Sustaining ecosystem services in a changing world requires enhanced understanding of their spatial-temporal dynamics, and responses to multiple drivers of environmental changes such as climate, land-use and management. To date, most research has quantified the magnitude of ecosystem service provisioning for a given snapshot. Yet few studies have taken a dynamic approach to investigating drivers and dynamics of ecosystem services simultaneously. In particular, variability of ecosystem services has been seldom studied. Such analysis of variability can yield insights into forces controlling ecosystem services. We evaluated indicators of nine food, water, and biogeochemical-related ecosystem services from 2010 to 2070 in the 1300-km2 Yahara Watershed (Wisconsin, USA) for four contrasting future scenarios. We asked: (i) Among space, time, and scenario, which factors explain most of the variability of ecosystem services? and (ii) Are there consistent patterns across different ecosystem services? Ecosystem services were simulated annually using linked biophysical models – Agro-IBIS and THMB, and summarized into spatial scale of subwatershed (N=100), and temporal scale of 8-yr running means over which temporal autocorrelation is eliminated. We performed hierarchical linear-mixed effects model and variation partitioning, where ‘scenario’ was treated as the random effects, and ‘space’, ‘time’ and their interactions as main effects nested with ‘scenario’.
Results/Conclusions
All ecosystem services demonstrated substantial spatial and temporal variability. Specifically, for crop and perennial grass production, ‘scenario’ explained >70% of the variation, suggesting that these two food-provisioning services are largely affected by future societal choices and human interventions associated with each scenario. For nitrate leaching and phosphorus yield, ‘space’ explained most variation (73% and 51% respectively), followed by ‘time’ and ‘space-time’ interactions. These results highlighted profound influences of processes associated with location (i.e., land use and management) and time (i.e., climate variability), as well as their synergistic interactions for water quality. For drainage, ‘time’ was the dominant factor, whereas for number of extreme runoff days, ‘space’ explained most of the variation (i.e., 76%), suggesting that these two water quantity services differed in their controlling processes. For net ecosystem exchange and sediment yield, ‘time’ and ‘space’ contributed to comparably high variations, underscoring influences of both time- and location-specific processes for these two regulating services. For soil carbon storage, ‘space’ was the single most important factor, implying the dominant role of location-specific factors in this service. Our results revealed factors or processes that are key controls for ecosystem services, and provided significant theoretical and practical implications for managing future ecosystem services.