Modeling extreme events requires the understanding of the spatial dependence structure in order to construct reliable statistical models. Assuming the stationarity of the dependence structure of the spatial process may not be reasonable, depending on the topology of the region under study for example. In environmental extreme events, different types of extremal dependencies could appear across the spatial domain. In this study, we present an adapted spectral clustering algorithm for spatial extremes by combining spectral clustering with extremal concurrence probability. This algorithm involves a heuristic method that can detect non stationarity in the dependence structure. In the case of a non-stationary dependence structure, the algorithm clusters the stations into k regions so that each region has a stationary dependence structure. To validate the proposed methodology, we tested it on different simulation cases using one or more max-stable models. The accuracy of the results encouraged us to apply it on two real data sets: rainfall data in the east coast of Australia and rainfall over France.