Comparison of the Performance of ALARM and SEBAL in Estimating the Actual Daily ET from Satellite Data

Abstract

The Analytical Land Atmosphere Radiometer Model (ALARM) has been tested in several countries with observed evapotranspiration (ET) obtained from different ground-based methods, but none from a lysimeter, which is considered the most accurate ground-based method of measuring ET. The first objective of this work was to test ALARM against lysimeter ET data. The second objective was to compare the performance of ALARM with Surface Energy Balance Algorithm for Land (SEBAL), the most widely used model for estimating ET from remotely sensed data. An experiment was conducted on a 1.5 ha potato field in the Jordan Valley that has a weighing lysimeter located midfield. Potatoes were planted on the January 14, 216, and harvested on the May 1, 216. Three images from Landsat 8 were used to compute ALARM and SEBAL ET. The measured daily ET of the potatoes ranged from 5 mm d−1 on March 22 to 7.5 mm d−1 on April 23. ALARM underestimated the daily measured ET by 8% on March 22, 16% on April 7 and 33% on April 23 while SEBAL underestimated the daily measured ET by 4% on March 22, 15% on April 7 and 37% on April 23. Some modifications were made to improve the accuracy of ALARM and SEBAL ET estimation. The modified ALARM overestimated ET by 1% on March 22 and underestimated ET by 6 and 22% on April 7 and April 23, respectively. The modified SEBAL underestimated ET by .7, 1, and 35% on March 22, April 7, and April 23, respectively. These results demonstrated that the modified ALARM provides reasonable ET estimates that are superior to those obtained from SEBAL or modified SEBAL.