Spatio-temporal variation of reference evapotranspiration from empirical methods in Chihuahua, Mexico

Abstract

<strong>Introduction:</strong> Evapotranspiration is key in the management of arid agricultural areas. In Chihuahua, the volume of irrigation water is based on reference evapotranspiration (ET<sub>o</sub>) calculated with empirical methods and extrapolated to the cropped area, which is inaccurate. The alternative is to calculate ET<sub>o</sub> variation by spatial interpolation.</br> <strong>Objective:</strong> To analyze the spatio-temporal variation of ET<sub>o</sub> using empirical methods and spatial interpolation in Chihuahua, Mexico.</br> <strong>Methodology:</strong> Records from 33 meteorological stations from 1960-2013 and seven ET<sub>o</sub> estimation methods were used. The results were compared with the Penman-Monteith method, modified by FAO (PMMF), ANOVA analysis (P ≤ 0.05), and homogeneous ET<sub>o</sub> surfaces built from the point values by spatial interpolation.</br> <strong>Results:</strong> The Hargreaves method (R<sup>2</sup> = 0.91, RMSE = 1.16 and ME = -0.69 mm-day<sup>-1</sup>) had a smaller bias with respect to PMMF. ET<sub>o</sub> values ranged from 2.5 to 7.1 mm-day<sup>-1</sup> in a west-east direction, with maximum values at low elevations and minimum values at high elevations, which showed the influence of the Sierra Madre Occidental on ET<sub>o</sub>. This characteristic was most noticeable in the warm months (June to September).</br> <strong>Limitations of the study:</strong> The use of estimated data needs field validation.</br> <strong>Originality:</strong> The ET<sub>o</sub> estimation with seven empirical methods and one spatial interpolation method to extrapolate values to areas with scarce meteorological data.</br> <strong>Conclusions:</strong> The Hargreaves method allows estimating the spatio-temporal variation of ET<sub>o</sub> in large extensions and areas with limited meteorological information.</br>

Estimation of Reference Evapotranspiration in a Semi-Arid Region of Mexico

“…However, considering the month with the minimum ET 0 (February) and the PM method, HS yielded an ET 0 12.8% higher vs. PM; BC, 6.8% higher; and PM_NP, 14.2% higher. It is observed that HS and BC underestimate ET 0 over most of the study period, consistent with the findings reported by some authors for an agroclimatic region similar to the study area [ 71 ].…”

“…In addition, MBE is negative in the three periods, pointing to an underestimation with the HS _NP method. Some authors report this same ET 0 underestimation effect in semi-arid regions during the winter–summer period [ 27 , 71 ].…”

“…The latter parameter returned values close to 1, indicating that the estimated values are statistically similar to observed or reference values [ 16 ]. Some authors reported similar RMSE values (1.1 mm d −1 ) when comparing ET 0 estimated by the HS equation and the PM method on a daily basis [ 71 , 76 ]. However, some authors recorded an RMSE (0.7 mm d −1 ) for 10-day mean data, which is similar to the RMSE value obtained in the present study for 5-day cumulative ET 0 [ 27 ].…”

Estimation of Reference Evapotranspiration in a Semi-Arid Region of Mexico

“…However, considering the month with the minimum ET 0 (February) and the PM method, HS yielded an ET 0 12.8% higher vs. PM; BC, 6.8% higher; and PM_NP, 14.2% higher. It is observed that HS and BC underestimate ET 0 over most of the study period, consistent with the findings reported by some authors for an agroclimatic region similar to the study area [71]. However, when considering total ET 0 (whole study period) and the PM method, HS recorded an ET 0 value 5.5% lower vs. PM; BC, a value 15.6% lower; and PM_NP, 10.6% higher; therefore, HS was the equation that yielded values closest to the PM method.…”

“…In addition, MBE is negative in the three periods, pointing to an underestimation with the HS _NP method. Some authors report this same ET 0 underestimation effect in semi-arid regions during the winter-summer period [27,71]. The 5-day mean ET 0 estimate recorded the best values in most statistical parameters relative to the mean daily ET 0 .…”