The NCEP/NCAR index dropped again in June, from 0.471°C in May to 0.369°C in May (anomaly base 1994-2013). The drop is smaller than previous months, and may be a sign of levelling. We are now back in average to about Sept 2015, the first small rise of the El Nino.
Cold in E Europe to W Siberia, most of US except E Coast, and a band of cold from Labrador/Greenland into the N Atlantic. Warm in Arctic and Canada, and the reanalysis still has the ENSO region fairly warm. Globally, the temperature rose somewhat at end of month.
In other news, UAH V6 also dropped considerably, from 0.55°C to 0.34°C. Arctic Sea Ice recovered somewhat (relatively), and 2016 is now not quite in the lead. Big drops in the last few days, but looking back through the record, this seems to be a feature of the end of our financial year.
Twenty years of blogging in hindsight
1 hour ago
Posts
It looks to me like there is a slight backing off on the La Nina forecast.
ReplyDelete"In its latest update (9 June 2016), El Niño conditions were considered replaced by ENSO-neutral conditions as of the end of May, and a "La Niña" watch was hoisted for boreal fall and winter 2016-17 with 75% odds. While I do not believe that the odds are that high, La Niña is certainly anticipated in most forecast models. " - Wolters
And this today from BOM:
ReplyDeleteInternational models have weakened in their outlook for La Niña compared to last month. While all models still indicate more cooling in the tropical Pacific Ocean is likely, only five of eight models suggest La Niña could form in the second half of 2016. The other three models remain neutral. If La Niña does form, models do not suggest it will be as strong as the near-record event of 2010-12.
UM CCI just posted their final daily CFSV2 global temperature anomalies for June from which I calculated a June average of +0.267C, down from +0.417C in May, and sharply down from the +0.72C peak in February (referenced to 1981-2010). These estimates compare well with the WeatherBELL CFSV2 estimates of +0.268C for June, +0.413C for May, and +0.701C for February (also referenced to 1981-2010). It will be interesting to see how much the GHCN based estimates drop for June.
ReplyDeleteNOAA just posted the Multivariate ENSO Index for May-June at 1.029, down from 1.699 for April-May, and well down from the peak of 2.308 for October-November 2015.
Thanks, Bryan - that's a drop of 0.15°C, vs the 0.1°C here. The current TempLS (mesh) estimate for June is down 0.067°C from May; the grid estimate, less sensitive to Antarctic, actually rose slightly. There are 4013 stations in, which is about the threshold for posting. The ocean component was up a bit on May.
DeleteBryan, there is a bit of a calibration problem with these numbers. I don't know what the original source for the error is, but it could be CFSR itself. I also don't know whether it is to do with the transition to CFSv2 or GFS forecast for that matter. But what I do know is that there is clearly identifiable discontinuity in the timeseries provided at your website (oz4caster.wordpress.com/monthly-trends)
Deleteas well as on Ryan Maue's website (models.weatherbell.com/temperature).
Here is Ryan's plot compared with NCEP Reanalysis 1 and GFS: NCEP_Rean1_vs_GISS_vs_GFS_vs_CFSR_WB.png
There's an inconsistent "jump" in April 2010. All data beyond that point are biased cold by approximately 0.15K. Hence the anomalies are too low by 0.15K in comparison to literally all observational estimates (wrt 1981-2010 reference period). Equally, trend estimates after 2010 based on these data are biased low. I calibrated my GFS anomalies wrt GISS (www.karstenhaustein.com/climate). Between 2011-2014 the GFS bias (compared to CFSR climatology) was +0.13K. After an GFS update in Jan 2015, the bias has gone slightly down to +0.07K. With these calibration factors, I'm usually pretty close to the final GISS numbers.
KarSteN, yes I've noticed that low departure of the CFSR/CFSV2 estimates versus the GHCN estimates. Interestingly, the satellite TLT estimates show a better match with the CFSR/CFSV2 during that period and all three types seemed to come closer during the big peak around February 2016 (when reference periods are matched). According to UM CCI, the switch from CFSR to CFSV2 in their data set occurred in April 2011 if I remember correctly, so that does not appear to be a factor.
DeleteIt's like we have three types of watches and they don't show an exact match, but my guess is that the full uncertainty of the recent estimates from all three types of sources may be on the order of plus or minus 0.3C and possibly more and the differences are within that uncertainty most of the time. The trends may be a bit more accurate than the absolute anomaly estimates, so I have more confidence in the larger ups and downs. The smaller ups and downs are probably in the noise range.
Nick, I just noticed that the HadCRUT4 May global temperature anomaly estimate has been posted. I get +0.386C when I convert the reference period to 1981-2010 and this is actually lower than the +0.417C from UM CCI CFSV2 for May. It is also much lower than the May NCEI +0.445C and GISS +0.51C adjusted to the same reference period.
DeleteThe UM CCI CFSV2 shows the daily Antarctic temperature anomalies dipping quite negative in June and that should account for at least part of the global drop from May to June. I suspect the CFSV2 has better coverage at both poles than any other temperature estimate source, but I have no way to verify that other than looking at the number of temperature reports from synoptic weather stations and buoys (I've been looking daily using the OGIMET synoptic weather data plots) and assuming that these reports are used in creating the CFSV2 estimates.
I would not say that CFSR/CFSV2/GFS is gold standard for Antarctica. It runs rather cool compared to most other indices.
DeleteI screened Knmi Climate explorer for temperature anomalies in Antarctic Winter months (NH summers) in the recent decade.
Giss is corroborated by BEST which to my knowledge use almost all data from every little weather station available.
Era-interim and NCEP/NCAR has a warm bias vs Giss and BEST.
The cool bias of the GFS/CFSR family vs Giss is supported by MERRA2, but AFAIK the MERRA reanalysis doesn't use weather station data..