ACOMPARISON OF AUTOMATICALLY AND MANUALLY COLLECTED PAN EVAPORATION DATA

Abstract

Pan evaporation is an important weather variable that has numerous applications related to decision makingin agriculture, forestry, ecology, hydrology, and other fields. The automation of pan evaporation measurements throughthe use of electronic sensors has the potential to increase the availability and resolution of measurements, while reducingthe overall cost of data collection. Information is needed with respect to the field performance of these devices. Theobjective of this study was to compare automated and manual pan evaporation measurements. Daily pan evaporationmeasurements calculated from 15-min averages of water height in Class A evaporation pans of the Georgia AutomatedEnvironmental Monitoring Network (AEMN) were compared to daily pan evaporation data collected at National WeatherService (NWS) cooperative stations. Collocated weather stations in Griffin and Watkinsville, Georgia were selected forthe comparison. Data from 1991 to 1996 at the Griffin location and data from 1993 to 1997 data at the Watkinsvillelocation were used. Data sets consisted of 733 and 808 daily evaporation totals from Griffin and Watkinsville,respectively. An estimate of potential evapotranspiration was also calculated for each daily record using the Priestley-Taylor equation. Daily pan evaporation amounts from the automated observations were generally less than theevaporation measurements from the manual observations. Average total annual pan evaporation from the manualobservations was 537 mm for Griffin and 1051 mm for Watkinsville. The average total annual pan evaporation from theautomated observations was, respectively, 414 mm and 676 mm for the same locations. The Priestley-Taylorapproximation of pan evaporation was generally closer to the manual observations than the automated observations.Average total annual evapotranspiration estimated by the Priestley-Taylor equation was 491 mm at Griffin and 842 mm atWatkinsville. The daily automated pan evaporation data included many low values for days in which considerable panevaporation should normally occur. Records of water height from the automated observations showed that mechanicalproblems with the sensor used in the automated pan evaporation system were responsible for much of the difference seenbetween the automated and manual observations. Improved maintenance of the automated observations is recommendedto justify replacement of the manual observations. A change in the design of the float mechanism might also be consideredby the manufacturer.