Geoscience Reference
In-Depth Information
Box 16.2
(continued)
the ultraviolet part of the spectrum could be larger than previously thought.
Furthermore, climate models simulate changes in the NAO in broad agree-
ment with observations if forced by these changes. Improved inter-annual
to decadal predictions might therefore be possible through better represen-
tation of solar forcing.
•
Improved understanding of physical processes
. Assessing likely forecast
skill is diffi cult because hindcasts only sample a few decades, and are
made using much sparser observations for initialization compared to Argo
fl oats. Understanding the physical processes that control regional climate
variability and change is therefore crucial both for gaining confi dence in
forecasts and for improving climate models to ensure that the relevant pro-
cesses are properly represented.
Initialization of the current state of the climate is essential for decadal forecasts.
Sustaining the present observing system, especially Argo, is therefore crucial.
Indeed, the skill of forecasts that benefi t from Argo could be signifi cantly higher
than that achieved in historical hindcasts. Recently launched satellites will also pro-
vide new observations of important quantities, including sea ice thickness
(CRYOSAT), soil moisture and sea surface salinity (SMOS). There is also consider-
able scope to improve initialization techniques, especially by developing coupled,
multivariate assimilation systems.
Climate model errors are a major source of uncertainty (see Fig.
16.5
). There is
therefore considerable scope for improved skill through model development aimed
at reducing biases and improving the simulation of teleconnections and the response
to external forcing factors. This will be achieved by increased resolution as comput-
ers become more powerful, improved parameterization of unresolved processes and
better representation of interactions between aerosols and clouds. Progress in model
development may accelerate by studying the development of errors in seamless sea-
sonal to decadal predictions. Furthermore, techniques for combining different mod-
els to optimize the skill are under investigation.
Idealized experiments suggest that the predictability of AMV depends on the
initial state (Griffi es and Bryan
1997
; Collins et al.
2006
). Regime dependence of
skill could therefore be exploited further to increase confi dence in predictions under
certain circumstances. These windows of opportunity during which very skilful pre-
dictions could be achieved could therefore be used to give forecasts with higher (but
conditional) skill. This could arise, for example, if the effects of several different
sources of skill align to produce a particularly strong signal.
Seasonal forecasts have been available for over 15 years, but their use in disas-
ter management has been very limited (Braman et al.
2013
). Possible reasons for
this include inability to predict exactly where and when extreme events might
occur; skill of predictions varies from place to place, season to season and year to
Search WWH ::
Custom Search