A Sensitivity Analysis of a Numerical Model for Estimating Evapotranspiration

A simulation approach has been used to study the sensitivity of evaporation, estimated with an energy and moisture balance model, to the meteorological data used to drive the model. The numerical model solves the surface energy balance equation. Daily evaporation is calculated by integrating the instantaneous fluxes. A Monte Carlo method is used to apply random errors to the meteorological data, and constant biases are also applied. This was done for both bare and wheat fields. Previous sensitivity studies have used analytical methods, ignoring the interaction of the meteorological measurements, their biases, and the feedback mechanisms which couple the energy and moisture fluxes. The method described here removes these restrictions and does not require linearity and differentiability in the evaporation estimation technique. The numerical method was shown to produce results generally similar to those expected analytically for random errors under near-potential conditions. Measurement biases produced larger errors in evaporation than random errors. The method will be applicable for dry conditions, where an analytical approach is not feasible.