January 11, 2006

Jumping To Conclusions: Frogs, Global Warming and Nature (Revised)

Filed under: Extinctions

(This is a revised version of our original posting. It is been changed to correct for a misinterpretation of our original reading of the Pounds et al. Nature article. Please see the * and the footnote at the end of the article for more details. Our original conclusions remain unchanged.)

Me and my apparently few friends have been ragging on the review process at Nature for some time, which was once the world’s most prestigious science periodical for all subjects. While it still may be the best for certain biochemical and genetic topics, it surely has lost it on global warming.

My antennae went up on this one in 2003 when my colleague, Robert Davis, and I submitted a paper to Nature showing that, as our cities have warmed, heat-related mortality declined significantly as people adapted to the change. They declined to even send it out for review; but after it was accepted in International Journal of Biometeorology it was awarded “paper of the year” by the Climate Section of the Association of American Geographers. Something is clearly amiss.

Nowhere is that more clear than in a paper, “Widespread Amphibian Extinctions from Epidemic Disease Driven by Global Warming,” by J. Alan Pounds, that appeared in their January 12 issue. We’ll put it simply: with regard to global warming papers, the review process at Nature is dead. Gone. Kaput.

As a concrete example, assume that Nature’s editors had sent me this manuscript for peer-review. Here’s what I would have responded.

“Thank you for asking for my professional opinion of the Pounds et al. manuscript. It suffers from a number of severe analytical problems, scientific overreaching, and clear political bias. Publishing this paper will severely harm the credibility of Nature.

The title of the manuscript, “Widespread Amphibian Extinctions from Epidemic Disease Driven by Global Warming” implies that the authors have proven a pervasive link between a large number of toad and frog extinctions and warming climate. They have done nothing of the sort.

The paper describes extinctions in Central and South America caused by fungal infection caused by a class of organism known as chytrids. The seminal paper describing these extinctions is from 2005, in another journal, Biotropica, published by La Marca et al. La Marca is the seventh author in the list of 14 listed on the Nature paper.

According to La Marca (2005), most of the toad and frog extinctions took place between 1984 and 1996 in the regions studied in the current paper by Pounds. This was shortly after the first discovery of the chytrid fungus in the region, which is described by Lips et al. in 2003 in the Journal of Herpetology. According to Daszak et al. (2003) in the journal Diversity and Distribution , the chytrid fungus was most likely introduced by humans, possibly by ecotourists and/or field researchers (Daszak et al., 1999).

It has been known nearly a half-century (see Charles Elton’s 1958 book, The Ecology of Invasion by Animals and Plants) that the introduction of exotic species produces genetic pandemics over a broad range of climates. The concurrence of human introduction of the chytrid fungus and amphibian extinctions cannot be ignored.

Temperature changes observed over the period of disappearance (1984-1996) were on the order of a half-degree. Yet the range of the daily maximum temperature for amphibian populations studied by Pounds et al. ranges from 34˚C to 12˚C.

Figure 1 is taken from the Pounds et al. manuscript. It shows the distribution of amphibian populations as a function of daily average maximum and minimum temperatures. It also shows the original range of the chytrid fungus.

Pounds et al. cite climate data from Columbia and Venezuela showing a decline in daily maximum temperature of 0.6˚C between the 1941-70 average and the 1981-90 average and a rise in daily minimum of 1.0˚C for the same period.

We have imposed these changes upon the Pounds et al. data in Figure 1. The number of new populations possibly effected by chytrid infestations as a result of the decrease in daily maximum temperature is two and it is four for the increase in daily minimum. There are 50 populations, meaning that 12% more of the populations are subject to chytrid infection.*


Figure 1. The effect of climate change in Central and South America, according to Pounds et al. (2006) is to slightly expand the range where chytrid infestations could take place. The original range is defined by the blue dashed lines, while the new range is illustrated by the light red box. In all, there are about an additional 12% more frog and toad populations that now fall within the chytrid thermal infestation range. Even if all these additional populations succumbed to chytrids, this value is far beneath the 69% or so loss identified by Pounds et al. Obviously there are much bigger factors at play than simply a small change in the climate.

This is the worst-case scenario using their hypothesis about climate change. Yet they claim 69% species lost. Obviously the “widespread amphibian extinctions” are being driven by something other than climate change (again a reference to Elton would be appropriate!).

The climatic hypotheses in this paper are inconsistent, incomplete, and untested. (I presume that other reviewers have noticed this; if they have not, you need to question your selection of reviewers).

The authors cite a warming of nearby ocean waters as driving local warming, but they find that daytime temperatures are in decline while night temperatures are rising. What causes the warming to be preferentially divided into the night, with an actual daytime cooling? Pounds et al. suggest that this is a result of an increase in cloud cover related to global warming (and, in particular, oceanic warming). This is a testable hypothesis. It was not tested.

Figure 2 is global cloudiness as measured since satellite records began, taken from the International Satellite Cloud Climatology Project (ISCCP) site (http://isccp.giss.nasa.gov). This is generally an era of planetary warming. Cloud cover increases from the beginning of the record in 1983, through 1987, falls from 1987 to 2000, and rises again slightly after 2000. The overall correlation between warming and cloudiness is negative.

More importantly, the ISCCP global data is an average of individual spatial measurements. Ellis et al. (2004) provides a global map of cloud cover trends from 1987-2001, which includes the period of maximum amphibian extinction. We include this, along with highlighting the study area of Pounds et al., as Figure 3. There is no change in cloud cover in that region.


Figure 2. Global total cloud cover, 1984-2004 (source: ISCCP, http://isccp.giss.nasa.gov/climanal1.html)


Figure 3. Annual cloud cover trends (from ISCCP data), 1987-2001 (source: Ellis et al., 2004). The red box indicates the general region of the Pounds et al. (2006) study.

In a previous Nature publication on amphibian extinction, Pounds et al. (1999) argued that warming was decreasing the frequency of mist, and that caused the species loss. They stated that it was a result of an increase in the elevation of the condensation level of local clouds. This would result in an increase in the daily temperature range (the opposite of what was documented in the recent manuscript) and is more likely to be associated with a decrease, rather than an increase, in total cloudiness. The current Pounds et al. (2006) explanation is seemingly in opposition to this initial explanation.

There are several other manifold problems with this manuscript, but I will finish with just one.

The “Abstract” of a paper is supposed to succinctly summarize the scientific content of the paper. Here is the last sentence from Pounds et al.’s abstract:

“With climate change promoting infections disease and eroding biodiversity, the urgency of reducing greenhouse-gas concentrations is now undeniable.”

As you can see from the above, the current Pounds et al. manuscript produces no defense of that hypothesis. Further, there is no policy analysis whatsoever in the manuscript. There is clearly no way that this statement can remain in the abstract.

I presume that you will not publish this paper for the many reasons given above. If it does appear in Nature in anything close to its present form, the credibility of your journal may be damaged beyond repair.”

* It has subsequently been pointed out to us that the points on our Figure 1 (Pounds et al. Figure 4c) are not actually amphibian populations but simply a sampling of weather stations in the regions studied by Pounds et al. (this was somewhat unclear from the original Pounds et al. text, but clarified in the Supplementary Materials). So, our calculation that climate change introduced an additional 12% of the amphibian populations to chytrid infestations was not precisely correct. Instead, we should have cast the climate-change induced chytrid range expansion simply in terms of an elevation shift. When doing this, we find that the range expansion is about 250 meters (from our Figure 1). Since the amphibians range from near sea-level to greater than 4,000 meters (Pounds et al., Figure 4d), the range expansion for optimal chytrid infestations cover only about 6.25% (250/4,000) of the total range. If the number of species were evenly distributed across the entire range, then this would mean that climate change has potentially introduced an additional 6.25% of the amphibian species to better chytrid-growth conditions. Taking into account that there are about twice as many amphibian species living in the middle of the range than the low and high extremes (Pounds et al., Figure 4d), we’ll up the number from 6.25% to somewhere around 12.5%–virtually the same number that we originally calculated. As it can be seen, no matter how you slice it, the small change in thermal climate should have an effect on species disappearance far beneath that reported by Pounds et al. Our overall conclusions remain unchanged—the primary reason that the amphibians appear to be disappearing in Central and South America bears little relation to temperature changes, but more likely is caused by the anthropogenic introduction of a virulent pathogen into the amphibian’s environment during the past 10 to 20 years.


Daszak, P., et al., 1999. Emerging infectious disease and amphibian population declines. Emerging Infectious Diseases, 5, 735-748.

Daszak, P., et al., 2003. Infectious disease and amphibian population declines. Diversity and Distributions, 9, 141-150.

Davis, R.E., et al., 2003. Decadal changes in summer mortality in U.S. cities. International Journal of Biometeorology, 47, 166-175.

Ellis T.D., et al., 2004. Evaluation of cloud amount trends and connections to large scale dynamics. 15th Symposium of Global Change and Climate Variations, Paper No. 5.7, American Meteorological Society.

La Marca, E., et al., Catastrophic population declines and extinctions in Neotropical harlequin frogs (Bufonidae: Atelopus). Biotropica, 37, 190-201.

Pounds, J. A., et al., 2006. Widespread amphibian extinctions from epidemic disease driven by global warming. Nature, 439, 161-167.

Pounds J. A., et al., 1999. Biological response to climate change on a tropical mountain. Nature, 398, 611-615.

No Comments

No comments yet.

RSS feed for comments on this post. TrackBack URI

Sorry, the comment form is closed at this time.

Powered by WordPress