The origin of the claim that the Syrian refugee crisis is partly caused by climate is a paper by Kelley et al in PNAS. This has picked up quite a lot of media attention, yesterday's Independent article being just the latest.
Kelley et al is a bit odd though. Consider at what they found. In the top panel of the following figure, they claim to have found a significant drying trend. They are using a significance level of P <0.05. (Questions, questions: why do they calculate trends since 1931 when they have data going back to 1900?)
However, if you wanted to examine climate as a cause of the Syrian uprising, you would imagine you would want to look at the climate in Syria. Kelley et al, however, have not done this, but instead have looked at the climate in the greater Fertile Crescent, an area that also takes in most of Iraq, half of Iran and the majority of Turkey.
When you look at their climate maps, you can see why this matters. In panel B, below, the big reductions in precipitation, show in darker brown, are not actually in Syria at all, but on the Turkey/Iraq border and in Iran. Referring to panel A, those biggest reductions are also taking place in the areas that already have the most rainfall. Panels C and D are about the recent drought, but are obviously over too short a period to be relevant to "climate".
Going back to panel B, the authors say they have found a statistically significant drying at several stations in the area, marked by circles. However, in fact as they note in the text only 5 of 25 stations showed this "significant" drying and only one of these was in Syria. Moreover, when they say "significant" in this case, they are using a significance level of only 0.1.
It looks to me as if what they have done is to use a minor drying trend in the wetter parts of Turkey, Iraq and Iran to make a completely spurious claim about the climate in Syria.
According to the paper, it was reviewed by Sir Brian Hoskins of the Grantham Institute.