Use of Deterministic and Empirical Models to Predict Potential Evapotranspiration in an Advective Environment1

Abstract

AbstractA field experiment was conducted for three successive summers in the arid southwestern United States to determine the applicability of traditional empirical and deterministic potential evapotranspiration formulas for predicting water loss under extreme advective conditions. Five models for predicting potential evapotranspiration were compared to hydrologic balance estimates of water loss from a cowpea (Vigna unguiculata L. Walp) crop over two summer seasons and to each other during a third summer season. Of the five methods, pan evaporation was the least reliable indicator of full‐cover crop water use whether considered on a weekly or seasonal basis. The Priestley‐Taylor correlation, which required local calibration could not describe extreme changes in sensible heat advection from year to year with a single calibration. A form of the solar radiation correlation which contains parameters that are functions of humidity and wind run, worked adequately in all years as did an advection‐modified Priestley‐Taylor equation which requires a single calibration and vapor pressure deficit values. The Penman combination equation, which did not require local calibration, corresponded most closely to measured values. All models which contained some facility for recalibration or correction in response to changes in vapor deficit worked adequately under changing advective conditions during the entire experiment.