Friday, March 31, 2017

Moyhu WebGL interactive graphics facility, documented.

I wrote a post earlier this month updating a general facility for using WebGL for making interactive Earth plots, Google-Earth style. I have now created a web page here which I hope to maintain which documents it. The page is listed near the bottom of the list at top right. I expect to be using the facility a lot in future posts. It has new features since the last post, but since I don't think anyone else has used that yet, I'll still call the new version V2. It should be compatible with the earlier.

18 comments:

  1. Nick, since this post, I have not been able to use the Latest Ice and Temperature data link.

    ReplyDelete
    Replies
    1. Seems to be something to do with the link changing domain and the security certificate not supporting it. Try deleting the 's' from 'https://' in the address bar after going to the page and it bringing up the error. That seems to fix it so it routes to the right domain.

      Delete
    2. It's something I did. Last year Blogger changed to require https: everywhere, and I went through and fixed everything. But I noticed yesterday that I had not changed the addresses in that table, so I "fixed" that and changed them to https:. I've no idea why it provokes this response, but anyway, I'll change it back.

      Delete
    3. I think I see the reason. The old address was http://www.moy... It seems https://wwww. is very bad, and what I have to say is https://moy... But I won't change anything now.

      Totally off topic, but I noticed something funny since I've been tweeting notices about posts. Twitter can't figure out moyhu. Last tweet, they offered to translate from Romanian. I took them up on it, but they failed. I saw another tweet offering to translate from Indonesian. The origin of the name is actually Irish.

      Delete
    4. Oops wwwww probably would be bad - I meant www.

      Delete
  2. Hey Nick, sorry for this being off-topic (I didn't see a thread for off-topic discussions), but I thought you might find this interesting since you've encountered issues with WUWT moderation before:

    http://www.hi-izuru.org/wp_blog/2017/03/a-new-ebook-coming-next-month/#comment-13577

    Anthony Watts has confirmed I am being moderated, refuses to let a comment pointing out factual errors in a post on his site through and implies I've behaved unreasonably in posting an innocuous comment under a different name to test to see if I was in fact being moderated (because Watts doesn't bother to tell people he's moderating them). I seem to recall "skeptics" having a thing or two to say about censorship.

    Oh, and Watts also posted my IP address for everyone to see. Because telling people who my ISP is is apparently cool.

    ReplyDelete
    Replies
    1. Well, as they say, if you lie down with dogs, you will get up with fleas

      Delete
    2. Brandon,
      Yes, I had an eventful brush with moderation recently, but it's OK for now. The mechanics can be interesting - he seems to put names in the list of words that sink posts, so that other people who mention you go into moderation too. A few people learnt to modify the spelling of my name. Sometimes the post goes into regular moderation, other times it makes a detour via the spam bin. In my case, at least, it usually got through in the end.

      Delete
    3. Hey Nick, you might find Anthony's response to my commentary interesting:

      http://www.hi-izuru.org/wp_blog/2017/03/a-new-ebook-coming-next-month/#comment-13579

      I think his self-pity is unwarranted and more of a way to dodge dealing with criticisms (something I have said several times in the past), but that doesn't matter. What I think is interesting is his only reason for putting me in moderation is I've said mean things about him. His site routinely attacks people so it's not that I said mean things. It's that I said those mean things about him that's the problem. In other words, he's put me in moderation because I upset him.

      Mind you, he openly acknowledges he saw the comment I posted that was factual, on topic and not out of line in any way but chose not to let it through. That is, he's acknowledging he'll censor factual criticisms simply because he doesn't like who is making them. I think that's an interesting development for anyone who has had to deal with WUWT moderation before.

      I won't argue if people want to say I'm mean and nobody should like me. I just think if a site routinely complains about censorship, it would seem strange to censor comments just because you don't like the person making them.

      Anyway, sorry for the tangent. I just remembered seeing discussions about whether or not you were actually being censored so I thought it'd be worth drawing your attention to this.

      Delete
    4. Brandon,
      The bill of particulars in my case was here.

      Delete
    5. Thanks for that link Nick. I had never realized he thinks you're a paid commenter. I wonder who he thinks is signing your checks.

      Delete
  3. Also off topic. C&W's paper from few months ago touched on the the difference between SAT and SST, and a Hansen mentioned it over a decade ago. But other than that, it doesn't seem to get much press, even though it makes a significant difference for model/observation comparison. Have you given this any thought? e.g. compare co-located (island/coast) SAT observations and SST observations? At the very least, the ratio would give an upper bound to the warming over the ocean.

    Just strange there hasn't been much work to quantify this.

    ReplyDelete
    Replies
    1. cce,
      Interesting timing. My next post, in a day or so, will be an update of my old "60 stations" series, where one of the exercises is to whittle away ocean stations more rapidly than land, so when you get down to 60 there are only land stations - or else do the same without preferring land.

      But in a way, GISS treats this on an ongoing basis with their Ts (met stations only) and Ts+SST (land/SST) series. Ts is different, but not that much. They are both in the interactive plotter.

      Another place to look is the globe plot pages, eg trends here. You can see the extent to which islands stand out relatively to the fairly smooth SST.

      But yes, it does need more quantification.

      Delete
    2. Thanks. Guess I shouldn't say "C&W" as there were many other authors.
      http://onlinelibrary.wiley.com/doi/10.1002/2015GL064888/abstract

      Hansen mentions it in this paper from 2006.
      https://pubs.giss.nasa.gov/docs/2006/2006_Hansen_ha07110b.pdf

      There are no doubt others.

      Delete
    3. The true global SAT is likely between that of Gistemp dTs and loti. An average of them should be fine, I believe..

      Nick, if you run your "60 stations", make also a 60 station apples-to-apples model comparison, by subsampling model data from the same 60 location, and run the global averaging code with them. By doing this you can also check how well the 60 points sampling and averaging represent the true global value (the globally complete model data)

      I have done this with RATPAC data. Originally there were 85 stations but today it is only 68 left according to the IGRA list. In the following comparison I have included the station dropout, and luckily the dataset seems very robust (yet), and changes of global average are very small due do station loss

      https://drive.google.com/file/d/0B_dL1shkWewaX3dRa0VDeU5fRnM/view?usp=sharing

      The Ratpac sampling and averaging method represents complete global data very well according to the subsampled model comparison. The trend 1970-2016 differ by about 0.01 C/decade (0.22 vs 0.21)

      The Ratpac surface data is not much noisier than the more spatially complete Gistemp dTs or NCEP/NCAR, and all observational datasets agree well with models.

      The simple Ratpac method seems to work well. Divide the world in 18 (30x120 lat/lon)regions, populate them as evenly as possible with stations, and calculate area-weighted averages stepwise through regions, zones, hemisphere to global.

      I have another example here were I have picked only 18 points of UAH data(0.17% of the spatial information) and compared with the original complete dataset. Quite enough to catch the global warming signal, and almost indistinguishable from the original dataset:

      https://drive.google.com/open?id=0B_dL1shkWewadExwM3VOaVJBUU0

      Disclaimer: 2016 is year to October only (when I did the chart) and I believe that the final annual averages became slightly lower..

      https://drive.google.com/open?id=0B_dL1shkWewadExwM3VOaVJBUU0

      Delete
  4. This recent chart from Olaf Reimer's twitter shows that CMIP5 does very well over land, both raw and adjusted. It maybe worthwhile to separate land and ocean when evaluating model performance.

    https://twitter.com/OleBoule/status/844523064884969473/photo/1

    Chubbs

    ReplyDelete
    Replies
    1. Yes, SST prior to 1950 is the weak link of global datasets. There is a lot of uncertainty in kinds of bucket/ship ERI, altered war time behaviour, etc.
      Sceptics commonly note that the fast temperature increase from 1910 to WW2 is poorly explained by models (which makes all models bad with their logic). However, it could as well be the SST observations that are bad.
      If we sea mask Gistemp dts (only use island an coastal met stations) the dubious SST features don't appear, or become less prominent. Thus, I believe that a SST dataset adjusted by co-located met stations would lead to an improvement, ie a better fit with models

      Delete
  5. We have a lot of different data sources that don't necessarily measure the same but are highly correlated. SAT vs NMAT vs ship/buoy SST vs AVHRR SST vs radiosonde vs MSU. That is a lot of relationships to compare, and a lot of ways to tease out problems that appear in one dataset at a given time/location but not others. Subsampling them all, including model output, is critical for any apples to apples evaluation.

    ReplyDelete