Accurate estimation of wetting front dimensions in surface and sub-surface irrigation systems is very important for optimal management of water resources as well as increasing the performance of irrigation systems. Analytical and numerical models need lots of computational efforts and costs, so their applications in practical issues might be difficult. Meanwhile, artificial intelligence models could be suitable substitutes for these models. The present study aimed at simulating wetting front (wetting bulb) dimensions in different soil types for surface and sub-surface irrigation systems using continuous and pulse applications, through the gene expressions programming (GEP) and random forest (RF) techniques. Different experiments were carried out to achieve the required data. A robust k-fold testing data scanning technique was adopted for assessing the applied models, where different criteria were considered for defining the "k" values per test procedure. The results obtained revealed that, with minor exceptions, both the applied GEP and RF models presented good ability in modeling the wetting front dimensions in all the conducted treatments. The results also showed that soil type and emitter installation depth were the best criteria for defining the "k" values of k-fold testing procedure for the surface and sub-surface irrigation systems, respectively.
