Future of the National Weather Service Cooperative Observer Network (1998)

Because observations of basic weather and climate variables differ with instrument exposure to nearby structures and terrain, sensor response characteristics, the time of observation, and the method of recording, care must be taken to understand fully the ramifications of changing an instrument, site, or routine in order to maintain the integrity of decades of observations. As a minimum requirement for new weather observing systems, the following general recommendations are offered:

Develop and apply standard procedures for collecting side-by-side overlapping measurements for all potentially significant changes made in observation and measurement techniques. This period of overlap should span at least one annual cycle.

Make routine assessments of ongoing calibration, maintenance, and climate record homogeneity problems for which corrective action can be taken. Such assessments and subsequent actions must be documented and archived with the data.

Along with routine transmissions of observations, regularly (as opposed to ad hoc) schedule transmissions of station observation and measurement practices, as well as local environmental conditions in the vicinity of the station, that are pertinent to the interpretation of the observations and measurements. Station histories should be a mandatory part of the permanent data archive along with the measurements and observations. They should be treated with importance equal to the data itself.

Ensure that network designers and instrument engineers are provided climate requirements at the outset of network design and instrument design.

Develop, wherever feasible, some level of "low-technology" backup to "high-technology" observing systems to safeguard against unexpected operational failures (power interruptions, lack of replacement parts, etc.).

Archive raw data sensed from the instruments prior to transformation into standard atmospheric variables or products along with the processed data and processing algorithms.

Restrict the number of station relocations to an absolute minimum.

Discontinue observations of atmospheric variables with a long historical record (spanning many decades) only after a thorough evaluation of the impact on the climate record.

Develop standard data packages that fully describe all algorithms, averaging procedures, quality control, homogeneity checks, and corrections that have been applied to the derived data. This now includes quantities such as temperature or precipitation, which can now be measured indirectly.

In addition to these general principles there are a number of specific recommendations relevant to existing and planned observing networks within the NWS.

Develop a policy to assess biases introduced by station relocation or changes in instrumentation, and develop and deploy a standard observing system to be operated by part-time volunteer observers that meets accuracy and reliability requirements for climate data.

Quantify the biases introduced by the Maximum-Minimum Temperature System relative to the liquid-in-glass thermometric measurements obtained in Cotton Region shelters. This is likely to be heterogeneous over various weather regimes and should be quantified on this basis.

Quantify the bias associated with unshielded precipitation measurements. This bias is likely to be heterogeneous over various weather regimes and should be quantified on this basis.

Wherever and whenever possible, conduct overlapping simultaneous measurements when there is a necessity to change observation sites. The simultaneous measurements could be discontinued when the impact of the change can be quantified. The National Weather Service Operations Manual (Section B-11 ) recommends overlapping observations for a period of one to three years. A rededicated commitment to this procedure is required. In an operational environment, this may not always be possible. For this reason it is advisable to operate a dense network of stations designed so that occasional station losses will not badly degrade climatic analyses. The trend over the past two decades of moving toward fewer and fewer temperature monitoring sites should be stopped or reversed.

Site stability needs to be a key criterion in the selection of new sites. National parks should be ideal candidates for sites not likely to undergo substantial changes.

Every effort must be applied to protect the sites and data sets in the network that have provided the crucial, long-term, consistent measurements utilized to assess climate change within the United States. The cooperative observer program should develop a prioritized list of network sites for preservation and continuation based on their contribution to climate change assessment.

Every effort should be made to implement the technology to retain maximum and minimum temperature measurements on a midnight-to-midnight basis.

Routine reports of each station's operations should be included with the monthly data sent to the archives. Ad hoc reporting of changes leads to questions regarding the quality of the station histories.

NRC. 1992. Toward a New National Weather Service--Second Report. Report of the Committee on National Weather Service Modernization, Commission on Engineering and Technical Systems, National Research Council. Washington, DC: National Academy Press.