6
Long-Term Observations and Analyses

Sustained, accurate, long-term global observations of key variables in the climate system are essential to achieving the objectives of GOALS. A robust description of seasonal-to-interannual variations and the decadal modulation of these variations can be built up only from such an observational base. Diagnosis of the causes of these variations is impossible without long, accurate records. Continued global observations are also necessary to initialize and evaluate the global and regional coupled models envisaged under GOALS. These observations need to be assimilated into climate prediction models in much the same way that observations presently are assimilated into operational weather forecast models. The manner in which the observed fields are used to provide a description of climate and to develop hypotheses is illustrated in Figure 4-1. Observations to test specific hypotheses are gained through process studies. The conclusions reached from these data may be used to redefine the long-term observational strategy in order to describe the climate system more completely.

To be useful, measurements of seasonal-to-interannual climate variability must be consistent, continuous in duration, and of sufficient accuracy and resolution that the climate signals can be distinguished from instrumental noise and from high-frequency geophysical signals unrelated to climate variability. Because the phenomena being studied are generally of a large scale and influenced by processes in distant regions and because the intent is to improve climate predictions in locations remote from phenomena such as ENSO, long-term global coverage is essential for many aspects of GOALS. Thus, a principal component of the global monitoring of the state of the ocean, atmosphere, land, and ice



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



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 20
--> 6 Long-Term Observations and Analyses Sustained, accurate, long-term global observations of key variables in the climate system are essential to achieving the objectives of GOALS. A robust description of seasonal-to-interannual variations and the decadal modulation of these variations can be built up only from such an observational base. Diagnosis of the causes of these variations is impossible without long, accurate records. Continued global observations are also necessary to initialize and evaluate the global and regional coupled models envisaged under GOALS. These observations need to be assimilated into climate prediction models in much the same way that observations presently are assimilated into operational weather forecast models. The manner in which the observed fields are used to provide a description of climate and to develop hypotheses is illustrated in Figure 4-1. Observations to test specific hypotheses are gained through process studies. The conclusions reached from these data may be used to redefine the long-term observational strategy in order to describe the climate system more completely. To be useful, measurements of seasonal-to-interannual climate variability must be consistent, continuous in duration, and of sufficient accuracy and resolution that the climate signals can be distinguished from instrumental noise and from high-frequency geophysical signals unrelated to climate variability. Because the phenomena being studied are generally of a large scale and influenced by processes in distant regions and because the intent is to improve climate predictions in locations remote from phenomena such as ENSO, long-term global coverage is essential for many aspects of GOALS. Thus, a principal component of the global monitoring of the state of the ocean, atmosphere, land, and ice

OCR for page 20
--> variability must be a comprehensive satellite system. For other potentially predictable elements, where the intent is to use oceanic variability in one region to predict climate variability on a global basis, additional long-term observational needs may be specific to particular regions. As an example, SSTs and winds in the tropical Pacific are essential as initialization data for models used to predict El Niño and its impacts. ENSO model predictions can also be evaluated from surface air temperature and precipitation data at specific locations remote from the tropical Pacific. During TOGA, a core in situ, regionally focused, long-term observing system was established in a research mode, for monitoring and forecasting seasonal-to-interannual variability of equatorial central Pacific SST (NRC, 1994b). This observing system provides a starting point for the in situ oceanic component of an observational system for GOALS. In conjunction with a comprehensive satellite observational program, the expansion of long-term oceanic observations under GOALS from the tropical Pacific to the tropical Indian and Atlantic Oceans and to higher latitudes is believed to be necessary to achieve additional improvements in seasonal-to-interannual predictability. This, of course, imposes significantly increased demands on observing systems as well as data management and data exchange systems. An important consideration is that the viability of some existing operational systems is seriously threatened by internationally imposed restrictions on data exchange. An internationally agreed upon policy for the free exchange of and access to scientific data for research purposes is critical to the success of GOALS, and indeed all other research endeavors. The deterioration of existing operational observing systems is of serious concern to the GOALS Panel. During TOGA, enhancements important to the monitoring and prediction of ENSO were implemented under the World Meteorological Organization (WMO) World Weather Watch (WWW) program. However the upper-air sounding system has continued to degrade since the First GARP Global Experiment (FGGE) in 1979. Currently, there are about a third fewer upper-air soundings per day because of reductions from two to one per day in many countries (including the United States), determined in some cases by the increasing cost of expendables. The viability of the upper-air sounding system is also threatened by external factors. While the quality of rawinsondes has improved in the past 15 years, the decommissioning of the Omega navigation system in 1997 has led to the loss of some wind soundings. Some systems that depend on Omega have transitioned to the use of the Global Positioning System, but rawinsondes using this technique are much more expensive. Notwithstanding the above situation, the greater long-term threat is the assigning of frequencies used for transmitting rawinsonde signals for the use of cellular phones. This means narrower available frequencies and, thus, more expensive transmitters. These issues are of concern because changes in the type(s) of upper-air rawin-

OCR for page 20
--> sondes or in their total number or frequency of deployment can alter the climate record as seen through analyses and reanalyses. Of special importance are the existing operational observational and data exchange systems. For the atmosphere, the WWW should be maintained and improved. WWW, a part of the operational system developed for weather monitoring and prediction, is composed of the Global Observing System (GOS), Global Telecommunication System (GTS), and Global Data Processing System. The WWW involves national, regional, and world centers. The International Oceanographic Commission (IOC) of the United Nations Educational, Scientific, and Cultural Organization jointly with the WMO, coordinates the exchange of real-time and near-real-time oceanographic data via GTS. Delayed-mode exchange is managed directly by IOC's International Oceanographic Data Exchange system, which includes national, regional, and world centers. Enhancements to observing systems, and the collection and exchange of data for climate purposes are being planned by the Global Climate Observing System (GCOS) (WMO, 1995a), which includes a significant array of oceanographic and land surface measurements. The full set of data required for oceanography is handled by the Global Ocean Observing System (GOOS). Both GCOS and GOOS rely heavily on existing operational systems and mechanisms. In contrast to the systems already in place for atmospheric and oceanic data, the international infrastructure does not fully exist for the collection and exchange of data to characterize the land surface (and vegetation) and land processes pertaining to climate variability. The promotion and establishment of such an infrastructure are being discussed under the Global Terrestrial Observing System (GTOS) (WMO, 1995b). Efforts to coordinate the requirements stemming from GCOS, GOOS, and GTOS are being undertaken under the Integrated Global Observing Strategy (IGOS) in order to identify, without redundancy to the extent possible, key enhancements to the GOS that serves the multiple needs of these programs. The panel feels strongly that GOALS should collaborate and coordinate with all of these programs. Another important consideration is that the data from operational observing systems (both space and surface based) suffer from various problems, among them accuracy, representativeness, calibration, and continuity over time, as well as inadequately documented changes in instrumentation, observing practices, and data processing procedures or algorithms. Although these problems are especially important for decade-to-century time scales, they are also important for seasonal-to-interannual time scale research such as represented by GOALS. Although the problems summarized above are serious concerns, the greatest observational challenge for GOALS involves the ocean variables that typically have a very short length of record, poor spatial coverage, and poor resolution in time. The fundamental problem is that many of the in situ oceanic observations (e.g., the TOGA-TAO array and surface-drifter observations of SST), as well as many of the satellite ocean observations that are crucial to GOALS (e.g., altim-

OCR for page 20
--> etry), have been developed and maintained largely in research modes. At present, no federal agency has committed to a comprehensive operational ocean observing program specific to the needs of seasonal-to-interannual climate forecasting, and more generally, to climate research. One of the objectives of GOALS must, therefore, be to promote the transition of observations from a research framework to an operational framework in order to ensure long-term continuity of the observations. Note that the term operational in this context does not necessarily mean the transitioning of a system to an operating agency (within the United States, a NOAA line organization such as the NWS or the National Ocean Service), but rather a commitment to both the infrastructure necessary to sustain observations and to maintain the quality of the observations. An example is the TOGA observing system and, in particular, the TOGA-TAO array of buoys moored in the tropical Pacific. The latter is maintained in a quasi-operational mode by a multinational research group spearheaded in the United States by the NOAA Pacific Marine Environmental Laboratory (PMEL). While it is crucial to maintain and enhance existing observing systems, GOALS will inevitably have to rely on and utilize the observations planned by the next generation of Earth observing satellites and programs. Space-based satellite observations will play an important role in providing global data coverage for many of the key variables required by GOALS with a spatial and temporal resolution, which cannot be matched by in situ measurements alone. In addition to providing much of the information necessary for coupled climate models, satellite observations of seasonal-to-interannual variability of ocean, atmosphere, and land variables will be used to assess and refine the models. Importantly, these observations will also play a vital role in monitoring the interactive processes within the purview of GOALS on a global basis. Space-based observations are already used as a powerful and complementary adjunct to in situ observing systems with the latter providing absolute calibration and satellites providing global coverage. Global scale fields of SST are best constructed through a well conceived combination of in situ and satellite data—a process that maximizes the advantages of satellite observations (e.g., global scale coverage and relatively accurate gradient information) while simultaneously compensating for deficiencies in situ observing networks even though they provide direct absolute value measurements. Continuous, uninterrupted satellite observations are required for at least the 15-year duration of GOALS, if the program is to achieve its objective of determining the predictability of climate (especially precipitation) over North America and other regions on seasonal-to-interannual time scales. Such long, continuous time series, exceeding the lifetime of individual satellite missions or instruments, are extremely difficult to acquire outside of operational programs. While many of the needed satellite-based observations of key atmospheric and solar forcing variables are already incorporated into U.S. and international operational meteorological satellite programs, few of the key oceanographic or land variables are

OCR for page 20
--> observed from operational satellites. Lacking an operational framework for these important satellite observations, the research community has worked hard over the past decade to maintain the satellite observational systems in a research mode. These satellite systems have been supported financially by the National Aeronautics and Space Administration (NASA) in highly successful collaborations with international partners (primarily the French space agency CNES and the Japanese space agency NASDA). Because of the long lead time required to build and launch dedicated satellites, firm commitments must be established in the near future to ensure continuous observations from these research satellites over the full duration of GOALS. The NASA Office of Earth Science sponsors the Earth Observing System (EOS) designed to observe atmospheric, oceanographic, and land variables of importance to climate change issues. The Office of Earth Science has provided a framework within which individual research-mode missions will be launched to provide high-quality measurements on a mission-by-mission basis with no assurance of continuity of measurements of any specific variable. The GOALS objectives require continuous time series through at least 2010. NASA's charter to demonstrate and extend technology and the budgetary pressures facing the agency raise questions about NASA's ability to continue to provide quasi-operational data for all of the important satellite instruments over the duration of GOALS. Given the importance of the satellite measurements for studies of seasonal-to-interannual variability of the ocean and atmosphere, as well as the requirement for high-quality satellite data for decadal and longer time-scale climate change, steps should be initiated now to incorporate mature satellite technology into U.S. and international operational satellite programs. In the United States, present plans are to merge the Department of Defense and NOAA operational satellites into the National Polar-Orbiting Environmental Satellite System (NPOESS). The first of the NPOESS satellites is not planned for launch until 2007. Since the lead time to achieve operational status of a satellite sensor is a decade or longer, a coordinated long-range plan must be developed at the earliest possible opportunity in order to provide benefit to the latter stages of the GOALS program and any future research programs that evolve from GOALS. Key Variables The challenging combination of accuracy, spatial and temporal coverage and resolution, and duration requirements of the observational component of GOALS is compounded by the range of variables needed to address the scientific objectives of the program. At this early stage of GOALS, all of the relevant variables and processes have probably not yet been identified. Expanding on the TOGA experience and other experimental programs that took place during the last decade and a half as well as recommendations made in reports such as NRC (1994b) and WMO (1995a,b), a list has been compiled of priority variables considered

OCR for page 20
--> essential for a sustained, long-term observational system to support seasonal-to-interannual climate variability research and prediction. Table 6-1 summarizes the variables presently considered necessary for each of the three components of the coupled ocean—atmosphere—land system. The proposed list is divided into state variables and external (or forcing) variables. State variables are defined as variables necessary to monitor the ''state'' of the physical system or subsystem, while "external" or "forcing" variables are defined as those variables that define the degree of interaction between system components. Though from an Earth-system standpoint, the partitioning of the system into subsystems and into state and forcing variables is somewhat arbitrary, for many purposes, such categorization is conceptually useful. Within each group, variables are listed in approximate order of importance to the program. Some variables could appear under more than one category. It is also worthy of mention that GOALS's data sampling requirements have not yet been completely established. The spatial and temporal resolution requirements for a global in situ observational system for research on short-term climate variability and prediction have not been reassessed in light of the knowledge gained from TOGA, the World Ocean Circulation Experiment (WOCE), and more than 5 years of TOPEX/POSEIDON altimeter data. Moreover, it is not apparent at present whether the expansion of moored ocean arrays such as TAO to mid-latitudes is the best way to obtain long-term, extratropical ocean observations. As mentioned above, observations from satellites presently in orbit or soon to be launched will, therefore, play a key role in achieving the global requirements of the observational component of GOALS, at least in the early stages of the program. However, satellites alone cannot meet the observational needs of GOALS. In situ observations such as the TOGA-TAO array, temperature and salinity profiles from buoyancy-adjustable subsurface floats, drifter observations, island-based tide gauge measurements, and Volunteer Observing Ships (VOS) observations of upper-ocean thermal structure and the overlying atmosphere also must be maintained to validate and calibrate satellite observations and to provide measurements of vertical structure and surface properties that cannot be obtained from satellite observations alone. An important and overriding principle for the GOALS observing system is that individual components should not be developed in isolation which, historically, has often been the case. Priority should be placed on obtaining fields of variables especially useful for model initialization and evaluation. The specific platforms used to obtain the required measurements could be of secondary consideration, although they are important from the standpoint of implementing specific enhancements to existing observing technology. It should be clear from the list of key variables tentatively identified in Table 6-1 that a composite GOS comprising both space-based and in situ measurements is required by GOALS. In the design and implementation of the GOALS observing system, the stringent

OCR for page 20
--> TABLE 6-1 Key "State" and "External (or Forcing)" Variables Necessary for the GOALS Program. Variables Are Listed in Order of Priority. Ocean State Variables External Variables   upper ocean temperature upper ocean currents sea level upper ocean salinity optical absorption sea ice extent, concentration, and thickness   wind stress net surface solar radiation downwelling longwave radiation surface air temperature surface humidity precipitation Atmosphere State Variables External Variables   wind structure thermal structure surface air temperature sea level pressure water vapor structure columnar water vapor and liquid water content cloud cover and height precipitation   sea surface temperature net radiation at top of the atmosphere land surface variables (listed below) Land State Variables External Variables   soil moisture snow cover and depth vegetation type, biomass, and vigor water runoff ground temperature   precipitation net surface longwave and shortwave radiation surface wind surface air temperature surface humidity evaporation evapotranspiration

OCR for page 20
--> sampling requirements of some variables of interest must be recognized explicitly. One of the major objectives of GOALS (especially its observational component) is to contribute to the development of IGOS being developed by the Committee on Environment and Natural Resources (CENR) of the White House Office of Science and Technology Policy (OSTP), within the framework of a comprehensive system of analysis to ensure that global-scale coverage of priority variables and calibrated products for research and applications can be provided to the scientific community. The GOALS Panel feels that the scientists participating in GOALS research programs should be involved actively in providing advice on various aspects of the long-term observing system for climate research and prediction as IGOS evolves. Some of the issues surrounding observational requirements are likely to be resolvable through statistical analyses or numerical experimentation. Limitations in the quality and quantity of previous observations and in the physics of numerical models will almost certainly stall the satisfactory resolution of some questions. The research agenda of GOALS should be able to make a strong contribution to the future development of global climate observing systems by conducting appropriate process studies (see Section 7), and by making specific recommendations on enhancements needed for observing systems. Through process studies in the regions selected for focused experiments by GOALS, even more specific recommendations are likely for enhancements needed to regional observational networks. Of course, the research conducted under GOALS would specifically target the data and observing system needs to support seasonal-to-interannual climate prediction. Notwithstanding the above, the observational responsibilities of GOALS should not be limited merely to the measurement of the variables for the duration of the GOALS program. As mentioned earlier, the seasonal-to-interannual focus of GOALS overlaps significantly with programs pertaining to both shorter and longer time scales. To be useful to other components of WCRP/CLIVAR and the USGCRP, long-term observations must be quality controlled, organized, and ultimately processed into gridded fields for use by scientists involved in GOALS research, as well as other programs such as deacon and GEWEX. The GOALS program should, therefore, be active in the construction of data sets as well as in objective analysis and data assimilation. A close Cupertino between the scientists involved in Goal's model development effort and those involved in the development of observational data assimilation models is highly recommended by the GOALS Panel, because numerical models are an integral part of state-of-the-art assimilation procedures. Measurements and Observing Systems The priority state and external observations identified in Table 6-1 will be

OCR for page 20
--> obtained from a combination of current and new observing systems. In support of these, observation facilities need to be developed for transmitting the data to regional and international data centres. The panel highlights several issues that need to be considered seriously in the implementation of the observing systems and networks supporting the scientific objectives of GOALS. They are also important for the programs complementary to GOALS such as GEWEX and deacon. Critical elements of existing observing systems should be maintained in order to obtain long, continuous, and consistent time series. In this regard, continuous and consistent in situ time series of key field variables should be given high priority. As noted previously, they are needed for model evaluation, empirical studies, ongoing calibration and validation of remotely sensed observations, and information that cannot be obtained from satellites. There may also be a need for new, long-term observational technologies to be developed and incorporated into operational systems following the findings of future process studies directed at improving models and their prediction skills. It is important that all measurements within observational programs are accompanied by sufficient "meta-data" in order to document changes in location, observing practices, satellite retrieval algorithms, instrumentation, and other characteristics such as exposure, space/time coverage, and accuracy. Where possible, use should be made of observational products from other programs, such as cloud and radiation products from the International Satellite Cloud Climatology Project (ISCCP) and precipitation estimates from the Global Precipitation Climatology Project (GPCP) and the Tropical Rainfall Measuring Mission (TRMM) (the NASA/NASDA satellite launched in November 1997 to measure tropical rainfall more accurately than ever before from a space-based system). To help define and develop appropriate ocean, atmosphere, and land observing elements for incorporation into the GCOS, the GOALS Panel recommends that close coordination should be maintained with the DecCen component of CLIVAR and other WCRP programs such as GEWEX and ACSYS (Arctic Climate System Study). Among other activities, this coordination between programs should include the evaluation and prioritization of observed fields and the articulation of resolution and accuracy requirements. Advantage should be taken of emerging new methods for real-time data transmission to enhance real-time data return and reduce the cost of transmission relative to the cost of making observations. Also, the GOALS Panel considers several other issues to be of importance. These are the: maintenance of the network of in situ sea-level observations; calibration and validation of paleoclimate records relevant to seasonal-to-interannual climate variability;

OCR for page 20
--> calibration of global water vapor fields; continuity of accurate satellite altimetry; continuity of accurate satellite scatterometers; and maintenance of the rawinsonde network throughout the tropics and the tropical Pacific, in particular. In addition, there are other important issues that need to be addressed by the international community. These are the: Reduction of costs of using the ARGOS system for transmission of remotely located in situ data; and Reduction of the high cost of obtaining observations and analyzed fields from international data centers outside the United States. For the coordination and implementation of the above issues, among other activities of importance to GOALS, the panel strongly recommends the establishment of a GOALS Interagency Project Office (discussed in Section 12). Data Analysis and Assimilation; Data Reanalysis GOALS requires the capacity to observe and use multi-disciplinary data covering the atmosphere, the oceans, the land surface, and snow and ice fields, among others. For observations from the combination of space-based and in situ platforms to be of most value to the modeling component of GOALS, they should be transformed to gridded fields. Consequently, methods to effectively combine the output from different measurement systems and platforms should continue to be developed. The combined use of altimeter sea-level data with thermal structure information to estimate the basin-scale field of upper-ocean thermal structure is an example of such a consolidation. New objective analysis and assimilation techniques should also be developed, especially for ocean and land surface variables to determine land surface processes, ocean—atmosphere interactive processes, and the hydrological cycle. Historical data analysis warrants special mention. Because GOALS is concerned with seasonal and interannual time scales, it is important to compile a data record as long as possible so that many samples of seasonal and interannual phenomena may be examined. To this end the panel recommends that continued efforts should be made to improve the quality and volume of the historical data base through data archaeology and support for the development and evaluation of data sets. Typically, there is a need to correct and adjust historical data in various ways to compensate for changes in the manner in which measurements were made in the past. These data are used to determine the historical climate record, which provides estimates of the variability on interannual and longer time scales. Such variability includes the impact or influence of both natural and anthropo-

OCR for page 20
--> genic forcing components. Where possible, a systematic and periodic inter-comparison of historical data sets should be carried out. The reanalysis of the data available from operational data assimilation schemes is an extremely important task for GOALS and currently an area of considerable activity. Much of the global gridded data fields used for research are based on past operational analyses techniques that are not ideal for climate studies. This is because the gridded fields contain inhomogeneities resulting from shifts in model and data assimilation/analysis methodology that have been tuned for operational weather forecasting purposes. These data fields have changed and will continue to change in time as forecast models and their corresponding data assimilation schemes are improved. For this reason efforts to reanalyse historical data should continue and should be applied to operational atmospheric analyses, all oceanographic data and analyses, and satellite-based products. The panel underscores that reanalysis needs to be a periodic and ongoing effort. That is, the reanalysis system cannot be frozen in time, because data assimilation and modeling sophistication will certainly improve and change with time. Even the present reanalysis efforts, while highly commendable, have been found to contain several problems principally resulting from assimilation schemes and the state of models, both of which are still developing. In summary, the panel suggests that data reanalysis should become a routine activity that is repeated at regular intervals of time, conceivably every 5 to 10 years, in order to take advantage of the latest state-of-the-art systems and techniques available. Observations Working Group To carry out a thorough investigation of various issues dealing with long-term observations and analysis, the panel recommends that a GOALS Observation Working Group (GOWG) be established. This group should develop the scientific basis for GOALS observations. The activity is considered crucial because the scientific basis for observational requirements for climate studies and model evaluation is continually evolving. Furthermore, during the past decade and a half, our knowledge of the climate system has improved, resulting in changes in resolution and data type requirements. In addition, other forms of data, especially satellite, are now available. An important role of the GOWG will be to provide guidance in the prioritization of satellite-based measurements to be incorporated in the operational satellite program. Furthermore, an observations working group, working in conjunction with GEWEX, DecCen, GCOS, and the CLIVAR Upper Ocean Panel, would be well placed to make specific recommendations about future satellite instruments and products. Fuller community participation should be encouraged by the organization of one or more workshops to address data requirements for the support of GOALS objectives.