Geoscience Reference
In-Depth Information
quantify trends in the intensity and frequency of extremes on different space/time
scales? Despite a continuous improvement in most observing systems, high-frequency
information (e.g., hourly precipitation) required to properly assess extremes is often not
made available and shared. New satellite observations and the synthesis of all
observations will help and may free up some data. Metrics for quantifying extremes
need to be assessed, and new ones should be introduced to improve diagnostics of
extremes and scale them to different areas. It is necessary to determine for which
regions (national observing systems) the requirements are close to being satisfied and
where they are not. There is an urgent need for research on design, development and
maintaining optimum observing systems, the regular analysis of their adequacy/
inadequacy for future investments in such systems.
• How can models be improved in their simulation and predictions or projections of the
magnitude and frequency of extremes? Current models have difficulty in simulating the
hydrological cycle, and they typically have problems handling the diurnal cycle. Model
resolution is insufficient in most cases to simulate many of the extremes of interest,
including floods with scales of a few kilometers and even drought whose worst-affected
areas are typically in areas only of order a few 100 km or less. Model parameterizations
addressing precipitation, convection and clouds are insufficient for accurate simulation
and timing of many extreme events. Models need to be confronted with the new
observational products in innovative analyses and with new diagnostics and metrics of
performance. This includes numerical weather prediction and climate models. There
are conceptual difficulties in validating model results against observations, first of all
associated with (but not limited to) co-location in space and grid cell data versus point
measurements. Many observational products are developed independent of models so
that gridding projections and associated error characteristics are often different from
model-derived data products thus making their direct inter-comparisons very difficult if
not impossible. Focused investments by space agencies (e.g., ESA and NASA) to make
the observational products consistent and inter-comparable is quite timely. Such efforts
facilitate research on observations and make inter-comparisons with models much
easier and enhance the use of observations by the modeling community.
• How can the phenomena responsible for extremes be better simulated in models? Many
phenomena that are responsible for extremes are not well simulated in models; some
because of resolution (such as tropical storms and highly localized precipitation
events), but also others that are resolved (such as blocking anticyclones). As well as
statistical analyses, studies should examine the phenomena responsible for extremes,
whether and how well they are depicted in models, and how to overcome incompatible
resolution requirements. Developmental needs should be used to focus field programs,
process studies and numerical experimentation.
• How can we promote development of applications for improved tracking and warning
systems arising from extremes? It is essential to develop ways to better assess changes
in risk of drought, floods, river flow, storms, coastal sea-level surges and ocean waves.
Such information has the greatest benefit to society for management of risks associated
with these events to reduce their adverse impacts. In most cases, such applications will
be done in conjunction with the CLIVAR and CliC projects and made available through
networks sponsored by GFCS and other regional climate information systems.
Prospects for advancements are excellent on this question because of new observations,
research, modeling and prediction activities already underway and planned. A number of
