Wondering Willis Eschenbach at WUWT gets lost in Arizona

Wondering Willis Eschenbach has got stuck into Arizona mountain plants.  He's written an article on WUWT about a new study of changes in vegetation.  He tries to claim they got it wrong.  Personally, I think he got stuck trying to come up with something new to write about.  Anthony Watts copied the press release yesterday and his followers got all upset by it then.  Probably gave Willis the idea to take advantage of the outrage.

Dramatic response to climate change

The study is by Brusca et al called: Dramatic response to climate change in the Southwest: Robert Whittaker's 1963 Arizona Mountain plant transect revisited. Ecology and Evolution. It's an intensive grid analysis of vegetation, comparing the changes with Robert Whittaker's 1963 plant transect in the Santa Catalina Mountains of southern Arizona.

The study found that many common plants had shifted.  The altitude range of many plants had contracted.  Some had moved to higher elevations and some to lower, with the changes being attributed to changing rainfall patterns as well as changes in temperature.

Here is a summary, comparing the species elevation with those found by Whittaker. Click the figure to enlarge it.

Figure 1. Summary of elevation range of the 27 most common upland montane plants along the Catalina Highway. White bars are 1963 elevational range data from Whittaker and Niering (1964), the two terminal (stippled) 1000-ft bands denoting Whittaker's upper- and lowermost 1000-ft vegetation bands. Black bars represent 2011 elevation data from this study. To be as conservative as possible, a change in a species elevation limit (high or low) was noted only if that species was found outside (above or below) the upper- or lowermost 1000-ft band. Thus, if anything, we underestimate the elevational change in the species since 1963 (see 'Materials and Methods'). Following this protocol, 15 species show an unambiguous increase in lower elevation, four show an increase in upper elevation, and eight show a decrease in upper elevation.

Willis has got himself into a dither because he reckons they messed up the study.  He found that they restricted their sampling to 3500 feet and above in altitude, but didn't bother to check the details.

First up, Willis thinks they are wrong to say that Muhlenbergia porteri was observed to stop at the elevation reported, because he says they didn't look at all below 3,500 feet. So I went to check.  First thing I noticed was that M. porteri was reported in Table 1 as being in the range of 3428–3845 ft.   That's below 3,500 feet I noticed.  Looking further, I investigated details of the actual survey.  Here is what they write:

Our study used the same Catalina Highway Transect as W–N, which runs from the base of the mountain to the top of Mt. Lemmon, the highest peak in the Santa Catalinas (9157 ft/2791 m). We established 33 0.1 Ha quadrats (10 × 100 m) separated from one another by an average of 50 m elevation. All plant species present in these 33 quadrats were recorded. Habitats within biomes on the southern slopes of the Catalinas are remarkably uniform (as noted by Whittaker and Niering 1964, 1965, 1975; Whittaker et al. 1968; and others), and our 0.1 Ha elongated quadrats traversed all the slope aspects and microhabitats in all the upland plant communities on the mountainside (described in Moore et al. 2013). Our sites were all farther than 0.25 km from the highway, to minimize possible road effects. The location, coordinates, plant community type, slope, aspect, and other features of all our quadrats are described in Moore et al. 2013. 

In the paper the authors refer to Moore et al for details of the sampling.  (I found a cached copy of the pdf file.  You can access it here, but it's missing the diagrams and maps.) A list of all the areas surveyed is provided in Table 6 on page 14 of the Moore et al pdf file.  (Table 6 lists more areas than were covered in this particular paper.) Here are two excerpts from Moore et al:

On the south side of the range we established five transects in Desertscrub (at elevations of 1045–1172 m), six in Oak-Grassland (at elevations of 1384–1433 m), seven in Pine-Oak Woodland (at elevations of 1803–2422 m), two in Chaparral (at elevations of 1923–2052 m), and eight in Pine Forest (at elevations of 2224–2463 m). On the north side of the range we established seven transects in disturbed Desert Grassland (historically grazed areas at elevations of 1323–1451 m), six in relatively undisturbed Desert Grassland (at elevations of 1330–1645 m), two in Oak Woodland (at elevations of 1939–2000 m), five in Pine-Oak Woodland (2032–2149 m), five in Chaparral (at elevations of 1845–1971 m), and four in Pine Forest (at elevations of 2218–2305 m). Ten transects in mixed conifer habitat were established in the Mt. Lemmon and Mt. Bigelow areas (at elevations of 2442–2777 m)...

...Plant surveys were conducted August 5-14, 2011. Surveys were made for each transect site in two ways. First, the transect line was walked and every plant recorded within one meter on either side of the line. This produced a 200-m² belt transect record of plant species and abundance. Then, the transect was walked again, this time scanning the broader area outside the 200-m² area, noting the presence but not abundances of any plants that might not have been in the belt transect itself.

Despite reading the paper and Moore et al, I could not work out just which were the 33 quadrats included in this research. Obviously it included Desertscrub and it most likely did not include the grazing-disturbed grassland.  The paper also refers to the "biomes on the southern slopes".

Back to Willis.  Seems to me that Willis may not be aware that the areas surveyed are not contiguous.  They are at different elevations.  The above excerpt shows the lower range of the plant in question and indicates they did survey beyond the transects to check the presence of plants.  The Desertscrub elevation (bold italic above) of 1045 m (3,428 ft) matches up with the observations of Muhlenbergia porteri in their Table 1.

Let's move on.  Willis' next argument is that the researchers should not have reported what they actually found.  He writes:

Since the new study only began at 3,500 feet, we cannot say that the lower range of this species is increased. They have incorrectly counted it as having an “unambiguous increase in lower elevation”. 

 We've covered his first point.  Let's look at his second point.  He claims that they cannot count it as not being lower.  It's true that they don't appear to have looked for it below 3,000 feet and that 3,428 feet was the lowest elevation sampled.  So even though I expect the scientists were correct, let's grant Willis that point.  While we're giving Willis a pass here, you'll notice the startling constriction of the range of Muhlenbergia porteri.  Whittaker found it up to elevations of between five and six thousand feet, whereas today it's only found below 4,000 feet.

Now to Willis' next point.  He writes:

They want to compare their results to W&N63, which is a good thing, they should do so. But we need to have an apples to apples comparison. To do that, the very first thing you have to do is to convert their data to the “elevational bands” intervals used by W&N63. Yes, you need to throw away information to do it, but if you want to compare the two studies, you have to do it. For an example of what this does, look at Agave schottii, the first succulent in the purple band. It looks like the elevational range has shrunk considerably. But if the proper method were used, counting only by elevational bands as in W&N63, we’d see that there is absolutely no difference between the two records. W&N found it in three bands, and so did the latest study. Now, to their credit, they count it as unchanged. But it is incorrect to present their data for comparison with W&N63 in absolute elevations, rather than elevational bands as in the original study. To compare apples to apples, they have to first convert their data to elevational bands. Their failure to do so has lead them to false conclusions.

If you can't follow that I'm not surprised.  Willis is telling the scientists to do exactly what they did and then complaining that because they did as he suggested they were led to false conclusions.  Doesn't make sense?  I don't think so either.

Look at the plant Willis refers to: Agave schottii.  In their Table 1, Brusca et al reported finding this plant at the elevation of 3801–5028 ft.  Looking at Figure 1, Whittaker reported finding it within the bands from  3,000 feet to 6,000 feet.  That's the same range of altitude.  Willis says they drew false conclusions.  But he has no basis for saying that. You'll notice that while he claimed they drew false conclusions (presumably because they reported the actual data and didn't "throw away information" like Willis wanted them to do, he contradicts himself, writing: Now, to their credit, they count it as unchanged.  

So I say to Willis - what's your point?

Next Willis gets his knickers in a knot over three other plants.  He writes:

With that in mind, look at the bottom row, the grass “Urochloa arizonica”. When we convert their data to the W&N63 elevational bands, we see that there is no evidence of an upward trend at the bottom end, because the new survey only started at 3,500 feet. In terms of elevational bands, all we know from the new data is that that grass is still found between 3,000 and 4,000 feet, just as in the original W&N63 survey. For all we know, the current authors may have found the exact same patch that W&N found in 1963. As a result, we have no new information establishing the bottom end of its range.

The same is true for the woody shrub Mimosa aculeaticarpa, and the grass Aristida ternipes. 

To help you follow all that, I've put the details of the three plants in the following table.  The new study results are in the column on the left under "Brusca" and the 1963 study findings are listed under "Whittaker".  You can see the difference in the ranges - with the "low end" and "high end" showing where Brusca differs from Whittaker.

Table of plant species

Plant	Brusca	Whittaker	Low_End_2013	High_End_2013
Urochloa arizonica	3806–4893 ft	2,500 - 4,000 ft	Low end is higher than in 1963	High end moved higher than in 1963
Mimosa aculeaticarpa	3801–4893 ft	2500-7000 ft	Low end is higher than in 1963	High end dropped lower than in 1963
Aristida ternipes	3801–6732 ft	2500-6000 ft	Low end is higher than in 1963	High end moved higher than in 1963

Willis is wrong and misleading when he writes: we see that there is no evidence of an upward trend at the bottom end, because the new survey only started at 3,500 feet. In terms of elevational bands, all we know from the new data is that that grass is still found between 3,000 and 4,000 feet, just as in the original W&N63 survey. 

First of all, Whittaker found all three of these species below 3000'.  They found them at between 2500' and 3000'.  While Willis' argument is that the plants were only surveyed down to 3,500 feet, the researchers found that the lowest extent of their range was around 3,800 feet in all three case.  If the lowest they found the above plants was as reported, then it's reasonable to assume they weren't found any lower.  The lowest extent was 300 feet or more than 90 metres above 3,500' altitude. Again, a list of all the areas surveyed is provided in Table 6 on page 14 of the Moore et al pdf file.  The only area that includes 3,800' is the Desertscrub, which ranges in elevation from 3,428' to 3,845'.  There were five transect sites coverage that vegetative zone.  It would have been obvious if the plants were recorded at the lower elevation.

Bear in mind that these 27 plant species are not rare plants that you could miss if you didn't look too closely.  The researchers describe them as:  the 27 most common upland montane plants along the Catalina Highway.  Nor are they like sheep or cattle which might just wander off into a corner of the paddock and get missed.  These are plants.  They can't walk away.  If the 27 most common upland montane plants aren't observed below 3,800 feet in altitude then it's likely they don't exist below 3,800 feet in altitude.

Remember, too that we're not talking distance here.  We're talking altitude.  The only way that altitude would equate to distance would be if they were growing on an absolutely vertical cliff.

Here's a map showing some of the higher elevations in the general area.  Not flat by any means but neither are the mountains made up of vertical cliff faces. Click the map to enlarge it and you'll find that at even at much higher elevations than Willis is talking about, the slope is far from vertical.

 You can click here to explore the area in Google Earth.

Seems to me this is just a slightly more elaborate denial by Willis Eschenbach at WUWT.  Must say, Willis does tend to favour elaborate denials.  I also suspect he imagines the surveyed areas as one continuous stretch of land.  Reading this paper and Moore et al it's obvious that the sites were not contiguous.


Brusca, R. C., Wiens, J. F., Meyer, W. M., Eble, J., Franklin, K., Overpeck, J. T. and Moore, W. (2013), Dramatic response to climate change in the Southwest: Robert Whittaker's 1963 Arizona Mountain plant transect revisited. Ecology and Evolution. doi: 10.1002/ece3.720

From the WUWT comments: rain, fire and damnation

Tim Ball ignores the paper completely.  Contrary to what he writes, in almost every paragraph of the discussion section of the paper there is considerable weight put on changing precipitation, at least as much as to changes in flowering time in response to changes in temperature.  He also ignores the obvious fact that whatever the conditions are today, some things have changed since 1963 or the plants would still be occupying the same niche as back then.  The aspect of the mountain slope hasn't changed so maybe it's the climate that's changing? Plants don't move around for no reason at all, Dr Ball!

August 15, 2013 at 1:10 pm  No sooner had I mentioned the neglected value of [paleontological ?] studies in a comment on another posting than Willis finds one – a poor one at best.  It virtually ignores the role of precipitation, especially in determining lower level boundaries. It also ignores the aspect of the mountain slope that is a south facing slope in the Northern Hemisphere have very different evapotranspiration rates. Also there is the effect of cold air drainage that can create cooler night time conditions in lower levels. These are just a few of the factors that can explain what is happening and apparently ignored.

Philip Bradley also doesn't bother with the paper itself, ignoring the fact that the authors particularly mentioned that all quadrats were all further than 250 m from the highway "to minimize possible road effects" and says:

August 15, 2013 at 1:12 pm  Alongside a highway? The highway itself could well be a significant factor, facilitating transport of seeds, diseases, competing plants, affecting browsing animals, insect pests, and creating a micro-climate alongside the road.

Rud Istvan ignores the paper as well, so he doesn't know that the authors made the point that: "To facilitate comparison with the W–N data, we did not sample in any areas of significant forest fire history since 1963, including the 2002 Bullock Fire or 2003 Aspen Fire." He says:

August 15, 2013 at 1:17 pm  As posted on the previous thread, and offered to Anthony in the form of a longer factually illustrated possible post, Willis quite valid critique is not the real problem. Almost all of the transect was burned by the 2003 Aspen wildfire, as the dead juniper below 5000 feet in the PR proves. When you compare mature wilderness with decade old regeneration, you are making only fruit salad, not science.

Theo Goodwin similarly ignores how the researchers deliberately avoided sampling areas that had been burnt, so he can be suitably outraged at the Alarmists and "shocked" and says:

August 15, 2013 at 2:33 pm  Rud Istvan says:  August 15, 2013 at 1:17 pm “As posted on the previous thread, and offered to Anthony in the form of a longer factually illustrated possible post, Willis quite valid critique is not the real problem. Almost all of the transect was burned by the 2003 Aspen wildfire, as the dead juniper below 5000 feet in the PR proves. When you compare mature wilderness with decade old regeneration, you are making only fruit salad, not science.”

I am shocked. I had thought they were mere Alarmists doing Alarmist work. But this information strongly suggests something far worse.

Damned if you don't and damned if you do!

When it's pointed out that the researchers avoided sampling areas burnt by fire, Willis tries another tack, just to show that nothing scientists do can be right - and says:

So as a result, if they are avoiding taking samples where it was burned, they’re missing out on sampling half the local ecosystem … and it’s known to be different from the half they are sampling. No bueno …

A few screws loose ...

Gotta love those crazy science deniers.  The contortions they have to go through.  It's no wonder people say they haven't got their heads screwed on properly.