Field scale spatiotemporal analysis of surface soil moisture for evaluating point-scale in situ networks

Abstract

Soil moisture is an intrinsic state variable that varies considerably in space and time. From a hydrologic viewpoint, soil moisture controls runoff, infiltration, storage and drainage. Soil moisture determines the partitioning of the incoming radiation between latent and sensible heat fluxes. Although soil moisture may be highly variable in space and time, if measurements of soil moisture at the field or small watershed scale are repeatedly observed, certain locations can often be identified as being temporally stable and representative of the an area average. This study is aimed at determining the adequacy of long term point-scale surface soil moisture measurements in representing local field scale averages which may ultimately serve as in situ locations for the calibration and validation of remotely sensed soil moisture. Experimental data were obtained by frequency-domain reflectometry (FDR) sensors permanently installed in two agricultural fields, AS1 and AS2 (2.23 and 2.71ha, respectively) at a depth of 5cm. Twenty additional FDR sensors, spaced 35m apart, were installed horizontally at a depth of 5cm in each field with automated data collection being transmitted every 30min from July 15 through September 20, 2009. Additionally, meteorological data were obtained from existing weather stations in each field. The FDR sensors revealed persistent patterns in surface soil moisture within each field and identified sites that were temporally stable. The locations that were optimal for estimating the area-average field water contents were different from the permanent sensor locations in both fields. Permanent sensor data showed approximately 4 and 10% mean relative differences for fields AS1 and AS2, respectively, with relatively large standard deviations. Thus, minimum offset values could be applied to the temporally stable field sites to obtain representative field average values of surface soil moisture. However, use of permanent sensor data for offset estimates gave poor results. These findings are of relevance for applications of geospatial surface soil moisture data assimilation in hydrologic modeling when only point-scale observations are available, as well as, remotely sensed surface soil moisture calibration and validation studies.