Geography Reference
In-Depth Information
Table 11.8. Cross-validation performance of estimating the 100-year flood in ungauged basins in Austria
as a function of catchment area (only gauging stations with more than 40 years of flood records were used)
Catchment area
Number of catchments
Rbias (%)
RRMSE (%)
< 100 km²
65
18.7
43.4
100 - 1000 km²
180
8.38
40.5
> 1000 km²
53
-4.56
19.6
All catchments
298
8.32
38.4
Figure 11.38. Estimates of the normalised specific 100-year flood from top-kriging (shown as width of the stream network) in the Danube region
in Upper Austria. The estimates of the gauged catchments are shown as circles. Units are in (m 3 /s)/km 2 . From Merz et al. ( 2008 ) .
enhancing the robustness and reliability of the estimates.
Overall, the estimates of the T-year floods for the stream
gauges of this study were deemed more reliable than stat-
istical estimates from the flood samples alone, based on the
judgement of the staff of the Hydrographic Services. The
regionalisation approach
characteristics of the catchments including precipitation,
catchment soil moisture (through mean annual precipita-
tion) and flood retention by reservoirs and lakes. Leave-
one-out cross-validation tests indicated that the accuracy of
the estimates is very good across Austria. However, the
accuracy differs in terms of catchment size (where floods
in larger catchments are estimated more accurately than in
smaller catchments) and mean annual precipitation (where
a combination of top-kriging
based on spatial correlations with additional catchment
attributes
-
-
was
able
to account
for hydrological
Search WWH ::




Custom Search