ABSTRACT Soil moisture is classified as an essential climate variable (ECV) and is relevant to understanding hydrological, agricultural and ecological processes. Yet, in spite of its importance, direct observations of soil moisture remain limited globally—those that exist are typically limited in duration and spatial extent. Consequently, alternative approaches for estimating soil moisture have been developed, including water balance (‘bucket’) models, the use of remotely sensed information and the application of land surface modelling techniques. Spaceborne and land surface modelling based methods offer significant potential for monitoring and modelling soil moisture at a variety of spatial scales; however, their resolution remains relatively coarse for global and continental scale applications. At country scale, land surface models have demonstrated their potential but they require access to computational resources to deliver high resolution products. With the advent of machine‐ and deep‐ learning and data fusion techniques, high resolution global and regional soil moisture datasets are increasingly becoming available. Here, we evaluated a statistical machine learning approach to downscale the European Space Agency's (ESA) Climate Change Initiative (CCI) combined passive and active soil moisture product for Ireland using covariates that included both static (e.g., topography) and dynamic (e.g., gridded rainfall and temperature) variables. The model was developed using in situ cosmic ray neutron sensor (CRNS) measurements obtained from a network of sites in the United Kingdom, justified on the basis that the United Kingdom is geographically similar to Ireland in terms of its climate, soil types and land cover management practices. The model was found to perform reasonably well when validated against limited in situ data obtained from available time domain reflectometry (TDR) measurements available from Ireland. The developed model was subsequently used to derive spatial estimates of soil moisture on a 1 km grid across the Republic of Ireland.
Fealy et al. (Thu,) studied this question.