Strategic Plan for the
Climate Change
Science Program
Final Report, July 2003

  This chapter also available as PDF file.

Authors of this Chapter

Final Report also is available as single PDF file (8.8 Mb)

See also:
 

Vision for the Program and Highlights of the Scientific Strategic Plan

Acronyms, Abbreviations,
and Units

The Global Change Research Act of 1990 specifically calls for "global measurements, establishing worldwide observations necessary to understand the physical, chemical, and biological processes responsible for changes in the Earth system on all relevant spatial and time scales," as well as "documentation of global change, including the development of mechanisms for recording changes that will actually occur in the Earth system over the coming decades." The program continues to respond to this call by following a strategy for the development and deployment of a global, integrated and sustained observing system to address science requirements and decision support needs at appropriate accuracies and space and time resolutions.

The CCSP strategy for observations and monitoring includes guiding principles, identification of priorities, and effective management of available resources. The purpose of this chapter is to describe how all the disparate observations described by the CCSP plan will be systematically organized and managed to improve our understanding of the climate system. The Data Management discussion that follows (Chapter 13) describes the plan for archive and distribution of these data.

The observations and monitoring component of the CCSP seeks to address the following overarching question:

The development of space-based and in situ global observing capabilities was a primary focus of the program's first decade. Several new Earth-observing satellites, in situ networks, reference sites, and process studies are now producing unprecedented high-quality data that have led to major new insights about the climate system. For the purposes of this document, surface-based remote sensing observations, as well as aircraft or suborbital measurements, will be considered as in situ measurements, and any references to in situ measurement networks should be considered as applying to networks using these techniques wherever appropriate. The observing system for the future will build upon this success.

The challenge for the coming decade is to maintain current capabilities, implement new elements, make operational the elements that need to be sustained, and integrate these observations into a comprehensive global system to address the objectives of the CCSP research elements and decision support activities. Fundamental questions about the climate system and societal benefits that can be addressed are described in Chapters 3 through 11, these provide the basis for the observing system design and implementation. Illustrative examples of CCSP observation needs and milestones that address the research questions are described in Appendix 12.2. In addition, the overall climate observing system must address the five basic integrating goals for CCSP outlined in the introductory chapter:

A system that integrates atmospheric, oceanographic, terrestrial, cryospheric, and cross-cutting observations does not currently exist per se. However, many components are available. For example, the Global Climate Observing System (GCOS) is a fairly well documented but not completely implemented international approach that has components in place to satisfy some climate requirements (see GCOS, 2003). GCOS is intended to provide a focused set of observations from a subset of established measurements sites that are considered to have a sufficient climate history and spatial distribution. The system discussed here, however, goes beyond GCOS. The CCSP has expanded the initial inventory of important climate observations (see Appendix 12.1) to encompass the needs of research on and applications to the global cycles of carbon, water, energy, and biogeochemical constituents; atmospheric composition; and changes in land use.

Building on the CCSP mission, the United States has also taken a leading role in fostering the development of a more broadly defined and integrated global observing system for all Earth parameters, for example including geological as well as climate information. The United States is hosting a ministerial level Earth Observation Summit in July 2003, with participation by many developed and developing nations as well as many intergovernmental and international NGO groups. This summit will initiate a 10-year commitment to design, implement, and operate an expanded global observing system that builds on the major observational programs currently operated by the U.S. and many other governments and international organizations. CCSP agencies have provided the leadership, definition, and support for the Earth Observation Summit, and the CCSP will closely integrate the U.S. observation and data management programs with the international programs launched at the Summit.

The global observation system needed to fully implement CCSP includes the evolution of the observing capability provided through the USGCRP and the operationally oriented monitoring systems routinely provided by several federal agencies. The latter were never included as part of the USGCRP, and may or may not be included as part of CCSP in budget inventories. In fact, as noted below, a critical issue associated with the implementation of the observing system for CCSP is the transitioning of research observations, typically made through the USGCRP, into operations, not currently included as part of the USGCRP (objective 1.3). Any consideration of budgets associated with global observations should be based on those for all the component programs, and not the historic USGCRP budget.

The basic elements of a global observing system must consist of:

These elements must be managed by an effective national entity, and coordinated at the international level. The management must have the capability to establish observing protocols, provide oversight, address deficiencies, and mobilize resources as required to maintain the integrity of the entire end-to-end system. A major challenge for CCSP in this decade will be the transition of many observation elements developed in a research mode to a sustained and operational environment (e.g., NASA research satellites to NPOESS and some in situ networks). The resulting global system must obviously engage many countries in a cooperative enterprise. Developing such a system presents a daunting challenge that must be met to provide essential information for decisionmakers.

In order to move toward a global observing system the CCSP will focus on six goals:

1. Design, develop, deploy, integrate, and sustain observation components into a comprehensive system.
2. Accelerate the development and deployment of observing and monitoring elements needed for decision support.
3. Provide stewardship of the observing system.
4. Integrate modeling activities with the observing system.
5. Foster international cooperation to develop a complete global observing system.
6. Manage the observing system with an effective interagency structure.

Goal 1: Design, develop, deploy, and integrate observation components into a comprehensive system.

A system is the integration of interrelated, interacting, or interdependent components into a complex whole. The path to an overall observing system must address a number of key issues in addition to the observation components themselves, including the priorities for development, the implementation strategy, assessment, utilization, cooperation with other providers of environmental measurements, data management, international development, and system management. Some of these activities are considered in this goal, but others are of sufficient complexity that they are a goal unto themselves and are discussed later in this chapter, or are planned to be addressed under the CCSP and are discussed under Goal 2.

Prioritization: Observing elements are in place within existing research and operational programs that partially fulfill the requirements for meeting these objectives. Other key sensors and observing networks still need to be developed and implemented. Priorities for these augmentations are required because resources are limited. The criteria include: benefit to society, scientific return, partnership opportunities, technology readiness, program balance, and implementation of the climate monitoring principles (Appendix 12.3). The CCSP research element questions and decision support goals provide the basis for determining priorities. In the CCSP implementation plans, the research and decision support elements will provide a linkage between research question or decision support goal, measurement requirements, and the observation elements that meet the requirements. The management of the program will recommend priorities in consultation with the scientific community using the management mechanisms outlined in Chapter 16.

Evaluation: A key lesson learned over the past decade is that observing systems and networks must be implemented in a way that allows flexibility as both requirements and technology evolve. Therefore, the program will regularly assess the science and decision support requirements and propose modifications to the observing systems required for the CCSP to execute its research plans. This process will involve the scientific community and program managers working on each research element, as well as those involved in modeling, scientific assessment, and other integrative activities (see Goal 6).

Cooperation: Many important observing systems are developed and operated by organizations that are not formal participants in the CCSP, making the development of strong cooperative relationships that extend beyond the current CCSP a necessity. The CCSP will work with observing system partners and the scientific community to identify requirements and set priorities in light of available resources and competing needs.

A requirements based design will be developed to identify baseline and minimum requirements that will address CCSP research element questions and decision support goals. US needs and contributions will be weighed in the global context of the requirements and contributions of the other nations of the world and coordinated with them in a manner that will lead to a global system. The most efficient and effective observation elements and networks will be selected that will meet the requirements.

The CCSP will maintain and improve basic research observing facilities, networks, and systems (both space-based and in situ). Climate quality data requires long-term continuous records (see Chapter 4, Figure 4-1). It is critical to maintain records required to answer research questions before they are lost for other reasons (see Figure 12-2). Long-term observations require a focus on maintenance and replacement to sustain the capability at a sufficient level of accuracy to detect climate change over decades.

[

Figure 12-2: The U.S. Geological Survey (USGS) Hydro-Climatic Data Network (HCDN) was compiled in 1988 and included 854 gauges that were active and had at least 36 years of record. Stations that are still active are shown in green. Stations discontinued since 1988 are shown in red. These stations would have had at least 50 years of record as of 2002 if they had been kept in operation. Source: William Kirby, Global Hydroclimatology Program, USGS. For more information, see Annex C.

Figure 12-3: For scientists to understand climate, they must also determine what drives the changes within the Earth's radiation balance. From March 2000 to May 2001, the Cloud and Earth Radiation Energy System satellite (CERES) measured some of these changes and produced new images that dynamically show heat (thermal radiation) emitted to space from Earth's surface and atmosphere (left sphere) and sunlight reflected back to space by the ocean, land, aerosols, and clouds (right sphere). Source: CERES Project

The transition of proven research observation elements to sustained or operational status will make these components more cost-effective and sustainable for long term benefit. The CCSP will work toward an effective integration for the planning and development of research and operational systems. The CCSP will develop a strategy for transition from research to operations that will adhere to the climate monitoring principles (see Appendix 12.3), including continuity, in order to have the most benefit to its objectives. Management structures will be developed to provide for a handoff from research management to operational systems management. The operation management will be responsible for continuing the research mission and producing and delivering day-to-day operational results and products.

Operational observation networks continue to be the backbone of climate system measurements. These networks, with only modest incremental costs will satisfy significant parts of the climate observing requirements as described in the Climate Monitoring Principles (Appendix 12.3). Scientific, as well as decision support, oversight of operational systems for climate research must be implemented using the concept of Climate Data Science Teams described in Goal 6 below.

The CCSP must maintain a sustained research and development program to address major deficiencies in observing systems and the paleoclimate record identified by the research elements in Chapters 3 to 9 (e.g., closing the budgets for carbon and regional and global water cycles, integrating the coastal ocean monitoring systems, and filling gaps in our records about decadal- and century-scale climate variability prior to the instrumental record). To the extent possible, new observational capabilities will be integrated into existing networks so as to minimize redundant operations and costs. Future global measurements from satellites will be developed that dramatically improve quality and vertical, spatial, and/or temporal resolution, especially to enhance regional coverage for decision support applications.

The science chapters have described and justified numerous field studies that integrate airborne (in situ), surface, and satellite observations over regional scales and durations from days to several weeks. These intensive observation periods provide valuable data for testing and validating satellite retrieval algorithms, and for the fine scale resolution necessary to test, validate, and constrain physical processes in climate models. These coordinated observation efforts will need to become even more sophisticated as satellites evolve towards formation flying, onboard processing, and smart sensor technology. Aircraft remain a valuable platform for the CCSP and mechanisms to make efficient use of these facilities across the CCSP agencies will be investigated.

A global effort of this magnitude must be managed as a system in order to be effective. For agencies to assess the performance of the observing system, an institutional mechanism must be put in place to monitor the status of the globally-distributed components and to evaluate their combined capabilities.

The CCSP, working with international partners, will develop a system architecture for monitoring distributed global operations that establishes and maintains linkages among the in situ and satellite elements and the data and modeling activities that are essential components of climate observation. The purpose is to provide a framework so that a nation can manage its observing system more efficiently and effectively and:

The guiding principles for development of this uniform interagency and international monitoring capability include:

Many countries have gathered climate system data to document climate system variability using many different instrument types during the past 150 years. In order to document and understand change from a historical perspective, we need to develop global, comprehensive, integrated, quality-controlled databases of climate system variables based on historical or modern measurements, and provide the user community with open and easy access to these databases. We must integrate these records as far into the past as is practical to reduce uncertainties in the climate trend estimates of individual parameters.

Our understanding of some of the changes in the physical climate system over the past 50 or more years has improved because of systematic reprocessing and integration of climate observations using state-of-the-art climate models and data assimilation technology. This must be continued through a routine and iterative improvement process that incorporates a rigorous research and validation component.

The integration of satellite and in situ observing networks for synthesis of the ocean, atmosphere, or terrestrial spheres in a systematic manner will be addressed through focused pilot programs. The Global Ocean Data Assimilation Experiment, while not specifically focused on climate issues, is an example of a pilot approach to the synthesis of ocean observations.

Across the range of research on human response and consequences to climate change there is a particularly strong need for the integration of social, economic, and health data with environmental data (see Chapter 9). Observations will be used to address gaps in understanding, modeling, and quantifying the sensitivity and vulnerability of human systems to global change and measuring the capacity of human systems. Using retrospective analyses of consequences of shifts in climate will also help model future ability of hazard and resource management institutions to respond.

The CCSP provides resources to develop observation systems and processing and support systems that will lead to reliable and useful products. These products will provide critical policy-neutral information for decision support and policymakers in areas such as climate and weather forecasting, human health, energy, environmental monitoring, greenhouse gases, and natural resource management.

The CCSP will enhance the existing long-term monitoring elements with accelerated focused initiatives to provide a more definitive observational foundation for determining the current state of the climate. Many shortcomings of the current climate observing system relate to understanding forcing of the climate. In addition, fundamental observations for characterization and understanding of the state of the climate are needed for the global ocean, land surface, and ice variables. For example, in its first few years of operation only half of the GCOS Upper-Air Network (GUAN), established for climate purposes, has been reporting regularly, and the GCOS Surface Network (GSN) for climate has had similarly disappointing results.

The ocean is poorly-observed below the surface, which limits our understanding about the ocean's response to a warming planet and its ability to naturally sequester greenhouse gases. Over land, the great spatial heterogeneity requires extremely detailed measurements and presents a major challenge.

The CCSP encourages partnerships with developed countries and support for developing countries in order to build the global observing system.

In the budget requests for the past 2 years (FY2003 and FY2004) progress has been made to address many of these deficiencies, as described in Appendix 12.3. The objectives in this goal outline specific elements of the CCSP that will expand on this progress and support a directed strategy to focus resources and accelerate observations for climate change and decision support.

Within the climate monitoring arena, atmosphere and ocean observation elements to measure the physical aspects of the climate system are the most complete, and are ready to be brought together as a system. For example, the detailed open ocean observing system has been designed at both national and international levels through numerous community workshops and implementation is about 40% complete (see Figure 12-4). The atmospheric system, which is a mixture of both climate and non-climate elements, has been recently examined (GCOS, 2003) and needs support for fixing degradations, sustaining current capabilities and adding improvements. An accelerated effort to complete these two subsystems will improve understanding of climate characterization, forcing, and prediction, as well as facilitate the implementation and testing of the interagency management of complete observing systems. Focused new data management practices are being developed and will promote efficient acquisition, validation, delivery, and archiving of the measurements and data products. In order to complete these systems the following steps need to be taken in conjunction with international partners:

Figure 12-4: The international global ocean observing system for measurements of the physical climate system is 40% complete. Source: Michael Johnson, NOAA Office of Global Programs.

There are two aspects of the paleoclimatic record that need to be addressed by accelerated and focused programs within the CCSP in order to improve our characterization of long-term climate change and provide a valuable benchmark for testing models that are used to project the future (see Goal 4). First, some important paleoclimate sites are disappearing (e.g., tropical glaciers). Second, there is a need for an improved global perspective of the historical changes in climate. The number of observations in the Southern Hemisphere is small and will be increased.

Changes in an environmental variable -- most often warming, but also changes in precipitation and air quality -- have often been related to observed changes in biological and ecological systems. Several examples were mentioned in the Working Group II section of the Intergovernmental Panel on Climate Change's Third Assessment Report (IPCC, 2001b), including thawing of permafrost, lengthening of the period of active photosynthesis in mid- and high-latitude ecosystems, poleward shifts of plant and animal species ranges, movement of plant and animal species up elevation gradients, earlier spring flowering of trees, earlier spring emergence of insects, earlier egg-laying in birds, and shifts in a forest-woodland ecotone (the boundary between the forest and the woodland).

These changes in ecosystems and organisms are consistent with warming and changes in precipitation, but the possibility remains that the observed biological and ecological changes were caused (in part) by other factors such as biological invasions or human land and marine resource management. Because of this, the attribution of the causes of biological and ecological changes to climatic change or variability is extremely difficult. Moreover, because many ecosystem-environment interactions play out over long periods -- ultimately involving evolutionary changes and adaptations within ecosystems -- long periods of study are needed in many cases to draw firm conclusions about relationships between environmental change, effects of that change on biological and ecological systems, and the significance of any observed biological or ecological changes for the functioning of ecosystems (see Chapter 8).

New research is needed to provide a significantly more complete picture of how biological and ecological systems may have responded to recent climatic change and variability, including possible biological or ecological responses to extreme events. New observational systems will also be needed to appropriately monitor potential future changes in the environment and accompanying biological or ecological changes (if any). A key challenge will be to provide organization, guidance, and synthesis for the emerging field of observed effects of climate change on biological and ecological systems.

The CCSP will initiate studies of early effects and indicator systems across diverse ecosystems and geographic regions. A substantial amount of existing climate and effects data, a variety of monitoring efforts, and comparisons to scenario-based effects studies can be marshaled in this effort. The CCSP will facilitate linked analyses of climatic trends and observed biological and ecological effects by supporting identification of appropriate past and ongoing monitoring efforts, design of needed new monitoring systems, and synthesis of results across ecosystems and regions. Research efforts will target those ecosystems that are subjected to the most rapid or extensive environmental changes and/or are most sensitive to possible environmental changes.

Long-term, spatially explicit, and quantitative observations of ecosystem state variables and concomitant environmental variables are needed. Initial activities and products will include:

Objective 2.4. Provide regular reports documenting the present state of the climate system components.

The CCSP will initiate a regular reporting program on the state of the climate system. Reports on various components of the system will highlight analysis products from the observing system to address the overarching question stated at the beginning of the chapter:

Observations of the climate require careful scientific oversight because climate signals are usually quite small relative to higher-frequency phenomena in the data. Consequently, the climate science community has developed a set of principles that can be uniformly applied to all relevant measurements. In addition, special rigor must be applied to the algorithms, which are used to translate fundamental physical measurements into useful geophysical products,so that consistent results can be obtained and improved. Finally, scientific oversight of the algorithm, instrument calibration, and data processing and validation is critical to understanding the differences between instrument error and climate signal, and reduction of the error bars that define the uncertainty. This approach is captured in the four objectives of this goal.

Efforts in the last decade to use current research and operational data sets in global climate change research have provided a critical set of lessons learned. These lessons have been gathered into two sets of climate monitoring principles (see Appendix 12.4). These principles and experience will be used to guide the major improvements needed for the observing system. Instrument calibration, characterization, and stability become paramount considerations. Instruments must be tied to national and international standards such as those provided by the National Institute of Standards and Technology (NIST). When observations cannot achieve sufficient absolute accuracy or changes in spatial or temporal sampling occur, overlapping observations with high stability are required to ensure accurate monitoring of global change. Agreed-upon measurement protocols and procedures (e.g., continuous data validation and inter-comparisons) are required to produce climate quality data.

To be policy-relevant, climate data must have reliable confidence intervals. Narrow intervals are most effective for optimal policy development and implementation. Experience from previous scientific assessments, the NIST, and other national standards laboratories have shown that actual accuracy is only known after comparison of independent measurements and analysis from multiple laboratories. Most climate observing systems are pushing the edge of capability in calibration. This suggests that a final climate observing principle that climate observations should strive to address: each key climate variable will be measured using independent observations, as well as independent data analysis.

Observation independence verifies instrument accuracy, while analysis independence verifies algorithms and computer code. For sea surface temperatures, examples of independent observations include satellite-derived measurements from infrared (Advanced Very High-Resolution Radiometer -- AVHRR), passive microwave (Advanced Microwave Scanning Radiometer -- AMSR), and in situ measures from ships, buoys, and floats. But many climate variables do not yet have independent sources. When surprises in climate data sets occur relative to current theory, independent confirmation is essential to ensure policy-relevant confidence in the results. A recent key example of this need is the air temperature record from radiosondes and the microwave sounding unit (MSU) satellite data, as well as the different results from two analyses of the MSU satellite record. Using a single measurement of a key climate variable is a high-risk approach to reducing uncertainty for policymaking decisions.

An instrument radiometric mathematical model, termed the "measurement equation" by metrology experts, is developed to define, describe, and represent the relationship between the sensor output and desired physical observable (e.g., volts to temperature, counts to radiance, etc.). Additional models are used to reference the results in space and time (geo-location), and finally, models are used to derive desired physical parameters (ozone profile, chlorophyll concentration, etc.). This process of characterization has been very successful in the past, especially for satellite programs. Its practice will continue to be encouraged, the results promulgated, and the instrument designs critically evaluated so that future measurements will produce the desired results.

It is important to understand that a sensor does not have to be ideal, but it must be well-characterized, so that systematic effects can be evaluated and removed, even years after the mission, if new effects are identified. If these data are not available, the utility of the measurements is greatly compromised. The measurement equation approach defines the problem and places the entire procedure on a scientific basis. In the future all climate quality observations need to be linked to their climate data records through an approved algorithm model that is described, for example, in an algorithm theoretical basis document (ATBD). These documents were pioneered by NASA in its implementation of the Earth Observing System and provide a valuable lesson learned for future measurements.

Achieving climate accuracy with global and decadal sampling represents a unique scientific and organizational challenge. Calibration and stability requirements for climate data are often an order of magnitude more stringent than weather data. This accuracy is often achieved in research in situ and satellite data sets, with the length of this accuracy being dependent on the lifetime of the platforms and rate of instrument degradation; while in the longest satellite case (Earth Radiation Budget Satellite) this may approach 2 decades, in other cases it is much shorter. Global decadal sampling is often achieved in operational data sets, but without the accuracy and stability. Scientific data stewardship addresses this challenge through the production of Climate Data Records (CDRs).

Ultimately, the production of CDRs includes setting climate accuracy and stability requirements, instrument calibration and characterization, algorithm selection and development, end-to-end data validation and error analysis, quality control, and data production. Long-term archiving and effective distribution are also essential and are addressed in Chapter 13. Experience in the last decade has shown that achieving CDR stewardship requires active and continual collaboration with the science community including both data producers and data users. Examples include Atmospheric Radiation Measurement (ARM) and Aeronet surface observations, the global surface temperature record, Tropical Atmosphere-Ocean buoys, and many forms of satellite observations for ocean, land, and atmosphere climate variables. Climate Data Science Teams (CDSTs) will be discussed in Goal 6 as a critical implementation strategy for the climate observing system.

Long-term CDRs will come about through the harmonious integration of highly accurate, research quality observing systems with longer-term operational and paleoclimate data. To reach the desired level of climate quality will also require additional efforts for current operational systems. We anticipate that CDSTs will be required to produce climate quality data and represent the organizational link that provides operational level continuity with research quality science direction.

The stewardship of scientific data is required to produce the climate quality data sets that are a critical resource for policymakers and a legacy for future generations. As the length of CDRs increase, the power of the data to narrow uncertainties will increase and provide a basis for future decisions.

There are two primary roles for observations in the improvement of climate models: evaluation and initialization. At the same time, climate system models provide important insights for the improvement and interpretation of measurements through the use of simulations. These two-way interactions of observations and models affect not only the design but also the improvement of the observing system.

Model testing uses observations to determine the uncertainties in current climate models and to prioritize needed improvements in future models. This is the most critical function of climate data records and depends primarily on two factors: a) accuracy and stability of the climate data; b) the intrinsic background variability or noise in the Earth's climate system. Both factors vary with temporal and spatial scale. Climate system noise is typically largest at short time scales and small spatial scales (e.g., seasonal/regional scales), and smallest at long time scales and large spatial scales (e.g., century/global scales). Useful climate model tests require that both the data accuracy and the climate system noise be carefully assessed. As a result, climate accuracy requirements cannot be specified as a single number per variable, but must also be defined at several relevant temporal and spatial scales. Examples of relevant scales include monthly/local, seasonal/regional, annual/zonal, and decadal/global temporal/spatial scales. Some climate data records will have sufficient accuracy or stability to resolve regional climate change but inadequate coverage to resolve global changes and vice-versa.

Consequently, observations and models of the climate system cannot make progress without continuous interaction between and understanding of the uncertainties in each research area (see also Chapter 10, Objective 1.3). Rapid progress requires regular interactions between the modeling and observation communities through CDSTs. One of the most critical collaborations between climate model and measurement scientists in the future will be the definition of prediction accuracy metrics capable of constraining key model uncertainties, such as cloud and water vapor feedback. Currently, such metrics are rather ad hoc because of the difficulty in unraveling the complex nonlinear climate system and in documenting climate data accuracy as a function of time and space. In the simplest sense, how would you recognize a perfect climate model if you had one? What tests would it pass? How will we know when the observing network is sufficient to characterize, attribute, and predict climate changes accurately? The answers are a function of both climate models and the climate observing system design.

The definition of climate model evaluation metrics will impact observing system design. As models and data improve in quality, these metrics will evolve in an iterative fashion over the next decade.

Climate model initialization is primarily effective only for seasonal-to-interannual time scales, and for initialization of the state of the world's oceans, soil moisture, and vegetation. These latter conditions reflect climate processes with sufficiently long time scales that their initial state can affect seasonal-to-interannual climate states. The prime initialization example is prediction of El Nino events using initial ocean conditions. To improve short-term climate forecasts, it will require further new and improved technologies in data assimilation and better utilization of in situ and remotely sensed global ocean, atmosphere and terrestrial observations of key physical variables, such as ongoing observations of temperature and new measurements of salinity and soil moisture (see Chapter 10, objective 1.4).

While climate models are not perfect, for some variables they can strongly support the estimation of accuracy requirements for CDRs. Consider the following simple paradigm of climate change: forcing => feedback => response => climate change. If the Earth were such a simple linear system, then designing observing systems and testing climate model predictions would be a relatively straightforward process. But the climate system is highly coupled and fundamentally nonlinear. Consequently, intrinsic internal variability is an inherent part of the real climate system. Climate change must be detected and understood. However, this signal is usually smaller than the background climate variability (e.g., year-to-year climate variability). CDRs will require accuracies at a temporal and spatial scale greatly below the level of natural variability.

Unfortunately, most quantitative observations of the climate system only exist for the last century and are a combination of background and climate variability, climate change, and observation error. Using the total variability of the system to define measurement requirements represents only an upper limit on required accuracy and is not likely to be sufficient. Climate model simulations, on the other hand, do not include the full complexity of the Earth's climate system. The models typically represent an underestimate of background variability. As a result, climate model ensemble simulations can be used to estimate the required accuracy for the observation system The ensembles can also specify intrinsic variability of the system as a function of temporal and spatial scale. These simulations can use varying climate forcing and can separate background variability from the climate signal by running large numbers of simulations with slightly varying initial conditions, but the same boundary conditions.

Increasingly, global and regional data sets derived from data assimilation models and retrospective reanalysis models are being used for climate research, monitoring (time series), diagnostics, applications, and impact assessments (see Objective 1.8 and also Chapter 10, Objective 1.5). Evaluation protocols must be developed to assess the accuracy and stability of these reanalysis products for climate applications. Acceptable error standards need to be established that will provide guidance on whether the data sets are usable or not, especially at regional scales and for model- derived/computed parameters.

The range of global observations needed to understand and monitor climate processes, and to assess human impacts, exceeds the capability of any one country. Cooperation is therefore necessary to address priorities without duplication or omission (see Chapter 15). Satellite missions and in situ networks require many years of planning. Observations of the state of and trends in planetary processes cut across land, water, air, and oceans. National programs need to fit into larger international frameworks. At the international level, participation in both science and operational oversight committees must be continued, and strengthened or developed in disciplines that have not yet developed a global system.

In 1998, Parties to the United Nations Framework Convention on Climate Change (UNFCCC) noted with concern the mounting evidence of a decline in the global observing capability and urged Parties to undertake programs of systematic observations and to strengthen their capability in the collection, exchange, and utilization of environmental data and information. The United States supports the need to improve global observing systems for climate, and the US will join other Parties in submitting information on national plans and programs that contribute to the global observing capability.

The United States is an active and leading partner in the development and support of a global observing system that assembles key elements from a number of observing networks under the aegis of appropriate international organizations. With regard to climate, the GCOS has fostered the integration of key elements including meteorological observations from the GOS of the WWW, atmospheric constituents from the GAW, hydrological observations from the World Hydrological Observing System, critical oceanographic climate variables from the GOOS, and several terrestrial variables from the GTOS. Coordination of global satellite observations is carried out through the Committee on Earth Observation Satellites (CEOS). This is particularly important as multi-national, multi-spacecraft constellations and increased use of higher-altitude orbits (e.g., geostationary) are used to improve the temporal resolution of global observations that may further enhance decision support.

Given the importance of validating satellite observations under a broad range of geophysical and biogeochemical conditions, participation of international partners -- in providing coincident and correlative information that can be used to test and improve satellite observations -- is especially important. International partners are also important in the implementation of field campaigns that are best carried out with the full scientific involvement and logistical support of the host countries.

The full implementation of a global system for climate will require enhanced international coordination and commitment. Components for atmospheric, oceanic, terrestrial, and satellite observations are supported at varying levels depending on scientific priorities, availability of national contributions, and the sophistication of the relevant observing technologies. A specific focus on climate variables is essential to provide an adequate database to meet climate needs.

Recently, these observing systems, their sponsors, and the satellite community have developed the International Global Observing Strategy (IGOS). The IGOS is a strategic planning process, uniting the major satellite and surface-based systems for global environmental observations of the atmosphere, oceans, and land in a framework for decisions and resource allocations by individual partners. The IGOS Partnership (IGOS-P) focuses specifically on the observing dimension in the process of providing environmental information for decisionmaking. This includes all forms of data collection concerning the physical, chemical, and biological environments of our planet, as well as data on the human environment, pressures on the natural environment, and environmental impacts on human well-being. The IGOS-P is currently focusing on identifying and gaining commitments for essential requirements for observing oceanic processes, global carbon, atmospheric chemistry, the global water cycle, geo-hazards, and coral reefs as part of a future coastal initiative.

Bilateral partnerships provide another strategy for achieving partnerships in observations and monitoring. Additionally, activities such as the U.S.-led Earth Observations Summit, the Subsidiary Body for Scientific and Technical Advice (SBSTA) meetings, and UNFCCC Conferences of the Parties provide opportunities to build international consensus -- both scientific and political -- for the implementation of a more integrated global climate observing system.

Developing countries provide key opportunities and challenges for observing systems. While many developing countries have the potential to make routine weather observations, and do so on a regular basis, many do not have the capability to collect and disseminate reliable observations of other variables that are critical for climate characterization and understanding. Further, many developing countries do not have adequate human resources to take full advantage of climate projections that would yield many benefits to their citizens.

It has been established that many developing countries lack either the capital or human resources to support high-quality observations or to sustain data and information systems. Countries often lack adequate capital for investment in equipment and supplies, trained technical staff, or maintenance capability.

The U.S. research community can point to numerous examples where it has contributed to measurements of key variables in developing countries. Many of these have provided valuable long-term data. The networks to measure atmospheric constituents through flask sampling and through vertical soundings have contributed a global data base of important information on greenhouse gases. The oceanographic community has successfully engaged many coastal nations to participate in one or more of its observing systems (e.g., sea level, drifting floats and buoys, volunteer ships, etc.). International programs addressing terrestrial processes have demonstrated similar successes in observing ecosystems, obtaining hydrologic information, and initiating cryospheric measurement.

The CCSP has specifically noted the key role that the United States can play in improving observational networks. In the President's June 2001 Rose Garden speech, he stated that " We'll also provide resources to build climate observation systems in developing countries and encourage other developed nations to match our American commitment.."

The United States can contribute further by:

It is essential for the United States to maintain a leadership role in those international programs, both research and operational, that support climate observations of importance to U.S. programs. The principal international research programs (e.g., the World Climate Research Programme, the International Geosphere-Biosphere Programme, and the International Human Dimensions Programme) invite members from the U.S. scientific community and often from federal agencies to serve on relevant steering committees and working groups. These committees and working groups provide a forum for planning future research campaigns and for developing observational components that often lead to continuing measurement strategies. Individuals participating in these groups provide key links between U.S. and international research programs.

With regard to observations, U.S. scientists serve on the steering committees and working groups of the global observing systems and provide continuing advice to them. The principal programs for atmosphere, ocean, and terrestrial observations of climate include the GCOS, the GOOS, and the GTOS as well as their working groups.

U.S. scientists also serve on more operationally oriented international committees that support ongoing observational activities. Examples include the Commissions of the WMO (e.g., Basic Systems, Climatology, Hydrology, etc.) and the Joint Commission for Oceanography and Marine Meteorology sponsored by the WMO and the Intergovernmental Oceanographic Commission. Individuals named to such groups serve as national representatives. These groups establish observing requirements, develop protocols for data and information exchange, and obtain international commitments.

The CCSP encourages U.S. scientists to continue their important roles in support of these diverse international programs.

The development of a national observing and monitoring system will require a coordinated management plan, which involves all federal agencies that conduct research and operational observations of the climate system. The management approach described in this goal follows the general guidelines outlined in Chapter 16. However, based upon the "best practices" learned over the history of the USGCRP, a more detailed management structure is described for the observation system within this goal. Experience has shown that scientific oversight is a key for the maintenance of climate quality data from observing systems; consequently the management plan will engage the science and user communities in key oversight and evaluation capacities. Finally, the observing system will be global and carried out with the nations of the world; therefore, cooperation at the international level for both operational and scientific oversight will need further development (see Goal 5).

At present, some of these management elements are in place, but a well-coordinated national management plan for observing the climate system is not in place and will be addressed in the coming decade. At the federal level, coordination at the program level will be strengthened to address implementation, and oversight by agency principals will be established. Scientific oversight will occur through a layered pyramid of working groups starting at the most basic level of instrument teams and building up to an observing system science advisory council. This structure has been demonstrated to be an effective strategy over the lifetime of the USGCRP and this goal expands the "best practices" learned during USGCRP to the observing system as a whole. These issues are addressed in the four objectives described below.

The management of the observing system elements will vary depending on the size and complexity. Space systems, both operational and research, require focused project management groups that are dedicated to monitoring health and safety, providing command and control, receipt and quality control of downlinked data, fault recovery, and ground tracking, if necessary.

In situ systems can vary from control and maintenance of a single convenient site on land to maintenance of a large international network at sea. While the former can rely on a simple management program, the latter requires a much larger national management group and international coordination. The CCSP will coordinate management groups of similar measurement techniques and logistics, coordinate with international management where applicable, work with international implementation coordination groups, such as the Joint Commission on Oceanography and Marine Meterology, and finally provide overall management and guidance for the national system as a whole through an interagency working group. The selection and oversight of management groups will depend on whether they are research or operational, federal, commercial or university.

The foundation of the management strategy is to obtain key CDRs through the use of CDSTs. Definition of the CDSTs is based on the last two decades of experience and lessons gained from previous earth observation systems focused on climate measurements. A clear message from this experience is the need for CDSTs. These teams are composed of a group of scientists and engineers whose purpose is to convert raw instrument data into CDRs, including calibration, algorithm development, validation, error analysis, quality control, and data product design. If the data volume is small, the CDST may also produce and distribute the data products. If the data volume is large, the CDST may interface to a separate data center for production, archiving, and distribution.

Examples of effective CDSTs include the production of climate versions of the surface air temperature records, most NASA Earth Observing System satellite data products, as well as those of precursor activities such as TOMS, SAGE, and ACRIM, the international Argo ocean profiling float program, the NOAA-led Baseline Surface Radiation Network (BSRN), and the DOE's ARM Program. The reason that CDSTs are required for most, if not all, climate data records is the extreme accuracy needed for rigorous climate records (see section 3.1). Because the methods of measurement vary so greatly, effective CDSTs focus on just a few of the climate data records, and some only on a single CDR. A minimal list of variables needed as CDRs is given in Appendix 12.1.

Most CDSTs are chosen by scientific peer review. They consist of a principal investigator, and a set of co-investigators. The co-investigators lead key instrument, algorithm, sampling, validation, and data management functions. CDSTs include members of the climate modeling or climate analysis community (i.e., data users). This is key to keeping the CDST focused on the most effective approach to meeting users' needs. Most CDSTs are currently funded by a single U.S. agency. CDSTs have their work and products peer-reviewed. CDSTs are often used in national and international assessments of the state of climate science and climate impacts. They document the climate data products and their quality through web documents, conference papers, and journal papers.

Observing system science direction will be provided by boards and councils. It is too long a step between a single CDST focusing on one or two climate variables and the CCSP. A logical structure will be needed between these two extremes. The structure chosen must be able to handle the following set of changes and prioritizations:

Addressing these trade-offs between resources, requirements, and climate variables will require a hierarchical science management committee structure. While there are many ways to provide such a structure, climate data and climate modelers will be best served by a structure that gives highest priority to tight integration of major climate system components (see Figure 12-6). Portions of such a system have evolved over the last decade, primarily at the international level as discussed under Goal 5. It is most fully evolved and effective for ocean observations, while the process for terrestrial observations is at a much earlier stage of development. More of this structure will be developed by the CCSP for the United States.

The CCSP cabinet-level management structure is at the top level, responsible for resource commitments to the climate observing system. The second level consists of a sub-group of the CCSP, the Observing System Management Executive Council (OSMEC), composed of agency executives capable of committing agency funding. This group will provide a formal interface with international coordination groups, such as the CEOS and the IGOS partnership.

Two groups comprise the third level. The first is an interagency program management group responsible for providing resources to implement the observations and support science investigations; the Observing System Operations Executive Council (OSOEC). The second is the highest-level science advisory group; the U.S. Observing System Science Executive Council (OSSEC). This science council makes recommendations to both the OSMEC and the OSOEC by prioritizing across the entire climate observing system. In addition, the OSSEC will develop the objectives and requirements for the observing system, based on input from the Climate Science Advisory Councils of the research elements and decision support groups, and will evaluate the ability of the observing system to meet these objectives. These Climate Science Advisory Councils will also provide the OSSEC with evaluation reports on the performance of the system in meeting the needs of its users.

At the fourth level, Climate Data Science Councils are responsible for climate observations to support each of the seven major CCSP research elements: Atmospheric Composition, Climate Variability and Change, Water Cycle, Land-Use/Land-Cover Change, Carbon Cycle, Ecosystems, and Human Contributions and Responses, as well as Decision Support activities. The Climate Data Science Councils could be a sub-group of the Climate Science Advisory Councils shown in Figure 12-6. Some of the Climate Science Advisory Councils already exist, such as the U.S. Climate Variability and Predictability (CLIVAR) team for the Climate Variability and Change theme. The Climate Data Science Councils address the complete range of climate variables within their theme.

A minimum set of climate variables is given in Appendix 12.1. Some variables will be relevant to several of the Climate Data Science Councils. Such overlaps are inevitable in any management structure given the tight coupling of processes in the climate system. It is the responsibility of the OSSEC to assign primary responsibility for each variable to one of the Climate Data Science Councils. The Climate Data Science Councils are responsible for setting CDR requirements for absolute accuracy, stability, and space/time sampling.

Figure 12-6: Schematic of the overall management structure for the observing system.

At the fifth level, Climate Data Science Boards will address data parameters grouped into the most natural subsets for climate processes and/or for the type of instrumentation. For example, one of the Climate Variability and Change science boards would cover radiative energy from the top of atmosphere to the surface of the earth. Some of these groups have already been formed by individual agencies (e.g., sea surface topography, ocean winds, sea surface temperature, and ocean color within NASA). Within each Science Board there are in situ, satellite, and field experiment observations. The advantage of this approach is that the linkages of surface, satellite, and in situ data types for key climate parameters are actively considered in any observing system trade-offs and will be valuable to the implementation decisions of the OSOEC. Each CDST will report to the relevant Climate Data Science Board.

The Climate Data Science Boards and Executive Councils will meet as required to assess progress in the data system, nominally once per year. These meetings are expected to be in-depth workshops sufficient to deal with the complex trade-offs and linkages between the observing system components. Every five years, a major assessment of the entire climate observing system would be produced, reviewed by the NRC, and coordinated with international assessments (e.g., the International Conference on Ocean Observing Systems for Climate, GCOS Adequacy Reports, IPCC Assessments).

These boards and councils must include members of the climate data user community (e.g., climate modelers, climate analysts), as well as CDST representatives and Data Management representatives. In addition, members of the community who participate in international science or operations groups will be ex officio members of their appropriate counterpart board or council. The most efficient organization will be to maximize the extent that the Data Management and Climate Modeling management structures parallel that of Climate Observing and Monitoring.

Observations and monitoring of the climate system are carried out by a number of federal agencies. Success will require a series of steps to be initiated jointly by the OSMEC, the OSOEC, and the OSSEC. These steps include: