Snow in Baghdad, and other ancient climates
Climate scientists often bemoan the imperfect data with which they have to work, particularly when it comes to building pictures of climates past.
If only Homo erectus had invented the weather station and distributed its invention evenly as the species expanded its footprint across Africa and then the world...
The shortness of the satellite record, going back only to 1979 or thereabouts, is a particular gripe.
One of the most fascinating areas of research that's emerged in the last couple of decades is historical climatology - trying to deduce evidence on past climates through written records.
Whereas palaeoclimatology measures tree-rings and stalactites and deposition patterns in sediment, historical climatologists scour the log books of ships, parish ledgers of grape harvests and the diaries of amateur naturalists for clues.
Perhaps the most fertile ground of all lies in ancient civilisations that established the equivalent of modern civil services early on - where records were kept for year after year, recorded as part of the dynasty's foundation.
Chinese records, for example, kept by court mandarins have yielded data on aspects of weather and climate including rainfall, temperature, thunderstorms and typhoons.
In the magazine Weather this month there's a fascinating paper doing the same thing with records from what is now Iraq going back over 1,000 years, using documents stored in Madrid and in Baghdad itself.
Founded in AD762, Baghdad quickly grew in size and importance as capital of the Abbasid Caliphate.
The Caliph's record-keepers didn't specifically write about weather and climate.
But Fernando Dominguez-Castro from the University of Extremadura, Spain, and his team of scholars have been able to extract a good deal through passing references in the writings of al-Ya'qubi, author of a major treatise on the region in 891, and his successors.
Back then, water was plentiful, judging by the network of canals constructed in the region.
The climate appears to have pleased al-Ya'qubi, with heat in summer, cold in winter, and mild in between.
It's hard to judge what "heat" and "cold" mean in this context, with nothing against which to measure them. But we get a better idea when he tells us that in summer, people spent most of the day in the coolness of rooms dug below ground-level, but their nights on open roofs under the stars.
And having established these habits, we get a better idea of the transformation wrought in AD920, when the city was caught in a snap so unseasonably cold that people took their blankets indoors in the height of summer.
Dr Dominguez-Castro calculates the temperature may have been about 9C below normal that month - related, he speculates, to a major volcanic eruption.
The records written by al-Ya'qubi and his successors also yield information on the frequency of droughts, floods, rainfall, hail, winds, hot and cold spells and locust swarms.
Snow was reported regularly in winters between 832 and 998. In 909, one writer records: "There were four fingers of snow on the ground, and the cold was intense. Water, vinegar, eggs and unguents froze."
Things sound even more extreme in 926, when "sherbet and rose-water froze, as well as vinegar.
"The scholar known as Abu Zakaria sat in the middle of the Tigris, on the ice, and gave lessons of the Prophetical Tradition."
Later scholars recorded no snow.
So here we have evidence of a change of climate, which appears to have occurred at the same time as some parts of the world were entering the Medieval Warm Period (MWP).
So this would support the idea that Baghdad, and by extension the rest of the Mesopotamian region, experienced the MWP as well.
There are clearly limits to historical climatology. A minority of ancient cultures kept records, and some texts in extinct languages (such as Etruscan) we cannot decipher.
While concrete events such as typhoons and snowstorms are faithfully recorded, less remarkable events are not, and there's little idea of gradual trends.
Nevertheless, for some parts of the world these records are among the best we have; and their use is likely to increase, through projects such as the German-funded Historical Climatology of the Middle East based on Arabic Sources back to AD800.
As Dr Dominguez-Castro told me: "A lot of Arabic sources remain unexploited with climatic objectives, with information of an epoch and places few studied until today." By studying more records, he says, it might become possible to discern slow trends as well as quick flips.
The field of inquiry also provides fascinating glimpses into people's lives during a time when weather was a bigger determining factor in their lives than it is today, and into the ways they found to adapt.