Guest post by Bob Tisdale
http://downloads.globalchange.gov/usimpacts/pdfs/climate-impacts-report.pdf
As noted in the title, it fails to address the multiyear effects of El Nino-Southern Oscillation (ENSO) events on global temperature.
Other than explosive volcanic eruptions, El Nino-Southern Oscillation events have the greatest impacts on global climate on annual and multiyear bases. The year-to-year global temperature impacts of ENSO events are clearly visible in a comparative time-series graph, Figure 1. Also visible are the overriding effects of the 1982 El Chichon and 1991 Mount Pinatubo volcanic eruptions.
http://i44.tinypic.com/144ag5f.jpg
Figure 1
The multiyear impacts of the 1986/87/88 and 1997/98 El Nino events on Northern Hemisphere Lower Troposphere Temperature (TLT) are clearly visible in the TLT Time-Latitude Plot available from Remote Sensing Systems (RSS). Refer to Figure 2 and 3, which are from my post “RSS MSU TLT Time-Latitude Plots…Show Climate Responses That Cannot Be Easily Illustrated With Time-Series Graphs Alone.”
http://i44.tinypic.com/16leq39.jpg
Figure 2
#########
http://i41.tinypic.com/2vwzmdj.jpg
Figure 3
A seldom-discussed, naturally occurring oceanic process called Reemergence (Refer to my post “The Reemergence Mechanism”) provides the mechanism by which the global oceans integrate the effects of ENSO events. And it only takes the cumulative effect of a very small portion (0.0045 or less than ½ of 1%) of the monthly ENSO signal, as shown in Figure 4, to reproduce the Global Sea Surface Temperature (SST) anomaly curve.
http://i42.tinypic.com/iom6ab.jpg
Figure 4
YET HOW MANY TIMES DOES THE USGCRP REPORT MENTION THE EL NINO-SOUTHERN OSCILLATION?
The USGCRP mentions “El Nino” nine times in the body of the 196-page report, but those references only pertain to global temperature on one occassion. The first reference, however, states that ENSO is independent of human activities.
On page 16, during a discussion Natural Influences, they wrote, “The climate changes that have occurred over the last century are not solely caused by the human and natural factors described above. In addition to these influences, there are also fluctuations in climate that occur even in the absence of changes in human activities, the Sun, or volcanoes. One example is the El Niño phenomenon, which has important influences on many aspects of regional and global climate.” [My emphasis.]
They acknowledged that ENSO is independent of anthropogenic influence. That’s significant.
On page 17, in the text of the comparative graph of “Global Temperature and Carbon Dioxide”, they wrote, “These year-to-year fluctuations in temperature are due to natural processes, such as the effects of El Niños, La Niñas, and the eruption of large volcanoes.” [My emphasis.]
Yet they fail to note the multiyear and cumulative effects of ENSO.
Page 36, during a discussion of Pacific Hurricanes, they write, “The total number of tropical storms and hurricanes in the eastern Pacific on seasonal to multi-decade time periods is generally opposite to that observed in the Atlantic. For example, during El Niño events it is common for hurricanes in the Atlantic to be suppressed while the eastern Pacific is more active. This reflects the large-scale atmospheric circulation patterns that extend across both the Atlantic and the Pacific oceans.” [My emphasis.]
That quote is important in many contests. Much can be inferred from it. Yet they fail to acknowledge the multidecadal epochs when El Nino or La Nina are dominant. These epochs are visible in a time-series graph of smoothed NINO3.4 SST anomalies, Figure 5.
http://i43.tinypic.com/33agh3c.jpg
Figure 5
On page 38, under the heading of Snowstorms, they wrote, “The northward shift in storm tracks is reflected in regional changes in the frequency of snowstorms. The South and lower Midwest saw reduced snowstorm frequency during the last century. In contrast, the Northeast and upper Midwest saw increases in snowstorms, although considerable decade-to-decade variations were present in all regions, influenced, for example, by the frequency of El Niño events.” [My emphasis.]
And again, they infer multidecadal influences of ENSO, but the USGCRP have failed to account for it in their attribution of global temperature change.
There are further references of El Nino and La Nina events on pages 81, 147, 148, and 152, as they pertain to tuna stock, droughts, coral reefs, and coastal currents. No need to repeat those in this post.
CLOSING
Like the IPCC, the USGCRP either fails to accept the significant multiyear and cumulative impacts of ENSO on global temperatures or they chose to ignore them in their presentation of the causes of global temperature change.
Posted by Bob Tisdale at 8:42 PM
Erl Happ: You wrote, “1. There is a sudden fall in OLR during the El Nino warming events (red rectangles) when the steepest rise in SST is experienced. For this to occur, precipitation (and therefore latent heat release that drives OLR and the cloud cover that would exclude sunlight) fails to increase as the ocean warms.”
Incorrect. The following is a graph of Southern Equatorial Pacific Precipitation Anomalies (10S-0, 150E-90W). There are significant increases in precipitation for all El Nino events that appear to coincide with SST anomalies. I could create a comparative graph (Precipitation anomaly vs SST anomaly) if you’d like.
http://i41.tinypic.com/15xkd21.jpg
Tallbloke: You wrote, “One curiosity is that the changes between positive and negative ENSO phases at 1915 and 1945 are preceded by reversals in the global SST trend by quite a few years on a plot smoothed over the solar cycle length. Perhaps Bob would comment on that.”
You’re basing the dates of ENSO phase reversal (1915 & 1945) on the point at which the curve of the 121-month (~10-year) smoothed data crosses zero. That graph is only intended to show that there are epochs when El Nino or La Nina events dominate, not necessarily the point at which they become dominant.
The point at which frequency and magnitude of La Nina events exceeds those of El Nino events (or vice versa) would be visible in the graph of the running-total of scaled NINO3.4 SST anomalies (Figure 4).
http://i42.tinypic.com/iom6ab.jpg
In order for the running total to stop rising, the frequency and magnitude of La Nina events would have to EQUAL the frequency and magnitude of El Ninos. For the running total to decrease, the frequency and magnitude of La Nina events would have to EXCEED the frequency and magnitude of El Ninos.
Bob Tisdale (00:55:07) :
Bob,
I have had another close look at it the rise in sea surface temperature that began in the equator to 10°S zone about November 1996 that went on to become the big El Nino of 1997-8. The increase in precipitation rate for the global zone between the equator and 10°S lags the rise in sea surface temperature by six months.
In that period of six months OLR increases along with SST. When precipitation starts to increase OLR rapidly and dramatically falls away
My interpretation was incorrect. It appears that the reason why OLR falls away so dramatically is because that latitude zone starts to cool via decompression. The latent heat release drives strong uplift.
So, I will have to look again at this issue. Presumably the loss of heat via decompression is matched elsewhere by an increase in temperature due to compressive descent. It may be worthwhile to do an analysis of each 10° latitude zone to check what happens in each case, right through to the subtropical high pressure cells and even the polar highs.
Thanks for the scrutiny.
One upshot of this is that I have a lot more faith on the validity of the OLR figures and that they do in fact enable latitudinal analysis.
Steven Wilde: You wrote, “Bob Tisdale’s ideas suffer from the same problem…”
They do? What problem? Maybe you’re looking for an answer that does not appear in this post or in the thread of comments, but your statement that my ideas are problematic is unwarranted, unjustified, and unnecessary.
You continued, “… which is why, above, I invited both of you to try and help me on that point.”
And since I did not respond to any of your comments on this thread prior to this one (you were expressing opinion and discussing matters with Erl), I do not understand how my ideas could have suffered from any problems.
Moving on, I assume it’s this request for help. You wrote, “Thus from our observations it must be changes within the oceans on the 30/60 year timecale (sic) that are in command and not the processes that you both describe. Unless that is either of you can suggest how the 30/60 year cycle within the oceans can arise from those shorter term phenomena. Can either of you help me on that?”
First note: A couple of “60-year” cycles over the past 100+ years do not necessarily mean the frequency remains constant. The 1650 to 1980 NINO3 SST Reconstruction shows low frequency ENSO oscillations that vary from 21 to 39 years, with an average of 27 years, not 60 year:
http://bobtisdale.blogspot.com/2009/03/low-frequency-enso-oscillations.html
Second note: Any “cycle” in the smoothed NINO3.4 SST anomaly (Figure 5) is obviously a function of the filtering. Smooth the data with a 61-month filter and the “cycles” would be there, but the graph would be noisier and would include more (but shorter) periods where SST anomalies were above or below zero; i.e., there’d be more oscillations.
I have no doubt that there are underlying cycles that impact the frequency and magnitudes of ENSO events and, therefore, create the epochs visible in the smoothed data, but I can’t put a finger on what drives what. And since GCMs haven’t yet fully grasped ENSO and all of the other oceanic processes, it’s doubtful there will be an answer soon.
Are the frequencies and magnitudes of ENSO events tied to feedback from AMOC? There are papers that describe the effects AMOC (or the AMO) has on the ENSO. In the other direction, I do know that ENSO impacts AMOC. During and after the 1997/98 El Nino, AMOC flow decreased noticeably…
http://bobtisdale.blogspot.com/2008/11/atlantic-meridional-overturning.html
…and SST anomalies in the North Atlantic increased in a step.
http://bobtisdale.blogspot.com/2009/02/there-are-also-el-nino-induced-step.html
The Humboldt Current carries waters from the ACC northward in the eastern South Pacific, so changes in Southern Ocean SST anomalies should impact SST anomalies in the tropical Pacific, and, therefore, should enhance or retard the buildup of heat in the PWP. There was a recent rise and fall in Southern Ocean SST anomalies in the portion south of the South Pacific that appears to agree with the underlying increase and decrease in SST anomalies in the tropical Pacific, but part of the rise in the Southern Ocean would be caused by the frequency and magnitudes of ENSO (more feedbacks). To add more complexity, in the Southern Ocean south of the Atlantic and Indian Oceans, it appears there could be a 100-year cycle. At minimum, there is a dip and rebound. Does it repeat? I have no idea. Refer to Figure 5 in the following post:
http://bobtisdale.blogspot.com/2009/04/closer-look-at-ersstv3b-southern-ocean.html
The California Current feeds eastern North Pacific waters south into the tropical Pacific. It, too, would impact the amount of heat in the tropical Pacific.
Changes in cloud amount would impact the amount of heat .
http://bobtisdale.blogspot.com/2009/04/did-decrease-in-total-cloud-amount-fuel.html
The PWP is also known as the Indo-Pacific Warm Pool, since it extends into the Indian Ocean. Indian Ocean SST anomalies would also provide feedback.
Is there an answer to your question about cycles? I’ve never run across one. But I am looking.
Erl Happ: The KNMI Climate Explorer website has a lot of data that you’d find useful.
http://climexp.knmi.nl/selectfield_obs.cgi?someone@somewhere
Thank you Bob. Those comments are helpful.
In the meantime, until the causes of those multidecadal phase shifts in the oceans are identified I will continue to consider it most likely that those shifts are caused by variations within the oceans rather than variations within the air.
Since your descriptions do essentially involve air driving ocean surface temperstures I see that as a problem only if your description is somehow meant to account for those phase shifts as well as the ENSO variations.
As you are content to accept that the cause of those phase shifts remains unclear then I agree that there is no problem for your basic ENSO scenario.
Erl, however remains sure that the air does drive the ocean SST changes and we will have to agree to disagree on that point until better evidence is available.
Steven Wilde: You wrote, “Since your descriptions do essentially involve air driving ocean surface temperstures…”
Really? Please quote something I’ve written that leads you to that conclusion. As far as I’m concerned, that’s 180 degrees from what I’ve illustrated and written to date. You must be thinking of something that Erl’s written. But, please, please, find and quote something I’ve written that led you to write that.
Bob,
I’ve had a closer look at your earlier post and think I should apologise.
You do accept that the energy is in the ocean in the first instance and that amongst other processes the Trade Winds just move it around so that it accumulates in warm pools from which the energy is then imparted to the air.
I have obviously confused your position with that of someone else.
Sorry for the confusion.
Bob Tisdale (01:45:03) :
The point at which frequency and magnitude of La Nina events exceeds those of El Nino events (or vice versa) would be visible in the graph of the running-total of scaled NINO3.4 SST anomalies (Figure 4).
http://i42.tinypic.com/iom6ab.jpg
Thanks Bob, it looks to me from that graph that the Nino3.4 index still generally lags the global SST changes, and I’m trying to square this observation with the general thrust of your posts regarding the way el nino particularly, spreads the warmth from the PWP outwards. Would you agree that there is some kind of two way process going on here, rather than a one way street for the heat to travel? Lots of different things that work on different cyclicities are all happening at once and it’s quite a puzzle to try to determine the path of causation.
Bob, I downloaded a nino3.4 dataset to do my own graphing from
http://climexp.knmi.nl/data/inino5.dat
and found a curious inconsistency with your data.
Here’s a side by side comparison:
http://s630.photobucket.com/albums/uu21/stroller-2009/?action=view¤t=ninos-datasets.jpg&t=1245524493535
Not only are the dates in disgreement but the data seems flipped or otherwise odd prior to 1920 on your graph or around 1908 on mine.
What’s up with that???!
Where is your dataset from?
Just to add, when I say I found a curious inconsistency with your data, I mean between my data and your data. Obviously, I have no idea which is right. 🙂
tallbloke: You asked, “Where is your dataset from?”
Sorry for not listing sources of the data. I prepared that graph for a recent post on the PDO but never used it. It starts in 1900 because JISAO starts their PDO index that year.
As you’re aware, that graph is NINO3.4 SST anomaly data from 1900 to present that’s been smoothed with a 121-month filter. The dataset is HADISST, Hadley Centre’s interpolated SST version with a 1 deg latitude and longitude resolution. Now, I don’t use any of the KNMI Climate Index data. I use the SST data and select latitudes and longitudes.
Since there may be someone else other than you, tallbloke, trying to duplicate that graph, someone who’s not familiar with the KNMI Climate Explorer, I’ll run through that process. Go to:
http://climexp.knmi.nl/selectfield_obs.cgi?someone@somewhere
Scroll down to SST and select the first dataset HADISST, then scroll back up and click on “Select Field”. On the next page, there are fields for Latitude and Longitude. The coordinates for NINO3.4 are 5S-5N, 170W-120W, so enter -5 & 5 for the latitudes and -170 and -120 for longitudes. (I also got into the habit of entering a zero in the “Demand at least” field somewhere along the line. It shouldn’t have any effect on the HADISST data though.) Click on “Make Time Series.” On the next page, scroll down to the third graph. It reads “Anomalies with respect to the above annual cycle”. On that same line, click on “raw data.” That next page is the raw HADISST NINO3.4 SST anomaly data. It’s in two columns that starts at “1870.0000 -0.804598” and runs through “present day”. For that graph (Figure 5 in the post) I deleted all data prior to 1900 and smoothed it with a 121-month running-average filter.
Now, if you’re trying to duplicate the running total graph (Figure 4), click back once to the webpage with the three graphs. Directly below the anomaly graph is a line that allows you to change the base years. That running total works with base years of 1950 to 1979, but doesn’t work with 1971 to 2000, so enter 1950 and 1979 and click “select.” On the next page, scroll down again to the anomaly graph and select raw data. That’s the data you need to create a running total that mimics global SST anomalies or global temperature anomalies, depending on the scaling factor you use.
Regards
Anthony: Thanks.
tallbloke: You wrote, “Just to add, when I say I found a curious inconsistency with your data, I mean between my data and your data. Obviously, I have no idea which is right.”
They probably both are, just different datasets and different start dates. HADSST2 and HADISST have different underlying curves than the ERSST.v2 and ERSST.v3 series datasets.
Bob, look again at the side by side comparison, the data is similar though not identical going back from the present as far as the small downtick at 1920 on your graph. All the main shifts line up. The same downtick is at 1908 on my plot of the KNMI data.
http://s630.photobucket.com/albums/uu21/stroller-2009/?action=view¤t=ninos-datasets.jpg&t=1245524493535
I think something is seriously wrong with the time scaling on one plot or the other.
The header info has this
# 1856-1949: Kaplan reconstruction
# 1950-now: CPC (Reynolds OI SST)
tallbloke: I “aligned” your two curves for my benefit.
http://i42.tinypic.com/6zvno8.jpg
I agree with your assessment, “Bob, look again at the side by side comparison, the data is similar though not identical going back from the present as far as the small downtick at 1920 on your graph. All the main shifts line up. The same downtick is at 1908 on my plot of the KNMI data.”
But I disagree with your conclusion, “I think something is seriously wrong with the time scaling on one plot or the other.”
Regards
They’re two different datasets. The one I used is the HADISST reconstruction and yours, if I’ve read what you’ve sent correctly, is based on the Kaplan reconstruction. Hadley and Kaplan employed different smoothing and different methods to infill missing data.
Consider this. How many ships passed along the equatorial Pacific before the opening of the Panama Canal in 1914? Not many. It wasn’t a major shipping lane. So a lot of the NINO SST data prior to 1914 is reconstructed: i.e., an educated guess. Between 1914 and 1950, there were still periods with missing data that needed to be infilled and that bucket adjustment that created the 1945 “discontinuity” is still in the SST data. That’s probably why two of the NINO3.4 Indexes, ONI and MEI, start in 1950.
If I remember correctly, Kaplan (the person) was (or is) an employee of Lamont-Doherty Earth Observatory. That SST dataset was created in the late 1990s. They stopped updating it in 2004/05, somewhere around that time. I can only speculate that ERSST replaced Kaplan as the “official” U.S. SST reconstruction data.
tallbloke: Sorry. I somehow got the concluding “regards” in the middle of my comment. Too many hours in front of this computer today.
Stephen Wilde
Can I repeat a question that you may have missed in my post back at erlhapp (20:27:00) :
What is it that accounts for: “variations in the rate of energy flow from ocean to air.”
Anna and others, tree ring data is a far better proxy for PDO than just about anything else people try to use tree rings for. When looking at the reconstructed PDO going back to the 17th century, which would include several SSN cycles, one will be hard pressed to find a correlation between SSN and the tree-ring/PDO reconstruction. And just to be clear, whatever the Sun is putting out that some are saying is yet unknown, the, shall we say, proxy measure of this unknown (SSN or TSI, or for that matter, any of the other measures known by Leif and others) would at least show some correlation if cycle variations are somehow coupled with weather pattern variation. But there is none to be found, not even a predictable lag.
http://horizon.ucsd.edu/maltmn/sasha/Biondi%20et%20al.pdf
Pamela – Another point to be made is that the PDO curve doesn’t follow a cycle that allows you to predict where it is headed. People are very excited that the PDO has entered a negative phase, so the Earth will begin to cool. However, that graph indicates that the PDO could go anywhere. Just because we’ve entered a negative period, it doesn’t mean it will stay there.
Bob Tisdale,
Thanks for the helpful comments including the detail about accessing and using the climate explorer site.
I have had a look at the OLR response to the 1997 El Nino and other warming events in each 10° latitude band between the equator and the South Pole. The patterns of response (wiggles, in Leif’s terminology) are of interest. Whereas there is a steep fall of OLR in the 0-10°S band as soon as precipitation gets underway there is an increase in OLR in all latitude bands between 10°S and 40°S with the steepest increase between 10°S and 30°S where OLR is always about 8% more than in the 0-10°S band. Beyond that latitude there is no apparent response to warming events at the equator in the OLR statistic.
Logically, the enhanced OLR between 10 and 40°S during warming cycles is due in part to enhanced compressive warming of the air in the downdraft zones (high pressure cells) in direct response to the cycle of uplift and de-compressive cooling associated with enhanced convection over the ITCZ. It may also be due in part to simultaneous or prior warming of the sea between 10 and 40°S that could be associated with cloud loss at these latitudes.
If this enhanced OLR were to impact surface temperature at 10-40°S it could do so in a couple of ways. Firstly, it is likely to be associated with an expansion of the cloud free area. Secondly, if there is a material greenhouse effect from the presence of carbon dioxide it should be enhanced at these latitudes during tropical warming events due to the increase in OLR.
So, it is of interest to discover whether the sea warms at 20-30°S earlier than it does at the equator. Looking at the data the equatorial zone leads 20-30°S on nearly all occasions but the lead is just a month or two. On the other hand, there are particular southern locations that consistently lead the equator and the interval is many months. That is the case with the south East Pacific off Chile. When the ocean warms at this location it is probably in response to change of cloud cover and unlikely to be related to warm pool dynamics or upwelling phenomena. In the South East Pacific atmospheric pressure and cloud dynamics are intimately related. There is an inverse relationship between 200hPa temperature and high cloud cover in a long strip of ocean between Queensland and Tierra del Fuego.
My conclusion is that change in cloud cover is important to both the initiation and evolution of tropical warming cycles. The change in cloud cover occurs in the main, outside the zone of the equator and is tied in with ozone dynamics in the upper troposphere/lower stratosphere. The timing of warming cycles at the equator is given by the very pronounced cycle of 20hPa temperature in the stratosphere that is in turn closely tied to prior temperature change at the poles. This too points to cloud dynamics as the initiator of tropical warming cycles.
One thing that must be recognized is that a lot of the real action is remote from the equator. I believe you have already pointed to this at http://bobtisdale.blogspot.com/2009/04/did-decrease-in-total-cloud-amount-fuel.html.
and also in this comment:
“The Humboldt Current carries waters from the ACC northward in the eastern South Pacific, so changes in Southern Ocean SST anomalies should impact SST anomalies in the tropical Pacific, and, therefore, should enhance or retard the buildup of heat in the PWP.”
This has been a very useful thread. Thanks for your diligent responses, useful suggestions and benevolent critique.
Bob Tisdale (17:50:56) :
tallbloke: I “aligned” your two curves for my benefit.
http://i42.tinypic.com/6zvno8.jpg
I agree with your assessment, “Bob, look again at the side by side comparison, the data is similar though not identical going back from the present as far as the small downtick at 1920 on your graph. All the main shifts line up. The same downtick is at 1908 on my plot of the KNMI data.”
But I disagree with your conclusion, “I think something is seriously wrong with the time scaling on one plot or the other.”
Bob, I really appreciate the time you’ve taken to work on this with me. Before we leave it, please take a look at this second side by side I’ve done with the time scales on our two graphs matched alongside the first one by opening them both in separate windows:
http://s630.photobucket.com/albums/uu21/stroller-2009/?action=view¤t=ninos-datasets2.jpg
http://s630.photobucket.com/albums/uu21/stroller-2009/?action=view¤t=ninos-datasets.jpg
As you can see, the dates now line up on the ‘datasets2’ image, but whereas the data visually agreed reasonably well before, it’s now ‘stretched and shifted’. To me, this seems more like a problem with the data and date collation than a difference between Hadley and Kaplan’s individual monthly temperature readings.
Pamela Gray (18:54:03) :
Anna and others, tree ring data is a far better proxy for PDO than just about anything else people try to use tree rings for.
This is probably true for American tree ring data, and we’d probably find British tree ring data gave a reasonable proxy for the AMO. One thing about trees I learned which interested me is that it’s been discovered that the internal temperature of leaves stays almost constant regardless of ambient temperature. It seems likely therefore that tree ring differences are more to do with changing precipitation than temperature.
Looking at Bob Tisdale’s graphs of temperature versus precipitation the data seems to run counterintuitively to anna v’s warm=wet cool=dry adage, though on different timescales maybe it’s a different story.
http://bobtisdale.blogspot.com/2008_07_01_archive.html
tallbloke: You wrote: “To me, this seems more like a problem with the data and date collation than a difference between Hadley and Kaplan’s individual monthly temperature readings.”
Both Kaplan and HADISST start with the same data, which is derived from ICOADS SST. Kaplan used an early version of Hadley SST data that had been corrected for the Folland bucket adjustment prior to 1945. Kaplan then applied their smoothing and infilling techniques to it.
The UCAR “Informed Guide to Climate Data Sets” on Kaplan is here if you’d like a more detailed explanation:
http://www.cgd.ucar.edu/cas/guide/Data/kaplan_sst.html
The UCAR write up about HADISST is here, but it’s incomplete:
http://www.cgd.ucar.edu/cas/guide/Data/hadisst.html
The Hadley Centre’s description is here:
http://hadobs.metoffice.com/hadisst/
Plotting the RAW Kaplan and HADISST NINO3.4 SST anomaly data shows that ENSO events appear at the same time. There are differences in magnitude, some minor, some major, but the timings of the ENSO events agree. Also note the minor differences in the ENSO-neutral years.
http://i42.tinypic.com/20iawzd.jpg
Smoothing the Kaplan and the HADISST NINO3.4 SST anomaly data with a 121-month filter brings out those underlying minor and major differences:
http://i39.tinypic.com/2lo0oly.jpg
Regards
Bob Tisdale (05:25:22) :
Smoothing the Kaplan and the HADISST NINO3.4 SST anomaly data with a 121-month filter brings out those underlying minor and major differences:
Bob, many thanks again for sticking with me through this, and for all the valuable information you’ve added about the background to the datasets. As an exercise, I’m going to plot both series on the same graph for myself at various smoothing levels and with various median lines and see how it affects the ‘look’ of the data. I’m sure I’ll learn a lot about ‘eyeballing the data’. :o)
erlhapp (18:29:37) :
What is it that accounts for: “variations in the rate of energy flow from ocean to air.”
I think Stephen is taking the weekend off Erl.
My guess is it will have quite a lot to do with night time air temperatures and wind speeds. I came across an old post of Willis Eschenbach’s earlier on CA that would be worth a look. It the last comment on this thread about long wave radiation entering and leaving the ocean surface. Even Gavin Schmidt joins in!
http://www.climateaudit.org/?p=213