and naturally results in a variety of research interests. The complex interactions between hydrology, chemistry and ecology have ensured that process studies remain a vital element of catchment studies. A greater understanding and synthesis of the processes and mechanisms responsible for streamflow generation, variation in flow components and cycling of the main nutrients in different physiographic and climatic conditions is needed.
The current phase of FRIEND has seen developments in three main areas. Firstly, the knowledge gained in experimental research has been used to validate and improve mathematical modelling of hydrological processes (Güntner et al., 1999, Warmerdam et al., 1999, Koivusalo et al., 2000). Secondly, the cycling of nutrients and transport of solutes (Andersson & Lepistö, 1998, Peschke & Seidler, 2000, Sambale et al., 2000) has been studied and, thirdly, changes in hydrological regime due to land use and climate changes have been estimated in small catchments (Buchtele et al., 1998, Querner et al., 2000).
FRIEND has provided a valuable framework for exchanging experience among participants.
As the research interests of group participants are close to those of members of the European Network of Experimental and Representative Basins (ERB), annual ERB conferences have been, and will in the future be, organised jointly with FRIEND (e.g. Liblice, Czech Republic, 1998; Ghent, Belgium, 2000; Demanovská dolina, Slovakia, 2002). Proceedings of the International ERB-NE FRIEND Project 5 Workshop held in St Petersburg, Russia, during the last phase of FRIEND have now been published (Herrmann, 1999).
Tracer Aided Catchment model (TAC)
As noted above, a variety of hydrological processes can be studied in small catchments.
Nevertheless, a common output of much of the research is rainfall-runoff modelling. Despite improved knowledge on runoff generation, gained from small catchments during the last decades, there are still gaps in the development of rainfall-runoff models. One example of where knowledge of runoff generation has been applied in development of a rainfall-runoff model is the conceptual Tracer Aided Catchment (TAC) model (Uhlenbrook & Leibundgut, 1999). Model development comprised several steps:
NE FRIEND
l Distribution of the catchment into zones with the same dominant runoff generation processes;
l Development of the runoff generation module, taking into account the dominant runoff generation process;
l Spatial delineation of zones with the same dominant runoff generation processes (Rutenberg et al., 1999) was based on a digital terrain model and maps of geology, forest habitat and saturated areas. Eight spatial units were defined including saturated areas, urban areas, moraines, boulder trains, flat areas, hilltop areas and gently sloping hilly uplands, steeper hillslopes and boulders on top of steeper units.
The runoff generation module was based on linear and non-linear reservoirs. It distinguished seven zones based on spatial delineation and tracer studies, each with a characteristic dominant runoff generation process. In such a way, different runoff components were calculated. All runoff components from the different reservoirs were added together to give the total catchment simulated discharge.
Good agreement was achieved between simulated and observed discharge for the period July 1995 to March 1997. The modelled fast, delayed and slow runoff components corresponded to the results given by 18O measurements (Figure 2.10). The simulated and measured silica concentrations were also compared. Although for the whole period investigated, a correlation coefficient (r2) of 0.36 was achieved, for shorter periods reasonable agreement could be found (see Figure 2.10). Model simulations appear plausible, as not only was the total discharge simulated reasonably well, but, modelled proportions of the runoff components were also correct (multiple response validation).
The TAC model is therefore a promising approach to rainfall-runoff modelling which takes into account the results of field studies. For hydro-chemical problems, improved representation of the response of different catchment units to rainfall, in a way that enables practical application of the model, is important, and may also contribute to the simulation of floods and justify assessments of climate and land use impacts.
Figure 2.9 Prediction bounds of flood frequency curve on Ryzmburk catchment, Czech Republic
Hydrological effects of forestry treatment
Many experimental catchments are situated in forested areas and forestry treatment can lead to major land use change in such areas. Seuna (1999) in a summary paper describes the effects of forestry treatments in Finland, namely forest harvesting and draining.
Forestry drainage has been carried out on about 20% of the land area of Finland. It was found to have two major effects: a decrease in groundwater levels and a change in the hydraulic properties of the basin. These result in depletion of groundwater storage, decrease in evaporation from the soil surface, increase in the water storage capacity of peaty soils and a change in transpiration. As a result, total runoff increased by 0.3-0.6% per 1% drainage during the first 10 years. This increase was higher during the first years after drainage and from very wet areas, and lower after 10 years, from dry areas or if the drainage percentage was above 50%. After 15-20 years the total runoff returned to its original level. Drainage often increased snowmelt and summer runoff maxima, and clearly increased minimum runoff and sediment transport.
Forest harvesting (cutting of trees) increased total runoff by 5-10 mm per 10 m3 timber/ha removed. The spring (snowmelt) maximum increased by 0.5-0.8% per 1% of clear-cut area;
the summer maximum increased by 0.3-1% per 1% of clear-cut area. The summer minimum increased strongly but only for a short period. Suspended solids increased by 0.4-1% per 1%
of clear-cut area, but generally remained low. Phosphorus loads increased strongly after forest harvesting, especially from peaty soils. Nitrogen loads also increased, but to a lesser extent and probably for a shorter period. Groundwater tables increased after forest removal.
As both forestry drainage and clear-cutting tend to have parallel effects on hydrological conditions, application of both treatments in the same area would therefore increase the risks of flooding and erosion (Figure 2.11).
NE FRIEND
Figure 2.10 Example of TAC simulation during validation period (left) and comparison of modelled and tracers-calculated runoff components (right), from Uhlenbrook & Leibundgut (2002)
RUNOFF COMPONENTS
Determined with 18O and 3H measurements
direct runof f (11%)
Discharge [mm.d -1] simulated discharge observed discharge
3-V-98 12-VI-98 22-VII-98 31-VIII-98 10-X-98
2
2.4 Conclusions
2.4.1 Achievements
Northern European FRIEND has evolved into a well-developed network of hydrologists, who are implementing an active research programme. Annex 5 lists about 150 publications that have been completed by Northern European FRIEND participants since 1997 and which illustrate clearly the breadth of research topics and the range of output, including journal and conference papers, contract reports, software and spatial and temporal databases. This research is being transferred to the user community mainly through national partnerships between research groups and operational hydrological agencies. Examples of practical applications include:
l Regionalisation of flood frequencies;
l Estimating uncertainty in flood prediction;
l Regionalisation of low flows;
l Pan-European drought analysis and monitoring;
l Abstraction management;
l Determination of small-scale hydropower potential;
l Impacts of climate change on water resources;
l Estimating the balance between water supply and demand at the European scale;
l Improved process understanding underpinning model development.
Northern European FRIEND has also supported capacity building in other FRIEND regions, particularly by contributing to the running of training courses in low flow hydrology in Southern Africa and HKH FRIEND.