Direct contact membrane distillation experiments were carried out under this work to study the influence of operational variables on membrane wetting. In the first part of this work, experiments were designed according to a Box-Behnken methodology and results were analyzed statistically using Pearson correlation coefficients, principal component/factor analysis and cluster analysis. The independent operational parameters were the temperatures of both the hot and cold streams (T-f, T-c) and their flow rates (F-f, F-c). The analyzed responses were the time and rate of wetting along with distillate flux. Statistical analysis showed strong evidence of a relationship between the selected variables and the wetting patterns. In general, parameters enhancing flux production led to suppression of wetting (both delayed wetting and reduced wetting rate). The second part of the work focused on reversing the wetting with minimal operation disruption by varying the operational parameters. The data generated helped in understanding the salt passage and wetting mechanisms. The wetting hypothesis developed herein is based on water bridging as a consequence of the weak hydrophobicity of the PVDF membrane and a net absolute transmembrane pressure. Data were analyzed through the Peclet number, the Poiseuille flow and a mass balance in order to understand the interplay between diffusion and convection/advection. High transmembrane temperature (Delta T) (Delta T = T-f-T-c) counteracts the build-up of a net absolute trans membrane pressure and reduces the viscous liquid flux. In this case, the diffusion of salt through the stagnant water layer in the membrane pores (a much slower mechanism) becomes more important and the wetting rate can be reduced and further reversed. (C) 2016 Elsevier B.V. All rights reserved.