Geoscience Reference
In-Depth Information
Figure 8.4
Subcatchments and observation sites in the Skalka catchment in the Czech Republic (672 km
2
)
(after Blazkova and Beven, 2009, with kind permission of the American Geophysical Union).
A modified version of TOPMODEL was implemented for the different subcatchments, driven by a
weather generator that was fitted to temperature, rainfall and snow accumulation data for sites in the
catchment, taking account of the elevation range in the catchment. The stochastic rainfall component
of the weather generator included joint distributions for both high intensity and low intensity events at
the subcatchment scale, and took account of the observed spatial correlation of rainfall amounts in the
subcatchments. Snowmelt calculations were made for 13 different snow accumulation areas spanning
the elevation range in the catchment using a degree-day method.
The purpose of the frequency estimation was to evaluate the safety of the dam at Cheb at the outlet of
the catchment. A particular feature of the study was that there were discharge measurements at Cheb for
60 years (1887-1959) before the dam was built. This includes 50 years prior to the earliest measurements
at any of the other sites, the oldest of which, at Marktleuthen on the Eger subcatchment, started in 1937.
The Cheb data were not used in calibrating the model but were retained as a validation check on the
frequency predictions.
The modifications to TOPMODEL were made because early runs of the model showed that it could
not reproduce the low flow characteristics of some of the subcatchments. A master recession curve was
derived from the observed discharges for each subcatchment, using the method of Lamb and Beven
(1997), and used to modify the subsurface store in the model. This would be even more important if it
was the distribution of extreme low flows that was of interest. The model was then applied within the
GLUE framework of Chapter 7, using the limits of acceptability approach in model evaluation described
