The catchment area of Timis-Bega (13,000 km2) is located in the western part of Romania. The orographical context, the Mediterranean climatic influence, the nature of the ground and the vegetation contribute to this relatively restricted geographical area’s great diversity of flows. This latter characteristic is based on a study of 19 sub-basins, ranging in size from a few km2 to several hundred km2. The regionalization of annual mean discharge and low flow in the Timis-Bega catchment was undertaken following similar work realized on the French Moselle catchment. This earlier work showed that the two-parameter Weibull distribution adequately described the spatial and temporal variability of annual mean discharge and low flow for the studied sub-catchments. The two principal results of the preceding study were the weak differentiation among the regional models established for annual mean discharge and low flow, and the recommended use of episodic gauging of low flows for the improvement of the regional models. Our intention was to verify these preceding considerations, established for the French Moselle basin, in the hydro-climatic context of Timis-Bega. The results obtained allowed us to confirm the weak differentiation among the regional models as well as the improvement of the regional models. More specifically, this improvement resulted from the better precision, compared to that obtained by traditional multiple regression methods, of the estimation of the local discharge descriptor, the median annual mean discharge or the median minimal daily low flow discharge. The use of low flow episodic discharge measurements, that are concomitant between the studied and reference sub-catchments, allows the estimation of the median minimal daily low flow discharge and the median annual mean discharge. This fortuitous result is explained by the weak differentiation observed between the regional annual mean discharge model and the regional average low flow discharge model for average to dry years. We believe episodic measurements of low flow should be considered when collecting data that will be used for the generation of regionalization flow descriptors by geostatistical methods. Lastly, the similarities between the regional distributions of annual mean discharge and low flow for both the Moselle and Timis-Bega regions were unexpected. This result should, however, be considered with caution, as it implies the existence of homogeneous hydrological regions.
Regionalization, annual mean discharge, low flow, statistical QdF modelling.
Unité de Recherche Hydrologie Hydraulique,
3bis quai Chauveau,
69336 Lyon Cedex 09,