Comparison of Mid-Century Temperature-Indexed and Change-Factor-Derived Potential Evapotranspiration in the Delaware River Basin.

The geospatial availability of historical and projected time series of precipitation and surface air temperature make them the variables most commonly used for estimating potential evapotranspiration (PET) in hydrologic-impact analyses of climate change.  Other relevant factors, including solar radiation, wind speed, and canopy measurements, are much less readily available.  The result is that simulations of soil-water storage and streamflow are frequently conducted using a temperature-indexed approach, often with incorporation of the well-understood relations among latitude, day-length, and angle of solar incidence.  Previous researchers have noted a high sensitivity of PET to temperature when calculated using a temperature-indexed approach in combination with daily temperatures indicated by climate projections.  This concern is relevant to the Hamon derived estimate of PET used in the Water Availability Tool for Environmental Resources (WATER), because that temperature-indexed estimate of PET is based on a historical, empirical relation between temperature and energy availability that cannot be assumed to hold under anthropogenic climate change, including a forecasted increase in temperature.  We introduce a multiplicative change factor (delta) based on a radiation-based PET (Priestley-Taylor) that can be applied in the same manner as the change-factor approach frequently applied for precipitation forecasts.  We will discuss the significant difference in PET forecasted using the temperature-indexed approach versus the change-factor approach for the Delaware River Basin for the 2030 and 2060 time periods relative to historic estimations of PET.