BAMOS Vol 31 Special Issue October 2018 Bulletin Vol 31 Special Issue 01 2018 | Page 23

BAMOS
Special Issue
23
Counting on Rain
Progress, challenges and relevance to Society
Jason P. Evans
University of New South Wales
What rainfall predictions are needed for societal
activities? What is our ability to make these predictions
today and what is the potential for improvements in
these predictions over the next decade?
Most water resource issues derive from rain, whether there is
too much or not enough. Defining “too much” or “not enough”
depends intimately on both the time and area over which the
rain falls. Too much rain often results in flooding, spanning flash
floods to widespread flooding that occurs over time scales from
minutes to weeks and space scales from kilometres to 1000
kilometres or more. Not enough rain results in droughts which
generally occur over seasonal to multi-year timescales and 100s
to 1000s of kilometres. Societal activities like infrastructure
design, agriculture and water supply are impacted by floods
and droughts and hence are impacted by rain-producing
phenomena across all these time and space scales. This spans
many phenomena from local thermals, through thunderstorms
and fronts, to cyclones, monsoons, and large-scale climate
modes such as ENSO.
Beyond the time and space scales of the phenomena of
interest there is also the prediction timescale to consider. For
infrastructure design we need to understand the short time
and space scales that flooding can occur over, and be able to
predict the rainfall at these scales for the next few decades
which is the life of the infrastructure. For agriculture, predictions
at seasonal to annual scales are needed for farm management,
while decadal scales are needed to inform the development
of improved crop varieties. For water supply, predictions at
annual scales are needed for reservoir management, while
predictions over the next century are needed to plan water
supply infrastructure such as dams.
Research shows that large space and time scale phenomena
are somewhat predictable at seasonal to annual timescales
(Figure 1). These prediction timescales are a focus of significant
research efforts around the globe and improvements are
likely over the next decade. Currently, longer climate scale
predictions can be reasonable but substantial improvement
is needed. Models have shown slow improvements over the
last decade and this is likely to continue. Seasonal predictions
are a focus of both the World Weather Research Programme
(WWRP) and the World Climate Research Programme (WCRP)
that provide international coordination and leadership. Models,
data assimilation and ensemble systems all continue to improve
and are expected to continue improving over the coming years.
Small time and space scale phenomena are quite predictable
at short, weather forecasting timescales but these predictions
have rarely been investigated at the needed decadal and longer
timescales. This relatively new area of investigation has good
prospects for improvements over the next decade as models
are being actively developed and improved at these scales
and WCRP is now providing international coordination and
leadership through the CORDEX initative. These predictions are
currently strongly constrained by computational power, hence
continued increases in computing power will also enhance the
potential for fast progress in this area.
Figure 1: Time, space and prediction
timescales for precipitation phenomena
and societal activities.