This is version 2 of Princeton University MEaSUREs Potential Evapotranspiration (PET) dataset, which provides a set of estimates of PET based on near surface meteorology and surface radiation data derived from a combination of reanalysis, satellite and gridded gauge data. The rationale of the project is to reduce the error from the input meteorological forcing and provide a variety of widely-used PET methods for different research and application purposes.
PET is estimated using three methods: Penman open-water method (Penman), Priestley-Taylor method (PT), Reference crop evapotranspiration using the UN Food and Agricultural Organization approach (FAO). The Penman equation assumes PET occurs from an open water surface and calculates PET based on observations of surface net radiation, near-surface air temperature, wind speed, and specific humidity (Shuttleworth, 1993). The PT equation calculates PET based on surface net radiation and near-surface air temperature and does not account for the aerodynamic component (Priestley and Taylor, 1972). The FAO equation is a specific application of the Penman-Monteith equation for crop and short-grass reference surfaces and is based on surface net radiation, near-surface air temperature, wind speed, and specific humidity (Allen, 1998). As a follow-on to PET_PU_3H025 Version 1, PET_PU_3H025 Version 2 seeks to continue to reduce the error from the input meteorological forcing and provide a variety of widely-used PET methods for different research and application purposes. The modifications include the addition of the FAO tall reference crop equation, an extension of the radiation data and temporal coverage, the usage of the GLASS satellite albedo to estimate upward short-wave radiation and the estimation of ground heat flux based on diurnal phase shift of net radiation.
This second version of the dataset is estimated at a 3 hourly temporal resolution and 0.25x0.25 degrees spatial resolution globally, spanning the 33-year period 1984-2016. Datasets are stored as a 3-dimensional array with dimension 720 x 1440 x 8 for each day, in NetCDF-4 format.
The DOI for version 1 of the data is: 10.5067/GPZDZYELYG1A.