Impact of Backscatter Variations Over Water Bodies on Coarse-Scale Radar Retrieved Soil Moisture and the Potential of Correcting With Meteorological Data

Abstract

The Northern Hemisphere is, to a large extent, underlain by permafrost, which is prone to thawing due to rapid warming in the Arctic during the 21st century. In this context, satellite-derived soil moisture data are valuable for modeling purposes. Assessing the applicability of such data at high latitudes is essential but has, until recently, been given little attention. Recent studies have pointed out that seasonal land cover variations and the presence of small water bodies, which are typical in the Arctic, can cause complications for soil moisture retrieval. Here, it is hypothesized that a bias related to water fraction is caused by variations in the water surface roughness. The impact is quantified for the Metop Advanced Scatterometer by investigations of the higher spatial resolution synthetic aperture radar (SAR) data acquired by ENVISAT Advanced SAR over 11 sites across the Siberian Arctic. The bias calculated as an average over time can be explained by the lake fraction: a water fraction higher than 20% causes a bias of more than 10% relative surface soil moisture. This can, to a great extent, be attributed to wind, based on which a bias correction was developed. The correction was applied and evaluated with in situ soil moisture data, which were available from one of the sites: the Lena Delta. Weak results are obtained because water surfaces correspond mainly to rivers at this site. Variations in discharge, water height, and streams may therefore also affect the water surface roughness.