More information about my Phd project

Runoff processes

Runoff formation in a catchment results from spatially and temporally varying runoff processes playing together. The most commonly distinguished processes are: runoff produced through limited infiltration capacity (Hortonian overland flow, HOF), through saturation of the soil (saturated overland flow, SOF), through fast infiltration and lateral flow along preferential pathways (subsurface flow, SSF). Deep percolation (DP) produces no fast runoff. On a specific site, several runoff processes might occur. The one process that contributes most to runoff is called the dominant runoff process.

Mapping of dominant runoff processes

The mapping of the dominant runoff processes is based on process knowledge acquired on experimental hillslope plots [Scherrer, 1997]. The dominant runoff processes on plot and hillslope scale are investigated by assessing of the storage capacity, the permeability and the layering of the soil. Besides, field experiments like infiltration and sprinkling experiments combined with tracer techniques are performed. These investigations allow, together with information about soils, topography, land-use, geology, etc., the mapping of the different dominant runoff processes in a catchment. The upscaling from plot to catchment scale is then done with expert knowledge. During the upscaling to catchment scale simplifications and generalisations have to be made. Adding to this, process interactions between neighbouring plots must be taken into account. All these points make it necessary to objectify and verify the upscaling procedure.

Pesticides as tracers

For the verification it is important to be able to differentiate between the different fast flowing runoff components. Systematic concentration measurements in the stream flow of pesticides, which have been applied to the fields under controlled conditions, are used to get spatially differentiated information of the catchment. In this way, the pesticides serve as artificial tracers. In the framework of the Greifensee project we investigate and map the dominant runoff processes in those catchments, where other groups of the project investigate and measure pesticide transport into rivers. Both partners will benefit in a co-operation between runoff processes research and the pollutant transport research.

Aerial Photography

The pattern of drying down of a catchment after a heavy rainfall event contains valuable information about runoff processes. On impermeable and deep percolation areas, the soil moisture content should decrease faster than on areas where saturated overland flow occurs. Direct airborne measurement of soil moisture changes in the microwave spectrum is technically demanding, expensive and not yet fully developed. Therefore, the potential of using a sequence of visible and infrared aerial photographs to map runoff processes will be assessed. A sequence of aerial photographs in the visible and infrared spectrum were taken after extensive rainfall events to document the dry down of the catchments.

Rainfall runoff modelling

One important aim of hydrological research is to improve the understanding as well as the prediction of catchment response to precipitation. The applications of this research reach from evaluating the effects of leaching of pollutants from a field into the groundwater to flood estimation in large basins. Vast efforts have already been made to develop reliable methodologies and models for these purposes. The huge increase in computing power has allowed to set up and run ever more complex models and optimisation procedures. However, such models require a large number of parameters, which must usually be found by calibration. This often leads to non-unique solutions for the parameter calibration.
A way out of this problem lies in the better understanding of the actual processes. In the IHW modelling approach the dominant runoff process map of a catchment serves as a base for a spatially highly differentiated conceptual model, where only a limited number of parameters has to be calibrated. If the spatial distribution of the dominant runoff processes in a catchment is known, separate models for each process can be combined. Using this a priori knowledge, the model is much simpler. This reduces the number of parameters and many of them can either be determined from measurements or at least their ranges can be limited. Only for such models multiresponse data (e.g. groundwater levels, soil moisture measurements, spatial distribution of mapped saturated areas, etc.) and not only discharge data alone can be used for calibration and verification. these results into rainfall runoff modelling. The benefits of basing the computations on actually observed processes are:

• An extrapolation beyond measured discharges is more reliable. This makes the methodology well suited for flood frequency studies.
• Land-use influences runoff processes. Certain, but not all runoff processes can be influenced by a change in land-use. It is not possible to evaluate the potential effects of land-use changes on the runoff without considering runoff processes.
• The runoff processes in a catchment influence pollutant transport from agricultural areas into rivers. For example, fields on which Hortonian overland flow occurs contribute heavily to pesticide transport into the river. Fields where deep percolation is the dominant runoff process do practically not contribute to pesticide leaching. To interpret pollutant measurements in the stream flow without knowledge of spatial distribution of the dominant runoff processes is not possible.

Soil maps enhanced for hydrological purposes

We investigate to what extend process knowledge can be extracted from soil and other maps. In most cases, soil maps are actually soil type maps, produced for agricultural and forestry use. Therefore, the summarising and classifying of units shown, does not adequately consider hydrological parameters influencing runoff formation. To enhance a soil map for hydrological purposes, all information contained in a normal soil map and other maps, will be re-evaluated with the help of runoff process knowledge. Then units will be classified according to hydrological parameters. The soil map used for this purpose is the digital soil survey map of the Kanton Zürich 1 : 50000 [FAL and VD-ZH, 1995 and 1998].

Literature

Scherrer, S. 1997: Abflussbildung bei Starkniederschlägen, Identifikation von Abflussprozessen mittels künstlicher Niederschläge. Mitteilung der Versuchsanstalt für Wasserbau, Hydrologie und Glaziologie, Nr. 147.