Abstract

Snowpack in the Sierra Nevada plays a critical role in natural and human systems. Snowmelt runoff from the Sierra is utilized for irrigation, power generation, recreation, in addition to urban consumption from nearby population centers, effectively functioning as a large surface “reservoir” on which millions of people depend. In addition, analysis of the snow-derived streamflow in the Sierra Nevada has been widely used to quantify the climate change and its impact on various aspects of society. Being able to accurately estimate large-scale snow water equivalent (SWE) and snowmelt-timing is an important objective that has both scientific and civil merit.

Estimating large-scale SWE in mountain environments is challenging, however, due to the significant variability of the snow properties caused by complex physiographic and atmospheric conditions. Remote sensing and modeling are arguably the only available methods for large-scale SWE estimation. However, both methods contain intrinsic limitations, and thus the accurate estimation of mountainous SWE remains elusive. Although weaknesses exist in each individual method, all methods contain complementary snow information; integrating the snow information from multiple datasets shows promise for improving SWE estimates in mountain areas.

In this study, we carried out an experiment to estimate SWE in the Upper Kern Basin, Sierra Nevada, by assimilating spaceborne observations of the snow microwave radiance into a coupled hydrology and radiative transfer model using an Ensemble Kalman Batch Smoother (EnBS). The data assimilation framework merged the complementary SWE information from modeling and radiance observations for an improved SWE estimate. Specifically, the coupled land surface model SSiB3 and radiative transfer model MEMLS first simulated a prior estimate of SWE and snow microwave radiance using the NLDAS2 meteorological forcing data; the prior SWE estimate served as a first guess of the true SWE, but could be biased and uncertain. With the EnBS, the prior SWE was updated based on the SWE information in the satellite observations; the updating adjusted the prior estimate and yielded a more accurate posterior SWE estimate.

The modeling was at a very high resolution (90m) and spanned a range of mountain environmental factors to better characterize the effects of the mountain environment on snow distribution and radiance emission. On average, the EnBS assimilation reduced the overall bias of the accumulation season SWE estimates by 84.2%, and reduced the accumulation season SWE RMSE by 35.4%. The assimilation also reduced the bias and the RMSE of the April 1^st SWE estimates by 80.9% and 45.4%, respectively. Future work includes assimilating visible band remote sensing observations and in-situ snow observations to improve the overall SWE estimates in the entire western U.S.

Bio

Dongyue Li is a UCLA Postdoctoral Researcher co-advised by Dr. Lettenmaier and Dr. Margulis. His research interests are in the areas of characterizing snow as a water resource, and of assessing the availability and vulnerability of water resources in the context of climate change. His research focuses on developing and applying data assimilation systems to combine the large-scale hydrologic information from modeling, remote sensing, and in-situ observations. He received the B.S. in photogrammetry and remote sensing in 2009 from the Information Engineering University, China, and received the M.S. and PhD in earth sciences from the Ohio State University in 2011 and 2016. Dr. Li is a recipient of the NASA Earth and Space Science Fellowship and the NSF-CUAHSI Pathfinder Fellowship, and was among the ten finalists of the student paper competition in the IEEE International Geoscience and Remote Sensing Symposium in 2014.