Repeating patterns in runoff time series: A basis for exploring hydrologic similarity of precipitation and catchment wetness conditions

Abstract

Runoff responses to precipitation at the catchment scale exhibit a high variability in space and time due to a complex interaction of numerous factors, e.g., topography, land use, soil properties, geology, and climatic conditions. To find similar patterns in the runoff response and examine the effects of these factors on runoff variability, previous studies have either compared and classified catchments in space or focused on grouping extreme events. Here, we analyzed runoff processes in three highly instrumented catchments in Germany and Austria individually and compared them to themselves in time. To this end, we used long-term time series of 10 to 13 years and classified runoff events as similar by performing a cluster analysis based on calculated goodness-of-fit criteria between each possible pair of runoff events. For each cluster, we examined the degree to which precipitation and catchment wetness conditions were similar to themselves at the respective times when similar runoff events occurred by calculating Spearman rank correlation coefficients (ρ) as well as their descriptive statistics. The similarities assessed varied among the three catchments, with the two catchments in western Germany with maritime climates showing a stronger correlation for soil moisture conditions (ρ = 0.76 and ρ = 0.74) for classified similar runoff events rather than precipitation (ρ = 0.26 and ρ = 0.36). The Austrian catchment with a predominantly continental climate showed an overall higher correlation for precipitation (ρ = 0.57) and a lower one for soil moisture (ρ = 0.53) for similar runoff events compared to the other two catchments. The proposed method assesses similarity of precipitation and wetness conditions under similar runoff responses, and gives an indication of possible influencing factors controlling runoff generation in the three catchments in relation to their respective wetness and precipitation patterns. The similarities investigated help identify similar catchment functioning and can be used, for instance, to develop enhanced catchment similarity indices.