Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 110
9
RECOMMENDATIONS
Over the past decade we have learned much about the complex natural
processes that influence climate variability and change, and our ability to
model climate has increased significantly. We have gained a better
appreciation for the important connections between physical, biological, and
human dimensions of the climate system. We have also begun to better
identify those parts of the climate system that are particularly important and
not well understood, and that therefore limit our ability to project the future
evolution of Earth's climate. A critical area where understanding is needed
is the role of feedbacks in the climate system and their role in determining
climate sensitivity.
This Panel believes that refining our understanding of the key climate
feedback processes and improving their treatment in models used to project
future climate scenarios is an effective way forward in the quest to better
understand how climate may evolve in response to natural and human-
induced forcings. An appropriate strategy for accomplishing this is to make
more rigorous comparisons of models with data and to focus particularly on
observational tests of how well models simulate key feedback processes.
This report highlights broad guidance on the key avenues of research
that need to be pursued to better understand climate feedbacks and is
intended to call attention to those areas where additional focus might be
productive in the near term. The key finding of this report is that an
enhanced research effort is needed to better observe, understand, and model
key climate feedback processes. Research on climate feedback processes
should be designed to
Integrate observational and modeling efforts toward understanding and
modeling of climate feedback processes;
· Integrate the subdisciplines of climate science for a comprehensive
study of the key climate feedback processes; and
110
OCR for page 111
RECOMMENDA TIONS
111
· Integrate different time scales of weather and climate variability into
studies of climate feedback processes.
KEY OBSERVATIONS NEEDED TO MONITOR AND
UNDERSTAND CLIMATE FEEDBACKS
Because climate feedback processes are so important to understanding
climate change, it is necessary to monitor the variables that characterize the
feedback processes as well as the variables that define the basic climate. In
addition to temperature and precipitation, variables such as clouds, water
vapor, aerosols, land surface properties, snow cover, sea ice, and radiation
budget quantities need to be monitored. Stable long-term measurements of
these variables can be used to monitor the feedback processes, to better
understand these processes, to identify the contributing causes to observed
climate changes, and to improve confidence in climate projections.
Recommendation:
An integrated global climate-monitoring system must include
observations of key climate feedback processes. Stable, accurate, long-
term measurements should be made of the variables that characterize
climate feedback processes. ~
As climate and greenhouse gases change, certain variables (see below)
must be adequately monitored to advance the objectives of climate change
feedbacks research and to define the state of the climate system including
feedback processes. Although some of these observations are made, there
are deficiencies. Some of these observations are made across insufficient
lengths of time or across too limited regions, or are not made routinely and
globally as required for global climate monitoring. Other measurements are
made globally, but lack the quality required for long-term climate
monitoring and analysis (see NRC, 1 999a, 2000b).
1. Observations with Insufficient Time or Space Resolution
Ice thickness;
· Temperature and salinity of the upper ocean and other portions of the
ocean that affect interannual to decadal climate change;
· Atmospheric trace gas concentrations (e.g., CO2, O2-N2, CH4, 03) and
ocean chemistry; and
Soil moisture profiles and snow properties (e.g., depth, moisture
equivalent, snow state).
OCR for page 112
112
UNDERSTANDING CLING TE CHANGE FEEDBACKS
2. Observations That Do Not Meet Quality Standards
· Temperature and humidity (particularly in the upper troposphere and
stratosphere), precipitation, and wind,
Global cloud and aerosol distributions and properties;
Sea ice margin characteristics;
Terrestrial vegetation, and snow extent;
Radiation budget at the top-of-tropopause and at the surface; and
Ocean color.
The collection and validation of these datasets will require international
collaboration and cooperation among U.S. agencies. As recommended in
several previous NRC reports, there are advantages to collecting these
observations in the context of an integrated global monitoring system (e.g.,
NRC, 1999a). Such a system is required for other aspects of climate change
research and applications not addressed in this report including climate
change attribution and detection and providing a broad range of climate and
weather services (NRC, 2001 e).
Details of these observation needs were presented in Chapters 2 through
8. As explained in Chapter 2, a water vapor observing system is needed that
has sufficient accuracy to measure decadal trends in the water vapor
distribution and sufficient spatial resolution to aid in understanding the
mechanisms by which the water vapor distribution is maintained. The water
vapor observing system should be closely linked to a global cloud, aerosol,
and precipitation observing system. As was discussed in Chapter 3, detailed
datasets for Arctic and Antarctic ice cover and albedo, and more
comprehensive sea-ice thickness data are needed that extend over long
periods of time to account for interannual variability. Techniques for
efficiently measuring sea-ice thickness over the globe need to be developed.
Improved definition of the basic temperature and salinity state of the
upper ocean also is needed, as explored in Chapter 4. This will require full
implementation of a system with the capabilities of the Argo global array of
profiling floats, plus a strategy for monitoring key regions of the ocean that
are important for the thermohaline circulation, such as the Labrador,
Greenland-Iceland-Norwegian, Weddell, and Ross seas. The need for
integrated datasets for soil moisture, skin and soil temperature, vegetation
properties and cover, and snow water equivalent are examined in Chapter 5.
A new suite of in situ and remotely sensed observations are needed to
provide information on a variety of aerosol characteristics and key
~ As defined in NRC (1999a, 2000b).
OCR for page 113
RECOMMENDA TIONS
113
atmospheric chemical properties that have been unavailable heretofore on a
temporally or spatially extensive basis. These characteristics, outlined in
Chapter 6, include the chemical composition of aerosol particles, optical
extinction, and scattering under a wide range of conditions. Chapter 7 notes
that a highly diverse set of observations is required to more tightly constrain
understanding of Earth's biogeochemical feedbacks. These observations
include O2-N2 ratio, ocean carbon in a variety of forms, ocean color, and
atmospheric CO2 and CH4 concentration.
EVALUATING PROGRESS IN UNDERSTANDING CLIMATE
FEEDBACKS
The simulation of individual feedback processes must be tested against
appropriate observations in order to measure our understanding, reveal the
reasons why climate sensitivity varies from model to model, and know
which models are most likely to be correct. To do this requires a well-
designed set of observed diagnostic tests, or metrics, that will measure our
understanding and provide standards against which models can be tested. A
sufficiently discriminating set of observational tests would lead to
improvements in individual models and better methods for objectively rating
the performance of climate models.
Recommendation:
Both global and regional metrics that focus on feedback processes
responsible for climate sensitivity should be used to more rigorously test
understanding of feedback processes and their simulation in climate
models.
An expanded set of data comparisons, or climate model performance
metrics, should be developed that focus on each of the key climate feedback
processes and include geographic and seasonal variations. In this context
metrics are robust statistics that can be derived from observations and
capture the essence of some fundamental aspect of a phenomenon or
process. A first step toward developing metrics to evaluate feedback
processes would be for the relevant agencies to organize a workshop or
series of workshops to define observational metrics. These workshops would
include scientists engaged in observation, diagnosis, and modeling of
climate and climate processing.
Effective metrics must be based on a basic description and at least a
rudimentary understanding of the phenomenon or process in question. They
OCR for page 114
114
UNDERSTANDING CLIMA TE CHANGE FEEDBA CKS
must be tailored to the variety of time and space scales on which the relevant
climate change feedbacks operate. They must be based on measurements
that are of good quality and can define the process or phenomenon well.
Metrics should evolve as our understanding and observations improve.
A good set of diagnostic tests, or metrics, will do much more than assess
the state of a system; it will also capture the co-variation or coupling
between the system's components. If effectively employed, metrics can be
an essential tool to help organize and stratify diagnostic analyses, as well as
to relate model simulations to the fundamental aspects of observed
phenomena. They can also be a useful tool to monitor the evolution of the
climate system and thus make important contributions to the field of climate
change detection and attribution. Utilization of observational metrics as
proposed in this report would increase the observational constraints on
feedback processes and could help to improve confidence in regional climate
projections. A broadly accepted suite of climate feedbacks metrics could
also provide a partial solution to the "need for uniform criteria with which to
judge climate models," which has been identified as a key issue in a
previous NRC report ARC, 2001c).
Examples of Metrics
Because the magnitude of the changes in feedback processes may not
yet have been sufficiently large to evaluate our understanding of the
feedbacks as they relate to long-term climate change, it is not possible to
fully evaluate an understanding of climate feedback processes by observing
long-term trends in global climate. An alternative strategy is to use
observations of shorter-term variability to test understanding and simulation
of climate feedback processes. Strongly forced variations such as annual,
diurnal, and ENSO cycles seem to provide valuable tests for understanding
the processes underlying climate change feedbacks.
Although there are benefits to using short-term variability as a
diagnostic tool for improving understanding and modeling of climate
feedback processes, it is recognized that success in simulating the role of
climate feedback processes in short-term climate variations does not
necessarily translate to success in simulating global warming. Nevertheless,
short-term climate variability provides a promising avenue for quantitatively
evaluating the feedback processes of a model used for the study of global
warming. In other words, a model may not be regarded as reliable for
climate change projections simply because it realistically simulates
feedbacks on short time scales. But if a climate model can accurately
OCR for page 115
RECOMMENDA TIONS
115
simulate climate feedback processes on interannual, annual, and shorter time
scales, a much stronger argument can be made that it should be reliable for
climate change projections.
An example of a large-scale regional metric is the response of
atmospheric temperature, water vapor, clouds, radiation fluxes, and
atmospheric circulation to sea surface temperature anomalies associated with
warm events in the tropical Pacific (e.g., Hartmann and Michelsen, 1993,
Ramanathan and Collins, 1991~. This approach can be expanded to examine
the covariability of sea surface temperature, tropical convection, upper
tropospheric water vapor, the vertical profile of atmospheric temperature,
and other observations over a variety of time scales, including the seasonal
time scale. These covariance metrics should then be applied to model
simulations to pinpoint those aspects of the models that appear to accurately
represent nature and those that require further work.
On the continents the global warming response of precipitation, clouds,
and soil moisture is important. A metric that might enable improvement of
feedback processes over land would be the simulation of the observed
diurnal variations of temperature, clouds, and precipitation, and the slow
evolution of the diurnal cycle on the seasonal timescale as, for example, soil
moisture decreases during the summer months. The snow and ice feedback
could be quantified by linking the regional climate sensitivity and the
amplitude of the seasonal cycle to the snow cover and surface energy
balance for latitude-longitude blocks of the North American and Eurasian
continents.
Many other possible metrics for testing the simulation of climate system
feedbacks can be envisioned. Individual disciplinary chapters in this report
give additional examples of metrics that might be used to diagnose specific
aspects of climate system feedbacks. The set of metrics will likely evolve
with time as understanding and simulation of the climate system evolves and
improves.
Climate Modeling and Analysis for Climate Feedbacks Research
A practical goal of climate feedbacks research is to provide information
necessary to support more reliable projections of future climates. To test
understanding and modeling of climate feedback processes using a set of
climate feedback metrics requires a substantial infrastructure and a
proportionate intellectual effort. To undertake a rigorous program of testing
the simulation of climate feedback processes in our most capable climate
models requires that the observations and the expertise in applying them be
OCR for page 116
116
UNDERSTANDING CLIMATE CHANGE FEEDBACKS
brought together with the modeling capability. Previous NRC reports have
stated the need for capable and effective climate modeling facilities (NRC
1998c, 2001 c), and have recommended the development of centralized
operations for climate predictions and ozone assessments (NRC, 2001 c). To
advance understanding of climate change feedbacks and their role in climate
sensitivity it is essential that U.S. climate modeling facilities also have the
capability and mandate to test climate feedback processes and their
interactions using the most discriminating observational constraints. Within
the context of climate feedback processes this will also address the need for
uniform criteria with which to judge climate models (NRC, 2001 c).
The Panel on Improving the Effectiveness of U.S. Climate Modeling has
previously noted "the need for strong interaction between observations of
the climate system, research into fundamental climate processes, and
integrative climate modeling" (NRC, 2001 c). That same panel recommended
enhanced resources for centralized operational activities addressed to short-
term climate predictions, to the study of predictability of climate on decadal
and century time scales, and to assessments of ozone depletion and climate
change (NRC, 2001 c).
A research program that uses observable metrics to test our
understanding and simulation of climate change feedbacks should be applied
most assiduously to the models that are most capable of both simulating the
complex interactions of the various feedback processes and making climate
change projections for planning purposes. Applying the most stringent
observational constraints and tests to the most capable integrated models
will benefit climate feedbacks research and increase our confidence in the
climate projections made with these models. For this reason it is important
that the research and operational facilities operating the most capable
climate models have access to the data and expertise necessary to employ
the most discriminating metrics of the feedback processes.
Recommendation:
Climate modeling facilities in the United States must be given the
capability and mandate to test understanding and simulation of climate
feedback processes and their interactions using the best observational
constraints on climate feedback processes. Periodic assessment of the
progress being made by major climate models should be conducted to
evaluate the ability of these models to simulate the processes underlying
key climate system feedbacks.
Testing and development of climate models can make more effective
use of existing datasets, and should rapidly incorporate new datasets as they
OCR for page 117
RECOMMENDA TIONS
117
become available. It is important that these efforts focus more directly on
issues related to the specific testing and quantification of feedback
mechanisms in climate models.
The Atmospheric Model Intercomparison Project (AMIP) and the
Coupled Model Intercomparison Project (CMIP) have been very useful for
evaluating the performance of GCMs in simulating the geographical
distribution of climate and its seasonal variation (e.g. Covey et al., in press;
Gates et al., 1998~. These analyses indicate relatively good matches between
some aspects of the observed climate and the climate produced by state-of-
the-art GCMs. It is highly desirable to develop a specific methodology for
the improvement and quantitative assessment of modeled feedback
processes and their effect on climate sensitivity. The approach recommended
here is a more specific focus on comparing observed measures of climate
feedback processes with the same measures produced by climate models.
Representations of critical climate processes in climate models are becoming
more realistic and must be rigorously tested against all available
observations. These efforts should remain focused on the objective of
producing more robust model representations of nature, and be prioritized
according to their impact on projections of future climates.
One approach for facilitating model improvements of key processes is
the Climate Process Team (CPT) concept, currently being developed by the
U.S. CLIVAR program. It has the potential to foster more comprehensive
investigations of climate change feedbacks. In this approach, teams of
scientists, including observationalists, process modelers, and global climate
modelers, are to undertake relatively comprehensive and integrated projects
focused on specific climate feedback processes and their treatment in
climate models. CPTs have the potential to make significant progress toward
reducing and better characterizing uncertainty associated with climate
change feedbacks, provided that they can develop and maintain a sharp
focus on the objective of improving model representations of the key
processes.
Representative terms from entire chapter:
climate feedback
