In arid and semi-arid regions, where meteorological infrastructure is limited, accurate estimation of reference evapotranspiration (ET0) is essential for efficient agricultural water management. The Penman-Monteith equation (PM FAO-56) is the reference method due to its physical robustness and accuracy; however, it requires variables such as relative humidity (RH), solar radiation (Rs), and wind speed (U2), which are not always available. In this study, the performance of the PM equation applied with missing data (PM-df) was evaluated in comparison with the standard version, using daily time series from six locations in the state of Chihuahua, Mexico. Scenarios involving the absence of single variables, combined variables, and all variables simultaneously were analyzed, along with their seasonal variability. The absence of a single variable resulted in acceptable performance in most cases, with the omission of Rs producing the largest errors, particularly during summer, when atmospheric demand is highest. The combination of two missing variables led to a further reduction in accuracy, especially when Rs was involved, highlighting its dominant role in ET0 estimation under dry climatic conditions. The greatest loss of accuracy occurred when all three variables were simultaneously absent, with coefficients of determination (R2) ranging from 0.75 to 0.92, and more pronounced errors in highly seasonal regions such as Camargo and Jim & eacute;nez, where the spring-summer contrast is more marked. In contrast, locations such as Aldama and Dubl & aacute;n showed greater stability in performance indicators. These findings emphasize that the sensitivity of the PM-df model depends both on the combination of missing variables and on the local climatic regime. The results support the use of PM-df under data-scarce conditions, while highlighting the need for future non-stationary approaches that better capture intra-annual variability.