Nonstationary Runoff Responses Can Interact With Climate Change to Increase Severe Outcomes for Freshwater Ecology

Climate change is projected to impact multiple components of the flow regime. However, changes in some ecologically important aspects of flow seasonality and variability are not well‐represented by global climate models. We used a stress testing method and global sensitivity analysis to investigate whether interactions between five different, but plausible, change “dimensions” (hydroclimatic variables or relationships) led to worse ecological outcomes than individual changes. The five dimensions include changes in long‐term average rainfall and temperature, low‐frequency variability of rainfall, seasonality of rainfall, and the rainfall‐runoff relationship. Our case study involved regulated and unregulated sections of the Goulburn River, Australia. We found that four different modeled ecological outcomes (condition of small bodied fish, large bodied fish, in‐channel vegetation, and floodplain vegetation) are most sensitive to changes in long‐term average rainfall. Sensitivity to changes in rainfall seasonality depends on river characteristics and appears to be heavily dampened by regulation and actively managed environmental water. Changes to the rainfall‐runoff relationship (which may be triggered by long‐term drying) were found to greatly influence ecological outcomes, but remain poorly understood. However, when considering the worst outcomes that are likely to present severe threats to ecological survival, all five dimensions were significant. These worst outcomes only manifest under certain combinations of changes with interactive effects. These joint interactions have implications for climate risk assessments that do not consider multiple dimensions of change, particularly those aimed at evaluating and mitigating severe threats or extinction probability. Plain Language Summary Different kinds of climate changes, such as lower rainfall overall, increased temperatures, or changes to the time of year when the most rain falls can all threaten aquatic plants and animals. By manually changing the inputs to our water resource models, we can answer questions like “what happens if rainfall decreases by 20% and temperatures go up by two degrees?” We can do this many times, and build a picture of how rivers respond to a range of changes. We used this concept to understand how river ecology responds to up to five different changes at once and asked whether the outcomes are worse when these changes are combined versus when they occur separately. Ecological outcome