An R Package for Implementing Multiple Evapotranspiration Formulations

The use of multiple evapotranspiration (ET) models is critical for exploring the ambiguity in the representations of ET processes. Although ensemble ET models are increasingly used to address this ambiguity, practical issues include: 1) the diversity of process representations, which require different input data and constants; 2) the diversity of nomenclature, terminology and units used in the literature; and 3) the complexity of some formulations, requiring significant time for coding and leading to potential user errors. We describe an R package that estimates both actual and potential ET from 17 well-known formulations. Results are presented as summary text and plots, and the package also can be easily coupled to rainfall-runoff modelling packages such as hydromad to estimate the effect of changing ET on runoff response. Additional plotting tools within the package allow users to visualise the association between estimated ET and climate variables such as temperature, solar radiation, wind and relative humidity. We provide a case study using Penman, Penman-Monteith FAO56 and Priestley-Taylor potential ET estimated using historical data from Kent Town weather station in Adelaide. The estimation from Priestley-Taylor formulation can be up to 20% lower than the estimation from the other two formulations which indicates the importance of the advection process at Kent Town.