Geography Reference
In-Depth Information
Figure 5.28. Normalised error (NE)
of predicting annual runoff in
ungauged basins as a function of
aridity (E PA /P A ), mean annual air
temperature (T A ) and catchment area
for different methods. Boxes are
40% - 60% quantiles, whiskers are
20% - 80% quantiles.
Global regr.
Regional regr.
Budyko
2.0
1.0
0.0
-1.0
-2.0
0.4
0.6 0.8
1.0 2.0
0.4
0.6
0.8
1.0
2.0
0.4
0.6
0.8
1.0
2.0
E PA / P A (-)
E PA / P A (-)
E PA / P A (-)
2.0
1.0
0.0
-1.0
-2.0
3
6
8
10
12
3
6
8
10
12
3
6
8
10
12
T A (
°
C)
T A (
°
C)
T A (
°
C)
2.0
1.0
0.0
-1.0
-2.0
0.05
Area(10 3 km 2 )
0.1
0.5
1
5
0.05
0.1
0.5
1
5
0.05 0.1 0.5
1
5
Area(10 3 km 2 )
Area(10 3 km 2 )
process-based approach and geostatistics perform best in
predicting mean annual runoff in ungauged basins. This
indicates that the use of regional data can improve the
predictions significantly beyond global methods. The
results also point towards the strength of the Budyko
model, which is very good even though it was not
calibrated.
Main findings of Level 2 assessment
The performance of all methods of predicting mean
annual runoff in ungauged basins decreases with
increasing aridity.
There is a tendency for the performance of the regres-
sion methods to decrease with air temperature but no
apparent dependence on catchment size.
 
Search WWH ::




Custom Search