The dataset contains the A and B sample matrices with all the parameters used in the paper "Current models underestimate future irrigated areas". It includes the model output, i.e. the empirical distribution of irrigated areas in 2050 [column Y, in Million ha (Mha)] jointly with all the other inputs (i.e., r, gamma) and triggers (i.e. X1, X2, ..., W1, W2) required to replicate the results of the manuscript.