Guest post by Steven Goddard
The GISS website shows the graph below, which indicates a steady, steep warming trend over the last 30 years. The monthly average anomaly for 2008 (0.44) is 0.26 degrees warmer than the monthly average anomaly for 1980 (0.18.) Data obtained from here: http://www.woodfortrees.org/data/gistemp/from:1980/plot/uah/from:1980
By contrast, the UAH monthly average anomaly for 2008 (0.05) is 0.04 degrees cooler than the UAH monthly average anomaly for 1980 (0.09.) Again, data obtained from here: http://www.woodfortrees.org/data/uah/from:1980
This 1980-2008 discrepancy between GISS and UAH is important, as it is nearly equal to the claimed warming trend since 1980.
Taking this one step further, I made a graph of the difference between the GISS and UAH monthly anomalies since 1980.
As you can see below, the discrepancy has increased over time. Using Google’s linest() function, the divergence between GISS and UAH is increasing at a rate of 0.32C/century. (GISS uses a different baseline than UAH, but the slope of the difference should be zero, if the data sets correlated properly.) The slope is not zero, which indicates an inconsistency between the data sets.
Click for larger image
Raw data from here: http://www.woodfortrees.org/data/gistemp/from:1980/plot/uah/from:1980 Calculations done here.
Factoring in the baseline
Some readers will undoubtedly again point out that the GISS baseline (“normal”) temperature is lower than the UAH baseline. This is true, but as I said above does not affect the slope calculation. The difference between the GISS and UAH monthly baselines is a constant, which affects the relative position along the y-axis – but it does not affect the slope. Subtracting a monthly constant from each point in a graph does not alter the slope over a large set of years. It only alters the y-offset.
The equation of a line is y = mx + b, where m is the slope and b is the y-offset. m and b are completely independent. The different baselines affect only b, not m. If the UAH and GISS data were closely tracking each other, the slope (m) would be close to zero. The fact that GISS shows 2008 temperatures much higher than 1980, and UAH shows 2008 temperatures lower than 1980, is also a clear indicator that the two data sets are divergent.
Steve McIntyre has coincidentally just done a similar comparison of NOAA USA yearly data vs. GISS USA yearly data, and came to the conclusion that the NOAA slope is even steeper than GISS, diverging from UAH by 0.39C/century.
This would imply that NOAA is diverging from UAH by an even larger amount than GISS is diverging from UAH.
Clearly, problems exist with both datasets.
JP,
The reason I picked 1980 as the start year rather than 1979, was because it was an interesting year in that it was warmer than 2008 in the UAH record, but much cooler than 2008 in the GISS record. The delta between GISS and UAH in 1979 was larger than in 1980, and would have made the divergence even larger, had I used it.
http://www.woodfortrees.org/plot/uah/from:1978/plot/gistemp/from:1978
There were also quite a few other years between 1979 and 1997 which were warmer than 2008.
http://www.woodfortrees.org/plot/uah/from:1978/plot/gistemp/from:1978
http://www.woodfortrees.org/plot/uah/from:1978/plot/rss/from:1978
As John Philip has pointed out, the comparison of the individual years 1980 and 2008 is essentially meaningless since the difference you get depends strongly on which exact year you use. I should emphasize further that UAH and GISS measure temperatures in a somewhat different part of the atmosphere…and it is in fact expected that the temperature that the UAH measures will vary more strongly with ENSO (El Nino – La Nina) fluctuations. Hence, since 2008 was a pretty strong La Nina year and it looks like 1980 was an El Nino year, we would expect to see exactly what you saw…that the temperature difference between 2008 and 1980 is considerably more negative for UAH than for GISS.
The difference in the full trend over the period is a little more relevant, however:
(1) It is a lot smaller (0.03 C per decade), with both showing a significant warming trend.
(2) There is no corresponding attempt to do error analysis of the trends to see if it is actually statistically significant.
(3) There is again the fact that they measure somewhat different things.
(4) As John Philip has noted, when you look at UAH, RSS, GISS, and HADCRUT, it is UAH that is the “odd” man out in the trends…with a trend about 20% lower than the other three.
Steve Goddard says:
I don’t think this keeps things in perspective at all. Rather, it completely blows things out of proportion. In fact, the difference between whether the trend is 0.13 C/decade or 0.16 C/decade has only a very small implication for public policy.
And, your fact about the behavior in the individual years of 1980 and 2008 has no implications for public policy and no direct implications concerning the trends.
‘Carrick (10:06:46) :
Just going to mention that this difference has been discussed over on Lucia’s blog.’
Thanks Carrick, by discussed I think you mean thrashed to within an inch of its life. I lost four hours out of my life reading it, but couldn’t drag myself away. It should be cautionary reading for anyone thinking of starting a climate blog.
The horror, the horror…
LOL. I just want to emphasize a key conceptual point here, which perhaps keeps things in perspective. The GISS graph shows 2008 warmer than 1979 by 0.35C. The UAH graph shows 2008 slightly less warmer than 1979 by 0.16C, Clearly the trends are marginally divergent, unsurprising when you consider the measurement uncertainties and the fact that they measure different physical quantities.
REPLY: Also unsurprising when there is a plethora of adjustments that start with NOAA/NCDC such as TOBS, FILNET, SHAP, and then there is the GISS homgenization, all of which tend to add amplitude to the surface temperature signal. UAH has no such post facto adjustments once the base data is derived.
See image below. – Anthony
http://www.ncdc.noaa.gov/img/climate/research/ushcn/ts.ushcn_anom25_diffs_urb-raw_pg.gif
JP, Joel,
2008 is not an outlier to be cast out. RSS shows a decline of 3.6C/century over the last six years, and even more interesting is that RSS is steadily diverging from GISS at 2.64C/century over the last six years.
http://spreadsheets.google.com/pub?key=pj0h2MODqj3ixuEoxE5xBVw&oid=3&output=image
If the GISS graph showed the same trend as satellites over the last six years, the appearance would be quite different. The entire steep period from 1980-2002 would be reversed.
It would be interesting to compare UKMet (HadCRU3) with GISS, since they are both land based. Also both are run by individuals who believe in anthropogenic glabal warming, but HadCRU is widely regarded as having some integrity whereas GISS is not.
Steve,
I also wonder why you have excluded 1979. That the UAH monthly average anomaly for 2008 (0.05) is 0.12 degrees warmer than the UAH monthly average anomaly for 1979 (-0.072) is not important at all. What is important is the trend of the difference between the GISS and UAH monthly anomalies since 1979.
The slope of this trend is indeed embarrassing. The question is: for which institution is it embarrassing? For GISS or for UAH? It is to expect that the same analysis comparing RSS and GISS gives a much lower slope. So, we have three institutions from which the figures have to be compared.
At this phase in the investigation, I trust UAH: “Ours is the only dataset that has been compared to non-satellite data,” said Christy. “This gives us confidence in its results. Several different radiosonde-based products have been compared to the satellite data and the results of those studies have been published.” (http://www.uah.edu/News/climatebackground.php) And since the year 1998, RSS and UAH have an almost equal trend in the temperature data.
Is it justifiable to suspect GISS of distributing improper data? It is clear that the last thing we should expect is that the cited difference between the GISS and UAH monthly anomalies would be growing! Of course, the GISS baseline temperature is lower than the UAH baseline. This can explain a difference of about 0.2526 degrees (average difference from 1979 to 2007). But we see some months with a difference of more than 0.4 degrees! I find 31 months with a difference of 0.40 to 0.49, 7 with a difference of 0.50 to 0.59 and two months with a difference above 0.6, viz. 0.637 and 0.666. I also detect that in 2008, there were 6 months with a difference above 0.4! This makes 2008 the year with the largest average difference that is registered since 1979. How can we explain this growing discrepancy? The answer can, in my opinion, not only be the result of “the different patterns of temperature variance at the surface and in the lower troposphere”( http://www.cgd.ucar.edu/cas/papers/jclim96/)
“A recent analysis of the surface and satellite datasets hints that the apparent disagreement might have as much to do with coverage as with differing trends at different altitudes.”, according Christy in the same UAH article. “In areas where you have high resolution, well-maintained scientific collection of temperature data, the satellites and the surface data show a high degree of agreement,” said Christy. “Over North America, Europe, Russia, China and Australia, the agreement is basically one-to-one.” “The greatest disagreement between the surface and satellite datasets is in the tropics, which includes regions where weather stations are sparse (including central Africa and South America), and the three-fourths of the tropics that are covered by oceans, where proxy information such as sea surface temperatures has been used in lieu of actual atmospheric temperature data.”
We can see the sparse implantation of weather stations in some areas: http://www.giss.nasa.gov/research/features/temptracker/global_weather_stations_map_532.gif
When the missing areas are cold, according to “Bonzo” , it is understandable that the averages of NASA are on the high side. This could explain differences but not growing differences. See (http://groups.google.com/group/sci.geo.meteorology/msg/5e621ec65ed0a480)
I cannot understand that a research centre as Goddard Institute of Space Studies with a budget of more than $ 1 billion, should not be able to supply reliable figures concerning the temperature of earth. Time has come to GISS to provide us the answers on the questions why there is a growing divergence between UAH and GISS and if it is true that the GISS warming trends are not consistent around the globe. If they cannot afford these arguments, that means they have lost their credibility.
UAH has no such post facto adjustments once the base data is derived.
See image below. – Anthony
Really? What then do you make of this UAH read me file:
http://vortex.nsstc.uah.edu/public/msu/t2lt/readme.03Jan2008
A random example –
Update 12 July 2007 *****************************
We are still relying on NOAA-15 for the current MT and LT products. We
have not instituted a diurnal correction on these, so they are likely
running a little on the warm side as NOAA-15 is “backing” into a warmer
period of the diurnal cycle. Regarding the long term trend, this will have
a very small impact. When the new MSU diurnal corrections as well as the AMSU
diurnal corrections are applied, we suspect the resulting trend will be
almost identical to the current trend, though there will be some changes
in the interannual variations.
Of course UAH makes post facto adjustments – the most notable having been the adjustments made in 2005 after their huge error in miscorrecting for diurnal drift was pointed out to them:
Update 7 Aug 2005 ****************************
An artifact of the diurnal correction applied to LT
has been discovered by Carl Mears and Frank Wentz
(Remote Sensing Systems). This artifact contributed an
error term in certain types of diurnal cycles, most
noteably in the tropics. We have applied a new diurnal
correction based on 3 AMSU instruments and call the dataset
v5.2.
An entirely reworked data set sounds like a post facto adjustment to me!
Steven Talbot,
There is no comparison between the satellite adjustments and the USHCN adjustments. The TOBS adjustment alone requires USHCN to go back and try to figure out for each thermometer on each date what time of day the max/min values were read and reset, not to mention all their other adjustments and problems with thermometers in parking lots, etc. which Anthony has been documenting.
By contrast, the UAH adjustments are relatively minor and consistent, having to do with well documented and understood issues like satellite roll.
If I’m correct, google LINEST uses ordinary least squares (it does in excel). OLS assumes no autocorrelation. If you want to properly estimate the best linear unbiased estimate, you need to correct for autocorrleation. It won’t change things much, but it will be statistically better.
Steven Goddard,
By contrast, the UAH adjustments are relatively minor and consistent, having to do with well documented and understood issues like satellite roll.
As memory serves, the 2005 correction to the UAH record resulted in a 40% increase in the warming trend recorded (may have been more, so by all means correct me). You call that minor?
The fact is that the satellite data is processed just as the land-based data is processed. The differences between UAH, RSS and other satellite records, despite using the same raw information, make it wholly obvious that this is so. It is entirely bizarre to suggest that ‘human judgment’ is only a concern with USHCN. Is this wholly satisfactory? Of course not – but let us at least be genuinely ‘sceptical’ of all such records rather than making out that the satellite records are somehow rock solid. The sheer scale of their past corrections makes it entirely obvious that this is not so.
RSS is steadily diverging from GISS at 2.64C/century over the last six years.
I fear you may be disappearing down a statistical rabbithole in pursuit of a fault in GISS, while losing sight of the physics. As Joel and I have stated the satellites (indirectly) measure the lower troposphere temperature which responds more to ENSO events – ie it warms more than the surface in al El Nino and cools more in a La Nina, in a six year period with a La Nina towards the end the divergence between the satellite based RSS and GISS is probably attributable largely to this effect. There was a similar divergence in the six years to 1998, only in the opposite direction, with RSS warming faster than GISS.
http://www.woodfortrees.org/plot/gistemp/from:1993/to:1999/trend/plot/rss/from:1993/to:1999/trend
Of course this period has an El Nino rather than a La Nina towards the end, which just goes to show the limited significance of such a short trend.
The rationale behind the various adjustments, with supporting papers is given on this NOAA page. But if, as implied, it is the adjustments that are introducing an incorrect warming bias and that explains why GISS rises more than UAH since 1980, we still have to explain the similar divergence between UAH and RSS, which also has no adjustments…
http://www.woodfortrees.org/plot/gistemp/from:1979/mean:12/offset:-0.206/plot/gistemp/from:1979/trend:12/offset:-0.206/plot/uah/mean:12/plot/uah/trend/plot/rss/mean:12/offset:0.02783/plot/rss/trend/offset:0.02783
“By contrast, the UAH adjustments are relatively minor and consistent, having to do with well documented and understood issues like satellite roll.”
The 2005 UAH adjustments increased their lower troposphere temperature trend from 0.086 to 0.12 deg/decade.
That’s an increase of 40% – hardly “minor”.
And the lower troposphere temperature trends between RSS and UAH still differ by about 0.05 deg/decade (despite the fact that they use exactly the same raw satellite data) so it is obvious that all of the “issues” with the satellite data are not completely understood.
One of the biggest difficulties with the satellite data is cross-calibrating between different satellites. Something like a dozen different satellites have been used since 1979, and every time a new satellite comes on line, it has to be calibrated against the existing ones.
It’s something like the calibration that takes place when a surface temperature station is moved- except the satellite calibration doesn’t just effect one of hundreds of temperature stations, it effects the data for the whole world! So even a tiny calibration error can have a big effect on the calculated temperature trends.
The raw satellite readings also have to be corrected for reflections and emisions from the satellite itself, from the antenae, from space, from the moon, orbital drift and decay…and more.
Then you have to remember that the lower trosphere readings are derived from measurements made of the same area from two different viewing angles, so all those potential errors are doubled.
In short, the adjustments/corrections made to the satellite readings are immensely more complicated than those applied to the surface temperature measurements. And any error in those adjustments effects not just a single temperature station, but readings for the entire earth.
Steve Goddard says:
Steve,
First of all, noone is saying that 2008 should be cast out; however, one should not do an analysis in a way that is completely not robust to the exact years that you look at, which is what your simple comparison of 1980 to 2008 is.
Your comparison of trends between 1980 and 2008 is a more reasonably thing to do, but then the resulting differences turn out not to be very dramatic, do they, especially when one considers that the satellite and surface data are not measuring exactly the same thing…and the fact that a more complete analysis with the HadCrut and RSS datasets shows that UAH is the outlier, with the HadCrut, RSS, and GISS trends all extremely close?
Now, you are desperately trying to look over shorter time intervals in an attempt to rescue the sort of conclusion that you want from your analysis. Sure, over short enough periods, the different data sets (particularly the ones that measure different things) will be quite divergent. That shouldn’t surprise anyone.
This statement is just bizarre to me. If the GISS graph showed the same trend as the satellites from 1980 to 2008, then the trend would be 0.13 rather than 0.16, a rather small difference…That is what your original trend analysis showed. I have no idea what you are trying to do now…cobble together a piece of this graph and a piece of that graph to somehow put together a different conclusion!?! You seem to desperately want to rescue something from your analysis despite the fact that it has showed that the difference in trend between the GISS and UAH datasets of the full length of the satellite datasets is small (and the difference between GISS and RSS over that length is positively tiny).
“By contrast, the UAH adjustments are relatively minor and consistent, having to do with well documented and understood issues like satellite roll.”
The 2005 UAH adjustments increased their lower troposphere temperature trend from 0.086 to 0.12 deg/decade.
That’s an increase of 40% – hardly “minor”.
And the lower troposphere temperature trends between RSS and UAH still differ by about 0.05 deg/decade (despite the fact that they use exactly the same raw satellite data) so it is obvious that all of the “issues” with the satellite data are not completely understood.
One of the biggest difficulties with the satellite data is cross-calibrating between different satellites. Something like a dozen different satellites have been used since 1979, and every time a new satellite comes on line, it has to be calibrated against the existing ones.
It’s something like the calibration that takes place when a surface temperature station is moved- except the satellite calibration doesn’t just effect one of hundreds of temperature stations, it effects the data for the whole world! So even a tiny calibration error can have a big effect on the calculated temperature trends.
The raw satellite readings also have to be corrected for reflections and emissions from the satellite itself, from the antenae, from space, from the moon, for orbital drift and decay…and more.
Then you have to remember that the lower troposphere readings are derived from measurements made of the same area from two different viewing angles, so all those potential errors are doubled.
In short, the adjustments/corrections made to the satellite readings are immensely more complicated than those applied to the surface temperature measurements. And any error in those adjustments effects not just a single temperature station, but readings for the entire earth.
These adjustments over the years have been so “minor” that in fact for many years, before the appropriate corrections were made, the satellite record was touted as showing that there was a cooling trend…i.e., they didn’t even have the freakin’ sign right!!! (Part of this may have also been due to the shorter record available at that time…although I don’t think that was the main effect and could easily be investigated by fitting trends over the current UAH satellite record starting at its beginning and ending in various years.)
And, while the issues in the satellite record may be “well documented and understood” now (although that isn’t entirely clear, particularly given the remaining significant divergence of the UAH and RSS records in the tropics), they were not for quite some time…which is why Spencer and Christy were the darlings of the skeptics for many years for supposedly demonstrating that we were cooling rather than warming.
Here’s a link to a power point presentation that discusses SOME of the calibrations and corrections done to the raw satellite data:
http://www.asic3.sdl.usu.edu/papers/weng.ppt
17-01-2009
DR (13:55:44) :
Has anyone seen David Archibald’s prediction for UAH data through May 2009?
http://icecap.us/images/uploads/oftheMay2009UAHMSUGlobalTemperatureResult12thJanuary2009.pdf
Whoa, he’s really sticking his neck out. It is 180 deg out of phase with Hansen and Met O, and most likely other govt. funded institutions.
*******************************
I recall that I made a similar prediction, probably recently somewhere here on wattsup, but I can’t find it. Usually trusty Google has failed me…
Could also be on climateaudit, but less likely.
I think my “guess” was somewhere in the -0.2C to -0.4C range for May or June 2009, based on similar logic to David (for the UAH LT anomaly).
I am sure we came up with this prediction independently – but we coud both be wrong – time will tell.
If someone happens to find my post please help me out by posting the url on this page.
Thanks in advance, Allan
Still no P value?
This site is moving from scepticism to denial.
REPLY: That is uncalled for, you are making an assumption that he has seen the request and discarded it. Your assertion is a jump to conclusions based on your own personal bias and an apology is in order. I’ll email him to make sure he has seen the request in the nearly 100 comments posted. In the meantime, perhaps you missed “Calculations done here.” Good manners are welcomed here, even if we disagree with the content, assumptions and subsequent denigration such as the one you just made are not welcome. – Anthony
REPLY2: I did indeed get in touch with Mr. Goddard via email, and he posted a response on this new thread, which is the follow up post to this one, but for the benefit of all on this thread, I am repeating his response here:
We have a 40% correction in UAH trend because UAH operates completely on the table, accessible to independent review of all sides in the debate.
That is not the case with the institutions that measure surface, though, regarding their adjustment methods.
USHCN-1 at least gave the amounts of each step of adjustment (if not the methodology). But the USHCN-2 adjustment page doesn’t even do that. Just a bunch of more-or-less meaningless words.
(Come to think of it, a 1.3 vs. 1.6 trend is itself a > 20% difference. I’m not sure that’s insignificant.)
Chris V,
You wrote: “The 2005 UAH adjustments increased their lower troposphere temperature trend from 0.086 to 0.12 deg/decade.Thats an increase of 40% – hardly minor.
A change of 0.034K is a change of less than 0.01% when measuring temperatures at 260K, which is what the satellites measure. Nice try.
The USHCN adjustments are based largely on WAG data. No one really knows what time of day some thermometer in Kansas was read on December 12, 1904, so many times they just take a guess. Do you think that is accurate within 0.034K? Do you think they have enough information about that thermometer to calculate trends with a hundredth of a degree?
With satellites, the information can be calculated. By contrast, USHCN simply does not have enough information to make the calculations, which would be of near infinite complexity due to an unimaginably large number of degrees of freedom.
Joel Shore (11:32:10) : writes
I don’t think this keeps things in perspective at all. Rather, it completely blows things out of proportion. In fact, the difference between whether the trend is 0.13 C/decade or 0.16 C/decade has only a very small implication for public policy.
I agree things are getting blown out of proportion particularly when we consider that the GISS record shows a 30 year trend (1915-44) of ~0.14 deg C/decade. And the early part of the 1915-1944 period didn’t have the ‘benefit’ of 2 volcanic eruptions to amplify the trend.
Remind me again what did cause the early 20th century warming?
Steven Goddard (10:19:21) :
That’s a rather astonishing statement. If the UAH adjustments were made in reference to “well understood issues”, one wonders how its compilers were unable to identify these “well understood issues” over a period of 15 years, and repeatedly had to have these “well understood issues” pointed out to them!
The UAH interpreters of MSU satellite data presented their first analysis in 1990 with a paper in which they suggested that their analysis produced a precise analysis of tropospheric temperature:
Spencer RW, Christy JR (1990) Precise Monitoring Of Global Temperature Trends From Satellites. Science 247, 1558-1562
We know that these analysis were anything but precise, and in fact have been confounded by a long series of errors and misanalyses that were repeatedly pointed out in the scientific literature. So already in 1991 it was pointed out that the statistical analysis of the data was flawed and that the absence of a temperature trend claimed by Spencer and Christy (SC) was unjustified:
B.J. Gary and S. J. Keihm (1991) Microwave Sounding Units and Global Warming Science 251, 316
In 1997 it was pointed out that the methods used to average the different satellite records introduced a spurious cooling trend:
J. W. Hurrell & .K E. Trenberth (1997) Spurious trends in satellite MSU temperatures from merging different satellite record. Nature 386, 164 – 167.
In 1998 it was pointed out that the failure by Spencer and Christy to properly consider satellite orbital decay introduced another spurious cooling trend in the UAH analysis:
F. J. Wentz and M. Schabel (1998) Effects of orbital decay on satellite-derived lower-tropospheric temperature trends. Nature 394, 661-664.
In 2004 it was pointed out that MSU-2 showed a spurious cooling trend due to spillover of stratospheric cooling into the tropospheric temperature signal:
Q. Fu et al. (2004) Contribution of stratospheric cooling to satellite-inferred tropospheric temperature trends Nature 429, 55-58.
And a little later still it was pointed out that the diurnal correction applied by Christy and Spencer was of the wrong sign and gave yet another spurious cooling trend:
.A. Mears and F. J. Wentz (2005) The Effect of Diurnal Correction on Satellite-Derived Lower Tropospheric Temperature, Science 1548-1551
Considering that the UAH compliers have consistently introduced spurious cooling contributions to their analyses for 15 years, and have only corrected these when these have been highlighted by more competent analyses, it seems entirely appropriate to question the accuracy of the UAH data…I would consider the RSS analysis to be more reliable based on the fact that it was the RSS group that identified some of the rather blatant errors in UAH.
In general the idea the satellite tropospheric temperature data is more accurate than the surface data is questionable. The satellite microwave sounding units (MSU) doesn’t actually measure temperature, which has to be constructed from tropospheric radiance. Second order effects of tropospheric warming such as increased water vapour, results in decreased atmospheric “brightness” which makes the troposphere appear cooler, and has to be corrected for. As time passes satellites undergo orbital decay (see above), and of course eventually fail and are replaced by new satellites. The data from the different satellites have to be independently calibrated and merged. Unfortunately there isn’t a completely objective means of doing this. And the potentially independent means of determining tropospheric temperature from radiosondes (weather balloons) for calibrating MSU temperature data doesn’t work since the radiosondes have their own serious artefacts:
Sherwood, S. C. et al (2005). “Radiosonde Daytime Biases and Late-20th Century Warming”. Science 309, 1556–1559
Steven Goddard (12:56:37) :
It’s the temperature TREND that’s important- that’s why all the temperature data sets show temperature ANOMALIES, and not ABSOLUTE temperatures.
“With satellites, the information can be calculated.”
But are the calculations correct?? If the satellite adjustments were well understood, then UAH and RSS would get the same results.
But the difference between UAH and RSS trends is about the same as the difference between UAH and GISSTemp trends.
John Philip
John, you clever sleuth. I bet you dug up that graph info from the lower right hand reference on the chart.
But if that graph isn’t your cup of tea, how about this one: click1 or this one: click2
Notice who is out of step? See Manfred’s explanation @ur momisugly 18:58:41 for the answer.
DJ:
See the ‘click2’ chart linked above, and rethink.