Evaluation of the Vapor Pressure Models in the Estimation of Actual Vapor Pressure and Evapotranspiration

The quantification of evapotranspiration (ET) is highly important in water resources planning. The most reliable method for determining ET is to measure it in the field using lysimeters. However, the use of lysimeters for measurement of crop ET (${\mathrm{ET}}_c$) is both expensive and time consuming, thus necessitating a cheaper yet reliable alternative method. One such method is to determine ET using the Penman-Monteith equation, which requires daily meteorological data as inputs. Three vapor pressure models (Models 1–3) are mentioned in Irrigation and Drainage Paper-56 of the Food and Agriculture Organization (FAO-56) for the estimation of actual vapor pressure ($e_a$) that is used in the estimation of reference ET (${\mathrm{ET}}_0$) by the FAO Penman-Monteith equation. Model 1, which uses daily maximum and minimum temperature and relative humidity (RH) data, is the preferred method to estimate $e_a$; other models are used in data-constrained situations. Another vapor pressure model (Model 5) is proposed in this study. The performance of the vapor pressure models along with two other models (Models 4 and 6) existing in the literature is compared with that of Model 1 in estimating daily $e_a$ and ${\mathrm{ET}}_0$ values. The $e_a$ values were underestimated by Model 2, whereas Models 3, 4, and 5 overestimated $e_a$. Results show that Models 4 and 5 (proposed model) performed better than FAO-56 Models 2 and 3 for estimating daily $e_a$ values for Roorkee station, India. However, Model 4, which uses only daily mean temperature and RH data, has lower accuracy in ${\mathrm{ET}}_0$ estimation when compared to Model 5. This is because Model 4 underestimated the daily saturation vapor pressure values. This suggests that daily mean temperature is not enough for the accurate estimation of daily ${\mathrm{ET}}_0$ values. The overall performance of Model 5, which uses the Lawrence dewpoint temperature-RH relationship, was more consistent and closest to Model 1 when compared to the other vapor pressure models for the study region in data-constrained situations. Furthermore, the effect of wind speed ranges on the performance of the vapor pressure models in ${\mathrm{ET}}_0$ estimation is also assessed. The effect of vapor pressure models in ${\mathrm{ET}}_0$ estimation is not considerable for light winds but becomes substantial for higher wind speeds.