Ilya Usoskin vs Leif Svalgaard : Oulu Neutron Monitor data quality

Posted: November 21, 2012 by tallbloke in Astrophysics, Clouds, cosmic rays, Dataset, Measurement, solar system dynamics

Ilya Usoskin says:

I have been pointed to this discussion by someone who got surprised by Leif Svalgaard’s claims on the drifts in Oulu NM data. I am the PI of the Oulu NM and am quite surprised that this issue, including direct cliams of my misqualification, are discussed here without contacting me first!
Oulu NM is regularly checked for the stabiity of electronics and counters and is regarded by experts as one of the most stable station of the world network. No aging is observed. Moroever, as Leif claims, Oulu is counting MORE cosmic ray than Thule, but this cannot be due to aging, unless this is aging of Thule. Aging can lead only to decreasing count rate!
Comparing Oulu to mid- and low-latitude stations is incorrect as the modulation during the cycle 23-24 is known to be more energy dependent than before. Moreover, Oulu data is totally consistent with most of the high-latitude stations (Apatity, McMurdo, Kerguelen, Terre-Adelia etc. – seehttp://www.nmdb.eu/nest/search.php ), except only two – Thule and even greater difference with the South Pole, the latter both showing a decreasing trend, absent in other stations. Moreover, McMurdo being counting even more than Oulu during the last years.
Thus, I consider Leif’s comments ungrounded and offensive as publicly discussed behind my back. I advice everyone to ask experts first if you think some data are wrong, not just claiming the data wrong because they don’t support someone’s idea.
Please don’t reply to me here, I am not reading this forum. If you have any questions, write to me directly (contact info is at http://cc.oulu.fi/~usoskin/ ).

Sincerely yours,
Ilya Usoskin

 

lsvalgaard says:

Olavi says:
November 20, 2012 at 11:57 am
I’ll think Leif should clean his face and ask apology. Leif who is the one that makes cherrypiking? Seems that Ilya is bit angry. :)
It is not a good idea to argue with angry people. There are only a handful of stations with very long records. The longest one is Climax. In this plot, I normalize several long-running stations to have the same mean as Climax [so they can be compared]:http://www.leif.org/research/Cosmic%20Ray%20Count%20for%20Different%20Stations-Oulu.png
Then I divide Oulu’s counts by the mean of all the other stations and plot them as the triangles. It is clear to me that there is an upward drift in the triangles and that Oulu therefore is not representative for the overall cosmic ray intensity [whatever the reason for the drift, be it instrumental or not] measured at the surface of the Earth [which is presumably what some people think has effect on climate]. BTW, one does not need permission to analyze data that is publically available.
It is possible that Oulu has seen more of the low-energy cosmic rays than other stations at mid-latitudes, but since the climate is supposed to be influenced by the high-energy cosmic rays, that possibility seems irrelevant.
It is instructive to use the site that Ilya linked to:http://www.nmdb.eu/nest/search.php to see for yourself what the variations at several stations have been.

___________________________________________________

http://cc.oulu.fi/~usoskin/personal/Sola2-PRL_published.pdf

http://www.leif.org/research/Kiel-Cosmic-Rays-and-Solar-Cycles.png
http://www.leif.org/research/Oulu-and-Thule.png the red curve is Oulu
http://www.leif.org/research/Oulu-and-Hermanus.png
http://www.leif.org/research/Neutron-Monitor-Thule-Newark.png

Comments
  1. tchannon says:

    When I read the Leif comment I imagined he was either not sober or angry so I ignored it.

    As someone knowing a little about electronics and sensors, “Aging can lead only to decreasing count rate!”, is a fair comment, that is what chambers do, it is loss of sensitivity. A rise in sensitivity if there was no defect in the first place is not going to happen with the sensor.

    Ignore him.

  2. Doug Proctor says:

    Presumably atmospheric thickness, content (pollution, other aeorsols), local geology (radioactive) and mechanical variations can result in different CR measurements. But they would be consistent for each site. Is Leif saying that the readings are changing realative to each other over time?

    The “defense” didn’t explain what was wrong with the graphical representation. It seemed only about feeling offended.

    Did I miss a technical explanation for the “observation” that Leif made?

    Oh, well.

  3. [snip]
    Leif and Ilya may both have points. Certainly as DP above points out Ilya has not answered Leifs Criticism.
    Leif’s plot shows Oulu rising above the others in his station plots and this is confirmed in his triangles. Oulu is measuring differently to others. Perhaps this should be investigated. Not dismissed as the rant of a drunk?

    [Reply] I posted the comment from Ilya and Leif without editorialising, and I expect comments to stick to the facts too.

  4. tallbloke says:

    In Fact Ilya did answer Svalgaard’s earlier criticicisms. I don’t understand how you think he could have answered what Leif wrote in the comment placed FOUR HOURS AFTER ILYA’s.

    “Moroever, as Leif claims, Oulu is counting MORE cosmic ray than Thule, but this cannot be due to aging, unless this is aging of Thule. Aging can lead only to decreasing count rate!
    Comparing Oulu to mid- and low-latitude stations is incorrect as the modulation during the cycle 23-24 is known to be more energy dependent than before. Moreover, Oulu data is totally consistent with most of the high-latitude stations (Apatity, McMurdo, Kerguelen, Terre-Adelia etc. – seehttp://www.nmdb.eu/nest/search.php ), except only two – Thule and even greater difference with the South Pole, the latter both showing a decreasing trend, absent in other stations. Moreover, McMurdo being counting even more than Oulu during the last years.”

  5. ATheoK says:

    “tchannon says:
    November 21, 2012 at 1:42 am
    When I read the Leif comment I imagined he was either not sober or angry so I ignored it.

    As someone knowing a little about electronics and sensors, “Aging can lead only to decreasing count rate!”, is a fair comment, that is what chambers do, it is loss of sensitivity. A rise in sensitivity if there was no defect in the first place is not going to happen with the sensor.

    Ignore him.”

    Um, ignore who?

    I’ve followed this thread to the sites linked above, back to WUWT and back here again. I didn’t find anywhere where Leif attributed anything to aging instruments.

    Catching Leif in error is nigh impossible as he tends to be meticulously thorough. I gather from Ilya’s post that he believes the same for himself and his crew, nor do I see any reason to doubt Ilya at this time. However, unike Tallbloke, I am not convinced Ilya answered or countered Leif on even terms.

    I do believe that this is a discussion well worth pursuing in detail and that possibly everyone can learn something.

    Instead of flashing graphs at us, I think we need to return to the basics and get an agreement on exactly which data is present and in calculations. Following the steps in Leif’s procedure should bring everyone on board evenly and at the end Leif and Ilya can discuss Leif’s triangles and trends rationally. Even if they agree to disagree at least Ilya should understand that Leif is not disparaging Oolu, just that he find their counts/trends outliers and therefore suspect. (Is this a reversal of the consensus trends?)

    And it may well be that Oolu is more efficient at counting non-high energy cosmic particles. Which is something we can all happily live with and perhaps utilize for even better research.

    A simple table of monthly, or better weekly, count totals for a few sites should let everyone see if the data at the beginning is identical. If not, then why? If yes, then next step…

  6. tallbloke says:

    Leif has made another comment which has been rebutted by Ilya Usoskin:

    lsvalgaard says:
    November 20, 2012 at 8:23 pm

    ed says:
    November 20, 2012 at 6:07 pm
    The appropriate (scientific) thing to do…
    One of the contributing reasons for the upwards drift of Oulu is that [as with so many time series in the climate debate] the values have been ‘adjusted’ over time, so that, for example, the value for the year 2008 is now [in 2012] higher than the value for 2008 published in 2009 and so on. Such upwards adjustments will naturally produce an ever rising series.

    ==========================

    Here is Ilya’s response:

    Ilya Usoskin says:
    November 21, 2012 at 3:10 am

    Leif said:
    “Such upwards adjustments will naturally produce an ever rising series.”
    This is simply a wrong bold statement. Oulu NM data has never been adjusted for anything but known changes of efficiency due to the change of the surroundings – see description at http://cosmicrays.oulu.fi/readme.html. It might happen that Leif discusses uncorrected Oulu data, but I cannot judge on that unless I know where and how exactly he obtained the data.
    If one compares Oulu data with other polar NMs, it is totally consistent – see plots at

    http://cosmicrays.oulu.fi/tmp/Oulu_vs_/

    where plots have been simply copied from NMDB NEST tool => just to avoid any further ungrounded claims of purposeful “adjustments”. Interestingly enough,while Apaptity and Kerguelen with the geomegnetic rigidity cutoffs Rc close to that of Oulu Rc=0.8 GV show count variability nearly identical to Oulu, polar stations with the lower cutoff of 0.3 GV (Fort Smith, Inuvik and Nain) show even stronger “upward drift” than Oulu. If someone makes a plot similar to that presented by Leif (ratio of Oulu to other mid-latitude stations), but for another polar station, the result would be the same. This is typical not for Oulu but for most polar stations. Is it an indication of the polar-station-mafia’s secret and simultaneous adjustment of data? Thule and South Pole show a decreasing trend instead.

  7. A C Osborn says:

    Leif is never wrong, ask Leif, he will confirm this.

  8. ATheoK says:

    Tracking graphs back and forth, I’ve just finished tracking locations back and forth. I think the latitude issue is moot. The locations are not mid-latitude versus artic latitudes as implied.

    Leif has primarily used NOAA locations. Outside of normalization after station moves/construction NOAA does not mention ‘adjustments’.

    Ilya used University of Delaware stations. Apparently all of these stations are also normalized and ‘adjusted’ for efficiency.

    Efficiency Correction factor

    Because of the changes of hardware/software during the station operation period, the efficiency of cosmic ray registration might have been slightly changing. This is carefully taken into account, and the efficiency correction factor is used to provide a homogenous long-term data series.

    The NM count rate in the database is normalized to the count rate before 1985 so that
    I (subset)normalised = I (subset)measured * F(subset)C

    where FC is the efficiency correction factor (see below). Period FC Reason
    01/1964 – 30/09/1985 1.00000
    01/10/1985 – 31/12/1994 1.00674 New automatic digital barometer
    01/01/1995 – 31/12/1999 1.01147 New data collecting system
    01/01/2000 – 31/05/2003 1.00914 replacement of section A high voltage system
    01/06/2003 – 31/07/2008; 1.00813 adjustment of section A high voltage system
    01/08/2008 – 31/10/2009; 1.0029 New data registration system
    since 01/11/2009 ; 1.0019 Change of the station environment (Muon detector is installed)
    ——————————————————————————–
    For any kind of comments/questions/requests please contact Ilya Usoskin via e-mail Ilya.Usoskin@oulu.fi

    What really puzzles me is why positive ‘efficiency‘ adjustments are used to normalize counts to pre 1985 periods. Am I missing something?

    Anyway, Ilya is using University Of Delaware, Bartol Research Institute stations with efficiency adjustments and Leif is using NOAA apparently unadjusted for efficiency stations. So Oolu should have higher counts given their efficiency adjustments. Efficiency?

  9. ATheoK says:

    My bad. I should have included the link for BRI’s information page. http://cosmicrays.oulu.fi/readme.html.

    NOAA’s page on stations.

    http://ngdc.noaa.gov/stp/solar/cosmic.html

    Have fun!

  10. vukcevic says:

    Svalgaard of Stanford knows a thing or two about correcting historic data. :)
    I had amusing exchange with Dr.S. I thought I found some evidence of the magnetospheric connection in the solar activity

    http://www.vukcevic.talktalk.net/Sub-cycle.htm

    claiming that ‘The variation of the surface field did go berserk in epoch 1940, time of SS cycle 17’.Dr. S claims data is wrong, didn’t fit what is expected .

    lsvalgaard says:
    November 19, 2012 at 2:01 pm
    vukcevic says:
    November 19, 2012 at 1:55 pm
    You may think so, but there is always strong possibility that there was no explanation why 1940-1950 epochs should be so much more active, and then someone has a ‘brilliant’ idea, “our models don’t look right lets correct the data”, …hmm, they still do it elsewhere. Fortunately the scientists at NOAA knew better and wisely left it on their data files.

    It is a mark of true pseudo-science that even after a correlation has been shown to be spurious because of errors in the data, the pseudo-scientist still maintains that everything is fine. The secular variation errors and the GSN errors are good cases in point.

  11. Doug Proctor says:

    ATheoK says: November 21, 2012 at 3:18 pm

    “Anyway, Ilya is using University Of Delaware, Bartol Research Institute stations with efficiency adjustments and Leif is using NOAA apparently unadjusted for efficiency stations. So Oolu should have higher counts given their efficiency adjustments. Efficiency?”

    Would that not result in a bulk shift situation? Changes through time mean equipment, environment or, Lord’s Alive! the subject being observed is changing.

    Ilya: the discrepancy appears real, so what is wrong with the data (other than Leif is commenting on it)?

  12. kuhnkat says:

    Doug,

    Ilya is not monitoring tallbloke’s.

    “Please don’t reply to me here, I am not reading this forum. If you have any questions, write to me directly (contact info is at http://cc.oulu.fi/~usoskin/

  13. tallbloke says:

    Leif has made further comment: also rebutted by Usoskin:

    lsvalgaard says:
    November 21, 2012 at 7:53 am
    I got the ‘official’ data from Oulu’s website back in 2008 and again now in 2012. There is a smooth, progressive upwards change in the data over the period 1964-2012 of 2%.
    The issue is not whether Oulu is right or wrong, but to what extent Oulu is representative of the modulation of the high-energy [10 GeV and up] cosmic rays that is thought be some to be active in climate change.

    He then backs off from the 2% claim:

    This was from eyeballing the graphs. A better determination using the actual counts shows a more jumpy change of 1.2%, with the main change in late 1985. So the efficiency changes were applied retroactively to 1985. Nothing wrong with that if the changes are justified.

    And here is Usoskin’s response:

    Ilya Usoskin says:
    November 21, 2012 at 8:37 am

    Leif wrote: “I got the ‘official’ data from Oulu’s website back in 2008 and again now in 2012″

    Now it’s cristal clear that what was ascribed to my ill-intentioned non-scientific “data adjustment” was in fact caused by Leif’s carelessness. the Oulu web-site provided, in 2008, data which were NOT corrected for the efficiency (changes of the local surrounding), and a user was expected to do it manually, as was explicitly stated at the web-page. In 2010, following numerous requests of users, careful enough to read the information on the web-page, we implemented this correction into the database. Thus, the data downloaded in 2012 are already corrected for the efficiency, as again explicitly stated at the web-site.
    See a description of the correction at

    http://cosmicrays.oulu.fi/readme.html

    in the section “Efficiency Correction factor”.
    Thus, what Leif compared was corrected-vs-uncorrected values (I_normalized vs I_measured), exactly as I suspected. Their ratio simply gives the correction factor as in Table at the

    http://cosmicrays.oulu.fi/readme.html.

    Leif, WHY DIDN’T YOU SIMPLY ASK ME ABOUT THIS, instead of blaming me???? The issue would be resolved within 5 minutes…

    Ilya

  14. tallbloke says:

    lsvalgaard says:
    November 21, 2012 at 7:53 am
    I got the ‘official’ data from Oulu’s website back in 2008 and again now in 2012. [There is a smooth, progressive upwards change in the data over the period 1964-2012 of 2%.]
    This was from eyeballing the graphs. A better determination using the actual counts shows a more jumpy change of 1.2%,

    You ‘eyeballed’ a 2% difference?
    This is not credible, and inaccurate as it turns out.
    Plus, you have blatantly ignored this statement made by Ilya Usoskin:

    “the Oulu web-site provided, in 2008, data which were NOT corrected for the efficiency (changes of the local surrounding), and a user was expected to do it manually, as was explicitly stated at the web-page. In 2010, following numerous requests of users, careful enough to read the information on the web-page, we implemented this correction into the database. Thus, the data downloaded in 2012 are already corrected for the efficiency, as again explicitly stated at the web-site….Thus, what Leif compared was corrected-vs-uncorrected values (I_normalized vs I_measured), exactly as I suspected. Their ratio simply gives the correction factor as in Table at the
    http://cosmicrays.oulu.fi/readme.html.”

    “I’m not blaming anybody, just using the data as it was presented”

    More like you were misusing the data and created (accidentally or not) the impression Ilya Usoskin has made unwarranted and undocumented ‘adjustments’ to the data.

    You got it wrong Leif. Admit it and apologise to Ilya.

    The email to the Stanford Ethics committee about your slurs against Abreu, Steinhilber et al is still sat in my drafts folder. I’ll add this to it and press send unless I see a fulsome apology forthcoming to all these scientists.

    I’ve had enough of the way you misrepresent other peoples honest work.

  15. Henry Clark says:

    A couple weeks ago, in an earlier thread, Svalgaard once again attempted to make readers believe that there was no significant difference between the cosmic ray count in the last minimum versus the one a couple cycles before. As can be seen there, although unfortunately probably few if any readers paid close enough attention to notice, he had no counter to my http://wattsupwiththat.com/2012/11/06/solar-cycle-24-continues-weakly-perhaps-weakest-of-the-space-age/#comment-1137964 reply debunking such, resorting to simply ignoring it (not leaving our argument from lack of time, as posting many times more in the thread). His strategy is the standard one of sheer repetition, knowing that many readers will assume anything shamelessly constantly repeatedly implied must be true, while others don’t have the time to repeat the same arguments every thread. His direction of argument is predictable; for instance, he always tries to discredit any WUWT article suggesting solar activity changes may cause future temperatures to substantially cool rather than warm.

  16. Tim Cullen says:

    Henry Clark says: November 23, 2012 at 9:10 am

    It’s a tough job being the Galactic Gatekeeper for the Team :-)

  17. Geoff Sharp says:

    Svalgaard is pushing the envelope lately, perhaps predicated on impending doom. Rog’s ethics email to Stanford is worthy of merit….waiting to see how it pans out on WUWT.

  18. GAI says:

    Svalgaard does his darnedest to make sure any mention of the sun as a variable star is squashed flat over at WUWT and often elsewhere. It does not matter what the evidence or who produced it.

    Has anyone else notice that WUWT has well educated Trojan Horses who target specific areas of expertise? Svalgaard for example always shows when ever the sun is mentioned and kills discussion of the sun as a variable star. If you have been around WUWT for any length of time I think you can spot the others.

    I also notice that WUWT does not seem to have any discussion of Judithgate that I could find. Yet that was one of the biggest IPCC oops since Climategate.

  19. Ilya Usoskin says:

    Just to put a full stop in the discussion of the “trend” in the Oulu NM data.

    There is a recent paper devoted to a thorough analysis of different NMs stability on the long-term scale (Ahluwalia & Ygbuhay, Testing baseline stability of some neutron monitors in Europe, Africa, and Asia, Adv. Space Res., 2013, in press)

    http://www.sciencedirect.com/science/article/pii/S0273117713000367#

    The last sentence (summary) of the paper states that
    “it is suggested that Oulu neutron monitor may be a good standard for future comparisons of the baselines over longer intervals of time.”

    Ahluwalia is one of the pioneers of cosmic ray studies by NM is definitely among the best experts in the field.

    [Reply] Thank You Ilya for taking the trouble to update this thread. I have posted your response as a new article to give it prominence. Please join discussion there, you are very welcome at the talkshop. http://tallbloke.wordpress.com/2013/02/05/ilya-usoskin-new-paper-clarifies-discussion-of-oulu-neutron-monitor-data-quality/