A comparison of different methods of estimating energy‐limited evapotranspiration in the peace river region of British Columbia

Abstract

The performances of eight methods for estimating daily energy‐limited evapotranspiration (E_e) were compared with reliable values for the Peace River region of British Columbia. They included the methods of Priestley and Taylor (PT), Jensen and Haise (JH), Hargreaves (H) and Makkink (M), the first and second equations of Baier and Robertson (BR1 and BR2), and the modified methods of Blaney and Criddle (BC) and Thornthwaite (T). The reference data were obtained from previous workers, who in 1977 and 1979 made micrometeorological measurements of E_e using the Bowen ratio method.

The relationships between estimated and measured E_e values excluding the T method in 1977 had correlation coefficients (R) ranging from 0.65 to 0.84. These were not significantly different at the 5% level. The Standard Errors of the Estimate (SEE) ranged from 0.43 to 1.69 mm d^?1. The PT, JH, BR1 and T methods had low SEE values, whereas the BR2, H, BC and M methods had high SEE values. Six of the eight methods (PT, JH, H, M, BR2 and BC) were calibrated for local conditions using 1979 data. After calibration, the methods were tested with the 1977 data. The results indicated significant improvement in the fit of four of the methods (H, M, BR2 and BC). Overall, it was concluded that after calibration, six of the eight methods had predictive power and fit that were not significantly different at the 5% level.