If there are any cases that you are interested in seeing please let us know. Post-processing using the latest algorithm is being done as time permits.
June 28, 2017 CIRA ATMS estimates added as a SATCON member. Estimates are only used when the TC eye diameter is > 40 km.
March 27, 2017 Complete failure of F-19
September 21, 2016 Made a change to ARCHER inputs to SATCON sounder members (AMSU, ATMS and SSMIS). Previously the ARCHER intensity score and eye size closest to the time of the satellite pass was used. However this
occasionally resulted in weakening that was too rapid when the TC eyewall weakened quickly, typically due to wind shear. Also during eyewall replacment cycles the ARCHER algorithm may abruptly shift to the outer convective ring while a remnant inner eye may remain. Because there is lag in pressure and Vmax response now the strongest ARCHER score and smallest TC eye size
within 3 hours is used.
June 20, 2016 CIMSS ATMS algorithm added to SATCON as a member. ATMS is treated as an SSMIS member but gets unique weights. This allows for improved interpolation of SSMIS with two "SSMIS" members instead of one.
March 15, 2016 CIRA ATMS estimates added to SATCON plots.
February 12, 2016 SSMIS F-19 has failed. No longer transmitting however direct broadcast data is still available. The only currently working SSMIS is now F-17.
September 14, 2015 Both SSMIS F-16 and F-18 have failed.
December 29, 2014 Added a basic SATCON history file to the storm pages. File is satcon_hist_YYYYNMB.txt where YYYY is storm year, NM is storm number and B is the basin ID. Format is:
YYYY NM ORD_DAT YYYYMMMDD TTTT00 LAT LON MEM MLP VMX (ORD_DAT is the decimal ordinal date, YYYYMMDD is year month and day, TTTT00 is time in hour minutes and seconds, LAT is latitude, LON is longitide, MEM is number of SATCON Members, MLP is the MSLP and VMX is Vmax (knots)
2014 23 362.312 2014DEC28 073000 7.95 -127.85 3 998 37
2014 23 362.333 2014DEC28 080000 7.96 -127.80 3 998 37
2014 23 362.354 2014DEC28 083000 7.98 -127.75 3 998 37
2014 23 362.375 2014DEC28 090000 8.00 -127.70 3 998 38
2014 23 362.396 2014DEC28 093000 8.02 -127.65 3 997 39
2014 23 362.417 2014DEC28 100000 8.04 -127.60 3 998 38
2014 23 362.458 2014DEC28 110000 8.09 -127.49 3 996 42
October 14, 2014 CIMSS ATMS estimates added to SATCON plots. ATMS is not used in the weighting for plots that note this. These will have a note that states "Experimental ATMS esimtates plotted for reference only".
Jul 21, 2014
1) Addition of SSMIS intensity estimates to SATCON. Previously we had simply been including the SSMIS estimates in the plots.
2) Interpolation of SSMIS and AMSU estimates. This addresses a couple of issues. By interpolating between the estimates we are able to match an interpolated SSMIS and AMSU estimate to every available ADT estimate. This allows for a 3-member consensus at least in the post-processing even if all three members may not be available for the current estimate. This results in a smoother estimate trace with less variance between estimates. It also allows for the ADT to influence SATCON even if the sounder estimates may be several hours old due to a gap in pass coverage.
3) Inclusion of a pressure-wind SATCON estimate. This estimate is produced using the statistically accurate SATCON MSLP estimate from all available members. The estimate is adjusted using latitude, storm motion, radius to outer closed isobar (ROCI) and eye size (from ARCHER if it is available or IR if it is not). That P-W estimate is combined with the SATCON member Vmax consensus using a 25% weight. Thus the final SATCON Vmax estimate is 0.75*Consensus_Vmax + 0.25*PW_Vmax
4) Even after all this there is still a bit of noise in the SATCON plots so a 5 point smoother is applied to remove the low amplitude (usually a 1-2 knots) noise.
5) Inclusion of error bounds for SATCON showing ± 2 standard deviations.
6) Bias correction to adjust SATCON during the first 36 hours of TC age
As an example here is the previous version of SATCON
And the new version
We have resumed sending SATCON emails. However note that SATCON is more dynamic now. In order to not spam email boxes with multiple estimates an estimate is sent as soon as a microwave pass is processed. However if new information becomes available (for example a microwave imager pass with a new eye size) SATCON will update on the web page. A new email will only be sent for the same estimate time stamp if a signfificant change has occurred. Also when a new pass comes in the previous SATCON estimate will be updated with that information. For example an AMSU/ADT match at 1100Z would be updated when SSMIS comes in at 1200Z or even 1400Z. Or if an ARCHER TMI processing indicates that the eye size has changed or the eyewall has become more robust or weakened that information is also incorporated. In addition Best Track information from JTWC/NHC is input to the Pressure-Wind component (ROCI, storm motion). So SATCON is continually updating. Thus an estimate sent an hour ago may be different now as new information used for the estimate so users are encouraged to check the web page for the latest estimates.
1) The loss of N16 means that AMSU estimates may have large gaps and thus there could be many hours between estimates. Right now estimate age is not included in the weighting scheme. We are working on a time-weighted approach that reduces the estimate weight for a member that is getting "stale".
2) ATMS from NPP will be included in the plots when available. That processing is not currently automated but will be we hope some time this season. In the interim we will be post-processing ATMS as we have time and those estimates will show up in post-processing of the SATCON plots.
3) We hope to get METOP-B added and that will help fill the AMSU gaps. Same with F-19 as soon as that data is released.
5) We will be working to post-process cases from previous seasons using the new approach.
6) Additonal diagnostics and history file information.
7) CIRA AMSU is not currently being used in the SATCON estimates but IS
plotted for reference. We are working to develop a new weighting scheme
to add CIRA back into the estimates.
As always comments are welcome.