Decision support experiments and evaluations using seasonal to ...

Decision-Support Experiments and Evaluations using Seasonal to
Interannual Forecasts and Observational Data: A Focus on Water Resources
Figure 2.20 The NCEP CPC seasonal outlook for precipitation in the Seattle Region Climate Division
(Division 75 in Figure 2.19) shown as the probability of exceedence for total precipitation for
the three-month target period .
evolving changes in ocean circulation that can
alter SSTs and thereby change the boundary
conditions for the atmosphere. Not all possible
sources of SI climate-forecast skill have been
identified or exploited, but contributors that
have been proposed and pursued include a
variety of large-scale air-sea connections (e.g.,
Redmond and Koch, 1991; Cayan and Webb,
1992; Mantua et al., 1997; Enfield et al., 2001;
Hoerling and Kumar, 2003), snow and sea-ice
patterns (e.g., Cohen and Entekhabi, 1999; Clark
and Serreze, 2000; Lo and Clark, 2002; Liu et
al., 2004), and soil moisture and vegetation
regimes (e.g., Koster and Suarez, 1995, 2001;
Ni-Meister et al., 2005).
In operational practice, however, most of the
forecast skill provided by current forecast
systems (especially including climate models)
derives from our ability to predict the evolution
of ENSO events on time scales of 6 to
12 months, coupled with the teleconnections
from the events in the tropical Pacific to many
areas of the globe. Barnston et al. (1999), in
their explanation of the advent of the first operational
long-lead forecasts from the NOAA
Climate Prediction Center, stated that “while
some extratropical processes probably develop
independently of the Tropics… much of the
skill of the forecasts for the extratropics comes
from anomalies of ENSO-related tropical sea
surface temperatures”. Except for the changes
associated with diurnal cycles, seasonal cycles,
and possibly the (30 to 60 day) Madden-Julian
Oscillation of the tropical ocean-atmosphere
system, “ENSO is the most predictable climate
fluctuation on the planet” (McPhaden et al.,
2006). Diurnal cycles and seasonal cycles are
predictable on time scales of hours-to-days and
months-to-years, respectively, whereas ENSO
mostly provides predictability on SI time scales.
Figure 2.21a shows that temperatures over the
tropical oceans and lands and extratropical
oceans are more correlated from season to
season than the extratropical continents. To the
extent that they can anticipate the slow evolution
of the tropical oceans, indicated by these
correlations, SCFs in the extratropics that derive
their skill from an ability to forecast conditions
in the tropical oceans are provided a basis for
prediction skill. To the extent that the multiseasonal
long-term potential predictability of
the ENSO episodes (Figure 2.21b) can be drawn
upon in certain regions at certain times of year,
Most of the skill
provided by current
forecast systems
derives from our
ability to predict the
evolution of El Niño–
Southern Oscillation
events on time scales
of 6 to 12 months.
49