Atmospheric pollution contributed to destabilizing a small mammal community amid four decades of environmental changes

In the recent decades, climatic and land-use changes paralleled major decreases in atmospheric pollution, but unclear is how these concomitant changes shape ecological dynamics. We analyzed 39 years (1986-2024) of semiannual abundances for six small-mammal species in the Black Triangle of Europe (Czechia), a region historically-marked by acidic rains causing forest diebacks and toxic lead (Pb) deposition. We combined these small-mammal data with monthly time series of atmospheric pollution, vegetation index (NDVI), meteorology and soil moisture. We used Empirical Dynamic Modelling (EDM) to infer chaos, recover species-species and environment-species causal relationships, recover time-varying interactions, and perform out-of-sample forecasts. Small mammal populations were causally affected by atmospheric pollution, and its abrupt decline in the late 1990s likely triggered a delayed critical transition of community composition. Pollution primarily affected the dominant omnivores and strengthened the interactions among species, a hallmark of instability. We identified a second route to chaos with a role of vegetation dynamics and soil moisture that was especially strong on shrews. Our feedback-loop analysis suggests how pollution and other environmental changes favored destabilizing, positive feedbacks. Forecasting of these chaotic dynamics improved when adding the relevant environmental predictors, demonstrating that the unpredictability of chaotic systems can slightly improve when we identify the main drivers of interactions in a network, offering hope for anticipating the future of complex ecosystems. Our study offers a unique retrospective of the destabilizing effects of pollution, an overlooked macro-ecological driver.