Subsurface flow processes at the hillslope scale
Runoff generation processes were investigated at several hillslopes in mountainous catchments
in Germany and Austria. Therefore mainly tracerhydrological investigations were carried out. The
processes and hydrological responses in mountainous landscapes can be very diverse, even on
relatively similar hillslopes. However, the importance of shallow soil water and groundwater systems
was demonstrated in different areas. The spatial heterogeneity of hillslope processes is closely
related to highly variable soil structure overlain by land use and vegetation patterns. Future changes
in hydro-climatic input (e.g. rainfall, temperature and snow melt), or land use and vegetation cover,
will have a significant influence on the recharge of groundwater and, consequently, on the discharge
and composition of runoff components and their hydrochemistry. This shows the vulnerability of
groundwater, which is often used for the local water supply in mountainous regions.

Key publications:

  • Tilch N., Uhlenbrook S., Didszun J., Leibundgut Ch., Kirnbauer R., Zillgens B., Merz B.. 2003:
    Entschlüsselung von Abflussbildungsprozessen mit Hilfe tracerhydrologischer Ansätze in einem
    Wildbacheinzugsgebiet (Identification of runoff generation processes using tracerhydrological
    approaches in torrent catchment; in German). Zeitschrift Österreichische Wasser- und
    Abfallwirtschaft, Heft 1-2, 55. Jahrgang, 9-17.
  • Uhlenbrook S., Didszun J., Leibundgut Ch. 2004: Runoff Generation Processes in Mountainous
    Basins and Their Susceptibility to Global Change. In: Huber U.M., Reasoner M.A., Bugmann B.,
    (Eds.), 2003: Global Change and Mountain Regions: A State of Knowledge Overview. Advances
    in Global Change Research. Kluwer Academic Publishers, Dordrecht, in press.
Recent Results of Hillslope Investigations at the Brugga Basin
by Jens Didszun
Problem, Objective and Method
A long-term hydrograph separation showed that the shallow groundwater contributes about two-thirds
to the total runoff (Uhlenbrook et al. 2002). This highlights the importance of this runoff component.
Two hillslopes that are drained by a spring at the toe of each hillslope were investigated in further detail,
to extend the knowledge of runoff generation processes at the hillslopes contributing shallow groundwater.
In mountainous areas springs at this location are often used for local water supply because they show
a persistent discharge. The central question to be examined in this respect is how the land use and the
structure of the periglacial debris covers influence the runoff behaviour of the hillslopes which is observed
at the two springs. The chosen springs are more or less similar regarding their altitude, mean discharge,
length and inclination of the hillslope. But there are differences in the land use and the structure of the
periglacial debris cover. Although the structure of the periglacial drift covers in the Brugga basin is
complex due to the heterogeneity of the topography and the formation processes (Tilch et al. 2002),
four main layers reflecting the different stages and processes of their formation can be distinguished
(REHFUESS 1990). The differences between the two hillslopes with respect to the structure of the drift
covers are that the hillslope above spring A has no significant coarse top layer with boulders, but at the
hillslope above spring B coarse material of the top layer is visible at the surface. Furthermore the land
use differs, as pasture land and conifer forest is dominant at spring A and B, respectively.

During autumn 1999 the discharge during three rainfall events and several low flow periods was monitored
continuously. Furthermore water samples were taken at each spring with a 4-hour time interval using
automatic samplers. The water samples were analysed for stable isotopes (18 O and 2H), dissolved silica
and major anions (Cl-, NO3-, SO42-) and cations (Na+, K+, Ca2+, Mg2+). In addition, classical meteorological parameters were observed at a near climate station and rainwater was sampled every 2 mm and analysed
for stable isotopes (18 O and 2H), reported in ‰ relative to V-SMOW.

The results show perspicuous differences between the two springs that were more obvious than expected.
There were clear differences between the spring's hydrographs and between the chemical and isotopic
reaction to rainfall events. Spring A is characterized by a slow and delayed runoff behaviour (Fig. 1).
Although the time lag between rainfall and rising runoff is only a few hours, the peak discharge is
reached not until two to four days after the beginning of the rainfall, depending on its intensity. In
contrast to this, the time lag at spring B is shorter, the peak discharge is higher and reached about
two days earlier, and the recession is considerably steeper than at spring A. But despite the quick runoff
reaction of spring B it also shows a fairly constant discharge of 0.3 l/s during summer droughts, which
suggests that the spring is fed by at least two runoff components, a long-lasting base flow component
and a dynamic storm flow component.

Figure 1: Precipitation and discharge of both springs during autumn 1999; dotted lines represent interpolated
data due to missing data caused by technical problems.
The hydrochemical variations are in good agreement with the discharge behaviour. At spring A the
concentration of dissolved silica remains fairly constant throughout the events (Fig. 2). Only towards
the end of the recession a slight decrease of the silica concentrations is visible. In contrast, at spring
B there is a typical decrease in the silica concentrations in coincidence with the peak discharge.
During the investigated events the silica concentration dropped to a minimum of about 70 % of
the pre-event concentration, but reached the pre-event concentration again at the end of the
hydrograph recession. The concentrations of the main ions show more or less the same type of reaction.
At spring A the concentrations stay quite constant during the events, whereas at spring B a decrease in concentrations to a minimum of about 50 % of the pre-event concentration was observed.

Figure 2: Dissolved Silica concentrations, set to 100 % at the beginning of the event.

The runoff behaviour is also reflected in the isotopic compositions of the spring waters. Spring A
shows only little and not systematic variations in the ²H composition. At spring B is a trend towards
a ²H signature of a third component, different from the event and pre-event signatures. This reaction
is similar during each of the three investigated events. Because of the high d-value of ²H, it can be
assumed that this component reflects water from the debris and drift cover, recharged several months
ago during summer times.

To quantify the contribution of the different runoff components hydrograph separations using dissolved
silica and ²H were calculated. The theory of hydrograph separations using natural tracers is discussed
for instance by SKLASH AND FARVOLDEN (1979) and BUTTLE (1994). At spring A a two-component
hydrograph separation using dissolved silica was carried out accounting for the fact that only little
variations of the hydrochemistry could be found. Assuming that the two components are direct runoff
(rain water) and groundwater (hypothesis one), the fraction of the direct runoff, causing the little
decrease in the silica concentration, is about 3 % during the observed events (Fig. 4). The little direct
runoff could reach the spring within the event by flowing along preferential pathway (i.e. root channels,
earthworm channel etc.).The rest is delivered from the shallow and deep groundwater, however a
distinction between those two components was not possible. A two-component hydrograph separation
can also be calculated assuming a variable contribution of deep and shallow groundwater (hypothesis two),
with an increase of the latter during the event. However, the silica concentrations of these two groundwater components could not be estimated uniquely with the observed data, thus an estimation of the fraction
of each component would be quite uncertain. Considering this and the little variations of the silica
concentrations spring A shows only a slight change in the fraction of the runoff components for both hypothesis.

Figure 3: ²H (D) signature of spring water and precipitation at both springs.

At spring B a two-component hydrograph separation was not feasible, as the ²H composition indicated
a contribution of a third component (Fig. 3). Thus a three-component separation using silica and ²H
was calculated. The three components are direct runoff with very little silica and the ²H composition
of the rainwater, and the shallow and the deep groundwater. The ²H and silica concentrations for the
deep and shallow groundwater were determined using an end member mixing analysis (EMMA) according
to CHRISTOPHERSEN et al (1990), since it was no possibility to measure the concentrations directly.
Shallow groundwater contributes little already during base flow prior to the events, and becomes the
major component during the peak of the event (Fig. 4). During the three investigated events the fraction
of the direct component was about 10 % whereas the deep and shallow groundwater made up
approximately 40 % and 50 %, respectively (Fig. 4).

Figure 4: Hydrograph separations using dissolved silica (spring A) and using dissolved silica and ²H (spring B).

Discussion and Interpretation
These results show that runoff generation at the catchment of the two springs differs more clearly
than it could be assumed from the similar location of the springs at the toe of the two steep hillslopes.
The reasons for these differences are attributed to the land use and the structure of the debris and
drift cover at the hillslope above the springs. From the measured data, it can not be determined absolutely
which of the reason is the most important one, but it is likely that the root zone with many macropores
of the conifer forest together with the stony and very permeable debris cover enables a quick delivery
of water to spring B. In contrast, the grass covered hillslope of spring A with a slower percolation is
accounting for the delayed reaction of the hydrograph and minor contribution of direct runoff. The fact
that a highly dynamic shallow groundwater component (including contributions from soil water) is
dominant during events at spring B, but can not be observed at spring A, must be caused by the
different permeability of the drift cover. However, to quantify this effect and to identify the source
area of this component more precisely, further geophysical investigations are needed to gain more
insight into the structure of the drift covers. In particular, combined geophysical measurements and
tracer tests are required to understand the triggering process responsible for dynamic contribution of
shallow groundwater in further details.

Buttle, J.M. (1994): Isotope hydrograph separations and rapid delivery of pre-event water from
drainage basins. Progr. in Phys. Geogr. 18, 1, 16-41.

Christophersen, N.C., Neal, C., Hooper, R.P., Vogt, R.D., Andersen, S. (1990): Modelling
Streamwater Chemistry as a Mixture of Soilwater End-Members - A Step towards Second-Generation
Acidification Models. J. Hydrol. 116, 307-320.

Rehfuess, K.E. (1990): Waldböden - Entwicklung, Eigenschaften und Nutzung. Pareys Studientexte, 29,
Verlag Paul Parey, Hamburg und Berlin. Sklash, M.G. and Farvolden, R.N. (1979): The role of
groundwater in storm runoff. J. Hydrol. 43, 45-65.

Tilch N., Uhlenbrook S., Leibundgut Ch., 2002: Regionalisierungsverfahren zur Ausweisung von
Hydrotopen in von periglazialem Hangschutt geprägten Gebieten. Grundwasser, in press.

Uhlenbrook S, Frey M, Leibundgut Ch, Maloszewski P (2002) Residence time based hydrograph
separations in a meso-scale mountainous basin at event and seasonal time scales.
Wat. Resour. Res. 38, 6: 1-14.

back to top
zum IHF