Landscape controls on hydrologic responses to long-term climate oscillations

点击翻译按钮获取中文摘要

This project seeks to understand and quantify hydrologic responses to long-term climate oscillations, such as the El Niño-Southern Oscillation (ENSO), for 2733 individual watersheds across the conterminous United States. These watersheds represent a wide range of terrain, land cover, land use, and climate conditions, and they include varying degrees of human disturbance. With such an expansive dataset, the project will provide insight as to how and why watersheds respond differently (or not at all) to ENSO. The study will improve scientific knowledge and understanding about climate impacts on the terrestrial water cycle, not only for multi-year climate oscillations but also for long-term climate change, which are both critical for managing water supplies, planning civil infrastructure, and preparing for natural disasters. The project will also provide geospatially-oriented tools to improve hydrology education for K12 and university students while providing postdoctoral training for an emerging hydrologist.Preliminary analyses suggest that watersheds filter ENSO signals differently, even after accounting for variations in precipitation responses to ENSO across the country. Spatial characteristics associated with internal watershed organization are hypothesized to explain a significant amount of the observed variability in watershed responses to ENSO. This hypothesis is rooted in geomorphological instantaneous unit hydrograph (GIUH) theory, which links the spatial organization of watersheds to their hydrologic responses. This project builds on the GIUH framework and applies it to long-term climate phenomena such as ENSO, using statistical learning methods along with hydrological modeling to test hypotheses. Overall, the work seeks to improve scientific knowledge and understanding about climate impacts on the terrestrial water cycle by bridging the gap between mechanistic, watershed-scale studies of hydrological responses and large-scale studies of climate-streamflow interactions that are predominantly statistical in nature. This work acknowledges the so-called "big data" nature of many geospatial research problems in the hydrologic sciences and brings statistical learning methods, currently under-utilized in the field, to bear on large, publicly available datasets.