Methodological differences in projected potential evapotranspiration

There is growing concern that the higher temperatures expected with climate change will exacerbate drought extent, duration and severity by enhancing evaporative demand. Temperature-based estimates of potential evapotranspiration (PET) are popular for many eminently practical reasons and have served well in many research and management settings. However, a number of recent publications have questioned whether it is appropriate to use temperature-based PET estimates for long-term evaporative demand and drought projections, demonstrating that PET does not always track temperature. Where precipitation changes are modest, methodologically driven differences in the magnitude or direction of PET trends could lead to contrasting drought projections. Here I calculate PET by three methods (Hamon, Priestley-Taylor and Penman) and evaluate whether different techniques introduce disparities in the sign of PET change, the degree of model agreement, or the magnitude of those changes. Changes in temperature-based Hamon PET were more significantly and consistently positive than trends in PET estimated by other methods, and where methods agreed that summer PET would increase, trends in temperature-based PET were often larger in magnitude. The discrepancies in PET trends appear to derive from regional changes in incoming shortwave radiation, wind speed and humidity -- phenomena simpler equations cannot capture. Because multiple variables can influence trends in PET, it may be more justifiable to use data-intensive methods, where the source(s) of uncertainty can be identified, rather than using simpler methods that could mask important trends.