Prediction of floods in ungauged basins - Runoff Prediction in Ungauged Basins: Synthesis across Processes, Places and Scales - page 226

stream network structure to generate improved flood
predictions in ungauged basins. Overall, progress in all
of these areas is helping to advance the field of flood
frequency hydrology generally.

assessment). Both Level 1 and Level 2 assessments
indicated that the geostatistics method has the best per-
formance (especially when data availability is high),
index methods work next best, and regression methods
relatively the worst. The Level 2 assessment also indi-
cated that

Deciphering these regional patterns, and juxtaposing
them against regional climate patterns, patterns of land-
scape organisation, and patterns of inundation obtained
from satellites during major floods, are examples of
strategies that can be adopted globally,

index methods do not work well in arid

regions.

Flood frequency hydrology deals with the extreme end
of the natural runoff variability, yet in a way that com-
bines process hydrology, comparative hydrology and
paleohydrology (including learning from the recent past
history of floods). Connecting these to patterns of other
co-evolutionary indices in the landscape can be exciting,
and has the potential to reveal much about all aspects of
catchment hydrology, eventually leading to improved
predictions.

including in

developing countries.

Comparative performance assessment of several flood
frequency estimation methods has indicated, as in all
previous signatures, that predictive performance
worsens with increasing aridity (both Level 1 and Level
2 assessments). Also, as expected, predictive perform-
ance increases with increasing catchment area (Level 2