This article presents a conceptual study towards establishing a new method for altering lateral boundary conditions in numerical model based estimates for Probable Maximum Precipitation (PMP). We altered an extreme event in a physically and dynamically consistent way in a regional convective-scale weather prediction model (AROME-MetCoOp) by applying fields from a global ensemble climate model approach based on EC-EARTH. Ten ensemble members are downscaled with the regional model, which results in 10 different realizations of an extreme precipitation event for the West Coast of Norway. We show how the position and orientation of the moisture flow is different between the individual ensemble members, which lead to relative large changes in precipitation values for a selected catchment. For example, the modification of the moisture transport on scales of several hundred kilometers impacts the extreme precipitation amount by about 75% among the model members. Compared with historical rainfall records, precipitation changes of 62 % and 71 % are found for two selected catchments. Although the present study is restricted to one particular extreme event which is modified 10 times with the ensemble approach, there is a considerable spread of the moisture transport compared to the spread of the moisture transport of extreme precipitation event of the past 40 years. We conclude that the described approach is a step towards a new method to derive PMP values for a given catchment, however, a larger amount of events and larger ensembles would have to be considered to estimate PMP values.