The spatial dynamics of potential evapotranspiration in Hawai'i: How driving variables are influenced by the trade wind inversion

Estimates of potential evapotranspiration (PET) based on data obtained from eight high elevation climate stations on the island of Maui, Hawai'i were derived using the Penman-Monteith method. These climate stations, part of the HaleNet climate network are located on both the leeward and windward slopes of Haleakalā Volcano and cover a 2030 m vertical elevation gradient (960 - 2990 m) representing points above, below and near the mean trade wind inversion (TWI, 2200 m). The mean annual cycle of PET as well as the variables that drive PET: relative humidity (RH), temperature (Ta), wind speed (Ws), and net radiation (Rn), were calculated at each station based on available data between 1988 and 2012. Linear regression was used to assess the effects of driving variables on PET across vertical and horizontal gradients. Net radiation is the primary control on PET estimates explaining between 74-97% of the PET signal. The relative importance of secondary controls, such as RH, differed for stations above and below the mean TWI level. For example, RH explained 45-83% of the variance in PET for stations below, and 1-14% above the mean TWI. Other secondary controls such as Ws and Ta also showed distinct dependence on elevation and differed between windward and leeward exposures of the island. Improved understanding of the spatial variations of different controls on PET will improve the ability to predict future changes in PET in response to changes in climate.