Correlation, Causation and Climate Change

By William K. Olsen | On Feb 20, 2012 | No Comments | In blog

So Colorado Gov. John Hickenlooper, geologist by training, expresses skepticism about Global Frying. And just in case we were not paying attention, according to this piece, Colorado is “…one of the states most affected by climate change.” The kid reporter goes on to state: “That year, approximately 100,000 spruce trees a day were killed by a spruce beetle infestation spurred by warming temperatures in Colorado.”

With my attention snared, and after having safely discovered the location of my snow shovel, buried and nearly melted under the newly fallen hot snow, simple math was in order. Let’s assume a mean stocking of 200 trees per acre, which at 100,000 dead spruce per day equals 500 acres of mortality per day, equating to 182,500 acres in a calendar year.

Let’s also discount for now my own observation — that for some years I have watched localized areas of Engelmann spruce defoliate each summer, starting with new foliage and subsequently said needle loss sometimes extends downward along the twigs and branches by autumn. Depending on location, the defoliation is complete and culminates in mortality, sometimes on a very large scale. In other locations, foliage loss is limited strictly to the current year’s needles and the spruce fully recover in the next growing season. Sometimes but rarely there is a weak association with the presence of beetle, but spruce beetle is not a defoliator; it is a bark beetle. Some entomologists, and this forester, have observed a symbiosis of sorts between mountain pine beetle (MPB), which typically and traditionally attacks lodgepole pine and ponderosa pine, with spruce beetle (SB), wherein the MPB attacks the lower spruce boles and SB attacks at higher points on the trees. Regardless, during casual walk-throughs, it has often been difficult to find evidence of bark beetle attacks on the defoliated trees. I am a forester, not an entomologist. I have yet to find a satisfactory explanation for the described defoliation and subsequent mortality. The most viable explanation, spruce budworm, is easily nixed because they are not present and the needles are not consumed, and instead simply fall off fully intact. The phenomena is striking because the forest floor is characterized by extensive and deep needlecast smack in the middle of summer.

The thinkprogress author linked to a 2011 New York Times piece as an authoritative citation for the spruce mortality rate. Mr. Rudolf of the NYT, in turn, while centering on MPB mortality in lodgepole pine forests, takes a brief hard left to mention spruce beetle:

“In 2010, an estimated 100,000 spruce trees a day were toppled by a massive spruce beetle infestation in Colorado, according to a study of the state’s forests released earlier this month.

The 2010 outbreak covered nearly 210,000 acres of spruce forests in Colorado, nearly double the size of the affected area the previous year. The state also suffered the loss of hundreds of thousands of acres of lodgepole pine and aspen forests.”

Here, 100,000 trees per day equals 36,500,000 dead spruce in a year across 210,000 acres, or 174 trees per acre. That is certainly within range of my casual 200 trees-per-acre estimate on the back of a napkin, above, particularly where mortality is incomplete and incremental over time. In his 2011 piece, Mr. Rudolf cited the annual aerial insect and disease survey at the Colorado State Forest Service site. Unfortunately, this is a “Feature Stories” page, and the 2010 report so cited is replaced now by notification of the release of the recent 2011 report. A search for the 2010 Colorado State Forest Service (CSFS) report yields this. The 210,000 acre dead-spruce figure cited by the NYT is close enough to the 208,000 acres reported on page 2 of the report.

Just to be clear, aerial surveys are just that, from the air. As the CSFS report writers caution with their survey map:

Many of the most destructive diseases are not represented on the map because these agents are not detectable from aerial surveys. The data presented on this map should only be used as a partial indicator of insect and disease activity and should be validated on the ground for actual location and causal agent. Shaded areas show locations where tree mortality or defoliation were apparent from the air.”

Thus I HGH am skeptical of the 208,000-acre figure for a beetle-killed spruce impact in 2010. Those defoliated spruce mentioned above? I walked circles around many trees, chipped at the bark, and cut twigs and needles apart for clues to the needle loss and mortality. There was simply no correlation between the existence of spruce beetle and/or spruce budworm, and needlecast combined with tree mortality. I am not criticizing the CSFS survey work, and spruce beetle are certainly boring away in Colorado. Rather, all reported numbers should be questioned because otherwise we are led to draw quick conclusions that may be incorrect. CSFS pointed this out, correctly and professionally.

This simple fact-checking excursion was done to demonstrate that cause and correlation are different concepts. Yes, there are dead spruce visible from the air, and on the ground it is known that spruce beetle are attacking and killing spruce. This is the correlation. But the actual cause of the observational or mensurational changes or differences may differ very much from surmised correlations.

Let’s complete the loop. The skepticism of Gov. Hickenlooper is wise. He is questioning the dogma of global warming. It is a grand leap to move from data and climate models to acceptance of global warming as fact. It is a further excursion to claim global warming is here – and from afar claim that Colorado is “…one of the states most affected by climate change.”

Add the final orphaned statement already quoted above:

“That year, approximately 100,000 spruce trees a day were killed by a spruce beetle infestation spurred by warming temperatures in Colorado.”

The phrase “spurred by warming temperatures in Colorado” is an inferential triple-jump: trees are dying in Colorado because the beetle population has increased because temperatures are warmer because we are spewing CO2.

Referring back to the thinkprogress piece, Gov. Hickenlooper seems to be experiencing a great turnaround from his previous stance in 2009. More importantly, the author cites spruce mortality as the glaring example of the impacts of global warming in Colorado, even though the areal impacts of mountain pine beetle in lodgepole pine forests have been four times greater than that in spruce-fir forests. Fear not, global warming is persistently reported as the cause of lodgepole pine mortality in Colorado as well.

Here’s the problem. We allowed lodgepole pine forests to become severely overstocked by creating, through inactive management, millions of acres of beetle bait at unprecedented levels across an unprecedented extent. Then the beetle show up just when stands are most susceptible, with mean diameters of 7 inches and larger,  very high stocking levels, poor vigor and suppressed growth. We created beetle gastronomical heaven, with gravy and dessert, across the landscape and then the distant and disconnected fear mongers wish to claim that global warming caused the epidemic in one of the most-impacted states in the nation. Meanwhile, us know-nothing foresters and entomologists have been warning for decades that without anthropogenic intervention, i.e. active management to reduce beetle susceptibility, it would be only a question of when the epidemic begins, not if.

There is a common claim that it has not been cold enough to kill mountain pine beetle larvae during the winter months, a consequence of global warming. The problem is that lab tests can determine what degree of cold will kill larvae in a controlled environment, which is rather distant from real-world conditions within tree boles and behind bark. Research I have been involved in fails to reveal differences in larvae mortality rates across the north and south sides of attacked trees where insolation effects could otherwise be hypothesized as being detectable. The primary motivator behind cold-killing inferences centers on the correlation between the end of a Colorado spruce beetle epidemic in the 1950s, and a cold snap the same winter. One could reasonably speculate that beetle population collapse followed an opportunistic virus attack of spruce beetles, or the beetle simply ran out of food, and either scenario could have coincided with the cold snap.

Correlation is not causation.

This nonsense drives one nuts.

And so do those who claim they know better through disingenuous citation and then tell us how to manage our natural resources.

Bill Olsen is a private consulting forester who lives and works in Colorado. Write to him at

Written by William K. Olsen

Leave a comment

Your email address will not be published. Required fields are marked *

Sign-up for email updates!

Want the latest news from Monkey Wrenching America? Sign-up below:

News from around the web