Arguments
The hockey stick divergence problem
Posted on 1 December 2009 by John Cook
Tree growth is sensitive to temperature. Consequently, tree-ring width and tree-ring density, both indicators of tree growth, serve as useful proxies for temperature. By measuring tree growth in ancient trees, scientists can reconstruct temperature records going back over 1000 years. Comparisons with direct temperature measurements back to 1880 show a high correlation with tree growth. However, in high latitude sites, the correlation breaks down after 1960. At this point, while temperatures rise, tree-ring width shows a falling trend (a decline, if you will). This divergence between temperature and tree growth is called, imaginatively, the divergence problem.
The divergence problem has been discussed in the peer reviewed literature since the mid 1990s when it was noticed that Alaskan trees were showing a weakened temperature signal in recent decades (Jacoby 1995). This work was broadened in 1998 using a network of over 300 tree-ring records across high northern latitudes (Briffa 1998). From 1880 to 1960, there is a high correlation between the instrumental record and tree growth. Over this period, tree-rings are an accurate proxy for climate. However, the correlation drops sharply after 1960. At high latitudes, there has been a major, wide-scale change in tree-growth over the past few decades.
Figure 1: Twenty-year smoothed plots of tree-ring width (dashed line) and tree-ring density (thick solid line), averaged across a network of mid-northern latitude boreal forest sites and compared with equivalent-area averages of mean April to September temperature anomalies (thin solid line). (Briffa 1998)
Has this phenomenon happened before? In other words, can we rely on tree-ring growth as a proxy for temperature? Briffa 1998 shows that tree-ring width and density show close agreement with temperature back to 1880. To examine earlier periods, one study split a network of tree sites into northern and southern groups (Cook 2004). While the northern group showed significant divergence after the 1960s, the southern group was consistent with recent warming trends. This has been a general trend with the divergence problem - trees from high northern latitudes show divergence while low latitude trees show little to no divergence. The important result from Cook 2004 was that before the 1960s, the groups tracked each other reasonably well back to the Medieval Warm Period. Thus, the study suggests that the current divergence problem is unique over the past thousand years and is restricted to recent decades.
This suggests the decline in tree growth may have an anthropogenic cause. A thorough review of the many peer reviewed studies investigating possible contributing factors can be found in On the ’divergence problem’ in northern forests: A review of the tree-ring evidence and possible causes (D’Arrigo 2008). Some of the findings:
- Various studies have noted the drop in Alaskan tree-growth coincides with warming-induced drought. By combining temperature and rainfall records, growth declines were found to be more common in the warmer, drier locations.
- Studies in Japan and Bavaria suggest increasing sulfur dioxide emissions were responsible.
- As the divergence is widespread across high northern latitudes, Briffa 1998 suggests there may be a large scale explanation, possibly related to air pollution effects. A later study by Briffa proposed that falling stratospheric ozone concentration is a possible cause of the divergence, since this observed ozone decline has been linked to an increased incidence of ultraviolet (UV-B) radiation at the ground (Briffa 2004).
- Connected to this is global dimming (a drop in solar radiation reaching the ground). The average amount of sunlight reaching the ground has declined by around 4 to 6% from 1961 to 1990.
- One study suggests that microsite factors are an influence on whether individual trees are vulnerable to drought stress. Eg - the slope where the tree is located, the depth to permafrost and other localised factors (Wilmking 2008). This paper amusingly refers to the divergence problem as the "divergence effect" so as "to not convey any judgement by the wording" (you wouldn't want to offend those overly sensitive Alaskan trees).
There is evidence for both local and regional causes (e.g. drought stress) as well as global scale causes (e.g. global dimming). It's unlikely there's a single smoking gun to explain the divergence problem. More likely, it's a complex combination of various contributing factors, often unique to different regions and even individual trees.
In past weeks (even predating "Climategate"), there has been a kneejerk reaction among skeptics to assume that climate scientists are conspiring to cover up the divergence problem. Unfortunately most critics fail to acquaint themselves with the actual science before voicing their conspiracy theories. A perusal of the many peer reviewed papers since 1995 (conveniently summarised in D’Arrigo 2008) reveal the following:
- The divergence problem is a physical phenomenon - tree growth has slowed or declined in the last few decades, mostly in high northern latitudes.
- The divergence problem is unprecedented, unique to the last few decades, indicating it's cause is anthropogenic.
- The cause is likely to be a combination of local and global factors such as warming-induced drought and global dimming.
- Tree-ring proxy reconstructions are reliable before 1960, tracking closely with the instrumental record and other independent proxies.
The strange thing is these papers suggest the trees should be experiencing enhanced growth over recent decades not struggling. Maybe there is a flaw in the interpretation of tree ring data?
The Nemani paper shows that growth in northern Eurasia seems to be primarily limited by energy/insolation-- so global dimming would likely have an impact there. Nemani et al's Fig 1 does show an increase in VPD which would decrease stomatal aperture (increase canopy resistance) which would reduce gross photosynthesis-- so it is perplexing that the NPP there has allegedly increased. Their Fig. 1C shows that N. Eurasia has cooled somewhat between 1982 and 1999. So this is clearly a complex problem. Fortunately, we have more reliable proxies for global temperature.
Soil moisture is a limiting factor for tree growth. What is critical is whether or not the regional climate is such that water stress is an issue, certain regions are obviously more prone to frequent drought than others. In fact, you probably know that scientists also use dendrology data as a proxy for drought. That said, how do we know that historical regional droughts did not affect the tree growth in the past?
Anyhow, I'm clearly not an expert, but perhaps someone in the know or John could address these issues. Thanks.
You can't conclude the cause of the decline is anthropogenic, either from Briffka 1998, or Cook 2004, as there is not enough reliable data to make a comparison prior to 1200 AD (most of the MWP) in Cook 2004 with the post 1960 records, and none in Briffka 1998 prior to 1880 to make a comparison with post 1960 records. Your inferance that the cause is anthropogenic does not come from tree ring science.
"For example, reconstructions based
on northern tree-ring data impacted by divergence
cannot be used to directly compare past natural warm
periods (notably, the MWP) with recent 20th century
warming, making it more difficult to state unequivocally
that the recent warming is unprecedented."
For the same reason you can not conclude that "tree ring proxies are not reliable during warmer periods". Here you're assuming that the problem is temperature; again, it's just a possible explanation. Also, you should consider that not all the trees in all locations show divergence, so you should have specified (like D'Arrigo did) tree rings affected by divergence.
But whatever the reason of divergence is, if the reconstruction of MWP gives a temperature lower than the '60 of last century (or whenever divergence starts), there's no reason of concern.
Couple of quotes from Briffa 1998
"It seems clear that a major, wide-scale change has occurred in the ecology of Northern Hemisphere tree-growth and temperature. As yet, the cause is not understood, but a number of factors such as increasing
atmospheric CO2, higher levels of pollutant (i.e. nitrates or phosphates) transport, other changes in soil chemistry or increased UV-B levels might be involved. There is also evidence that increasing atmospheric opacity has resulted in a notable reduction in the amount of solar radiation reaching the earth's surface since the middle of this century (Bradley & Jones 1992). Several factors could be acting together, and possibly interacting with changes in climate itself."
And from the Q+A session at the end
"K. R. BRIFFA. At this time, we do not know why maximum late-wood density declines in relation to large-scale changes (recent warming) in temperature."
There are many possible explanations/suggestions put forward to explain the "divergence problem" but there appears no nailed explanation for it. In most cases the sorts of work that is required plant physiology, field ecology (as well as climatology) is not done by dendrochronologists. Until this sort of work is done it can only ever be speculation. In fact Briffa 1998 lists many possible factor this appears to be speculation (I count nine if you include ones in the first quote).
"potential factors such as changing CO2, O3, and UV-B levels, and widely experienced alterations in the other atmospheric and soil balances, must be considered"
"higher nitrate levels or perhaps tropospheric ozone"
I skimed through your document, and I am not sure what that was supposed to show me.
I guess, I would like to see evidence that there is a correlation between some other climate data (other than temp) that would suggest that Tree Ring data is valid.
1)Fully explain the divergence and do the appropriate compensation to the data. This would maintain the high correlation and high value to the data.
2)Remove the divergent period, another way to put this is "hide the decline". Many would see this as intrinsically devaluing the data. I would struggle with doing this to my own data.
With a failure to fully explain 1) then it appears there is a move toward 2)
John the value of data is an important aspect of science I hope you don't think this is off topic!!
as you correctly say, we all need a full explanation of the divergence problem and this is exactly what scientists are researching actively.
In the meanwhile we are forced to stick to possibility #2 "devaluing" (i'd say put on hold) part of the data, which is all but the most desirable solution.
I hope there's no #3, throw away all the data. This would really be giving data no value at all.
The reason I ask is some of the tree ring proxies might respond non-linearly to temperature as a simple natural matter of fact. If the agreement between tree rings and other proxies is only between the end of the MWP and the beginning of this most recent warm period, then there is a real possibility that the proxy is only valid under a certain (low) temperature range and beyond a cuttoff growth rate becomes a proxy for precipitation and not temperature. If the response to the high temperatures of the MWP was linear as expected, while the response to current high temperatures is not, then there's good reason to think some recent phenomena is causing the divergence problem.
Now, perhaps you can point me to other independent proxies that don't exhibit the decline. I know for sure that speleothems and the Tijlander series (used upside-down) do!
The temperature record is consistent with the broad rsnge of real world warming consequences (sea level rise, sea ice retreat, polar and mountain glacier recession, biologicsl cosequences etc.).
I haven't read any analysis that indicates that there is a problem with "the land based record". What did you have in mind?
Some issues with the instrumaent data
1) there has been a bit of a divergence between the HADCRUT and GISS data, the two main global records, over the past decade. THis has lead to the arguement about whether the last decade has shown cooling.
2) What is presented is not what has been noted down from the thermometer. The data goes through manipulation to account for changes in instrument, position, local changes, urbanisation etc. This process can be contraversial. As an example you can read about how this process upset some sceptical scientists in New Zealand if you wish. http://nzclimatescience.net/index.php?option=com_content&task=view&id=550&Itemid=32
I'm sure I could list more.
As an aside. If people like little gizmo's to play with online then NOAA have graphs on there front page where you can play with a slider. Hours of fun. www.climate.gov
If so, this would, for example, lead one to conclude that the Medieval Warm Period is not well represented in the proxy data.
And if the MWP is underestimated in the proxy data, well, the hockey stick disappears. This is one of the reasons the proxy data was adopted in the first place.
- The divergence problem doesn't occur with all tree-ring proxies but mostly in higher latitude sites with low latitude sites showing less or no divergence
- When you compare high latitude tree-ring proxies that show divergence to the low latitude tree-ring proxies that have no divergence, the two track each other back to the Medieval Warm Period. This indicates the divergence problem is unique to recent decades.
- Most importantly, when you exclude tree-rings and only use other proxies, the same result is found - that the last few decades are warmer than any period over the past 1300 years.
Of course, I explain all this in the post above. But I will make this further point - there seems to be this notion that if the Medieval Warm Period was warmer than now, then we have nothing to worry about. On the contrary, if this was the case, it would be a cause for great concern. It would mean climate is much more sensitive than we currently think. Therefore, the climate response to the current radiative forcing from CO2 will be even higher than predicted. Arguing for a warmer MWP to debunk man-made global warming displays a misunderstanding of how climate responds to energy imbalance.ice cores temperature reconstructions are based on isotopic ratios; sea and lake sediment reconstructions use some biological proxy; neither is precipitation dependent.
Trees, ice sheets and sea typically happen to be in different locations; you may find some lake sediments and tree rings from the same location.
The solution of what is mathematical more than observational?
It's not clear to me what you're asking.
Yes, ice cores measure isotopic ratios, but the deposition of ice and the entrapment of those isotopes depends on a certain minimum level of humidity in the area. Too little humidity and there won't be any deposition, and in extremely arid conditions, ice can even sublime, along with its entrapped isotopes.
Sedimentary deposits also depend on a certain level of water movement to cause erosion and transport of silt and other materials.
All I'm saying is that there is some difference between the experience of low latitude trees and high latitude trees to cause the divergence problem in the proxy record.
If the difference is not temperature, what is it? Could it be humidity/precipitation? Could it be cloud cover? Something else?
I haven't seen any papers trying to find a real physical explanation, only re-normalizations and adjustments to old data to bring them into line with the instrumental record.
right, the annual thickness of ice definitely depends on precipitation while keeping the same isotopic ratio. This is the very basic of temperature reconstructions from ice cores.
You ask what may cause the divergence problem; there's the whole dendroclimatology community asking itself the very same question. Unfortunally, no one came up with a solution yet; indeed, when temperature stops being the limiting growth factors, many other may come into play, including local environment. It's not rare that only some trees in a region show divergence while others do not.
I'd suggest reading the review (D'Arrgigo 2008) linked by John in this post.
P.S.
There's no way to use tree rings as temperature proxy without intesive data analisys, re-normalizations, "adjustments" and the like.
No, quite the opposite. McIntyre was aware of the divergence problem. Actually he has a commented on ipcc 2007:
"Show the Briffa et al reconstruction through to its end; don’t stop in 1960. Then comment and deal with the “divergence problem” if you need to. Don’t cover up the divergence by truncating this graphic. This was done in IPCC TAR; this was misleading. (Reviewer’s comment ID #: 309-18)"
It was rejected: "Rejected – though note ‘divergence’ issue will be discussed, still considered inappropriate to show recent section of Briffa et al. series."
Can anyone explain me, why it is 'still considered inappropriate ' to show a graphic in a honest way?
http://www.realclimate.org/index.php/archives/2007/05/the-weirdest-millennium/#comment-34017
Well and good. I believe this thesis is established. But what I would like to see and do not see, is proof that similar divergences could not happen in the past. Simply proving that today's divergence is man-made does not prove this, though the suggestion is rather strong.
At some point, the skeptics will think to ask this, once they figure out there are not making much headway with their false accusations of cover-ups. So it is important to do.
The hard part, of course, is figuring out what kind and level of proof is even possible. I don't know trees well enough to come up with helpful suggestions there.
as far as i know the final word on the divergence problem has not been said. Anyone is open to the possibility that the same problem showed up in the past but at the moment there's no evidence of it given the overall agreement of many different proxies.
The most logical explanation would seem to be that the relationship between tree ring width and temperature changes above a certain temperature level - the level experienced in the 1960s. That means that tree rings are a useful proxy for temperatures up to that level. It also means that tree rings are not useful beyond that level - i.e., it means that when it's warmer than it was in the 1960s, it's not going to show up in tree rings, thus the fact that tree rings don't show higher-than-1960s-temperatures during past periods doesn't mean it wasn't warmer during those periods.
And when you take the tree rings out of the equation, voila, the pre-hockey-stick climate history re-emerges.
Does the inability to rewrite the climate history preclude man's involvement in the climate changes that have occurred over the last century?
No.
But the effort to rewrite the climate history adversely affects the credibility of those arguing that man has been deeply involved in those recent climate shifts.
So why go through the effort?
Tree lines were higher in mountain ranges around the world, including the Sierra Nevadas and Alps. The droughts were so severe in North America that they were a factor in forcing the Anasazi to abandon their elaborate cliffside dwellings. American Indian legend holds that the buffalo migrated far to the north to richer grassland. The Vikings sailed the North Atlantic in wooden boats on ice-free waters (that would during the 14th century become choked with icebergs - i.e., from sea ice breaking off much like today) and maintained a colony on Greenland - and the settlers' diets were for 200 years 80% land-based. They maintained vineyards in England (and the fact that after 900 years of breeding new varieties of grape and improving technology for cold-hardiness British wine production has reemerged is really not a counterpoint). Olive trees in Cologne. Glacial retreat in the Alps is revealing archaeological finds from the MWP - items left behind by traders using mountain passes only now resurfacing. Lake Naivasha in Kenya dried up for 200 years. Further evidence has been found in Mongolia, Japan, the Arctic and Antarctic. And don't forget the countless contemporaneous observations. These have not been explained away as having occurred for reasons other than warmer temperatures, and they cannot be dismissed as "merely anecdotal evidence." These examples all occurred during the 1000 AD - 1200 AD timeframe. That there were a few cold years in between is about as relevant to the "geosynchronous" nature of the MWP as the fact that the United States had its third coldest October on record last year is relevant to geosynchronous warming now.
Even if some combination of boreholes and other indirect evidence indicates an MWP slightly cooler than late 20th century temperatures, we still have an MWP plateau that lasted for a few centuries within a few tenths of a degree of the peak modern warmth to date.
Rather than try to downplay this or write it out of the climate history altogether, why not simply accept it and try to make the case that the 20th century warming is different because its suspected cause is something that won't reverse itself, which means that it will continue? Why not also point out that since what caused the MWP isn't fully understood, there's every reason to expect that it could reoccur - and that the combined effect of increased CO2 plus whatever naturally caused the MWP would be severe? You might be less able to make a case that we're only a few years away from a "point of no return" but nobody believes that and it's a risky argument anyway - when those few years pass but the point of no return doesn't, it will simply fuel more skepticism.
There are a lot of skeptics who would be more open minded about AGW if it weren't for some of the overstatements and misstatements made by the more zealous of AGW supporters, including the efforts to rewrite the climate history. The Hockey Stick has done the AGW proponents more harm than good - when you mess with the established history, you create skeptics of your theories as to the present and future.
Tree ring width is affected by many factors besides temperature. Tree growth is obviously affected by many factors besides temperature - predators, sunlight, water, etc.... You can throw up these other remote possibilities but the obvious explanation is that the linear relationship between tree ring width and temperature breaks down above certain temperatures, which means that the absence of wider tree rings during a given period doesn't mean it wasn't warmer then. Why go to all the effort to string together an alternative thesis just so you won't have to budge on the "hockey stick" when you don't need the hockey stick? Sometimes you're wrong. It's better for your credibility to admit it and move on. It's clever, it's lawyer-like, to come up with "well the divergence must be anthropogenic too" but it's transparent. Don't risk losing credibility over a sub-issue that isn't necessary to make your larger point - - don't lose the forest for the trees!
- The hockey stick is not dependent on tree-rings. There are a number of proxies independent of tree-rings that give the same result.
- The hockey stick is not the smoking gun proof of human caused global warming. It shows past temperature change, not what's caused it. Sure, it's suggestive and visually persuasive but from a scientific point of view, it does not constitute proof.
- The reason tree-rings are considered reliable before 1960 is because there is close correlation between tree-ring proxies and the instrumental record for the periods where the two overlap (Briffa 1998).
- Personally, I have no emotional investment in the hockey stick. If it turned out it was erroneous and past climate change was greater than currently thought, this would mean climate is more sensitive than we currently think. This means the climate response to the extra heat trapped by CO2 will be even greater.
- Lastly, the peer review science doesn't say the divergence MUST be anthropogenic. It says the "decline in tree growth may have an anthropogenic cause". As the cause is yet to be definitely defined, it's speculation at this point. Similar to the hockey stick, it's not hard proof but it's suggestive.
This is not a case of "hockey stick" versus "not a hockey stick". This is a case of considering all the various proxies together and determining what is the most accurate temperature reconstruction.2) None of the direct evidence for the MWP has been addressed (except for realclimate's thesis that, because after 900 years of breeding grapes for cold-hardiness and advances to growing techniques, they now again have vineyards in England, somehow that means that it must be as warm now as it was when they had vineyards 900 years ago). We're talking about acknowledged history from nearly every corner of the world. If someone were to tell you that based on examination of the Delaware River or the soil underneath it or some other indirect physical proxy, Washington and his men couldn't have crossed it with their horses and gear that night, would you accept this without an explanation as to how it is that the next morning they were on the other side?
3) Right - tree rings are reliable within the temperature range experienced between 1880 and 1960. Above those temperatures they're not reliable - the most obvious explanation is that once you get above the temperatures experienced in 1960, the correlation between tree ring width and temperature falls off. There are other possibilities but there's one clear, obvious explanation that you really have to want to reject in order to reject it.
The mental gymnastics engaged in to defend the notion of an anthropogenic explanation for the tree ring divergence is a prime example of what I'm talking about. You fall back on logical contortions to defend aspects of your thesis that could be discarded without jeopardizing your main point. This tendency turns critical thinkers into skeptics.
If you just said yep, the MWP was warmer or comparably warm for 200 years but the CO2 we're emitting won't go away and so eventually we're going to have a problem, especially if whatever caused the MWP happens again, I'd be inclined to believe you. There are a lot of skeptics who are skeptics primarily because of this and other needless exaggerations (such as "it's still getting warmer" rather than "it's still warm and it shouldn't be given that natural forcings are have been on the cool side for 3-4 years"). I think the overzealous AGW proponents do their cause more harm than good.