The_Global_Warmer Posted July 20, 2014 Share Posted July 20, 2014 Global ssta drop hard in 2010 as well after the NINO. there is no lag for surface data with ssta dropping hard and the bias that came in from updates like in the GFS 2m high temp bias this could explain the drop pretty well. Link to comment Share on other sites More sharing options...
chubbs Posted July 20, 2014 Share Posted July 20, 2014 Global ssta drop hard in 2010 as well after the NINO. there is no lag for surface data with ssta dropping hard and the bias that came in from updates like in the GFS 2m high temp bias this could explain the drop pretty well. Most of 2010/11 looks reasonable with a small decrease in March and a larger decrease in the fall into 2011. The sharp drop in April however is not in other datasets. Link to comment Share on other sites More sharing options...
dtk Posted July 21, 2014 Share Posted July 21, 2014 The Weatherbell CFS series is not a "robust" long-term data set because of a spurious drop in April/May 2010 which is not matched by observations. This causes the post 2010 Wxbell values to be biased low by 0.2 to 0.3. http://www.karstenhaustein.com/reanalysis/info.php http://models.weatherbell.com/temperature.php I cannot explain the drop in 2010. However, January 2011 serves as a breaking point in the CFSR/CDASv2 as the transition from a quasi-frozen reanalysis to a real-time initial condition generation mechanism. As part of the CFS2 implementation, the resolution of the analysis was increase from T382 (this is what was used from 1979-2010) to T574 (from Jan 2011 to present). I believe that there were also some model changes to make the model that drives the CDAS more like the GFS. The original intent was to eventually merge the GDAS/CDAS into a single, unified system....i.e., the GFS would be in part driven by the CDAS system. This may happen some day but it has yet to pan out. This change to the model that is used in the DA cycling has likely led to a change in certain biases (especially if they went from enthalpy to virtual temperature for the thermodynamic variable, relaxed the physics toward the GFS configuration, etc.). This is manifesting itself by what appears to be a slightly different analyzed anomalies. It should be pointed out that the actual CFS prediction model (the one that was used in the reforecast) has always been frozen and will remain frozen. Only the model that is used in the data assimilation cycling as changed. As an aside, there is no team at EMC properly updating and maintaining the CDASv2 with the evolving observing system, and this may be leading to some further degradation. In summary, the 2010 drop may be some real but needs further evaluation. Everything after Jan 1 2011 is suspect because of the change in the model used to drive the CDASv2 (as well as issues related to the deterioration due to the lack of upgrades to handle the changing observing system since that time). Link to comment Share on other sites More sharing options...
Snow_Miser Posted July 22, 2014 Share Posted July 22, 2014 The global temperature anomaly from the CFS shot up from +0.06 to +0.36 in a matter of only four days for the month so far. I'm willing to bet that it's a spurious change. Link to comment Share on other sites More sharing options...
ORH_wxman Posted July 22, 2014 Share Posted July 22, 2014 The global temperature anomaly from the CFS shot up from +0.06 to +0.36 in a matter of only four days for the month so far. I'm willing to bet that it's a spurious change. I'm either missing something or you are looking at the wrong numbers. I still see +0.052C for the month of July on the CFS/weatherbell maps...with the latest daily also hovering around +0.05 Link to comment Share on other sites More sharing options...
Snow_Miser Posted July 22, 2014 Share Posted July 22, 2014 I'm either missing something or you are looking at the wrong numbers. I still see +0.052C for the month of July on the CFS/weatherbell maps...with the latest daily also hovering around +0.05 Through the 22nd: Link to comment Share on other sites More sharing options...
Snow_Miser Posted July 22, 2014 Share Posted July 22, 2014 Looks like the update at 0z 7/22 is no longer available on Weatherbell. Link to comment Share on other sites More sharing options...
The_Global_Warmer Posted July 22, 2014 Share Posted July 22, 2014 Ignore the ouput below in terms of raw data because of climo. But the AAO has tanked and Antarctica has dramatically warmed up. Combine that with near record or record ssta attm pending which data set you use + the general warmth over the continents a warm up would be par for the course the last 3-4 days. On my screen weatherbell hasn't updated since 12z on the 18th. Climate re-analyzer does not support the big warm up in Antarctica. At least on it's baseline. Link to comment Share on other sites More sharing options...
blizzard1024 Posted July 22, 2014 Author Share Posted July 22, 2014 Ignore the ouput below in terms of raw data because of climo. But the AAO has tanked and Antarctica has dramatically warmed up. Combine that with near record or record ssta attm pending which data set you use + the general warmth over the continents a warm up would be par for the course the last 3-4 days. On my screen weatherbell hasn't updated since 12z on the 18th. Climate re-analyzer does not support the big warm up in Antarctica. At least on it's baseline. The operational data I know from this site is not good for analyzing short term climate. They even say it on the PSD site that the operational data is not as "stable" as the reanalysis.... this is a quote from their website "Surface Temperatures: (see reanalysis, more stable climo available)" That's why I like the CFS because it has an excellent measure of short term climate. The anomalies etc could be in question as others have brought out on this which I think was needed since everyone typically discounts it because it is colder. If a dataset shows colder prove why its wrong. That's real science. thanks. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.