Finding 2.2) (Gaffen, 1993, 1994; Trenberth, 1995), the next-generation system should be implemented sooner rather than later and with adequate testing and overlap in order to establish a reliable, global, in situ sounding system.
3. CLIMATE DATA RECORD ISSUES
3.1 NOAA's Climate Mission
NOAA's traditional role regarding observations has been to ensure the operational utility of environmental data. With the importance of long-term climate issues on the rise, operational data streams are now vital for understanding these issues. Operational data today will be climate data tomorrow, but the utility of the climate data will depend on the care that is taken with the operational measurements (NRC, 1999).
NOAA should reinvigorate its efforts to “ensure a long-term climate record” (NOAA, 1995). This perspective should permeate the full range of activities related to climate observation, including instrument design and specification, instrument siting, specification of observing methods, data and metadata archiving, production and validation of CDRs, data analysis, and dissemination of products.
3.2 Production of CDRs
Some significant CDRs have been generated by individuals or small teams, often with little or no funding, who were mostly driven by an interest in solving an important scientific problem. Some of these groups continue to perform the routine functions of production, dissemination, validation, and evaluation of their CDRs (mostly unfunded), while others have discontinued these activities. It is often difficult to separate “production,” which may be viewed as “turning the crank,” and continuous evaluation, which requires experience, skill, and scientific insight.
NOAA should take responsibility for identifying proven CDRs and ensuring that the construction of these be maintained. In addition, NOAA should assume responsibility for supporting and developing the required scientific expertise, documenting the CDR construction methodology, and ensuring that the scientific expertise has been institutionalized, rather than merely residing with individual scientists. It is also important that the time series can be reproduced by future investigators.
3.3 Network Monitoring
NOAA does not have a network performance monitoring system or a system to routinely inspect space-based, upper-air, and other in situ observations for systematic, time-dependent biases. As a result, time-varying biases can go undetected for long periods, thereby degrading time series without a possibility of taking early corrective action, e.g., new calibration, changes in procedures and processing (NRC, 1999). Operational assimilation schemes have difficulty in identifying shifts and drifts of a magnitude important to climate (Basist and Chelliah, 1997; Trenberth et al., 2000).
A network performance monitoring system to identify both random errors and time-dependent biases in both space-based and in situ observing systems would enable NOAA and the scientific community to identify and correct errors as soon as possible in these critical observing systems. These diagnostics should become part of the metadata associated with the observations.
3.4 Inclusive Dialogue/Information Exchange
The panel notes that its two meetings 4 sparked lively interaction among a group of academic, private sector, and government scientists who welcomed the opportunities to exchange both general and highly technical information about the MSUs. Investigators who generate CDRs from operational instruments often do so in relative isolation and often have difficulty securing necessary information from design engineers, other investigators, operational satellite controllers, etc. It has been recommended for both the NPP and NPOESS missions that specific instrument science teams be designated to ensure that the construction of CDRs will be unhindered by communication problems.
NESDIS should establish for each POES operational instrument a structure by which the communication of information may be assured as CDRs are developed and refined. This could be implemented with the establishment of an ad hoc group of individuals who are involved in some way with the development of the instrument and the CDRs. Sponsored meetings or workshops, possibly with published proceedings, would help ensure that the right mix of people have access to one another. Another approach could be the formal establishment of, for example, an MSU/AMSU Climate Science Team which would afford the members the opportunity to deal with issues of calibration, validation, long-term stability, and future requirements for deep-layer atmospheric temperature (as well as other microwave-based products). The team could also advise NESDIS/NCDC on issues of data storage, data access, and which significant products to archive.
Panel on Reconciling Temperature Observations
The panel's first meeting was on March 8-10, 1999 in Asheville, North Carolina. Its purpose was to provide an opportunity for the panel to hear information and conduct deliberations pertaining to the panel's first report (NRC, 2000a). The panel's second meeting, held on June 8, 2000 in Washington, DC, was part of the study process leading up to this letter report.