|
History of changes in the RTG_SST_HR analysis
|
Description of changes effective November 1, 2016
The analysis now uses satellite SST retrieval data from the GOES satellites. In addition, the values of "SST" over land areas are now generated by an SOR Laplacian solver, which replaces the legacy "weaver" code.
Description of changes effective August 26, 2016
Handling of satellite SST retrieval data in the analysis was modified so the analysis will proceed in the absence of satellite data.
Description of changes effective July 15, 2015
Hot spots in the Sea of Azov and Caspian Sea were replaced by more temperate values of SST. Cool Congolese waters were replaced using a climatology from Robert Grumbine.
Description of changes effective June 2, 2014
Anomaly plots were changed from using a one-degree contour interval to employing contours based upon the frequency distribution of the SST anomaly.
Description of changes effective March 17, 2014
The web page displaying the RTG_SST_HR analysis was revamped. Previously, the page offered a choice of portions of the globe. With this revision, a full global plot of the SST analysis and the SST anomaly is displayed on the web page for the RTG_SST_HR analysis. Further, clicking on the global plots leads to a display of various portions of the globe. This includes polar-stereographic plots of the polar regions.
Details of the RTG_SST_HR analysis were previously available on the main web page. Now, they are available by clicking on the links near the bottom of the main web page.
Description of changes effective October 2, 2012
A "hot spot" had developed in the Beaufort Sea near the Alaska-Canada border in late June 2012. It intensified until early August, at which point, very slow cooling began. The feature was still quite intense though the month of September.
On October 1, after the analysis was performed, the SST field in the vicinity of the "hot spot" was replaced by a value of 4.0 deg. C wherever it exceeded that value. This became the first guess for all subsequent Physical SST satellite retrievals for October 2, as well as the RTG_SST_HR analysis performed on October 2.
Description of changes effective October 12, 2011
A number of changes were introduced into the RTG_SST_HR analysis:
(1) NOAA-18 physical retrievals were replaced by NOAA-19 physical retrievals.
(2) Procedures were modified to allow the use of AMSR-E retrievals in the analysis. Unfortunately, AMSR-E failed on October 4 (a week before the implementation), so the changes will not have any effect.
(3) Code RTGSSTHR_SUPERINS was corrected to allow all ice superobs to be used in the bias correction of the satellite SST retrievals.
Description of changes effective March 3, 2010
An error in the February 17, 2010 implementation was corrected. While the changes introduced on that day were effective in removing hot and cold spots in the RTG_SST_HR analysis, another error was introduced which caused the production of meaningless bias corrections to the satellite retrieval data. The errors introduced were as much as 2.0 deg. C in some areas. The statistics of the RTG_SST_HR analysis' fit to buoy observations were quite noticeably degraded. With this correction, the February 17, 2010 implementation is fully effective.
Description of changes effective February 17, 2010
Several changes were introduced into the RTG_SST_HR analysis:
(1) The physical retrieval procedure no longer produces SST retrievals over land. The handling of time-stamp and latitude-longitude were improved to facilitate the matching of Navy SST retrievals with physical SST retrievals in the newly introduced retrieval merging procedure (see point (2)).
(2) The remaining physical retrievals are being merged with Navy retrievals over inland waters and near coastlines. Over waters within 35 km of land, Navy retrievals are being used instead of physical retrievals. Over waters farther than 200 km from land, physical retrievals are still being used. Over waters in a transitional zone between 35 km and 200 km from land, the Navy retrievals are blended with the physical retrievals.
(3) A new land-sea mask is being introduced into the RTG_SST_HR analysis. It has more water points (and fewer land points) than the previously
used land-sea mask.
(4) The RTG_SST_HR analysis is no longer being smoothed as the last step
in the analysis program.
(5) Small errors in calculating the index to the land-sea mask are being corrected in the satellite retrieval and in-situ data handling codes. In addition, an error in the satellite retrieval code which allowed the re-use of some data from the preceding day has been corrected.
Description of changes effective June 25, 2009
NOAA-17 physical retrievals were replaced by METOP-A physical retrievals.
Description of changes effective June 16, 2009
The archive of the RTG_SST_HR analysis on the POLAR server was extended from 366 days from the present
to the first RTG_SST_HR analysis, on Sept. 27, 2005.
Missing dates are: October 1, 2005;
October 18, 2006; January 19, 2007; January 25, 2007; and Febrary 7, 2007
Description of changes effective March 2, 2009
The RTG_SST_HR analysis web page now includes SST and SST anomaly plots for the northwestern Atlantic Ocean and the northeastern Pacific Ocean.
Description of changes effective February 17, 2009
The RTG_SST_HR analysis state was replaced by interpolation from a modified 1/2 degree resolution RTG_SST analysis. This was done to remove hot spots from the RTG_SST_HR analysis. These were most prominent in areas around Japan, although hot spots existed in other areas as well, usually adjacent to land areas. Modifications were made so as to maintain the proper generation of ice in the sea-ice analysis.
Description of changes effective September 23, 2008
The filling of latitudes 88 N - 90 N with -1.8 deg C. was removed from the RTG_SST_HR analysis.
The satellite bias correction in the RTG_SST_HR analysis now uses ice data thinned from 1/12 degree resolution to 1/2 degree resolution. This will reduce the occurrence of too-cold satellite bias corrections caused by over-emphasis of the ice data.
Description of changes effective June 10, 2008
The physical (JCSDA) SST retrieval was modified to use to most recent version of the Community Radiative Transfer Model.
Description of changes effective February 15, 2008
The RTG_SST_HR analysis web page now includes SST and SST anomaly plots for the Arctic Ocean area (Polar Stereographic) and the Indian Ocean area (Cylindrical Equidistant).
Description of changes effective November 14, 2007
The archive of the RTG_SST_HR analysis on the POLAR server was extended from 60 days back to
366 days back from the present.
Description of changes effective April 7, 2007
The RTG_SST_HR analysis state was replaced by the RTG_SST analysis at 1/2 deg.
interpolated to 1/12 deg. in the Northern Hemisphere north of the continents.
The sea ice SST observations, which had been at 1/12 deg. resolution, are
now being superobbed to 1/2 deg. resolution. The observation errors for the
sea ice were increased to those for surface ship observations.
The QC criteria in the physical (JCSDA) SST retrieval were changed to allow
the generation of warmer SST retrievals.
Description of changes effective September 7, 2006
The RTG_SST_HR analysis state was replaced by the RTG_SST analysis at 1/2 deg.
interpolated to 1/12 deg. in the Northern Hemisphere north of the continents.
Cold conditions had persisted in Arctic regions of the RTG_SST_HR analysis
since the melting of the ice in May. The palliative step was taken while
investigation continues.
Description of changes effective March 8, 2006
NOAA-16 physical retrievals were replaced by NOAA-18 physical retrievals.
The
land, ice, and coastal checks were removed from the physical retrieval algorithm.
The lower limit of the correlation length scale was reduced from 100 km to
50 km.
The land/sea mask was modified to removed small, isolated lakes.
Description of changes effective January 24, 2006
Input brightness temperatures will no longer be eliminated from elevations greater than 100 meters. The condition is being eliminated to allow physical retrievals to be performed over inland lakes.
Description of changes effective September 27, 2005
The 1/12 deg. Real-Time Global SST (RTG_SST_HR) analysis was introduced.
This new analysis uses physical SST retrieval data from both NOAA-16 and NOAA-17
satellites. It also uses sea ice at 1/12 deg. resolution.
Description of changes effective May 6, 2004
The source of NAVY SST retrieval data was changed from NOAA-16 to NOAA-17.
Description of changes effective May 8, 2002
Array sizes for one week of in situ data were doubled from 90,000 to 180,000
in the program which performs quality control of in situ data.
The program in question was generating large numbers of messages indicating
that buoy data were being dropped. This was due to insufficient array sizes.
The effect was that up to 25% of buoy data were being dropped from the the most
recent day's worth of data (i.e., the real-time buoy data). The buoys dropped
were in the northern areas (off of Western Europe, the U. S. and Canadian East
Coast, the Gulf of Alaska and the West Coast off the U. S. and Canada).
The northernmost buoys started being lost around April 22.
Description of changes effective April 23, 2002
Instituted a smoother to eliminate 2-delta grid length waves, while leaving
larger scale phenomena unchanged.
The reason for making this change is that small-scale noise in the RTG_SST analysis
was growing noticeably with time. There was very little noise as of Febrary
2001, but it was increasing throughout the past year. The growth of the
noise has been attributed to the growth of round-off error in the RTG_SST analysis.
Description of changes effective October 23, 2001
Change in the operational run-time from 2045 UTC to 2240 UTC; and change the
window for admitting data to the analysis system, in order to ingest all SST
data received up to analysis run-time.
The reasons for making these changes are: to synchronize the daily RTG_SST
run-time with the 00 UTC EDAS and produce more timely SST fields for use by the
00 UTC Meso Eta model run. The previous operational run-time was 3 hours, 15
minutes before the 00 UTC EDAS, and most recent data in the previous operational
RTG_SST was already 20 hours old.
Description of changes effective June 5, 2001
Switch from NOAA-14 MUT SST retrievals to NOAA-16 SEATEMP SST retrievals.
The reasons for making this change are: (1) switch to a high-resolution,
higher quality SST retrieval from the Navy; (2) switch from the older
satellite (NOAA-14) to the newer satellite (NOAA-16).
|
|