Please join us on May 19, 2017 at Noon for a research presentation with Laurel Larsen, Assistant Professor, Physical Geography, University of California, Berkeley. 

Abstract

A prime challenge across many types of environmental systems is that climate change and other anthropogenic impacts generate novel conditions that go beyond the system’s past variability and may trigger nonlinear change. Previous statistical approaches to predicting system behavior are often inapplicable to these new conditions. Rather, generating meaningful projections requires having a detailed understanding of the inner workings of environmental systems—how different variables drive and limit each other, and how perturbations or disturbances propagate over space.

In practice, the trademark complexity of hydrosystems often complicates this challenge; flow paths are often unseen, and multiple potential drivers of water quality or ecosystem functioning, which are individually dynamic, make it difficult to resolve the dominant controls and limiting factors. Hydrologists have typically approached the challenge of resolving connections in space and across drivers through hypothesis-driven modeling. Alternatively, emerging datasets, computational tools, and analysis strategies make it possible for hydrologists to observe these connections in a top-down manner, directly from data. In several case studies, I highlight emerging data-driven strategies for resolving the “inner workings” of complex hydrosystems to strategize and evaluate the effects of restoration strategies and plan for climate adaptation. The case-study applications include the establishment of ecological flows in the Everglades, assessing effects of stream restoration in the Chesapeake Bay watershed, optimizing coastal restoration in Louisiana, and planning for drought adaptation in the western U.S. I argue that complementary hypothesis-driven and data-driven strategies for resolving hydrosystem connectivity will be essential for producing meaningful hydrologic forecasts and assessing baselines and hydrologic change.

 

Figure caption: Information transfer from climate indices to U.S. precipitation gauge stations. Information transfer maps the functional connections between geographically distributed climate indices (sensitive to local sea surface temperatures) and localized precipitation in the U.S. Redder colors refer to stronger and more significant transfers of information. Notably, regions with frontally dominated precipitation exhibit a stronger relationship to climate indices than those with orographically or convectively dominated precipitation. This figure implies both the potential for hydrologic big-data analyses to advance hydrologic forecasting (links contain both a quantifiable strength and an associated lag time) but also a classic challenge (many significant connections).