4km NODC/RSMAS AVHRR Pathfinder v5.1 5-day Data for 1982 (NCEI Accession 0051826)

The following information refers specifically to the Version 5.0 Pathfinder algorithm. The Version 5.1 Pathfinder algorithm, used to create the data in this accession, is different only in that it replaces the Reynolds Weekly Optimally Interpolated SST Version 2 (OISSTv2) with the Reynolds Optimum Interpolation 1/4-degree Daily SST Analysis Version 2 (DOISSTv2) as the first guess SST and in the quality control procedures. The primary effect of this change is to retain more data in the high gradient regions and in regions where meandering or feature advection is present; effect on the retrieved SST is minimal.

The Version 5.0 Pathfinder algorithm was used to produce twice-daily, 5-day, 7-day, 8-day, monthly, and yearly global SST and related parameters (described below). This algorithm is an improved version of the previously most successful of many methods used to derive SST from AVHRR observations (see Barton (1995) for a review of earlier techniques), the University of Miami Pathfinder version v4.2 SST algorithm (described fully in Kilpatrick et al., 2001).

The v4.2 algorithm offered marked improvement over operational retrieval algorithms such as MCSST and was applied to AVHRR data to maximize accuracy and to minimize artificial fluctuations arising from the sequence of AVHRR instruments flown on NOAA's polar-orbiting satellites during the past 2 decades. The 9 km v4.2 Pathfinder SSTs were shown to be the highest quality product then available for the construction of global climatologies (Casey and Cornillon, 1999) and longer-term SST trend determination (Casey and Cornillon, 2001), and have been demonstrated to be accurate within about 0.3 degrees C under optimal conditions (Kearns et al., 2000). Relative to the older 9 km v4.2 Pathfinder data, the new, ~ 4 km resolution Pathfinder Version 5.0 global SSTs increase detail roughly by a factor of four simply by virtue of the increased resolution. The increase in detail over widely used but relatively coarse SST datasets such as Optimally Interpolated SST Version 2 (OISSTv2; Reynolds et al., 2002) and the Hadley Centre's Global Sea Ice and SST (HadISST1; Rayner et al., 2003) is far greater.

In addition to the increased resolution, significant improvements have been made in the Version 5.0 which enhance the usefulness of the SST fields. Currently, these enhancements include the use of sea ice in the quality level determination scheme, inclusion of many inland water bodies, and the use of a greatly improved land mask. The greatest improvements are seen in coastal zones, marginal seas, and boundary current regions where SST gradients are often large and their impact on operational or research products is greatest. Separate SST products for daytime and nighttime AVHRR retrievals are made to better understand the differences in skin and bulk temperatures, since mean differences between AVHRR-measured skin temperatures and bulk temperatures of 0.1 to 0.2 degrees C (Schluessel et al., 1990) and locally varying differences of up to 1.8 degrees C (Minnett et al., 2000) have been observed.

In addition to SST values, the Pathfinder V5.0 Project makes six other parameters available, for a total of seven per time step:
1. All-pixel SST - The all-pixel SST files contain values for each pixel location, including those contaminated with clouds or other sources of error. The Overall Quality Flag values may be used to filter out these unwanted values. The SST value in each pixel location is an average of the highest quality AVHRR Global Area Coverage (GAC) observations available in each roughly 4 km bin.
2. First-guess SST - The Version 5.0 Pathfinder algorithm uses a first guess SST provided by the Reynolds Optimally Interpolated SST Version 2 (OISSTv2) product. The OISSTv2 is also used in the quality control procedures. Version 5.1 Pathfinder uses the Reynolds Optimum Interopolation 1/4-degree Daily SST Analysis Version 2 (DOISSTv2) both as a first guess SST and in the quality control procedures.
3. Number of observations - This parameter indicates the number of AVHRR GAC observations falling in each approximately 4 km bin.
4. Standard deviation - This is the standard deviation of the observations in each 4 km bin.
5. Overall quality flag - The overall quality flag is a relative assignment of SST quality based on a hierarchical suite of tests. The Quality Flag varies from 0 to 7, with 0 being the lowest quality and 7 the highest. For more information regarding the suite of tests, see the Kilpatrick et al. (2001) paper cited above.
6. Mask 1 - These files contain a mask code, which along with Mask 2, can be used to determine the tests in the hierarchical suite that were passed or failed, resulting in the Overall Quality Flag.
7. Mask 2 - These files contain a mask code, which along with Mask 1, can be used to determine the tests in the hierarchical suite that were passed or failed, resulting in the Overall Quality Flag.

This dataset also includes a collection of value-added, cloud-screened products in HDF4. These products contain daytime, nighttime, and day-night averaged data of quality flags 4-7 only, as well as standard deviation and pixel count information. They have been produced from the original Pathfinder Version 5 data by NODC for Daily, 5-day, 7-day, and monthly temporal averages at 4km and one degree spatial resolution, as well as five degree spatial resolution for monthly products.

This dataset also includes data that were processed using an improved reference field for use in the cloud screening process based on the Harmonic Climatology methodology; these data are to be considered experimental. The files in the directory in question (the "newref" directory) were the output of an experiment that used a new reference field based on the harmonic climatology approach to determining the "typical state" of SST. Since there were known weaknesses in the original reference fields in those early years of the Pathfinder record, we were hoping the new approach would yield better results. Quality levels in Pathfinder were determined by conducting a series of tests, like spatial homogeneity tests, comparisons to realistic bounds, and importantly a comparison against a reference field. The results of each of the tests feed into a decision tree, the output of which is the overall Pathfinder quality level. In this experiment, we changed only the reference test, which tests the SST field to see if it is within plus or minus two degrees C of the reference. If it is within those limits, the test is passed, and if it is outside the limits, the test fails. The results of these tests were promising but were never fed into the operational Pathfinder processing systems.

Note on Pathfinder Program: The Pathfinder program was jointly created by NASA and NOAA through the Earth Observing System (EOS) Program Office in 1990. The focus of the Pathfinder Program was to determine how existing satellite based datasets could be processed and used to study global change. The datasets were designed to be long time-series data processed with stable calibration and community consensus algorithms to better assist the research community. For more information, see: National Aeronautics and Space Administration (NASA). 1993. Earth Observing System (EOS) Reference Handbook, ed. G. Asrar and D. J. Dokken. Washington, D. C.: National Aeronautics and Space Administration, Earth Science Support Office, Document Resource Facility. The 4 km Pathfinder Project effort at the National Oceanic and Atmospheric Administration (NOAA) National Oceanographic Data Center (NODC) and the University of Miami's Rosenstiel School of Marine and Atmospheric Science (RSMAS) is an extension of and improvement on the sea surface temperature (SST) fields from that original NOAA/NASA AVHRR Oceans Pathfinder program.

In this accession, NCEI has archived multiple versions of these data. The latest (and best) version of these data has the largest version number.