The Role of Multi-sensor Land Data Assimilation in Runoff and Streamflow Estimation
Abstract
Runoff and streamflow are critical components of the water cycle and, reflecting integrated information on dominant hydrologic processes over the entire watersheds. Streamflow is precisely measured by gages, but the large-scale spatiotemporal variation in streamflow is poorly understood. We evaluate the impact of multi-sensor land data assimilation on intraseasonal-to-interannual availability of runoff. Eight experiments with assimilation of different combinations of satellites datasets are conducted using Community Land Model version 4 (CLM4) and Data Assimilation Research Testbed (DART). Different land states are updated upon the assimilated satellites observations (AMSR-E, MODIS, and GRACE) for 2003-2009. We compared the eight experiments to open loop. Assimilation of MODIS snow cover fraction affects simulated runoff in mid- and high-latitudes. Assimilation of lower frequencies AMSR-E brightness temperature plays an important role in soil moisture and therefore runoff in tropics. Generally, assimilating different satellite observations shows a spatially different impacts on runoff, and the combination of multiple satellite observations shows a largest spatial extend. We also quantify the impact of data assimilation by evaluating results with observation-based runoff and streamflow from GRDC. Our results suggest that land data assimilation of any of these datasets improve runoff estimation over the Arctic. This study indicates the limitation of modeling the large-scale hydrological cycle and shows how data assimilation can help improve streamflow estimation.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.H51V1628W
- Keywords:
-
- 1655 Water cycles;
- GLOBAL CHANGEDE: 1816 Estimation and forecasting;
- HYDROLOGYDE: 1847 Modeling;
- HYDROLOGYDE: 1855 Remote sensing;
- HYDROLOGY