Downscaling satellite soil moisture using geomorphometry and machine learning

Abstract

Annual soil moisture estimates are useful to characterize trends in the climate system, in the capacity of soils to retain water and for predicting land and atmosphere interactions. The main source of soil moisture spatial information across large areas (e.g., continents) is satellite-based microwave remote sensing. However, satellite soil moisture datasets have coarse spatial resolution (e.g., 25–50 km grids); and large areas from regional-to-global scales have spatial information gaps. We provide an alternative approach to predict soil moisture spatial patterns (and associated uncertainty) with higher spatial resolution across areas where no information is otherwise available. This approach relies on geomorphometry derived terrain parameters and machine learning models to improve the statistical accuracy and the spatial resolution (from 27km to 1km grids) of satellite soil moisture information across the conterminous United States on an annual basis (1991–2016). We derived 15 primary and secondary terrain parameters from a digital elevation model. We trained a machine learning algorithm (i.e., kernel weighted nearest neighbors) for each year. Terrain parameters were used as predictors and annual satellite soil moisture estimates were used to train the models. The explained variance for all models-years was >70% (10-fold cross-validation). The 1km soil moisture grids (compared to the original satellite soil moisture estimates) had higher correlations (improving from r2 = 0.1 to r2 = 0.46) and lower bias (improving from 0.062 to 0.057 m3/m3) with field soil moisture observations from the North American Soil Moisture Database (n = 668 locations with available data between 1991–2013; 0-5cm depth). We conclude that the fusion of geomorphometry methods and satellite soil moisture estimates is useful to increase the spatial resolution and accuracy of satellite-derived soil moisture. This approach can be applied to other satellite-derived soil moisture estimates and regions across the world.

Statistical Exploration of SENTINEL-1 Data, Terrain Parameters, and in-situ Data for Estimating the Near-Surface Soil Moisture in a Mediterranean Agroecosystem

“…We chose our terrain parameters (Table 4) for incorporation into the modeling framework because of their capability to reflect the site-specific soil hydraulic behavior. Therefore, the incorporation of the "hydrologically meaningful" terrain parameters (Guevara and Vargas, 2019) is in line with studies that employed geomorphometry in empirical approaches, such as for downscaling RS-based soil moisture estimates (e.g., Guevara and Vargas, 2019;Mascaro et al, 2019;Zappa et al, 2019). For instance, the SAGA wetness index (SWI) simulates hydrological processes based upon the topography.…”

“…Subsequently, terrain features enable approximating of spatial patterns θ from an ecohydrological perspective (e.g., Western et al, 2002;Robinson et al, 2008). Guevara and Vargas (2019) successfully employed topographic information in an ML approach to downscale RS-based soil moisture products over the conterminous USA and labeled several parameters as "hydrologically meaningful." In this way, it was possible to adequately address the landscape's capability to physically constrain water inputs (e.g., rain and irrigation water, overland flow) and support the link between spatial variability of soil moisture θ and topography (Guevara and Vargas, 2019).…”

A restructured and updated global soil respiration database (SRDB-V5)

“…There is much more potentially useful information that could be included in future SRDB updates, although it is important to remember that every additional piece of information comes with a never-ending cost (in terms of data entry time, quality assurance/quality control, etc). 1) Number_of_collar: The 355 number of collars within a certain study area is important information to evaluate the representability of the RS measurements; 2) Soil organic carbon (SOC) from regional or global estimates (Guevara et al, 2020;Hengl et al, 2017); 3) currently, Site_ID in SRDB-V5 are only comparable with Site_ID of MGRsD and HGRsD, further updates to Site_ID so it can connect with more external datasets [e.g., FLUXNET, COSORE, and AmeriFlux and a global database of forest carbon stocks and fluxes (ForC) (Anderson-Teixeira et al, 2018b)]; Annual_soil_moisture to 360 include a mean value of soil moisture or intra-annual soil variability derived from remote sensing (Guevara and Vargas, 2019) when this variable was not measured at the site . In addition, some meta information can be improved.…”

“Gap-Free Global Annual Soil Moisture: 15km Grids for 1991–2016” by Mario Guevara et al. Responses to anonymous reviewer #1

“…We believe most of the concerns could be addressed by editing the text to improve clarification and performing new analyses about variable performance, using the recently released version 4.5 of the ESA-CCI, and demonstrate the applicability of our methods at higher temporal resolutions (months) and across smaller areas and spatial extents. Please note that our previous work has demonstrated the effectiveness of our approach at the continental scale of CONUS using 1km grids (Guevara and Vargas, 2019).…”