The International Surface Temperature Initiative

TitleThe International Surface Temperature Initiative
Publication TypeJournal Article
Year of Publication2013
AuthorsThorne, P, Lawrimore, JH, Willett, KM, Allan, R, Chandler, RE, Mhanda, A, de Podesta, M, Possolo, A, Revadeker, J, Rusticucci, M, Stott, PA, Strouse, GF, Trewin, B, Wang, XL, Yatagai, A, Merchant, CJ, Merlone, A, Peterson, TC, Scott, EM
JournalAIP Conference Proceedings
Volume1552
ISSN0094-243X
Abstract

The aim of International Surface Temperature Initiative is to create an end-to-end process for analysis of air temperature data taken over the land surface of the Earth. The foundation of any analysis is the source data. Land surface air temperature records have traditionally been stored in local, organizational, national and international holdings, some of which have been available digitally but many of which are available solely on paper or as imaged files. Further, economic and geopolitical realities have often precluded open sharing of these data. The necessary first step therefore is to collate readily available holdings and augment these over time either through gaining access to previously unavailable digital data or through data rescue and digitization activities. Next, it must be recognized that these historical measurements were made primarily in support of real-time weather applications where timeliness and coverage are key. At almost every long-term station it is virtually certain that changes in instrumentation, siting or observing practices have occurred. Because none of the historical measures were made in a metrologically traceable manner there is no unambiguous way to retrieve the true climate evolution from the heterogeneous raw data holdings. Therefore it is desirable for multiple independent groups to produce adjusted data sets (so-called homogenized data) to adequately understand the data characteristics and estimate uncertainties. Then it is necessary to benchmark the performance of the contributed algorithms (equivalent to metrological software validation) through development of realistic benchmark datasets. In support of this, a series of successive benchmarking and assessment cycles are envisaged, allowing continual improvement while avoiding over-tuning of algorithms. Finally, a portal is proposed giving access to related data-products, utilizing the assessment results to provide guidance to end-users on which product is the most suited to their needs. Recognizing that the expertise of the metrological community has been under-utilized historically in such climate data analysis problems, the governance of the Initiative includes significant representation from the metrological community. We actively welcome contributions from interested parties to any relevant aspects of the Initiative work.

DOI10.1063/1.4821418
Refereed DesignationRefereed
Author Address

NERSC