A Spatially Consistent Bias Correction Method for Hydro-climatically Diverse Watersheds
Abstract
Water resources management and planning require the ability to generate statistically unbiased simulations of streamflow. However, all watershed models have systematic errors due to uncertainties in data sources, model structures, and parameter values. Statistical post-processing, or bias correction, is often used to remove distributional biases in model predictions for water resources management. Most existing methods apply bias correction methods independently to sites within a streamflow network, ignoring the inherent connectivity that the stream network provides and potentially introduces non-statistical errors such as inconsistent incremental flows that violate mass conservation. To incorporate the spatial relationships imposed by the channel network, we have developed a spatially consistent bias correction method that addresses these errors. We apply our method to watersheds across the United States that are key for water management. We describe how modeled historical streamflows before and after bias correction compare to observed historical streamflows. We compare performance tradeoffs between our method and traditional bias correction methods across various key watersheds. We show how statistical bias is reduced at hydro-climatically diverse sites while preserving modeled dynamics, demonstrating that the spatially consistent bias correction method can adapt to various watersheds.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMH004.0015S
- Keywords:
-
- 1847 Modeling;
- HYDROLOGY;
- 1871 Surface water quality;
- HYDROLOGY;
- 1879 Watershed;
- HYDROLOGY;
- 1880 Water management;
- HYDROLOGY