NOAA's National Climatic Data Center
Veach-Baley Federal Building
151 Patton Avenue
Asheville, NC 28801-5001
Telephone: +1 828.271.4302
Dr. Richard W. Reynolds has specialized in analyzing sea surface temperatures (SST) since he graduated from the University of Hawaii in 1975 with a PhD. In 1980 Dr. Reynolds began his career at NOAA as the lead scientist responsible for the development, implementation and operational production of sea surface analyses and associated. He has been active in improving the accuracy of the SST analyses by optimizing the advantages of in situ (ship and buoy) and satellite data.
Three analysis versions are currently available to users:
Dr. Reynolds has received the following awards:
Dr. Reynolds joined CICS-NC as a senior scientist in November 2009.
Continuously monitoring operations of GHCN-M version 3. Using the International Surface Temperature Initiative, new stations will be available for observing global climate variability and change. This will lead to the development of GHCN-M version 4.
Since the early 1990s the Global Historical Climatology Network-Monthly (GHCN-M) dataset has been an internationally recognized source of information for the study of observed variability and change in land surface temperature. Version 3, which marks the first major revision to this dataset in over a decade, became operational in May 2011 and provides monthly mean temperature data for 7280 stations from 226 countries and territories. This version introduces a number of improvements and changes that include consolidating “duplicate” series, updating records from recent decades, and the use of new approaches to homogenization and quality assurance. Since its initial release, GHCN-M version 3 has undergone minor updates to incorporate monthly maximum and minimum temperature, as well as improve processing run time.
Efforts are underway to upgrade GHCN-M and release version 4. One of the major priorities of this update is to add more stations to improve spatial and temporal scales. The International Surface Temperature Initiative (ISTI), a working group of climate scientists, meteorologists, and statisticians, consists of an effort to create an end-to-end process for land surface air temperature analyses. The goal is to create a single, comprehensive global databank of surface meteorological observations. The databank will be version controlled and seek to ascertain data provenance. There are multiple stages of the databank, including the original paper record, keyed data in its native format, and a merged dataset with duplicate source data reconciled. All data, along with its underlying code, will be made public free of charge, in order to be open and transparent.
Once the databank is released, more stations with better provenance will be added into current GHCN-M processing to create a more robust dataset. Other goals of GHCN-M v4 include improving quality assurance, as well as addressing uncertainties within the surface temperature record.
Multiple sources of data, on monthly and daily timescales, have been submitted to NCDC and uploaded to the Databank FTP site. Some sources contain digital images of the original paper copy (known as Stage 0), while all should contain keyed data in its native format (Stage 1). Data is then converted into a common format and data provenance flags are added to every single observation (Stage 2). These flags are determined based upon information provided by the data sender. Some examples include whether an original paper copy is available, and if any quality control or homogenization was applied to the dataset prior to submission. Currently, there are 20 daily sources and 18 monthly sources that have been converted to Stage 2. In addition, all daily sources were converted into monthly averages, in order to make the monthly version of the databank more robust.
Figure 1. Location of Stage 2 monthly station records in the databank, along with their period of record.
Currently there are 134,758 station records in Stage 2, which include the monthly sources, as well as the “monthly averaged” daily sources (see Figure 1). The algorithm to merge these sources into a single merged Stage 3 dataset with duplicates removed is currently under development. The process occurs in an iterative fashion through all the sources, and comparisons are made to determine if a station match exists and a merge should occur. This process is designed to be Bayesian in approach and based upon metadata matching and data equivalence criteria. Metadata comparisons include the geographic distance between the stations, elevation difference, and an algorithm for station name similarity. Data comparisons include “goodness-of-fit” measures for overlapping data, and neighbor comparisons for non-overlapping data. Weighted priors are formed based upon bootstrapping techniques, and are then recombined to form a posterior probability of station match. Validation of each technique is being applied using stations with known bias and noise and comparing output.
The algorithm is being designed to be modular, and easy to read by researchers. If an analyst determines that certain techniques are inadequate, or if they wish to apply their own techniques, the modularity of the algorithm should facilitate this with minimal effort.
Lawrimore, J.H. et al (2011) The International Surface Temperature Initiative’s Global Land Surface Databank. Submitted to the 9th International Temperature Symposium
Rennie, J.J. (2011), The International Surface Temperature Initiative: Global Land Surface Databank Development, CICS Science Meeting, National Climatic Data Center, Asheville, NC, 3 Nov 2011.