Prediction of low flows in ungauged basins - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 187

Which method performs best?
Figure 8.21 summarises the performance for different
regionalisation approaches, stratified by the aridity index.
The top, middle and bottom panels show the performance
for all catchments in Table A8.2 , and catchments with an
aridity index below and above 1, respectively. Overall, for
all catchments the performance of the global regression is
much lower than that of any other method. This is consist-
ent with the Level 1 assessment. In the arid catchments the
performance of the global regression is particularly low
and the absolute normalised errors are, on average, around
1.1. The poor performance of the global regression
methods is partly related to the biases of the methods but
the random errors are also large ( Figure 8.20 ). In the humid
regions the short records perform better than any other
method. This is, again, consistent with the Level 1 assess-
ment. However, this is no longer the case for the arid
catchments. For the arid catchments, the performance of
the short records is in fact lower than those of the geosta-
tistical methods and regional regression. It appears that in
arid regions the variability of the low flows between years
may be larger than in other climates, which makes the
record augmentation and other methods based on short
runoff records at the site of interest perform less well.
Another possible explication of the lower performance of
short records in arid regions is that gauging networks are
typically sparser in arid regions, so that donor gauges are
further away, less similar, and in final consequence less
appropriate for record augmentation. Methods may be
needed in arid regions that specifically acount for the run-
off generation processes in the region, and preferably are
based on proxy data that account for these processes.

0.0 All catchments

0.4

0.8

1.2

1.6

2.0

Humid catchments (aridity index<1)

0.0

0.4

0.8

1.2

1.6

2.0

Arid catchments (aridity index ≥ 1)

0.0

0.4

0.8

1.2

1.6

2.0

Geo-
statist.

Global
regress.

Regional
regress.

Short
records

Main findings of Level 2 assessment
-

Figure 8.21. Absolute normalised error (ANE) of predicting low
flows in ungauged basins for different regionalisation methods,
stratified by aridity. Lines connect median efficiencies for the same
study. Boxes are 40% - 60% quantiles, whiskers are 20% - 80%
quantiles. After Salinas et al.( 2013 ).

The performance of all methods for predicting low
flows in ungauged basins worsens with increasing
aridity and air temperature.

- There is a tendency for the performance to improve

with catchment elevation.

- Performance improves with catchment size, with the

exception of methods that use short runoff records at
the site of interest, which may be more dependent on
the temporal variability of low flows than on the
spatial variability.
- The global regression methods always exhibit lower
performance than other methods, particularly in arid
catchments.

catchment elevation. The bottom panels in the figures show
the performance as a function of catchment scale. For all
methods the performance increases with catchment scale.
This may be related to both data availability and space-time
aggregation of runoff processes in the catchments, which
will increase the predictability. The exceptions are methods
that use short runoff records at the site of interest. In these
cases, the performance dependence on catchment size is less
pronounced than for the other methods. These types of
methods may be more dependent on the representativeness
of the short runoff record to the temporal variability of low
flows, so the dependence on the spatial variability and
therefore catchment size may be lower.

- In humid conditions, methods that use short runoff

records at the site of interest perform much better
than any other methods. However, this is no longer
the case in arid regions when regional regressions
may perform better.