Human activities pose a major threat to tropical forest biodiversity and ecosystem services. Although the impacts of deforestation are well studied, multiple land-use and land-cover transitions (LULCTs) occur in tropical landscapes, and we do not know how LULCTs differ in their rates or impacts on key ecosystem components. Here, we quantified the impacts of 18 LULCTs on three ecosystem components (biodiversity, carbon, and soil), based on 18 variables collected from 310 sites in the Brazilian Amazon. Across all LULCTs, biodiversity was the most affected ecosystem component, followed by carbon stocks, but the magnitude of change differed widely among LULCTs and individual variables. Forest clearance for pasture was the most prevalent and high-impact transition, but we also identified other LULCTs with high impact but lower prevalence (e.g., forest to agriculture). Our study demonstrates the importance of considering multiple ecosystem components and LULCTs to understand the consequences of human activities in tropical landscapes.Tropical forests host two-thirds of all terrestrial biodiversity (
1), account for one-third of terrestrial productivity and evapotranspiration (
2), and store half of all terrestrial carbon (
3,
4). Despite their global importance, tropical forests are being severely affected by human activities (
5). Deforestation is a key driver of change—more than 100 million ha of primary tropical forests have been converted to agriculture and silviculture in the last 40 y (
6,
7). Many of the remaining primary forests are also degraded. Between 2000 and 2005 at least 20% of tropical forests were selective logged (
8), while other anthropogenic drivers and extreme droughts are increasing forest fires, with 54 million ha burned annually in the 1990s (
9). Even deforested landscapes are changing; agricultural abandonment is a key driver of secondary forest regrowth, and now these forests account for at least half of all tropical forests globally (
10), including 28% of deforested land in the Amazon (
11). Other areas are undergoing agricultural intensification, with pastures being converted to croplands (
12,
13). Hence, many tropical landscapes are now a mosaic of nonforested land uses, regenerating secondary forests, and variably degraded primary forests (
14).Several studies have quantified the different land-use and land-cover transitions (LULCTs) in tropical forest regions, including in the Amazon (e.g., refs.
15 and
16). We know, for example, that conversion of forests to pastures and degradation of primary forests are the main LULCTs in the Amazon (
15,
16), while secondary forest recovery is still very limited compared with deforestation (
11). In addition, the value of human-modified tropical landscapes has been examined in detail for above-ground carbon storage (
17), soil condition (
18), vegetation structure (
19), and biodiversity (
20). The ecological literature shows that conversion transitions, for example deforestation, are expected to cause the greatest impact on forest ecosystems (
21), but degradation of primary forests can also be as harmful as deforestation for biodiversity when it occurs at scale (
22). Other studies discuss how fast biodiversity and carbon can recover in regenerating secondary forests, suggesting, for example, that up to 80% of primary forest tree species could be present in 40-y-old Amazonian secondary forests (
23), while a meta-analysis suggests that, even after a century, plant species diversity does not recover to undisturbed tropical forest levels (
24).Although the scientific community has been building a solid knowledge base on LULCTs in tropical forest regions, the existing studies do not reveal the full extent of changes in landscape condition for four key reasons: First, a compilation of the rates of different types of LULCTs, including deforestation, regeneration, and forest degradation, is still lacking, impeding a comprehensive view of LULCT dynamics that allows a quantitative comparison between all transitions. Second, region-specific studies often focus on one or two ecosystem components (
24–
26) or compare changes with an undisturbed forest baseline rather than evaluating the full range of transitions (
23,
27). Third, meta-analyses have focused on 1) a single ecosystem component (
18), 2) a single type of LULCT (
28), or 3) comparing all changes with an undisturbed baseline, without exploring the transitions that occur between human-modified land uses (
21). Finally, ecological changes have not been linked to prevalence (e.g., measured as annual rates) of the land-cover transitions to date. Quantifying transitions in terms of prevalence and impact on ecosystem properties is a key step toward understanding the relative importance of changes across whole landscapes and could provide evidence-based scenarios for policymakers to decide how to better protect and benefit from tropical forest biodiversity and ecosystem services.Here, we quantified the prevalence and the ecological impacts of 18 LULCTs in the Brazilian Amazon to make a comprehensive assessment of the relative risk of the major LULCTs. We focused on the following questions: 1) What are the rates of LULCTs? 2) What are the impacts of LULCTs on different ecosystem properties (e.g., biodiversity, carbon, and soil)? 3) Which transitions impose the greatest magnitude of change on the ecosystem? Ecological data encompassed 18 variables sampled at 310 sites in two distinct regions of the Brazilian Amazon, which we grouped into three ecological dimensions that reflect policy or management levers: biodiversity (vascular plants, birds, and three invertebrate groups), carbon pools, and soil properties. These ecosystem components were surveyed in seven land-use or land-cover classes: undisturbed primary forest, logged primary forest, logged-and-burned primary forest, old secondary forest (>20 y since abandonment), young secondary forest (≤20 y old), pasture, and mechanized agriculture. The impacts of 18 LULCTs on ecosystem variables were then assessed against the annual LULCT rates estimated between 2006 and 2014 (
29) or 2006 and 2019 (
15), providing an understanding of both the magnitude and extent of changes in land use and land cover in the Brazilian Amazon.
相似文献