A Global Synthesis on Land-Cover Changes in Watersheds Shaping Freshwater Detrital Food Webs

Abstract

Anthropogenic land-cover changes are among the most pressing global threats to both aquatic and terrestrial ecosystems, jeopardizing biodiversity and the critical connections between these systems. Resource flows and trophic interactions intricately link aquatic and terrestrial ecosystems, with terrestrial-derived detritus playing a fundamental role in supporting aquatic food webs. These detrital inputs form essential cross-ecosystem linkages, underpinning key ecological processes and providing vital resources for aquatic communities. Yet, little research has focused on how land-cover changes cascade across this linkage. To better understand how land-cover changes in the watershed influence freshwater detrital food webs, we conducted a meta-analysis of field studies reporting the effects of vegetation changes on freshwater detrital consumers and organic matter decomposition. The results from 144 studies, reporting 1235 comparisons, showed that, overall, land-cover changes in the watershed vegetation, especially through harvest and land-use conversion, have negative effects on aquatic biodiversity and ecosystem processes. These vegetation changes reduced diversity, abundance, and biomass across multiple trophic levels in freshwater detrital food webs. Studies examining multiple organism groups most often observed negative responses across multiple trophic levels, suggesting that these land-cover changes negatively affected multiple detrital food-web components simultaneously. Our results also show that outcomes of restoration of watershed vegetation were context-dependent, and no clear trend of improvement was visible. Therefore, conservation of natural riparian and catchment vegetation is key to maintaining freshwater ecosystem processes and aquatic biodiversity worldwide, and more efficient and evidence-based restoration measures are urgently needed. As our global synthesis shows that direct human-induced alterations of vegetation in watersheds have significant negative effects on freshwater detrital food webs, there is a pressing need to consider cross-ecosystem consequences of land-cover changes in conservation and ecosystem management.

Keywords: allochthonous matter processing; aquatic‐terrestrial linkages; decomposers; decomposition; deforestation; detritivores; detritus; meta‐analysis; shredders; stream biodiversity.

FIGURE 1

Overview of methods used in this review. (a) PRISMA flow diagram with the number of studies located in the literature search and study selection and (b) Main moderators extracted for the assessment of the effects of vegetation change on freshwater detrital food webs.

FIGURE 2

Geographic distribution of studies included in this review and histograms of data distribution of different moderators. (a) World map showing the study locations with each circle representing a study. Map lines delineate study areas and do not necessarily depict accepted national boundaries. Panels (b–e) represent histograms for each level in the moderators: (b) type of vegetation change, (c) spatial scale of vegetation change, (d) metric, and (e) trophic level.

FIGURE 3

Impacts of human‐induced alterations in the watershed vegetation on freshwater detrital food webs. The global response (all data) is shown on the first row (a) and is separated by moderator level in the following rows for (b) climate, (c) type of vegetation change, (d) spatial scale of vegetation change, (e) metric, and (f) trophic level. The numbers in brackets after each moderator level represent the number of comparisons. For each moderator level, the circle represents the mean effect size (standardised mean differences Hedges' g between altered and reference vegetation conditions) with 95% confidence intervals (CIs) computed from the random effects model. Filled circles indicate statistically significant effects, whereas empty circles have CIs that cross the 0‐line and are thus statistically non‐significant. Negative effects sizes (Hedges' g < 0) indicate that the values in the altered conditions were lower compared to the reference conditions. Moderator levels sharing the same letter do not significantly differ as their CIs overlap.

FIGURE 4

Moderator analyses separated for the subsets of each trophic level with (a) detritus, (b) microbes, (c) shredders, and (d) omnivores. For each moderator level, the circle represents the mean effect size (standardised mean differences Hedges' g between altered and reference vegetation conditions) with 95% confidence intervals (CIs) computed from the random effects model. Filled circles indicate statistically significant effects, whereas empty circles have CIs that cross the 0‐line and are thus statistically non‐significant. Negative effects sizes (Hedges' g < 0) indicate that the values in the altered conditions were lower compared to the reference conditions. Moderator levels sharing the same letter do not significantly differ as their CIs overlap.

FIGURE 5

Quadrant graphs of median effect sizes of one trophic level against another in all possible combinations: (a) detritus‐microbes, (b) deritus‐shredders, (c) detritus‐omnivores, (d) microbes‐shredders, (e) microbes‐omnivores, and (f) shredders‐omnivores. Each point represents a study, and the symbols indicate the quadrant in which the effect sizes of the study lie in the x/y‐axes (triangles (−/−), circles (−/+), diamonds (+/−) and squares (+/+)). The quadrants are separated by a dashed grey line going through 0 on both axes. Negative effects sizes (standardised mean differences Hedges' g between altered and reference vegetation conditions < 0) indicate that the values in the altered conditions were lower compared to the reference conditions. For panels (a) and (f) we found statistically significant positive relationships indicated by the regression with the grey shaded area representing the confidence area around the linear regression.