Evapotranspiration Estimation Uncertainty from Wireless Temperature Sensor Data.

Field-based wireless sensor networks can improve crop water management by providing near real-time observations of plant canopy temperatures. By combining these temperatures with remotely sensed plant density maps and weather data, evapotranspiration (ET) can be modeled at hourly time steps. However, the process of merging single-station time series data with remote sensing data introduces dynamic model uncertainties that reduce confidence in ET estimation accuracy. Sources of uncertainties include sensor temperature errors, inaccurate plant cover models, and poorly constrained energy balance model parameters. Thus the value of wireless sensor systems for water management is closely tied to capability to identify and constrain the largest uncertainties. Using data collected from irrigation field studies at Maricopa, Arizona from 2012 to 2014, the effects of uncertain temperature data upon ET from wheat, camelina and cotton crops were evaluated. Wireless-based radiometric temperature data were collected, combined with estimates of canopy density maps, and then incorporated into a two-source energy balance model. Comparisons between model results and independent soil moisture observations will be discussed.