Predation and spatial connectivity interact to shape ecosystem resilience to an ongoing regime shift

Ecosystem regime shifts, characterized by abrupt and persistent changes in community structure, pose significant ecological and economic challenges. These shifts, similar to cascading failures observed in other complex systems (e.g., financial crises, social revolts, power grid failures), are influenced by stressors, individual unit resilience, and connectivity between units. While connectivity can sometimes facilitate the spread of alternative states, it can also enhance resilience by promoting the spread and persistence of organisms maintaining the original regime. However, empirical evidence on the role of spatial connectivity in preventing or reversing ecosystem regime shifts is limited. This study focuses on the Baltic Sea, where a spatially propagating regime shift is occurring, with predatory fish (European perch and northern pike) being replaced by the three-spined stickleback, an opportunistic mesopredator. This shift involves a predator-prey reversal; at high densities, predators suppress stickleback, while at low predator densities, stickleback suppress predator recruitment. The resulting cascading effects impact lower trophic levels, increasing filamentous algae and reducing water and habitat quality. The "stickleback wave," a gradual expansion of the stickleback-dominated regime, provides an ideal system for assessing the roles of connectivity and local environmental drivers in resilience to regime shifts.

Existing literature highlights the theoretical and conceptual understanding of spatial regime shifts and the potential for connectivity to influence resilience. Studies suggest that connectivity can affect the dynamics of shifts by influencing the movement of species and resources. While strong connectivity could accelerate the spread of an alternative regime, theory also posits that it could enhance resilience by supporting the persistence of organisms upholding the original regime. However, empirical studies directly testing these predictions in real-world ecosystems remain scarce. This gap underscores the need for empirical research to understand the interaction between spatial connectivity and local environmental drivers in shaping resilience to regime shifts, particularly given their substantial ecological and economic consequences.

The study utilized a large dataset (>7000 samplings) of juvenile fish communities collected over two decades along the Swedish Baltic Sea coast. The data, sourced from various research projects and monitoring programs, covered a heterogeneous island-rich archipelago (ca. 680 km). The researchers focused on juvenile fish densities to reflect the outcome of interactions during the spawning season. Data were corrected for various factors, including variations in sampling methods and detonation strength to standardize measures to the number of fish within a roughly 80 m² area. The analysis period spanned 2001-2020, excluding data beyond 40 km from the open sea due to sparse sampling. The influx of stickleback was represented by distance to the open sea, offshore stickleback densities (using acoustic survey data), and wave exposure. Connectivity was calculated using two approaches: a network-based method considering connected predator spawning habitat and a distance-weighted sum of available habitat within a 10 km radius. Predator spawning habitat was delineated based on depth and wave exposure cut-offs derived from existing studies of perch spawning preferences. Dispersal probability was modeled using data from a Finnish tagging study of adult perch, assuming distance-dependent dispersal. Local environmental drivers included predation pressure from seals and cormorants (estimated from counts and foraging range data), fishing pressure (combining commercial and recreational catch data), and water temperature (degree-day sums above 10°C). Generalized linear mixed models were used to analyze the data, incorporating random effects for year and spatial structure. Resilience was inferred from the positive effect of a driver on the probability of predator dominance.

The baseline model, considering only incoming stickleback, showed that predatory fish dominance increased with distance to the open sea and decreased with offshore stickleback density and wave exposure. The subsequent models, incorporating connectivity and local environmental drivers, revealed several key interactions. Connectivity had a positive effect on predatory fish dominance and densities at low and medium predation pressure from seals and cormorants but this effect disappeared or even became negative at high predation pressure. Fishing pressure negatively affected predatory fish dominance but showed no effect on absolute densities. Spawning season temperature positively affected predator dominance and densities, particularly near the open sea, suggesting that warmer temperatures boosted larval growth and reduced vulnerability to stickleback predation. The variance explained by the full model (including connectivity, local environmental variables, and interactions) was considerably higher than the baseline model for predatory fish densities (0.38 vs 0.23), indicating the importance of local dynamics in driving absolute predatory fish densities. Predation from seals and cormorants explained the largest amount of variation in predatory fish densities. The study also suggests a five-level trophic cascade, with top predators influencing predatory fish, which in turn influence stickleback, grazers, and filamentous algae.

The findings support theoretical predictions that both connectivity and local environmental factors influence resilience to regime shifts. The positive effect of connectivity at low predation pressure suggests that, when predators are abundant, dispersal facilitates redistribution and recolonization, enhancing resilience. The disappearance of this effect at high predation pressure likely reflects the scarcity of adults to redistribute. The negative effect of top predator densities on predatory fish juveniles, likely reflects a reduction in the adult population. The lack of a clear effect of fishing on absolute predator densities, compared to the stronger effect of seals and cormorants, indicates that fishing pressure may have a relatively smaller impact than top-down predation. Warmer temperatures likely increase resilience by enhancing larval growth and shortening the period of vulnerability to stickleback predation. The study highlights the importance of both the stressor (increasing stickleback densities) and the loss of resilience (predation by seals and cormorants) in driving the regime shift.

This study empirically demonstrates the interactive roles of predation, connectivity, and temperature in shaping resilience to a spatially propagating regime shift in the Baltic Sea. Habitat connectivity enhances resilience but only when top predator pressure is low. Temperature positively impacts resilience, potentially by boosting predator growth and recruitment. These findings underscore the need for spatially explicit management strategies that consider both the stressor and the local factors influencing resilience. Future research should focus on refining connectivity measures, improving understanding of dispersal patterns, and investigating the individual and combined impacts of various management strategies.

The study's large spatial scale necessitated a relatively coarse connectivity metric, potentially overlooking finer-scale habitat heterogeneity and dispersal barriers. The analysis relied on correlational data, limiting the ability to definitively establish causal relationships. The limited availability of data on adult predator densities and some environmental variables (e.g., eutrophication) constrained the scope of the analysis. The assumption of distance-dependent dispersal based on adult perch tagging data may not fully capture juvenile dispersal patterns.