Latest Ice and Temperature data

Latest Sea Ice and Global Temperature contents:


Introduction
Monthly Global temperature indices
Table: Latest Global Anomaly data
Report: Latest Monthly TempLS Report (mesh weighted)
Graphs: Global Temperature indices - graphs of recent months
Active Graph: Global Temperature indices - Also backtrends. Many user choices.
Table: Months ordered in descending surface temperature anomaly
Table: Graphics about highest month and year temperatures
Daily Reanalysis Temperatures
Table: Recent days NCEP/NCAR reanalysis surface global average anomaly temperature
Active WebGL map: Recent days NCEP/NCAR reanalysis surface global average anomaly temperature
Active map and plot: Recent days NCEP/NCAR reanalysis regional Arctic temperatures
NH and SH Sea Ice data
Table: Latest Sea Ice Extent - JAXA, NSIDC NH, NSIDC SH
Active graph: Latest Sea Ice Extent - JAXA, NSIDC NH, NSIDC SH
Miscellaneous
Table: Log of recently downloaded data files

This automatically updated page has an embedded html window in which the numbers will appear. I check for data hourly. There is a log (bottom) which shows when data arrives, and also gives latest links.

Temperatures anomalies in the tables are as stated by the providers, with different anomaly bases. They have been converted to the same base (1981-2010) for plotting. The active plot has this common anomaly base

Sources and glossary:


Monthly global temperature anomaly data

Here is the latest global anomaly temperature data, as supplied by the sources, with their varying base periods. An archive of this data, by month back to Dec 2016, is indexed here.


Here is the latest report from TempLS FEM weighted version. A reference on the Moyhu global temperature index (least squares) is here, with links to earlier posts. There is a Moyhu post each month on about 7th based on this report.


Here are some plots of temperature indices in recent times, set to a common anomaly base (1981-2010). Use buttons below the image to cycle through longer periods.



Here is an active plotter for global temperature indices, starting with to a common anomaly base of 1981-2010. Details of its operations are in a post here. You can vary the selection of plot data with the checkboxes. The legend bottom left is moveable, and intended to be placed on the plot for screen capture and image editing. Dragging in the plot area just translates the plots, but dragging below the x-axis stretches x, and similarly for y.

The plot also allows you to switch (toggle button "Trendback") to a mode which shows on the y-axis not the anomaly, but the trend from the x-axis time to the most recent data. In this mode you can't regress or smooth, and anomaly has no effect.




Table of months in descending order of temperature

Next is a table of monthly anomalies from various sources, listed (with annual) in descending order of temperature. The warmest 10 are shown. Each temperature is preceded by its year (last two digits).









Graphics showing warmest recent months and years

In this set of active plots, you can see up to date versions of plots of average global T anomaly of months and years in order of warmth. The first is a stacked plot of months, with whole year at the end. This gives a bar for ech month, colored according to year, but stacked by month, in some order. The first lays down warmest first, then next and so on. Each month is visible (unless there is a tie). The second lays them in time order, most recent first. Then the bars you see are only those which were a record in their time, and shows how the rcords built up.

The third is a progress of annual records. It is really just a bar plot of annual averages, but coored to mark which periods followed a record year before the record was broken.

The arrow buttons below each plot allow you to cycle through a number of data providers, including HADCRUT, GISS, NOAA, UAH,RSS and TempLS.








Daily Reanalysis Temperature data

Here is the latest NCEP/NCAR reanalysis (see post) surface temperature (°C) anomaly area-weighted average for recent months, and for days of the current month. The anomaly base years 1994-2013. You can find here an updated zip file of the daily and monthly values since 1994, with a readme file.

Here is a WebGL active map of this year's daily temperature anomalies. It has the usual trackball facilities. To get a specific day, let the mouse hover over the colored squares on the right; each square is a day, and the text top right tells you where you are. Click when you have the day that you want. The black squares at the top will give you a month average map. Details are at this post.
Update I have added buttons above the date selector which allow flipping through consecutive days to make a kind of movie.




Here is a tool for looking at daily Arctic temperatures. It is described in a post here. There is a lat/lon rectangle that you can manoeuver, and then press the "plot new" button to see a plot for that region of daily average temperatures (not anomalies), along with the 1994-2013 average. The domain is initially set to the region N of 80°.


NH and SH Sea Ice data


Here is the latest Arctic sea ice extent data. You can choose from JAXA Arctic Ice Extent, or from NSIDC NH or SH:

Here is an active polar plot of the Jaxa Ice extent data. Units are million sq km. You can click on the current region circle to get a magnification, or on any sector to see that part expanded. Clicking on the legend marks the year in dark; clicking on the central picture makes it go away. Details here.

Miscellaneous


And here is a log of the most recent data files to be posted. The links show the URL that was actually downloaded. The date on the left is the time marked on the file at origin, and the delay is the difference between that and the time of download and processing here. Sometimes that includes a delay between the date marked and when it became visible. NCEP/NCAR, for example, seems to have a six hour delay between the date marked and when it becomes available for download.


97 comments:

  1. Hi Nick,
    The NSIDC graphs for 2015 (both NH and SH) are missing, and the tabular values for 2015 are all "NA".
    Is this a result of the satellite problems that NSIDC had earlier this year, or..?

    ReplyDelete
    Replies
    1. Olof,
      No, it's a fault at my end. NSIDC produces annual files, so I keep an amalgamated copy of the old ones and merge that with the new one, now 2016. Somehow 2015 is falling through the cracks. I'll fix it.

      Delete
  2. Regarding the NCEP values, you state: "Here is the latest NCEP/NCAR reanalysis (see post) surface temperature (°C) anomaly area-weighted average for recent months, and for days of the current month. The anomaly base years 1994-2013."
    But is this really correct? On Climate Reanalyzer homepage it is stated "Temperature refers to air temperature at 2 meters above the surface. The temperature anomaly is made in reference to a 1979-2000 climatology derived from the reanalysis of the NCEP Climate Forecast System (CFSR/CFSV2) model. This climate baseline is used instead of the 1981-2010 climate normal because it spans a period prior to significant warming of the Arctic beyond historically-observed values. For context, see this timeseries plot showing how various climate baselines compare against the NASA GISS 1880-2014 global land-ocean temperature index." Or are these completely different data?
    As far as I can see, your values are not systematically lower than the data given on CR, in contrast to this page, where a 1981-2010 reference period is used: https://oz4caster.wordpress.com/cfsr/
    Best regards, Ole Klemsdal, Oslo, Norway

    ReplyDelete
    Replies
    1. "But is this really correct?"
      The data supplied is as simple temperature. Users like CCI (and Moyhu) have to make their own anomalies. I make an average of 1994-2013 from that data and subtract it. Climare Reanalyzer uses different years. My reason for choosing that unusual period is that it corresponds to be period of best data in NCEP/NCAR. They started assimilating data in real time in about 1994; before that they get what they can, but there are gaps and an obvious decline in quality.

      Delete
    2. OK, thanks a lot for rapid response and explanation!. I still don't understand how you when using a 1994-2013 reference can end up with figures that are very similar to anomalies based on the 1979-2000 reference period? These anomalies should normally differ by some 0,12 C, according to the figure that is referred to on Climate reanalyser home page http://cci-reanalyzer.org/DailySummary/GISS_land+ocean_1880-2014.png

      Delete
  3. I notice that the NCEP/NCAR numbers haven't updated for a few days. Here's what I have at home for the last while, on my own version. Note that it goes in ascending date order.

    2016/12/24 0.390
    2016/12/25 0.365
    2016/12/26 0.339
    2016/12/27 0.444
    2016/12/28 0.552
    2016/12/29 0.565
    2016/12/30 0.613
    2016/12/31 0.617

    For the month +0.392

    Walter Dnes

    ReplyDelete
    Replies
    1. A possible cause. Last time you had a problem, an unexpected change in the UAH download propagated to the NCEP/NCAR data script. I noticed today that trying to download RSS data now gets a dialogue asking for a login and password. Could that be throwing your scripts for a loop?

      Walter Dnes

      Delete
    2. The new working RSS URL is...

      http://data.remss.com/msu/monthly_time_series/RSS_Monthly_MSU_AMSU_Channel_TLT_Anomalies_Land_and_Ocean_v03_3.txt

      Note the uppercase/lowercase. I wonder if they've switched to a unix/linux/bsd server where case is important.

      Walter Dnes

      Delete
    3. Walter,
      Thanks, I've fixed it now. The problem was on coming to the end of a leap year, when some vectors have to have length 366. I was setting the dimension according to current year (2017).

      Delete
  4. No updates for January? My data shows...

    2017/01/01 +0.675
    2017/01/02 +0.674
    2017/01/03 +0.622

    Average for January 2017 so far +0.657

    Walter Dnes

    ReplyDelete
    Replies
    1. Thanks, Walter
      Yes, I'd left in a part of the end leap year fix that should have come out. Fixed now, I think, and in perfect agreement with yours.

      Delete
    2. Thanks Walter. The anomalies you refer to- are those based on a 1979-2000 reference period?
      best regards OleKlemsdal, Oslo.

      Delete
    3. The anomaly base period is the same 20-year period, 1994-2013, that Nick Stokes uses.

      Walter Dnes

      Delete
    4. OK, Thanks. OleKlemsdal

      Delete
  5. Nick, would you consider adding the RSS TTT (V4.0) dataset to your active plotter? I'm getting tired of pointing out to people that the current RSS TLT dataset is (by admission of the authors/maintainers of it) erroneous due to drift issues, but since RSS have still not updated it, the TTT is probably the best data they provide. Its also interesting to note that in revising UAH to version 6.0, one of the changes was to alter the atmospheric weighting profile so that is is now quite close to RSS TTT.

    ReplyDelete
    Replies
    1. Dave,
      Yes, I'll do that. I expect that data for V3.3 will soon dry up anyway. RSS have just announced the requirement to register before accessing data. I'm not sure what the effect o that will be.

      Delete
  6. Where are the latest NCEP/NCAR numbers?

    ReplyDelete
    Replies
    1. Too big to fit on the information superhighway.

      Delete
    2. Well, it is still warm. Fixed, thanks for warning.

      Delete
    3. Thank you, Nick.

      (Yes, it is still warm!)

      Delete
  7. Doesn't seem to be updating daily Nick?

    ReplyDelete
    Replies
    1. NCEP/NCAR I mean :-)

      Delete
    2. Dave,
      Not my program, it seems, this time. The NCEP/NCAR directory is here. File is air.sig995.2017.nc, last updated there 20 Feb. It happens sometimes, usually only a few days.

      Delete
    3. Nick, you've created a bunch of addicts.

      Delete
    4. Totally addicted to real facts. Speaking of which, Roy Spencer has a link on his blog to their new paper on UAH 6: Published in a fairly new Korean journal... I'm a little surprised to see no citation within it to RSS 4.0. I find that a little odd, considering both purport to address similar issues with their respective deprecated data sets....

      Delete
  8. Hi Nick,
    I believe you have the wrong source for RSS TTT v4. The values are too high, and I suspect they are land only.
    I have not (yet) got a personal login to RSS ftp, but the open data via the time series trend browse tool are different from the ones you post
    http://data.remss.com/msu/graphics/TTT_v40/time_series/RSS_TS_channel_TTT_Global_Land_And_Sea_v04_0.txt

    ReplyDelete
    Replies
    1. Olof,
      Thanks. You are right - I was using land only. Fixed.

      Delete
    2. Nick,
      Still wrong source file, I think it is TTT v3.3 land/ocean now.
      Third time's a charm.... :-)

      Delete
    3. Third time? I hope so. There can't be too many more files to get wrong. I hope it's fixed this time.

      Delete
  9. Nick:

    It looks like May 1st data is missing in the NCEP/NCAR reanalysis netcdf file. My script does a dump to an intermediate text file, for further processing. The first and last entries for May 1st are...

    time[120]=9.96920996839e+36 lat[0]=90 lon[0]=0 air[1261440]=9.96921e+36

    time[120]=9.96920996839e+36 lat[72]=-90 lon[143]=357.5 air[1271951]=9.96921e+36

    May 1st is Julian day 121, but netcdf is zero-based, i.e. January 1st is day 0, not day 1. I also get +0.271 for May 2nd, so that appears to work. Today I deleted all lines with "time[120]=9.96920996839e+36" with a text editor. I'll have to add an additional check to automatically throw out lines containing the string "=9.9692".

    Hopefully, this was a temporary glitch, and the missing data will be inserted soon.

    Walter Dnes

    ReplyDelete
    Replies
    1. Thanks, Walter,
      Yesterday, it didn't have a May 1 number at all. I'll put in a filter for numbers out of range.

      Delete
  10. Hi there,

    Is it possible to get temperature anomalies on countries level?

    ReplyDelete
    Replies
    1. I take it you mean country average anomalies. I used to try to do that, eg here. The problem is that you really just have some number of stations in and near, and it is hard to add up their effect observing the boundaries. It's easy enough just to average the stations, and for small countries, it isn't even that important to have area weighting. Now I prefer to stick to near-continent scale, as with the breakdown in the TempLS report above.

      Delete
  11. I have received your email regarding the GISS anomaly for September. I haven't payed attention to these maps in recent times, and note that the 'cold blob' that was near Greenland in 2015 seems not to be there any more. Could you tell me when this vanished, and whether it was anomalous or a regularly returning feature.

    ReplyDelete
    Replies
    1. Do you mean the region south of Greenland in the Atlantic? Yes, there was a cool spot for quite a while, not currently appearing. I don't think it was an artefact. There is a movie here which shows SST in the Arctic (but has N Atlantic too) which shows how it comes and goes.

      Delete
  12. "...
    - Click on the legend; the history of that year will show in a dark color
    - Click on the picture to make it go away
    ..."

    Does not work for 2016 in any of the three polar plots

    ReplyDelete
    Replies
    1. Now it works for JAXA data, but not for NISDC NH and NISDC SH

      Delete
    2. Erik,
      I think NH is OK now - I'll try to get SH fixed tomorrow.

      Delete
  13. Looks like your NSIDC charts have stopped working around 19 October? JAXA still OK though.

    ReplyDelete
    Replies
    1. Thanks. I think the NSIDC source is down. There is a log at the bottom of the post which says that it isn't getting data, and clicking on the URL gives an error. I'll investigate.

      Delete
    2. Thanks again. It seems there is a new version with a new URL. Hopefully will be OK soon.

      Delete
    3. NSIDC is fixed. Meanwhile, Jaxa went off the air. I hope it will be back soon. Normally it doesn't matter; it just shows the last data until it next comes on, but this time I overwrote the old data. Should be OK within 24 hrs.

      Delete
  14. Could I make a suggestion, Nick? In your graph of temperature anomalies over the maximum length of time, could you please use more distinct, contrasting colors for the six different indices? Those reds are hard to tell apart, as are the two pea greens. The two graphs above that particular one are much easier to read. Thanks.

    ReplyDelete
    Replies
    1. I take it you mean the 4-year graph. I currently use equally spaced colors from a rainbow scale, which probably isn't ideal. I'll experiment.

      Delete
    2. No, I mean the active plotter for global temperature indices.

      Delete
    3. I see. The colors as they originally show are rather earthy, and I could fix that. But generally here the remedy is that you can choose your own. Just click a color square in the box lower left. A set of choices will popup; choose one. Because there are so many choices, I can't usefully pre-determine the colors.

      Delete
  15. You will not get any more values from NOAA for your "Monthly global temperature anomaly data" table if you continue to monitor their 1880-2017.csv file. You have to switch to the 1880-2018.csv file. NOAA reported their January temperature anomaly long time ago but your table got stock in December.

    ReplyDelete
  16. Superb job done here, thank You! Please consider adding ISTI dataset into your plot graphs: Noaa Global Databank Monthly Stage3 Recommended

    ReplyDelete
    Replies
    1. Thanks, Simo,
      I wote quite a lot about ISTI when it first came out in beta in 2012. This was the last in that series; I should do more. I also wrote about GHCN V4, and again, I'm sure there will be more. There are more posts listed if you press the right buttons on the Moyhu index page.

      The main reason I stick with GHCN V3 for now for TempLS is its prompt reporting, updating almost every day early in the month. ISTI would give better coverage in areas that are already fairly well covered by V3, but does not add a great deal where coverage is sparse. I looked at the coverage comparison of GHCN V4 here.

      Delete
    2. Thanks for the links Nick, I pretty much agree after some reading, and noticing the use of Jones&Moberg(2003) in the ISTI-study you might actually have even closer to reality data in use. Keep up the great work you're doing!

      Delete
  17. Nick, something is wrong with TempLSmesh. Data and maps look distorted. It has been like this for a week, or so..

    ReplyDelete
    Replies
    1. Olof,
      I think I have messed up the spherical harmonics (which I have been tinkering with). It shouldn't affect the numerical results. Anyway, I'll investigate. The stations plot isn't showing either, for some reason.

      Thanks for pointing it out

      Delete
    2. Olof,
      I see the results are wrong too, so it isn't the SH. I'll check more tomorrow

      Delete
  18. Access denied errors in the fresh data import

    ReplyDelete
  19. Why are the GISS figures presented in the "Monthly global anomaly data" table above so different from the figures presented on the GISS website?

    ReplyDelete
    Replies
    1. Can you quote an example? There are small differences between the GISS tabulated numbers and the numbers shown on the GISS plots, for example, July 2019, the map showed 0.92, the table 0.93. I copy from the table, here.

      Delete
  20. The November TempLSmesh temperature anomaly in the table (+0.14) is not consistent with the value in the graph (-0.28). (They both appear to be unphysical low)

    ReplyDelete
    Replies
    1. Wondering why I got an email notification for this comment on May 24 ... almost 6 months after the event.

      Delete
    2. Hi Bob,
      Mishaps at my end, sorry about that. I've been getting intermittent blizzards of spam; there is one at the moment, and there was another at the time of your comment six months ago. When it gets too bad, I send all comments to moderation, and try to keep up. I missed this comment in the moderation queue initially, but in the recent outbreak, I was going through the queue and found it, and published it.

      Delete
  21. Dec 19 temps now charging toward an all time high. This going to take some serious 'splainin' on the denier blogs. The 2016 EN excuse is getting a bit smelly, but I guess it will have to be ladled out at least one more time...it's all they've got left.

    ReplyDelete
  22. Excellent and informative data helps me to move forward with my research work.
    Thanks a lot for the post and the pictorial graph.
    Keep up the good work.
    Let's hope for the best.

    ReplyDelete
  23. Say, great work!
    Given the Recent month of December's low temperature anomalies vs Dec 2016 (0.152 vs 0.391 respectively), what is your best prediction on how NASA's official anomaly will be for December?
    If greater than .61 it should beat the record set in 2020 (index of 1.01) for warmest year, given Nasa's rounding to the hundredth's spot).
    Welcome everyone's thoughts.

    ReplyDelete
    Replies
    1. Good question, CC. December was quite cool. I see the betting odds have dropped to about 70%. I think that should be lower. It will be a near thing.

      Delete
  24. Hi Nick,
    I see your globe has not been updated since January 15, and some dates earlier in January are also missing values. Do you intend fixing this?

    ReplyDelete
    Replies
    1. Thanks for noting. I had to revamp the NCEP programs because of end year troubles, so that is the likely cause. Yes, I'll fix it.

      Delete
    2. Thanks Nick. I notice you have updated it, but it is now displaying nonsense anomalies since Jan 15 .... 8, 9 or 10 instead of zero point something.

      Delete
    3. Your map is still broken. It is currently showing February 31, so even the dates are wrong. And the global average is 24 instead of zero point something.

      Delete
  25. Feb 31st? Daily reanalysis temperature should be 28th maybe?

    ReplyDelete
    Replies
    1. Yes, for some reason it has renamed 28th as 31st; that is an error in the filename, which I will track down. Other dates should be OK.

      Delete
    2. I've checked on the dates. They were actually running one day behind, so 1 march was written as 31 Feb. It was an error caused by switching between one system counting from 0, and the other from 1. I think it should be corrected when the program runs again.

      I'll chase up the incorrect average shown; I may just remove it, since I originally had it just so it could be checked against the text list to ensure the date was right.

      Delete
  26. 1) NCEP/NCAR reanalysis appears to be still on holidays or stopped working after the last day of 2022.
    2) The previous zip file (1994-2015) of historical data you have is out of date (has data upto 31-12-2020). This could be updated sometime this year.
    Thanks.

    ReplyDelete
    Replies
    1. Yes, sorry about that. Moyhu has been a bit distracted lately, but should be OK now. The calendar dependence of the NCAR process is such that it has never made a fully automatic start to the new year. Hoping for it to work next year.

      Delete
    2. Hi Bob,
      Sorry about the moderation. I'm not sure why the file terminated - I hope I can update it. And yes, I'll check the March 28 story.
      Thanks

      Delete
    3. Thanks Nick. I'll check back later. I hope you won't mind a reminder in a week's time if the changes have not been made.

      Delete
    4. It seems the problem is in the NCEP/NCAR data for 3/28 at 1800 UTC. In my implementation using the Python netcdf4 reader, that data is flagged as invalid. I have also an implementation using the average daily data and that data set has 3/28 flagged as invalid as well.

      In my code, at least, I don't see a trivial fix because of implicit assumptions about no missing data. This was one of the nice features of the data set as per Nick noted in his original essay on the subject of using this data set to calculate daily anomalies.

      A non-trivial fix to handle the error case will take longer . . .

      Delete
    5. Thanks. I've put in a fix for out-of-range data, which I think is the best that can be done.

      Delete
    6. tygrus
      I've updated the zip files - I hope that now they will update monthly

      Delete
    7. Your zip file for 'Daily Reanalysis Temperature data' is stated to be 'invalid'.

      Delete
    8. Thanks, Bob
      I'll look into it.

      Delete
    9. Hi Bob,
      I've rerun the program, and the result seems OK. I hope it was just a one-off faulty upload, but let me know of any further trouble.

      Delete
  27. The Daily Reanalysis Temperature Data is jammed up, Nick.

    ReplyDelete
    Replies
    1. THanks, fixed I hope. I've been changing over computers

      Delete
  28. The Daily Reanalysis Temperature Data is stuck again, Nick.

    ReplyDelete
    Replies
    1. THanks. I've checked; I think the source is stuck.

      Delete
  29. The Reanalysis Temperature Data is stuck again, Nick.

    ReplyDelete
    Replies
    1. Thanks. The program is running correctly. I think the NCEP source is lagging.

      Delete
    2. No, the source is OK, and the program ran on 15-11. But the data displayed hadn't been updated. I'll check.

      Delete
  30. Fixed. My uploading connection had disconnected itself. Should be OK now.

    ReplyDelete
    Replies
    1. Great, Nick, thanks! And thanks for the great site. It is invaluable.

      Delete
  31. It looks like the Reanalysis Temperature Data is lagging again, Nick.

    ReplyDelete
  32. The Daily Reanalysis Temperature Data is showing up as an empty box, Nick.

    ReplyDelete
    Replies
    1. Thanks, Anon
      Fixed. I'm having intermittent troubles with the automatic upload facility which only seems to affect these files. Hope to get it finally sorted soon.

      Delete