.1038/s41559-020-01332-9 (2020).Schulze, K., Malek, Ž. & Verburg, P. H. Towards better mapping of forest management patterns: a global allocation approach. For. Ecol. Manage. 432, 776–785 (2019).Article 

Google Scholar 
Harris, N. L. et al. Global maps of twenty-first century forest carbon fluxes. Nat. Clim. Change https://doi.org/10.1038/s41558-020-00976-6 (2021).Maxwell, S. L. et al. Area-based conservation in the twenty-first century. Nature 586, 217–227 (2020).Article 

Google Scholar 
Williams, A. P. et al. Large contribution from anthropogenic warming to an emerging North American megadrought. Science 368, 314–318 (2020).Article 

Google Scholar 
Blackman, C. et al. Leaf hydraulic vulnerability is related to conduit dimensions and drought resistance across a diverse range of woody angiosperms. N. Phytol. 188, 1113–1123 (2010).Article 

Google Scholar 
Liang, J. et al. Positive biodiversity–productivity relationship predominant in global forests. Science 354, aaf8957 (2016).Article 

Google Scholar 
Wieczynski, D. J. et al. Climate shapes and shifts functional biodiversity in forests worldwide. P. Natl Acad. Sci. USA 116, 587–592 (2019).Article 

Google Scholar 
Tomppo, E. et al. National Forest Inventories: Pathways for Common Reporting (Springer, 2010).Chirici, G. et al. National Forest Inventories: Contributions to Forest Biodiversity Assessments (Springer, 2011).Magnussen, S., Smith, B. & Uribe, S. National Forest inventories in North America for monitoring forest tree species diversity. Plant Biosyst. 141, 113–122 (2007).Article 

Google Scholar 
Lesiv, M. et al. Global forest management data for 2015 at a 100 m resolution. Sci. Data 9, 199 (2022).Article 

Google Scholar 
Vicente-Serrano, S. M., Beguería, S. & López-Moreno, J. I. A. Multiscalar drought index sensitive to global warming: the Standardized Precipitation Evapotranspiration Index. J. Clim. 23, 1696–1718 (2010).Article 

Google Scholar 
Hansen, M. C. et al. High-resolution global maps of 21st-century forest cover change. Science 342, 850 (2013).Article 

Google Scholar 
Forest Resources Assessment 2015 (FAO, 2015).Lyapustin, A. et al. Scientific impact of MODIS C5 calibration degradation and C6+ improvements. Atmos. Meas. Tech. 7, 4353–4365 (2014).Article 

Google Scholar 
Didan, K. & Brreto, A. NASA MEaSUREs Vegetation Index and Phenology (VIP) Phenology EVI2 Yearly Global 0.05Deg CMG, NASA EOSDIS Land Processes DAAC, https://doi.org/10.5067/MEaSUREs/VIP/VIPPHEN_EVI2.004 (2016).Olson, D. M. et al. Terrestrial Ecoregions of the World: A New Map of Life on Earth: a new global map of terrestrial ecoregions provides an innovative tool for conserving biodiversity. Bioscience 51, 933–938 (2001).Article 

Google Scholar 
Kline, T. J. B. Sample issues, methodological implications, and best practices. Can. J. Behav. Sci. 49, 71–77 (2017).Article 

Google Scholar 
Gourlet-Fleury, S. et al. Tropical forest recovery from logging: a 24 year silvicultural experiment from Central Africa. Phil. Trans. R. Soc. B 368, 20120302 (2013).Article 

Google Scholar 
Obiang, N. L. E. et al. Spatial pattern of central African rainforests can be predicted from average tree size. Oikos 119, 1643–1653 (2010).Article 

Google Scholar 
Plotkin, J. B. et al. Predicting species diversity in tropical forests. P. Natl Acad. Sci. USA 97, 10850–10854 (2000).Article 

Google Scholar 
Graham, M. H. Confronting multicollinearity in ecological multiple regression. Ecology 84, 2809–2815 (2003).Article 

Google Scholar 
Tukey, J. W. Exploratory Data Analysis (Addison-Wesley,1977). More