Towards soil property retrieval from space: An application with disaggregated satellite observations

Soil moisture plays a key role in most environmental processes, as evaporation and transpiration are heavily dependent on soil moisture variability. While it is one of the few important hydrological variables that can be directly observed, the high spatial and temporal variability makes it difficult to measure globally or even regionally. Reliance is therefore placed on land surface models to predict the evolution of soil moisture using low-resolution soil property information or typical values. But to make predictions with the required accuracy, more reliable and detailed soil parameter data are required than those currently available. This paper demonstrates the ability to retrieve soil hydraulic parameters from near-surface measurements, using Soil Moisture and Ocean Salinity (SMOS) observations disaggregated to 1 km resolution for a demonstration area the size of a single SMOS footprint. The disaggregated soil moisture product was first assessed against in-situ soil moisture observations, before testing the retrieval methodology using the disaggregated soil moisture data for individual soil columns co-located with three long-term monitoring sites in the Murrumbidgee Catchment. The retrieval methodology was then applied to the entire 40 km 40 km demonstration area at 5 km spatial resolution. The results suggest that spatially variable soil hydraulic properties exist in the study area, while published soil texture maps show only a single soil type, meaning that a single set of soil hydraulic parameters would normally be used in soil moisture prediction models for this region. Use of a single set of soil hydraulic parameters, rather than the spatially variables ones, was estimated to have an approximate 0.06 m/m impact on the soil moisture prediction. 2015 Elsevier B.V. All rights reserved.