Decision support experiments and evaluations using seasonal to ...

data and forecast products that support
hydrologic and water supply forecasting
efforts in the United States. In Section
2.4, we provide a more detailed
discussion of pathways for improving
the skill and utility in hydrologic and
climate forecasts and data products.
Section 2.5 contains a brief review
of operational considerations and efforts
to improve the utility of forecast
and data products through efforts to
improve the forecast evaluation and
development process. These efforts
include cases in which forecast providers
and users have been engaged
in sustained interactions to improve
the use and utility of forecast and
data products, and have led to many
improvements and innovations in the
data and forecast products generated by national
centers. In recent years, a small number of
water resource agencies have also developed
end-to-end forecasting systems (i.e. forecasting
systems that integrate observations and forecast
models with decision-support tools) that utilize
climate forecasts to directly inform hydrologic
and water resources forecasts.
2.2 HYDROLOGIC AND WATER
RESOURCES: MONITORING AND
PREDICTION
The uses of hydrologic monitoring and prediction
products, and specifically those that are
relevant for water, hazard and energy management,
vary depending on the forecast lead
time (Figure 2.1). The shortest climate and
hydrologic lead-time forecasts, from minutes
to hours, are applied to such uses as warnings
for floods and extreme weather, wind power
scheduling, aviation, recreation, and wild fire
response management. In contrast, at lead
times of years to decades, predictions are used
for strategic planning purposes rather than
operational management of resources. At SI
lead times, climate and hydrologic forecast applications
span a wide range that includes the
management of water, fisheries, hydropower
and agricultural production, navigation and
recreation. Table 2.2 lists aspects of forecast
products at these time scales that are relevant
to decision makers.
Decision-Support Experiments and Evaluations using Seasonal to
Interannual Forecasts and Observational Data: A Focus on Water Resources
Figure 2.1 The correspondence of climate and hydrologic forecast lead time to user
sectors in which forecast benefits are realized (from National Weather Service Hydrology
Research Laboratory). The focus of this Product is on climate and hydrologic forecasts with
lead times greater than two weeks and up to approximately one year.
2.2.1 Prediction Approaches
The primary climate and hydrologic prediction
approaches used by operational and research
centers fall into four categories: statistical,
dynamical, statistical-dynamical hybrid, and
consensus. The first three approaches are objective
in the sense that the inputs and methods
are formalized, outputs are not modified on an
ad hoc basis, and the resulting forecasts are
potentially reproducible by an independent
forecaster using the same inputs and methods.
The fourth major category of approach, which
might also be termed blended knowledge, requires
subjective weighting of results from the
other approaches. These types of approaches
are discussed in Box 2.2.
Other aspects of dynamical prediction schemes
related to model physical and computational
structure are important in distinguishing one
model or model version from another. These
aspects are primary indicators of the sophistication
of an evolving model, relative to other
models, but are not of much interest to the
forecast user community. Examples include
the degree of coupling of model components,
model vertical resolution, cloud microphysics
package, nature of data assimilation approaches
and of the data assimilated, and the ensemble
generation scheme, among many other forecast
system features.
Climate and
hydrologic leadtime
forecasts
range from
minutes to years.
33