Spatial R-vine copula for streamflow scenario simulation

As there are spatial and temporal dependencies of streamflow at hydroelectric power plants (HPPs), we develop a high dimensional statistical model for streamflow scenario simulation which considers both. We use a pair-copula-based model reparametrized in terms of spatial variables derived from the river network, the distance between HPPs and precipitation measurements. This approach reduces the complexity of the model by reducing the number of parameters, preserves however the flexibility introduced by the pair-copulas. Based on simulations, we demonstrate that our model can capture spatial and temporal dependence between HPPs and thus, generate multivariate scenarios that reproduce historical features as shown in an empirical exercise with 39 HPPs.