On Fri, Aug 14, 2009 at 12:40:03PM +1000, Terry Bishop wrote:
> How long before we can put the Eagle in the Budgie cage I wonder?
> http://www.sciencealert.com.au/news/20091408-19562.html
The referees for the paper seem to have done a poor job. SE Australia has
warmed too little too small to explain the reduction in bird size they've
reported. My back-of-the-envelope calculation is warming can explain
less than 20% of the decrease in size, leaving 80+% unexplained. And
if there is a major unknown factor, maybe it explains 100% of the bird
shrinkage and warming caused 0%. For those interested in the details:
They reduced their data set to two categories 1860-1950 and 1951-2001.
The BOM at http://www.bom.gov.au/cgi-bin/climate/change/timeseries.cgi
doesn't have data pre-1910, but it tells me for SE Australia 1910-1950
mean temperature was 0.31C cooler than 1951-2001. The Hadcrut3 Southern
Hemisphere index has 1860-1910 ~0.05C cooler than 1910-1950. Put them
together and say SE Australia 1860-1950 was 0.35C cooler than 1951-2001.
My very-approximate calculation in SE Australia is that a degree of
latitude north/south changes mean temp up/down by about 0.4C. SO the
warming between 1860-1950 and 1951-2001 is about the same as moving
north 1 degree of latitude
But the authors report the size change between their 2 datasets is
equivalent to 7 degrees of latitude - based on latitutinal variation
within the species. Suggesting warming can explain ~1/7 or less than 20%
of the size change in birds reported.
Obviously these are quick&dirty calculations but better than nothing - the
authors provide no comparison of warming between data sets to latitudinal
change. The referees should have picked this up and not allowed the
inference that global warming was the probable cause without such a
comparison, even worse is the misleading title indicating a correlation
with global warming.
I'm also surprised the referees let them provide so little explanation for
the puzzling reduction of the data set to two categories (1860-1950 and
1951-2001). This reduction throws away seemingly important information
- e.g. it lumps the 1950s, the coldest decade in SE Australia (on mean
temps) with the 1990s, the warmest. Its also strange to split before
not after the coldest decade of the century.
An analysis of year versus size (wing length) would be simple and
informative, it should have been included or a very convincing explanation
for its absence given. An analysis of size versus estimate temperature
at the collection location/time or at least SE Australia (say averaged
over the previous decade) would be trickier but throw a lot more light
on the role, if any, of temperature.
Andrew
|
