This was a little experiment that turned up some interesting results. The idea was to pick a paragraph from the IPCC reports and look at its provenance, just to see if anything interesting turned up. It did.
Unfortunately it turned up so much, that I've decided only to analyse the first sentence of the paragraph. I've got a life you know.
Here's the paragraph. It's from WG2, Chapter 10, and its the start of section 10.2.4.1 which is about the effects of climate change on food production.
10.2.4.1 Agriculture and food production
Production of rice, maize and wheat in the past few decades has declined in many parts of Asia due to increasing water stress arising partly from increasing temperature, increasing frequency of El Niño and reduction in the number of rainy days (Wijeratne, 1996;Aggarwal et al., 2000; Jin et al., 2001; Fischer et al., 2002; Tao et al., 2003a; Tao et al., 2004). In a study at the International Rice Research Institute, the yield of rice was observed to decrease by 10% for every 1°C increase in growing-season minimum temperature (Peng et al., 2004). A decline in potentially good agricultural land in East Asia and substantial increases in suitable areas and production potentials in currently cultivated land in Central Asia have also been reported (Fischer et al., 2002). Climate change could make it more difficult than it is already to step up the agricultural production to meet the growing demands in Russia (Izrael and Sirotenko, 2003) and other developing countries in Asia.
First step was to look at the citations. I've linked PDFs where I have them.
So by the end of the first sentence, none of the cited papers that I can lay my hands on support the text as written, namely that "Production of rice, maize and wheat in the past few decades has declined in many parts of Asia".
Some of the reasons for this change became clear when I looked at the Second Order Draft. Here's the equivalent paragraph.
The increasing pressures of changes in climate and its variability would make it more difficult than it is already to step up the agricultural production to meet the growing demands in Russia. Increasing surface air temperature along with increasing intensity and frequency of El Nino, and the reduction in number of rainy days have heightened the water stress in many agricultural areas in China lowering the production of wheat especially in the Changjiang river, the yield of pasture areas in Qiaghai [sic] Province and the southern part of Gansu province and corn production in central China (Tao et al 2003a, Jin et al 2001, Tao et al 2004. Rice and wheat production in the Indo-Gangetic plains of South Asia, which increased in the 1970s and 1980s have in recent decade stagnated, attributed to the rising (minimum) temperatures during the growing season, resistance to weedicides, and decrease in soil productivities that appeared in large areas under rice-wheat sequence (Agarwal et al 2000). In Sri Lanka 30cm of soil has been eroded from upland tea areas over the years while in the lowland tea plantation production has been adversely affected by soil erosion and increased soil moisture deficit brought about by increase in temperature and drought (Wijeratne 2002).
Blimey. It's a remarkable change they've made there. Notice how we change from talking about what might happen to what has happened. Hasn't the whole purpose of the sentence changed? Doesn't this make a bit of a mockery of the review process? It's only one sentence, I know, but all the same.
Some things, however, become clearer.
Finally to the expert and government review comments. These should help us understand who made the changes.
Or they should do, anyway. As far as I can tell, this radical rewrite between the second order draft and the final published piece was driven entirely by the authors. None of the reviewers requested the inclusion of Fischer et al. Nobody requested that tea be left out of the equation. And, nobody pointed out that Tao 2004 didn't support the text.
It's not right, is it?