[ << Band/volume 18 ] [ Freiburger Schriften zur Hydrologie ] [ >> Band/volume 21 ]

Band/volume 19: DIDSZUN J. (2004):

Experimentelle Untersuchungen zur Skalenabhängigkeit der Abflussbildung

The objective of this study was to investigate the scale dependence of runoff generation proc-esses and runoff components with discharge samplings on an event basis in catchments of low mountainous regions. The motivation results from the increasing number of applications of process-oriented catchment models to mesoscale basins, the process knowledge, however, implemented in these models was achieved by experimental research in microscale catch-ments. This research was part of the cooperative project "runoff generation and catchment modelling", funded by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG). This offered the possibility of research activities not only in the Dreisam basin (258 km², Black Forest Mountains), but also in the Rotherdbach catchment (0,09 km², Ore Mountains) and the Brachtpe catchment (2,6 km², Sauerland, Central West Germany). Be-sides some microscale experiments the research focused on numerous event samplings, i.e. simultaneous runoff sampling during precipitation events in several nested catchments. The samples were analysed for the major ions, dissolved silica and stable isotopes (Oxygen-18, Deuterium).
The experimental investigations in the microscale Rotherdbach catchment supported the pre-vious perceptions of runoff generation in this basin. Dye-tracer experiments provided evi-dence of fast lateral flow paths in the vicinity of the channel, delivering either pre-event or event water to the stream during precipitation events. The dominant runoff components of two sub-catchments were quantified by the analysis of the natural tracers and the application of hydrograph separation techniques. It became evident that one sub-catchment was dominated by fast interflow and surface saturation excess flow originating in the riparian zone. The inter-flow in the other sub-catchment was significantly slower, peaking one to three days after the precipitation maximum, the saturated areas in this sub-catchment being variable in size and of different vegetation structure.
In the Brugga basin, a sub-basin of the Dreisam, spring water of the elevated regions, hill-slope water and stream runoff could clearly be distinguished by their hydrochemical composi-tion. However, it was not possible to verify spatially and hydrochemically uniform hillslope water components. Neither the position on the hillslope nor the depth of the sampling location had systematic influences on the composition of the natural tracers. This can be explained by the significant heterogeneity of the subsurface flow paths in the unsorted and permeable mate-rial of the periglacial drift covers. The event sampling of a small forested catchment (1.5 ha) drained by a steep channel showed clear differences in comparison with the hydrochemical event reaction of the downstream catchment (1.1 km²) and the basin outlet (40 km²). The small hillslope channel is representing only one hydrotope which is dominated by fast inter-flow. In contrast, the typical sub-catchments of the Brugga basin (1?2 km²) include already the majority of runoff generation processes and runoff components respectively and therefore show a similar event reaction as to that of the basin outlet.
The event sampling in the Brachtpe catchment included four microscale sub-catchments, each about 0.5 km² in size, and the catchment outlet. A hydrological characterization of the sub-catchments and a quantification of the dominant runoff components was achieved by the analysis of the natural tracers and the use of the hydrograph separation technique. Differences between the sub-basins exist e.g. in the velocity of the runoff reaction caused by the fraction and the position of the saturated or quickly saturated areas. Normally, these areas also pro-duce increased fractions of event-water and surface runoff respectively. Only minor influ-ences on the runoff generation were shown for the landuse. This can be explained by the fact that the fast reaction is dominated by components originating in the swampy, saturated or quickly saturated areas near the streams especially in the headwaters. But these areas are more or less independent from the landuse. Overall, these results support the perceptions of the re-search in the Brugga basin: In the microscale sub-catchments it was possible to detect differ-ences in the hydrological and hydrochemical reaction. But at the basin outlet these differences could not be attributed to the respective sub-catchments anymore.
In both, the Brugga and the Dreisam basin, the spatial heterogeneity of the natural tracers was determined by synoptic sampling during low flow periods. The small sub-catchments showed the highest heterogeneity which decreased with increasing catchment size. On the one hand detectable thresholds are caused by the typical sub-catchment size in the Dreisam basin. On the other hand they are controlled by the distribution of the sampled sub-catchment sizes. A simulation of the tracer concentrations by using catchment properties (e.g. size, mean altitude, geology) and the landuse fractions was not possible.
Three event samplings in up to eight sub-catchments of the Brugga basin showed only minor differences in the systematic event reaction except for the effects of runoff concentration. Al-though the hydrological answer of the small sub-catchments (ca. 1.5 km²) was partly more pronounced than that of the rest, the reactions were not systematic. This means that there were differences in the hydrometric and hydrochemical reaction between the small sub-catchments during a single event and the reactions also varied between the events. The differences be-tween the bigger sub-catchments originate basically from the heterogeneity of the events (pre-cipitation, antecedent moisture). Regional distinctions in the runoff generation processes could not be found because the size of the investigated sub-catchments was big enough to comprise the majority of the runoff generation processes and therefore the sub-basins show similar hydrologic reactions. But it has to be considered that a prerequisite for the analogous reactions is the similarity of the catchment properties (topography, geology, precipitation) in these sub-catchments. In contrast, two event samplings in seven sub-catchments of the Drei-sam basin showed significant differences in the runoff generation of those sub-basins which differ from the Brugga basin with respect to geology and topography or exhibit a bigger frac-tion of urban areas (settlements, roads).
Catchments with these characteristics are mainly big-ger in size (> 40 km²), which results in a scale dependence of runoff generation in the Drei-sam basin. But this scale dependence does not originate from the basin size itself. It is rather caused by the scale dependent catchment properties. Although the effect of urban areas be-came evident only in the bigger catchments it can be treated as a hydrotope as well. The ap-plication of the statistic method principle component analysis on the event data from the Brugga and Dreisam basin on the one hand confirmed the systematic similarity of the differ-ent data sets, on the other hand, however, pointed out the variability of precipitation events. It was furthermore possible to statistically highlight the influences of urban areas and of runoff concentration on the hydrochemical event reaction.
In this work the effect of catchment size on the spatial heterogeneity of natural tracers and the systematic hydrological event reaction was presented. The work also offered the possibility of experimentally detecting the runoff generation processes, which take effect only in the mesoscale, and of comparing larger sub-basins in this regard. Thus, for the application of process-oriented catchment models to mesoscale basins both, microscale and mesoscale proc-ess knowledge is necessary. The limitations of quantifying runoff components highlighted in this work accentuate the need for future research in the field of new tracers and improved quantification strategies.

PDF-Download (4,0 MB)