Passive Microwave Remote Sensing of North Carolina Soil Moisture for Hydrologic Assessment and Forecasting.

Soil moisture data aid understanding and modeling of regional and global water cycles, ecosystem productivity, and processes linking water, energy, and C cycles; they can improve: weather and climate forecasts, flood prediction, drought monitoring, estimation of CO₂ uptake by vegetation; and are applied in construction and irrigation planning. However, most ground-based soil moisture monitoring systems are limited in extent and density, and it is not feasible to install enough sensors to adequately model soil moisture spatially over large areas. NASA's Advanced Microwave Sensing Radiometer for the Earth Observing System (AMSR-E) aboard the AQUA satellite is a passive microwave instrument yielding global microwave brightness temperatures that are processed to estimate soil moisture. We examined selected AMSR-E data for spatial and temporal correlation and for correlation with data from North Carolina's in-situ soil moisture network, ECONet, to determine the data's utility for statewide estimation. On dates examined, there was no correlation between AMSR-E soil moisture and that from 29 ECONet stations. Spatial correlation of AMSR-E soil moisture was strong (e.g., semivariogram nugget and sill, 37.9 and 903 cm⁶ cm^-6, respectively; nugget:sill=0.054; range=617 040 m: R²=0.98),  but values and temporal variability were low compared to that of ECONet data in all retrievals examined. Semivariograms indicated that ECONet data had relatively weak spatial correlation (mean semivariogram parameters: nugget=0.006, sill=0.014, nugget:sill=0.435, range=413 000 m,  R²=0.50), likely due to the distance between monitoring sites (9198 to 557 000 m), thus limiting their utility in statewide soil moisture estimation via interpolation (kriging). Continuing goals are to: develop and validate a soil moisture model using statewide precipitation and evopotranspiration records; evaluate temporal and spatial correlation for more sites and years and between AMSR-E brightness temperature and derived soil moisture; and improve AMSR-E-based soil moisture estimation by incorporating auxiliary data.