|Make a Smilebox slideshow|
|Make a Smilebox slideshow|
I made the slideshow above so that I could see if cirrus cloud distribution behaved signifcantly differently during the large El Nino of 1997/98.
Below is a graph showing the progression of the El Nino in the Pacific.
In the Pacific, it is apparent that there was (for the season) a strong cloud-free zone developing in the run up and very beginning of the SST rise. During the middle of the rise, some problem with the collected data makes it difficult to see what the trend is, but it appears to be average. During the end of the rise, the peak, and then the fall, there is an overwhelming amount of cirrus cloud distributed more South than would normally be seen during that season. So, there may be evidence for a relationship between SST (sea surface temp) and cirrus cloud distribution, though the largest portion of of rise in SST occured with average cirrus cloud cover. This would require a lag between changes in cloud cover and changes in SST, and this requires more work to determine if this is the case. Also, I need to spend more time looking at years with no La Nina or El Nino in order to determine the extent to which cirrus cloud distribution fluctuates regardless of tropical warming/cooling events. If the distribution remains extremely stable, except during El Nino/La Nina events, we would have more evidence to pursue the link between clouds and SST. However, if distribution varies significantly on its own, it will be very hard to pursue to the possible connection.
Additionally, I will look at what cirrus cloud cover does during La Ninas. Unfortunately, the data starts in July 1983, and there have not been many strong La Ninas since then.
|Make a Smilebox slideshow|
The above slideshow is intended to show how cirrus cloud distribution changes each year. I am chieflyconcerned with the tropical regions.
Cirrus clouds have a yearly path Northwards and then Southwards. From January to March, cirrus clouds are maintained South. From April to June, they move North. From July to August, they are maintained North. And from September to December, they move South. This annual pattern has a significant impact on cirrus cloud-free zones in the Pacific, Indian, and Atlantic oceans. As the cirrus clouds move Northward, regions near the equator in the Pacific ocean become very cloud-free along with the coast of S. America just South of the equator. A cloud free zone forms slightly south of the equator in the Atlantic, by Africa. And from October through April, the strong cirrus cloud cover over the Northern Indian Ocean (stretching outwards into the Pacific) dissipates, becoming less intense and less organized.
With these annual trends in cirrus cloud distribution, we are led to a question concerning our postulation that decreases in cirrus cloud cause El Ninos. Does it matter that inter-annual fluctuations in cirrus cloud cover are larger than changes during El Ninos? But that is a question to be asked down the road. Right now, I am focussing on determining if changes in cirrus cloud cover accompany tropical warming events.
One assertion made by Erl Happ’s theory of climate is that changes in cirrus cloud cover create tropical warming events (El Ninos). Cirrus cloud is a cause of albedo, reflecting sunlight back into space. Therefore, less cirrus cloud would lead to a warming event, while more cirrus cloud would lead to a cooling event.
After a lot of searching, I finally found this site: http://isccp.giss.nasa.gov/products/browsed2.html.
It is run by the International Sattelite Cloud Climatology Project, which describes itself:
“The International Satellite Cloud Climatology Project (ISCCP) was established in 1982 as part of the World Climate Research Program (WCRP) to collect weather satellite radiance measurements and to analyze them to infer the global distribution of clouds, their properties, and their diurnal, seasonal and interannual variations. The resulting datasets and analysis products are being used to study the role of clouds in climate , both their effects on radiative energy exchanges and their role in the global water cycle. ”
I have two pictures to show you, and they have something in common.
My point in showing the second image is that the behavior of cirrus clouds in the Pacific during October of 1996 is identical to changes in temperature seen in the Pacific during an El Nino.
This similarity between cirrus cloud and temperature trends (with the 97/98 El Nino) illustrates that cirrus cloud cover and sea surface temperature might be closely related. Yet, we can go a step farther. Below is a graph of sea surface temperature in the region in the Pacific that I have used before and with no smoothing. This graph should show the first indications of an El Nino.
Note that the El Nino does not start until the first month of 1997. This means that the decrease in cirrus cloud cover that was so consistent with sea surface temperature trends during El Ninos occured three months before the El Nino began!
The big question to answer now is this: Was the decrease in cirrus cloud responding to atmospheric changes related to the El-Nino that preceded the actual sea surface temperature rise? Or was the El Nino responding to the decrease in cirrus cloud?
Erl Happ’s theory is truly a theory of climate and not just “global warming”; it is a way of describing terrestrial temperature trends, and it does so very well. Some steps of the theory require more evidence than others, so I’ll continue posting as I try to confirm the various steps.
After looking at more months and more years, it seems that the region in the Pacific of varying cirrus cloud cover always holds the same shape (with decreased cirrus cloud) as an El Nino SST trend does for some circulation-related reason during the months July through January. Here’s a graph of the mean cirrus cloud cover for October (1983-2006).
With that said, my Oct 1996 map does hold somewhat more of a clear El Nino shape, though not significantly. This does not mean that there is no correlation; it just means that this is more complex than I assumed when I made the post because I will have to take into account monthly means when looking for an El Nino signature.
Due to a lack of material on the internet concerning how El Nino and La Nina events manifest themselves in the Indian Ocean, I extracted data from the NOAA site mentioned in my previous post. I broke up the tropical Indian Ocean into 9 blocks in order to find where the signal from a El Nino or La Nina first appears.
Listed below are the regions used.
|Region 1||4 to 12||52 to 66|
|Region 2||-4 to 4||52 to 66|
|Region 3||-12 to -4||52 to 66|
|Region 4||4 to 12||66 to 80|
|Region 5||-4 to 4||66 to 80|
|Region 6||-12 to -4||66 to 80|
|Region 7||4 to 12||80 to 94|
|Region 8||-4 to 4||80 to 94|
|Region 9||-12 to -4||80 to 94|
I extracted data for each of these regions, graphed them, and found that the Western Indian Ocean was the first to react to the 1997/1998 El Nino event, followed by the Central Indian Ocean, and finally the Eastern Indian Ocean. The region reacted simulataneously, regardless of latitude, although latitude did effect the intensity of the variations in sea surface temperature.
Below is a graph of the Western, Central, and Eastern reaction to the El Nino.
And here’s a graph of the Western Indian Ocean at different Latitudes.
And just to show that this isn’t unique to the 1997/1998 El Nino:
So, after this discovery, I went back to compare my Indian Ocean record in the last post with this new Western Indian Ocean record. Remember that I wasn’t looking for an accurate description of an El Nino/La Nina event; I was looking for the region that first expressed an event, which turned out to be the Western Indian Ocean.
Because this analysis is so dependent on determining the exact time that the El Nino starts, I decided to forgo the 12-point mean smoothing that the website offered and that I had been using. Instead, to rid the dataset of the influence of seasons, I calculated each month as anomoly from the average temperature of that same month from 1978 through 2007.
So, when I went to compare my new Western Indian Ocean dataset to my old, smoothed Indian Ocean dataset, I was surprised to find that my old dataset actually responded to the El Nino first. Out of curiosity, I applied my new data handing technique to the Pacific temperature box that I used in my previous post, and found that my smoothed Pacific preceded my new Pacific, too.
The new, un-smoothed data shows a much smaller lag-time (perhaps only a month) than the old, smoothed data. This may mean that the apparent inconsistent lag time between the Atlantic and Pacific Nino is an artifact of smoothing rather than an actual phenomena. I’ll need to revise my last post to use the un-smoothed data.
UPDATE (10/06): As I have written in more recent posts, this analysis is flawed. In fact, as I have begun to understand cirrus cloud behavior in the tropics better, I’m beginning to find this analysis unnecessary. Soon, I will make a post that will hopefully conclude my studies on cirrus cloud in the tropics.
A component of Erl Happ’s theory of climate change, which I will eventually post on, is that El Nino and La Nino events are not internal oscillations. He claims that these tropical warming events are caused by changes in tropical albedo, which is caused by a change in 200 hPa cirrus cloud cover, which itself is caused by changes in solar activity.
So if tropical warming events are due to an increase in the amount of UV radiation reaching the tropics, then the tropics should respond globally – not just in the Pacific.
Using regions defined as the first place that ENSO events become apparent in the Pacific, Atlantic, and Indian Oceans, I retreived Sea Surface Temperature (SST) for each of the oceans. (Source: http://nomads.ncdc.noaa.gov/. See this post for instructions.)
Indian Ocean: Lattitude: (-5, 5) Longitude: (60, 94)
Atlantic Ocean: Latitude (-5, 3) Longitude: (-15, 5.5)
Pacific Ocean: Lattitude (-5, 5) Longitude: (-132, -82)
Once again, these are not the entire areas where El Nino/La Nina events occur in the tropics, nor are they the entire areas where heating from a decrease in tropical albedo should occur. These are the regions that the El Nino/La Nina signal first becomes apparent in each ocean. The point of this is not to accurately describe the tropical warming events, but to determine when they each begin.
Here is a graph of each of these regions from Jan 1978 to Aug 2007.
To more clearly see the different response times in the oceans, here is a graph of the 1997/1998 El Nino event as exhibited in the three oceanic regions that show the first response to an El Nino/La Nina event.
For this tropical warming event, the Atlantic saw a 7-month lag behind the Pacific, and the Indian Ocean saw an 8-month lag behind the Pacific.
This large lag between the Indian and the Pacific ocean appears to hold relatively well during the period 1978-2007. However, the lag between the Atlantic and the Pacific is less clear. In fact, since 2000 it seems that changes in the Atlantic have preceded changes in the Pacific, though the Atlantic seems to have missed the past year’s large La Nina.
In 1982, the largest recent El Nino event occured, though its effects on temperatures were dampened by the eruption of El Chichon that same year. In this case, the Indian Ocean saw about the same lag, though the Atlantic Ocean lagged by almost two years!
So what does this mean? Seemingly, the Indian Ocean exhibits warming with a consistent lag time (regardless of the intensity of the El Nino/La Nina). Does this suggest that warming in the Indian Ocean is only caused by warming in the Pacific? Or is it still possible that changes in albedo are also impacting the sea surface temperature in the Indian Ocean, and that it merely takes longer for this warming to express itself than in the Pacific? If so, why would this be the case?
It also seems that Atlantic Nino events do not consisently lag behind Pacific Nino events. For smaller events, Atlantic warming might actually precede the Pacific warming. And even if small events in the Atlantic do not actually precede small events in the Pacific, they do not lag behind. But for larger events (like the 1982/3 El Nino, the 1997/8 El Nino, and the 2007/8 La Nina), they lag up to two years behind the Pacific. If the Atlantic Nino is actually preceding the Pacific Nino for small events, then conventional wisdom about the cause of El Nino/La Nina events is wrong, and Erl Happ’s theory stands a chance. Yet, with such large independent oceanic variation, it is difficult to actually say if this is actually the case.
I’ve emailed Erl Happ to give him a chance to respond to my findings. After that I’ll post a link to our discussion on the CA forum dedicated to this subject.
I’ve spent a lot of time tracking down and graphing this data. It’s part of my attempt to make a record of all terrestrial climate data since 1900.
Below are oscillations that I managed to find on a very long time scale.
(D’Arrigo, R., et al. 2005. Pacific Decadal Oscillation Reconstruction. IGBP PAGES/World Data Center for Paleoclimatology Data Contribution Series # 2005-020. NOAA/NGDC Paleoclimatology Program, Boulder CO, USA. ORIGINAL REFERENCE: D’Arrigo, R., R. Villalba, and G. Wiles. 2001. Tree-ring estimates of Pacific decadal climate variability. Climate Dynamics, Volume 18, Numbers 3-4, pp. 219-224, December 2001. )
The data that I have been using is the work of Dave Thompson, and is being used by the Joint Institute for the Study of the Atmosphere and Ocean (JISAO) between the University of Washigton and NOAA. Unfortunately, the data ends six months in to 2003. NOAA, though, separately maintains AO data that begins in 1950 and has continued through th first four months of 2008. For some reason, the JISAO datset lags NOAA by exactly one year, but once that is adjusted for, the two datasets fit nearly perfectly.
The large jump in the JISAO index in 2003 is not apparent at all in the NOAA dataset. In the JISAO index, 2003 contains only six months, and although the first six months of NOAA don’t match up to JISAO, it may be that the massive monthly variability (see previous post) has effected JISAO data due to a lack of datapoints in 2003.
For this same reason, I left out 2008 from the NOAA dataset. Altough, at this point in the year, it is trending upward.
So here is the combined record of JISAO (1900-2002) and NOAA (2003-2007). I’ve also included a polynomial in the graph to highlight the difference between the trends in this graph and the trends in the graph presented on the JISAO site that I’ve been discussing.
And here’s the graph presented on the JISAO site once more, for comparison.
If you’re researching the Arctic Oscillation, you’re bound to come upon this site: http://jisao.washington.edu/ao/.
The first graphic presented on the website is the one I posted in my previous post.
Along with it, the description: “Fluctuations in the AO can be seen in the time series of SLP anomalies for the North Pole”. SLP stands for sea level pressure. The paper referenced for the graph is listed as “Hodges, G., 2000. The new cold war. Stalking arctic climate change by submarine. National Geographic, March, 30-41.” Athough I could not gain access to the article, the abstract reads, “In an attempt to better understand changing climate conditions in the Arctic, where water temperatures are rising and ice cover is both thinning and receding, the American navy has made nuclear submarines available to scientists to help them conduct their research in this inhospitable and remote environment.” I’ve never seen data (or in this case “screw the data, we’re makin’ a graph!”) taken from a National Geographic article, and I am very suspicious.
The graph appears to show a steep trend towards lower-than-normal SLP over the past few decades. Before then, there doesn’t appear to be much variance. (In fact, if you look closely, you can see MBH98 with the Medieval Warm Period just below “normal” in the 1920s and with the mid-century global cooling scare in the middle of the 1990s.) No comment is made about why the Arctic Oscillation looks more like a hockey stick than an oscillation. The intention of the original author was to illustrate the effect of climate change on the Arctic, and perhaps that was also the intention of its use on the site.
Alone, this situation of using a graph from a National Geographic article, implying that the Arctic Oscillation is being significantly altered by climate change without ever directly stating it, would be confusing. But once you throw in the fact that the site then provides monthly data from David Thompson that does not seem to match up to the Hodges 2000 data, things get fishy. Why did they show Hodges data, but not Thompson’s? Thompson’s data, graphed both with monthly anomolies and yearly averages, is shown below:
So, I have a few questions.
Why is Hodge’s data being used at all? National Geographic? How was it smoothed? Why do they use his graph, without access to his data or his ascetic massaging of the data, when they have the raw data of Thompson on the same website? Why did they decide to use Hodge’s data over Thompson’s data? Should they not have provided an explanation or used both?
I doubt it’s intentional manipulation to prove a point; it’s just reflective of an attitude indifferent to accuracy.
There are many projected effects of an increase in GHG content under the assumption of a high climate sensitivity. One of which is a cooling lower stratosphere, a topic that I addressed in a recent post. Today, I ran into this we-knew-it-would-happen-all-along after-the-fact claim projection of climate models – an altered Arctic Oscillation with higher atmospheric pressure over the North Pole.
The claim is awkwardly outlined here at a University of Washington site: http://www.washington.edu/newsroom/news/1999archive/12-99archive/k121699.html.
The graph below is of atmospheric pressure over the North Pole, an indicator of the Arctic Oscillation. It is taken from from Hodges (2000).
My first reaction is how in the world do we know atmospheric pressure over the North Pole in 1900? I’ll have better informed comments on the issue soon.