Spatial Scale Issues in Environmental Flow Modeling

Environmental flow assessment typically includes modeling, and a fundamental issue in ecological modeling is spatial scale: What are the best spatial scales for modeling a particular organism, habitat, and management problem? If we are modeling fish response to flow, what cell size is most appropriate for a particular species and lifestage? The realization that ecological systems and models behave differently at different spatial resolutions was one of the most important advances in ecology but this understanding is not yet ingrained in instream flow modeling. Very few environmental flow studies have started with what is typically the first step in ecological modeling: selecting an ecologically appropriate spatial resolution (cell size). Key questions for this selection include: What species and life stages are being studied? What activities of these animals does the model represent? Over what distances do the animals undertake these activities? A drift-feeding trout may use a territory several times its body length in diameter, so a spatial resolution of one to several square meters—but no smaller—is appropriate. In contrast, modeling the feeding habitat of a bass that cruises one or several pools each day can only be modeled meaningfully at the scale of at least a pool. When several activities are important but occur at different scales, they often can be accommodated by modeling “tricks”. Our trout model represents feeding habitat using a cell size at least as big as an adult fish territory. Velocity shelter for drift feeding, though, occurs at a smaller scale and hence is represented via a cell variable representing the fraction of cell area providing shelter. Hiding cover can be useful even if it is far outside the feeding territory, so it is represented as a cell variable for distance to hiding cover. Incorporating modern spatial scale considerations in models is one of the simplest and most important ways we can improve instream flow science. It also eliminates many of the conceptual inconsistencies and technical challenges of current instream flow models.