Analysing spatio-temporal process and parameter dynamics in models to characterise contrasting catchments

The relevance of hydrological processes varies in space and time resulting in typical temporal patterns for catchments. Contrasting catchments moreover differ in their catchment metrics. Hydrological models claim to be able to reproduce typical temporal patterns of dominant processes using site-specific model parameters. Thus, patterns of temporal dynamics in dominant modelled processes and their corresponding dominant parameters are a fingerprint of how a model represents the hydrological behaviour of a catchment and how these process patterns vary between contrasting catchments. In this study, we demonstrate how catchment metrics, modelled processes and parameter dominances can be jointly used to characterise catchments. We assess how catchment characteristics are represented in spatiotemporal process dynamics in models and how to understand the reasons for hydrological (dis)similarity among catchments along a landscape gradient. For this purpose, catchment metrics which characterise contrasting landscapes (lowland, mid-range mountain and alpine catchments) are related to dominant processes and parameters which were provided by a temporally resolved sensitivity analysis (TEDPAS) and simulations of a hydrological model. Our study shows that the applied model is able to represent the different processes and their seasonal variability according to the specific hydrological conditions of the study catchments. By analysing catchment metrics, modelled processes and model parameters jointly, we show that the largest differences are identified for the alpine catchment, whilst similarities are found among the other catchments. Following a landscape gradient, high flow phases are dominated by different flow components. In contrast, the model shows groundwater dominance in low flow phases in non-alpine catchments while in the alpine catchment low flows in winter are mainly controlled by snow processes. The joint analysis of catchment metrics, temporal dynamics of dominant processes and parameters can therefore be used to better disentangle similarities and differences among catchments from different landscapes.