Leaf wetness duration variable plays an important role in crop protection warning systems and in studies of gas exchange in plants. The main objective of this study was to estimate the leaf wetness duration in a bean crop (Phaseolus vulgaris L.) by modeling under full and deficit irrigation and to compare it with field measurements. Four models were tested: (1) the model of the number of hours with relative humidity above a threshold; (2) the model of dew point depression; (3) the model of classification and regression tree, and (4) the model of Penman-Monteith. The first two models presented the lowest statistical errors when the empirical thresholds were equal to 83-84% and 2.6-2.7 ºC. The statistical errors of the third model were nearly the same as those obtained for the other models since it uses three empirical thresholds. The errors for the Penman-Monteith model were reduced when the air temperature and humidity input data were measured in the deficit irrigation treatment. For the Penman-Monteith model simulations with the fully irrigated treatment data, the errors remained close to those observed for the simulations with the weather station data.
dew; empirical models; physical model
