bdgwx

Here is the NSIDC's update regarding the sea ice data problems.

Quote

NSIDC continues to investigate errors in our sea ice processing, and we are upgrading software to address the errors. Daily Sea Ice Index/Arctic Sea Ice News and Analysis values after February 19 are erroneous. We will post new data as soon as the software upgrades are implemented.

https://nsidc.org/arcticseaicenews/2021/03/nsidc-continues-to-investigate-sea-ice-processing-errors/

2 hours ago, etudiant said:

The data on that link also stop at Feb 19th and the scan map shows that one sector was unscanned, which was the cause of the sudden downturn.

Maybe a data processing issue with the satellite or the ground station?

Oh...you're right. Between both Chrome and Edge cutting off support for ftp and having to switch to WinSCP to download the data I didn't even notice that it hadn't updated.

44 minutes ago, etudiant said:

No expert, but the data at :  https://cryospherecomputing.tk/NRT2.html  show some sort of glitch since Feb 19th. Is there a better source?

Separately, I do agree that the recent ice trends have been suggestive of an early peak, but we do need a better confirm.

I usually just go to NSIDC directly. https://nsidc.org/arcticseaicenews/charctic-interactive-sea-ice-graph/

I did see that glitch on that site earlier today. I was assuming the glitch was isolated to that site. The data on the NSIDC site looks good.

1 hour ago, cny rider said:

Can you explain how that could happen in February?
Thanks.

 

We've seen a pretty substantial decline in the NSIDC sea ice extent the last few days.

The last time the winter max occurred in February was 2015. So there is recent precedent.

We cannot eliminate the possibility that we've seen the winter max for 2021.

On 2/6/2021 at 1:52 PM, Vice-Regent said:

Disappointing news if you thought aerosols could cool the Earth significantly via geoengineering.

I don't know. That 0.03C response was for a single year. I think it is at least consistent with the hypothesis that aerosols have a big impact and that at current levels they are likely masking a lot of the GHG warming potential. Imagine if that 0.03C rate of change persisted for 10 years. That'd be a cumulative 0.3C change.

I'm not really endorsing aerosols as a means of geoengineering here though. The safest thing to do would be to minimize human influence altogether. That way we aren't trying to fight one influencing factor with yet another influencing factor.

We are starting to see studies regarding just how much the pandemic influenced the global mean temperature in 2020.

https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020GL091805

About +0.03C per the study above and is attributable to reduced aerosol optical depths.

I believe it will be in March.

https://www.emc.ncep.noaa.gov/users/meg/gfsv16/

39 minutes ago, skierinvermont said:

Haha I knew it, the posting style was so similar! What an interesting market that was. Great job with your model. I think I was the one who suggested removing UAH and RSS from the model. Might be remembering that wrong and I don't remember if that actually worked out for you. If I remember you had one model using early temp data like CFS/NCAR (and UAH/RSS) to predict the monthly result, but were you also running the GISS code on the GHCN files like 'takeyourmoney' was?

Oh yeah. I totally remember. I did remove UAH and RSS from my model...just didn't add any skill no matter what weighting I gave it. In fact once I started adding other inputs UAH and RSS became more of a liability than an asset. 

Yes. I was running the GISTEMP code on the GHCN files. The land only index from GISTEMP actually added some skill to my model with 20% weighting. Getting the ERSST data plugged in proved very difficult at least for me. I just didn't have the time to spend on doing it.

The input that mattered the most for me was Nick Stoke's TempLS dataset. I gave this input 50% weighting. When it was all said and done my model could predict the GISS update within 0.05 with 95% confidence. 

The guys posting as takeyourmoney and James Davis were clearly very smart. They had the modeling thing figured out long before I made my attempt. I wish those two would make an appearance on here. They were always respectful and their posts were packed full of relevant to the point information. They would be fun to engage with here.

9 hours ago, skierinvermont said:

Hey, I was on there too! Are you casimireffect?

That was me!

James Hansen says in his December update (just out today) that he thinks the warming may be accelerating as well.

http://www.columbia.edu/~jeh1/mailings/2021/20210114_Temperature2020.pdf

2 hours ago, chubbs said:

Per my #: 1.022 vs 1.018, a difference that could disappear in future updates. Did you collect your bet?

Yes I did. It was pretty nuanced though. The rules said 2020 had to exceed 2016 by 0.01 after rounding to 2 decimal places. The quirk was that 2016 had been getting reported as 1.01. But I (along with several other people) had figured out that the recent addition of observations into the GHCN repository was going to likely flip 2016 back to 1.02. And my model had predicted that GISS would revise Nov down to 1.11 and report 0.83 for Dec. GISS officially reported 1.11 and 0.81 respectively so I had already seen the 2020 round down to 1.02 coming as well. I exploited that situation as well. In the end I learned a lot from this exercise. First...I learned that prediction markets aren't that good. Second...I learned a lot of details about GHCN, ERSST, how the GISTEMP code works, and how to create a model for predicting GISS updates with publicly available information with up to 4 weeks lead time. It was really fun.

BTW...your comment above about 2010 being a good analog to 2020 kept me on my guard

Berkeley Earth released their 2020 annual report. 2020 was the second warmest year behind 2016.

http://berkeleyearth.org/global-temperature-report-for-2020/

2020 has eclipsed 2016 as the warmest year in the GISS record strictly speaking. But it was of the thinnest of margins and qualifies as a statistical tie.

Dr. Spencer posted documents containing the official seal of the Office of Science and Technology Policy on his blog under the direction of David Legates. These documents represent the views of a whos who list of contrarian scientists.  

http://www.drroyspencer.com/2021/01/white-house-brochures-on-climate-there-is-no-climate-crisis/

These documents are disinformation at best. They certainly don't represent "the current state-of-the-science on various topics of climate change" as Legates claim. In addition the OSTP did not grant approval to disseminate these documents. It is not even clear they were aware of their existence until now.

https://www.washingtonpost.com/weather/2021/01/11/controversial-climate-skeptics-release-papers/

Edit: Apparently the documents are also hosted by the Center for Environmental Research and Earth Sciences.

https://www.ceres-science.com/content/climate_change_flyers.html

Edit: David Legates and Ryan Maue have both been relieved of their duties in regards to the OSTP over this.

https://www.washingtonpost.com/weather/2021/01/11/controversial-climate-skeptics-release-papers/

Edit: NOAA has issued a statement regarding the matter.

https://www.noaa.gov/media-release/statement-on-climate-change-flyers-falsely-attributed-to-white-house-office-of-science

https://arstechnica.com/science/2021/01/attempt-to-red-team-climate-research-comes-to-a-pathetic-and-confused-end/

Copernicus released their final monthly report for 2020.

https://climate.copernicus.eu/surface-air-temperature-december-2020

2020 is the second warmest year in their record after 2016.

Per NSIDC the annual mean extent in 2020 was 10.160e6. This breaks the previous record of 10.163e6 set in 2016.

2 hours ago, Vice-Regent said:

That was a good read. Full text here.

There is some interesting commentary about how we've probably underestimated aerosol forcing. That's probably not a good thing since it means the warming may continue to accelerate as we clean up our aerosol emissions.

GISTEMP reported +1.13C for November easily beating out November 2015 for the #1 spot.

2 hours ago, chubbs said:

Could have gotten better than even money a couple of days ago on the site you linked above - I would have taken those odds

I got in at 35c a couple of weeks ago when my model was suggesting it should have been valued at 55c. My model is suggesting 80% odds now so there's not a lot of premium/discount to exploit anymore IMHO. I've found this to be a fun and challenging exercise. I've learned a lot in the last couple of weeks. I'm honing my statistical analysis skills (which are still subpar) and got the GISTEMP code running on my Ubuntu install. I also discovered that Nick Stokes' TempLS dataset which updates early makes a great predictor. He processes GHCN and ERSST directly.

I was hoping NASA would release the November GISTEMP update today. They didn't. My model is predicting the November value to come in at +1.10C +/- 0.11. At +1.10C December only needs to come in at +0.75C for the annual mean to hit +1.02C which would top 2016's value of +1.01C rounded the nearest hundredth. With observations + GFS forecast through December 15th my model is predicting a value of +0.99C and if this cooling rate persists through the second of half December as well that takes us down to around +0.93C for a final annual mean of +1.03C. We'll see what happens.

For a fun perspective on a crowed-sourced prediction regarding whether 2020 will beat out 2016 as the warmest year in the GISTEMP record check out this betting market.

ERA reported the warmest November on record by a wide margin.

https://climate.copernicus.eu/surface-air-temperature-november-2020

IMHO this positions GISTEMP favorably to report the warmest year in their dataset. ERA couples reasonable well with GISTEMP.

NCAR reanalysis is running warm too. In fact as of the 18th it is warmer than both Oct and Sep. It could be a photo finish on whether 2020 beats out 2016 on the GISTEMP record even despite the ongoing La Nina.

I will say that GISTEMP came in at +0.90C for October which is 0.09C cooler than September. This lowers the odds of 2020 being the warmest. It's probably below 50% now.