Since the “Daily Updates”, “Monthly Trends”, “ENSO”, “MLMW”, and “Paleo Climate” pages are not open for comments, I have added this page for comments and discussions regarding those pages as well as for general blog-related comments and discussion.

Please note if you post here for the first time, your comment will automatically go into moderation for approval.  Once your first comment has been approved, subsequent comments will post immediately.


32 responses to “Comments

  1. Hi,

    on the ENSO page you have this text: “A similar pattern can be seen in comparing the University of Alabama at Huntsville (UAH) Temperature for the Lower Troposphere (TLT) change over 12 months versus the MEI as shown below.”

    In the graph the UAH line does not match with the global UAH 6.0 graph here:

    This makes me think that the line (TLT) represents Tropical Lower Trophosphere and not global lower trophosphere. Perhaps this was the intention all along – to show Tropical Lower Trophosphere?


    • Ilkka, on the ENSO page I am graphing the 12-month change for the UAH TLT, which is the current month minus the same month a year ago. This approach detrends the data somewhat to make it easier to compare to the MEI which is a unit-less index. I have the actual UAH TLT monthly averages plotted in the first graph on the Monthly Trends page.

      • Ilkka Ponkanen


        the word “change” was there but I did not get its meaning. This might make the sentence and the graph easier to understand:

        “A similar pattern can be seen in comparing the University of Alabama at Huntsville (UAH) Temperature for the Lower Troposphere (TLT) change over 12 months versus the MEI as shown below (12-month change for the UAH TLT is the current month minus the same month a year ago).”


  2. Bill Rudd - Dallas, Texas

    Looked but didn’t find your copyright policies. I’m a retired Electrical Engineer, BSEE Texas A&M, MSEE SMU. Spent 40 years helping develop intelligence, surveillance and reconnaissance systems for our country’s national defense. I retired in 2002 and am pursuing the arts (painting, drawing, etc.). Recently, I became annoyed by the mass of climate change misinformation being publicized and determined to do my own search and review of credible published climate reports. Hence, my question about your copyright policies. You have some consolidated plots that I have considered creating of published data, and I would appreciate having permission to copy them for possible use, with appropriate credits, in a future blog. I am not going to write or publish any reports.

  3. This is great work..
    Thanks for spending the time to compute and report..not looking like a record year for the NH Is it?
    However NASA/NOAA may find a way..🤔👍

    • Thanks. The NH monthly GMSTA of 0.21C for both May and June was the lowest since that same level in April 2015. So even the NH temperature is showing signs of trending downward.

  4. Watching the Arctic influxes (h/t to Caleb’s blog ‘Sunrise’s Swansong’) on DMI maps etc., one sees swift uplift to space, of energy. I expect including the large latent heat release from snowing. Thus the c.2 trillion tonne increase of arctic ice in the last two years noted by tony Heller and others.. On Greenland especially, and also thicker seaice. Holder’s inequality should put the ‘broiling’ into perspective, which commenters generally fail to do. It is like the hot exhaust pipe which fails to affect the nearby petrol tank…..

  5. Hey Brian,

    Appreciate these updates. The updates seem to no longer agree with the Climate Reanalyzer, while it seemed like they did before. The updates are running lower. What does this mean?


    • Ned, the Climate Reanalyzer daily output is actually a forecast for the day based on the Global Forecast System (GFS) model output from the first run of the UTC day (0000 UTC). Recently I switched to directly using the NOAA Climate Forecast System Reanalysis (CFSR) Version 2 output instead of the Climate Reanalyzer for providing current daily and monthly temperature anomaly estimates, as described in the “Methodology” section under “Daily Updates”. The CFSR data that I use are based on reanalyses of the model input data associated with the 4 GFS model runs each day and do not include any forecast data. I back-loaded the CFSR daily data for all of 2018 so far and over the next few months will be back-loading the 2017 daily data at higher resolution than what was provided by Climate Reanalyzer.

      There have always been discrepancies between the Climate Reanalyzer GFS-forecast based daily temperature anomaly estimates posted each day versus the CFSR daily estimates that are eventually posted on Climate Reanalyzer, often many months later. The CFSR daily estimates for 2018 have not yet been posted there and estimates for 2017 were only posted to the nearest tenth of a degree Celsius, which is very coarse resolution.

      I still privately track the Climate Reanalyzer GFS-based temperature anomaly estimates and compare them to the CFSR estimates that I derive from the NOAA output. Recently, the GFS temperature forecasts have been running a bit to the high side, especially in the Northern Hemisphere and that appears to be why the CFSR temperature anomaly estimates are at present lower. You can see a graph of the GFS temperature performance for the 168-hour forecasts, as well as a graph of the model run initial temperatures plus the latest forecast here (referenced to 1981-2010 and adjusted to better match GISS results):

  6. Hi Bryan,

    I have noticed some changes to the UM CFSR data over the past four months. To highlight them I took one of your figures from November 2017 and another from March 2018 and overlaid them. March 2018 data is in blue and green as you post it, while November 2017 data is grey and pink. The baseline was changed from 1981-2010 to 1979-2000. This change doesn’t make much sense as it is going from 30 years to just 22, and towards comparing with a more distant past. Also the vertical scale went from 1.6 °C to 2.2 °C compressing the changes. But I have adjusted the vertical scale so they both match in this figure:

    What can be observed is that the data from 2016 was slightly changed and for example the minimum of 2016 in mid-June is now slightly lower, while the data of 2017 was changed upward quite significantly, so the cooling since February 2016 has been reduced by 21% from the November to the March data.

    Do you know the reason of this rather drastic change? Why does it take place at a time the baseline, the scale, and the resolution were changed, so it is more difficult to spot? If we are to trust UM CFSR data it is necessary that it is not subjected to the type of arbitrary changes that have made other datasets unreliable, and that always appear to go in the direction of warming the present and cooling the past.

    What is your opinion on this?


    • Javier, if you look at the first graph in the Daily Update page, it compares the UM CCI preliminary GFS estimates with the UM CCI final CFSR estimates. This is the only change to the UM CCI daily estimates that occurs. The 2014-2018 graph uses a combination of the CFSR estimates available plus the GFS estimates extending beyond that time. So when UM CCI reports new CFSR estimates, I replace the corresponding GFS estimates with CFSR estimates.

      On the Daily Update page I also compare the UM CCI GFS month-to-date and recent monthly averages with the WeatherBELL CFSR estimates for the same periods (see the paragraph just below the first graph). I have found that usually the final UM CCI CFSR estimates are much closer to the WeatherBELL CFSR estimates than are the preliminary UM CCI GFS estimates. Thus the WeatherBELL CFSR versus UM CCI GFS differences are a good indicator of what to expect when the UM CCI CFSR estimates are released. For 2018 January, February, and March so far, the preliminary UM CCI GFS estimates are running about 0.1C above the WeatherBELL CFSR estimates. Consequently, I expect the final UM CCI CFSR estimates to drop by about that amount. You will be able to tell when this change occurs in the graphs when the red line for the UM CCI CFSR jumps forward, closer to the preliminary GFS line in blue, on the first graph in Daily Updates. You will also see a slight change in the 2014-2018 graphs for the affected period, if you compare graphs before and after this update.

      Also, I just recently started calculating the global surface temperature anomaly from the hemispheric anomalies and retaining the result to the nearest 0.05C for graphing, which adds a bit more detail to the daily global changes on the graphs. Previously I had been graphing the UM CCI global estimates rounded to 0.1C as provided by UM CCI.

      • Thank you for the explanation, Bryan. It just looks like a 20% deviation from GFS to CFSR is a lot.

      • Javier, I calculated the difference between monthly averages of the daily estimates and below is the result in Kelvin (or Celsius) for final CFSR minus preliminary GFS for 2017:
        +0.047 Jan
        +0.068 Feb
        +0.074 Mar
        +0.087 Apr
        +0.047 May
        +0.102 Jun
        +0.040 Jul
        -0.053 Aug
        -0.098 Sep
        -0.105 Oct
        -0.063 Nov
        -0.032 Dec
        As you can see, early in 2017 the estimates went up from GFS to CFSR, but since August have been going down from GFS to CFSR. Individual days can have larger swings. So far for 2018, it looks the the estimates will continue to drop from GFS to CFSR and possibly as much as about 0.1C for monthly averages, based on the UM CCI vs WeatherBELL comparisons. If they continue to decrease as expected, that will also cause the running 365 day average to show a larger drop than at present using the preliminary GFS estimates since 2018 January 1.

        As near as I can tell, the preliminary UM CCI GFS estimates are actually forecasts for the current day based on the 0000 UTC model run at the beginning of the day and this may account for much of the differences.

      • Thank you Bryan, that explains a lot. I have also noticed that on the monthly chart UM has become an outlier, showing 0.2°C higher anomaly than the rest, which is a lot.

  7. Ilkka Pönkänen

    A much needed site! I wonder when you can update Mauna Loa and Mt Washington info from end of 2017 to present day (or end of Feb)?

    • Thanks Ilkka. There is about a 1-month lag beyond the end of the month for availability of the data. I will probably update the MLMW graphs quarterly, so the next update would be in early May for data through March.

    • Ilkka, I have updated the MLMW monthly graphs through 2018 March and two had data for April which I also included.

      • Ilkka Pönkänen

        Thank you. Will you update the “detailed views” as well? They still say “through 2017”.

      • Ilkka, I updated all of the graphs displaying monthly data. The remaining graphs display annual data and cannot be updated until all of the 2018 data are available early next year. I will provide another update of the monthly graphs in the early fall to show the summer data.

      • Ilkka Pönkänen

        Ok, thanks.

  8. I like your graphs but, as nature does not do straight lines, putting straight lines through everything is simply muddying the water.

    A smoothing curve is nice, which would suggest the cyclic nature of most of these. Why do people think that a downward part of a cycle and an upward part deserve a straight line? The slope of the line is clearly dependent on the time frame and thus is a waste of time and distracting from the picture.

    • higley7 I agree that trend lines are not likely to be an accurate predictor beyond about 10 to 20 percent of the time period for which there is data and thus are likely to be misleading for longer time periods for the reasons you cite. However, since they are commonly used to analyze what has happened over a measurement period I include them. I like running averages to eliminate seasonal effects seen in daily and monthly temperature data and for smoothing over longer periods like 5 years or more and have included these on some of the graphs.

  9. It’s too bad the daily graphs on UM CCI GFS/CFSR with actual values over the 1981-2010 period have been discontinued. I liked them better than the graphs with values rounded to the nearest decimal value and 1979-2000 baseline. Easier to get a feeling of how temp is evolving.
    Congrats for your fine site, and thank you for making the info available.

    • Javier, UM CCI recently began reporting the daily GFS-based global and zonal temperature anomaly estimates in tenths rather than hundredths as previously reported, beginning 2017 November 30. Consequently, the graphs I present here based on the UM CCI daily estimates must also be presented in tenths. For consistency and simplicity I also converted the older data to tenths and I am now presenting the data with the same 1979-2000 reference period used by UM CCI. To shift to the 1981-2010 reference period, the annual offset is -0.12C and monthly offsets range from -0.09C for June-August to -0.16C for November.

      In reality, these GFS-based estimates are probably not accurate to even a tenth of a degree Celsius at best, so effectively some of the random “noise” has been eliminated. However, I have to admit I like to see the hundredths detail as well, but I am not willing at present to commit the much greater time expenditure to go and get the GFS and CFSR output from NCEP and prepare my own analyses. WxBELL used to have time series graphs of the daily CFSR global estimates on their public model output web page, but since Ryan Maue left, they have removed them (although they still are providing a month-to-date global map).

      • Thank you for your answer. Of course it doesn’t make sense that you put extra time to change the display.

        Perhaps it is just me, but I am under the impression that whenever the data stops going the alarming way, displays are changed or discontinued so no fair comparison can be done. I have seen this happening with Arctic sea ice age, that they first changed the display and then stopped reporting altogether, and with the length of the Arctic melt season, once it stopped growing and started to shrink. I guess it gives them the chills to think about skeptics using their own graphs against them.

  10. This is the first time I’ve seen these charts. Do you know why the differ from the anomaly charts I often see i.e. , GISS, HADCUT, UAH, RSS. For example the 1998 spike isn’t as prominent.

    • Tony, the CFSR estimates are based on the same weather data used to initialize the global weather forecast models four times every day, which is a lot more data with much better spatial coverage than what goes into the GHCN-based estimates like those from GISS, NCEI, and HadCRUT. The CFSR estimates have actually compared better with the UAH and RSS satellite derived TLT estimates than the GHCN-based estimates.

  11. Ren, thanks, and yes the graphs I prepare from daily GFS/CFSR data posted by UM CCI are very similar to those compiled independently by Ryan Maue for WxBELL. I check the WxBELL graphs every day to make sure the UM CCI data are consistent. I agree that the reanalysis approach is probably the best for estimating regional and global surface temperatures and anomalies, although I am not certain the current methods are optimal. I especially like that it is available in near real-time to monitor global weather pattern changes.

  12. Your daily updates chart is remarkably the same as Ryan Maue’s chart he plots from the NCEP CFSR / CFSv2 data generated every 4 hours to feed the GFS / ECMWF and other weather models (WeatherBell – paywalled). I always thought this was the best to use since it can’t be homogenized or the weather models will be worse then they currently are.

    Good work.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.