Gap Filling of High-Resolution Soil Moisture for SMAP/Sentinel-1: A Two-Layer Machine Learning-Based Framework

Abstract

As the most recent 3 km soil moisture product from the Soil Moisture Active Passive (SMAP) mission, the SMAP/Sentinel-1 L2_SM_SP product has a unique capability to provide global-scale 3 km soil moisture estimates through the fusion of radar and radiometer microwave observations. The spatial and temporal availability of this high-resolution soil moisture product depends on concurrent radar and radiometer observations which is significantly restricted by the narrow swath and low revisit schedule of the Sentinel-1 radars. To address this issue, this paper presents a novel two-layer machine learning-based framework which predicts the brightness temperature and subsequently the soil moisture at gap areas. The proposed method is able to gap-fill soil moisture satisfactorily at areas where the radiometer observations are available while the radar observations are missing. We find that incorporating historical radar backscatter measurements (30-day average) into the machine learning framework boosts its predictive performance. The effectiveness of the two-layer framework is validated against regional hold-out SMAP/Sentinel-1 3 km soil moisture estimates at four study areas with distinct climate regimes. Results indicate that our proposed method is able to reconstruct 3 km soil moisture at gap areas with higher Pearson correlation coefficient (47%/35%/20%/80% improvement of mean R, at Arizona/Oklahoma/Iowa/Arkansas) and lower unbiased Root Mean Square Error (20%/10%/7%/26% improvement of mean ubRMSE) when compared to the SMAP 33 km soil moisture product. Additional validation against airborne data and in-situ data from soil moisture networks is also satisfactory.

Deep multi-modal satellite and in-situ observation fusion for Soil Moisture retrieval

“…In order to capture the information encoded in time-series (LSTM) networks were explored for SM estimation from Brightness Temperature measurements from SMAP, MODIS Vegetation Water Content and soil temperature in [7]. A work similar to the one reported here is the method proposed by Mao et al [8] where the authors employed machine learning, random forests in particular, for estimating the high-resolution SMAP/Sentinel-1 estimation given lowresolution SMAP radiometry data. An earlier version considered CNNs for downscaling SMAP radiometer brightness temperature measurements, focusing only on the period when both SMAP radar and radiometer were operational [9].…”

Multi-Temporal Convolutional Neural Networks for Satellite-Derived Soil Moisture Observation Enhancement

“…In order to capture the information encoded in time-series (LSTM) networks were considered for SM estimation from Brightness Temperature measurements from SMAP, MODIS Vegetation Water Content and soil temperature in [8]. A work similar to the one reported here is the method proposed by Mao et al [9] where the authors considered machine learning, random forests in particular, for estimating the high-resolution SMAP/Sentinel-1 estimation given low-resolution SMAP radiometry data. A earlier version considered CNNs for downscaling SMAP radiometer brightness temperature measurements, focusing only on the period when both SMAP radar and radiometer were operational [10].…”

A Multiscale Deep Learning Model for Soil Moisture Integrating Satellite and In Situ Data

“…Some downscaling methods are based on data assimilation (Dong et al., 2019; Kim et al., 2021; Lievens et al., 2017; Mascaro et al., 2010; Reichle et al., 2014), but most data assimilation systems are only designed to handle a single data type (Reichle, 2008). Statistical and machine learning methods directly build relationships between coarse‐scale observations, finer‐scale attributes, and in situ observations (Alemohammad et al., 2018; B. Fang et al., 2021; Im et al., 2016; Kolassa et al., 2018; Mao et al., 2019). Regardless of the mechanism, however, these downscaling methods depend on real‐time satellite‐based observations and thus cannot project long‐term into the future.…”