Regression models that estimate daily pan evaporation for inland and coastal regions of Southeastern U.S. were developed using observations of wind speed, solar radiation, minimum relative humidity, and maximum temperature. These weather elements are collected in numerous locations where measured pan evaporation records are not available, allowing an estimation of pan evaporation across large regions with higher point pattern density than is available using pan evaporation sites only. Sixteen models were developed and tested. An innovative model selection metric was developed, employing R square, Pearson's correlation coefficient, average difference between estimated and measured evaporation, root mean-squared error, and mean absolute error. Models selected included two validated for use in inland environments and one validated for use in coastal environments.