Processing High-frequency Soil Moisture Sensor Data to Characterize Drought Stress Experienced In Large Cotton Variety Trials.

Although the price of collecting high-frequency sensor data has declined in recent years, techniques to analyze and translate this data into actionable information have not progressed as quickly.  The potential of soil moisture sensors to characterize drought within a given field or initiate irrigations hinges upon the ability of the instrument to characterize soil moisture at the sampled point, extrapolate that information across the field, and convert the high-frequency data into crop water status.  Therefore, the objective of this study was to determine if a limited number of soil moisture sensors deployed into a dryland variety trial could accurately characterize the volumetric water content (VWC) at a given point within the field and if this measurement could be extrapolated out to the field scale while still correlating strongly to crop water status.  During the 2013-2015 growing seasons soil moisture sensor deployments were made into numerous dryland cotton variety trials across the Mid-South and Southeastern Regions of the U.S.  Inference on volumetric water content of each monitored profile was determined by (4) Decagon EC-5 sensors (Decagon Devices Inc., Pullman, WA) installed at depths of 7.5, 22.5, 45, and 75 cm.  Sensor reported VWCs related well to soil water content measured at installation (r²=0.617).  Relationships between within-location nodes varied but were typically moderate to strong (0.646<r2<0.988).  Mean relative difference analysis indicated a minimum of one node in six of seven instrumented trials was characterized as temporally stabile (σ<5%).   Further research should be conducted to see if trends in temporal stability remain through multiple seasons and if in-season measurements could be replaced by extrapolated weather data and radar-based modeling of rainfall as a means of deriving crop water status.