I know none of you noticed, but I wasn’t around the blogosphere much last week, because I was in York, at a workshop. I had left Helsinki in the depths of winter, with snow and ice on the ground.
Helsinki, a couple of weeks ago
The north of England, by contrast, was gearing up for spring. The trees were starting to produce buds, and over breakfast we could watch a female blackbird carrying the materials to build her nest.
And thus the hay-fever season begins
The difference is, primarily, one of latitude: Finland is just much further north, and hence much colder. Spring starts much later, and the growing season is shorter overall than in the UK. But the effects of latitude can be seen even within a country. Spring will arrive soon in Helsinki, but when it does Lapland in the north will still have a few weeks to wait before its snow disappears. In Sweden, the frogs in the north of the country are adapted to have a much shorter developmental period, because they can’t flop around as tadpoles eating and getting fatter – they have to metamorphose and start doing adulty things before the pond gets cold and freezes.
Trees also have to react to the season: in a classic experiment, pine tree seedlings were transplanted up and down Sweden1. The trees would start to produce buds according to the latitude where they had been taken from: they were adapted to that season, and unable to change. This control is tight: moving a tree 100km north or south is enough to get it out of sync with the seasons. Indeed, the time of budset can even be used to tell where a tree came from. The later a tree starts to produce buds, the more time it can do other things (like grow), but if it starts too late, the buds are liable to get hammered by the frost.
The effects of seasonal variation can be more subtle. In the north of the UK, the common blue butterfly))/species106/description.htm can only manage one generation per year. In contrast, in the south the season is longer enough that it can squeeze two generations into a season. Somewhere in the middle, it shifts between these two.
It’s blue, and at the right time of the years there’s lots of ‘em
The reason I’m in York is that I’ve been attending a workshop,a continuation of the one I attended last autumn. We are looking at distributions of species, and how they will be impacted by climate change. A warming world is already making the seasons longer in Finland and elsewhere, and species will have to respond to this. Of course, this is a large area of research at the moment, thanks (in part) to the perceived sexiness of the subject2. A first step in this is to know where species are right now: only then can we work out were they will move to. This requires some form of inventory of biodiversity. Whilst systematic surveys can be done, it is difficult to be exhaustive. But there is also a lot of other data out there, for example from amateur naturalists (records from butterfly chasers and birders, for example). This data will be more extensive, but not collected in as systematic a manner. Hence, we need to think more carefully about the biases created by the way the data are collected.
With butterflies we have a curious and important problem. They are usually caught as adults, because they are easier to see when they are flying. So, people go out on a walk (or ‘a transect’ if they are being serious) and record what butterflies they see. But they obviously can’t record a species if it isn’t ‘in season’. So when they report back their species list, a some species might not be recorded because it was the wrong time of year. Hence, the absence of a species may not be true, simply because of timing.
If we want to produce an accurate map of where a butterfly is, we want to be know how reliable our absences are. This then leads us to back to phenology, i.e. the annual timing of events in the life cycle. We can can estimate the flight times of teh butterflies (e.g. from organised repeat samples), and use this to work out if the species was flying during any particular survey. Making the link between the two forms of data requires a bit of statistical jiggery-pokery.
But some butterflies just want to make life complicated; they change their phenology. The Common Blue and its friends are going so far as to vary the number of generations they have in a year. Of course, this all has to be including that into our estimation of likelihood of observation. Hence, much of last week was taken with discussing phenology and how to estimate its effects, and how important it is in the end3.
I realised a few years ago that the reason I work as a statistician is because I love doing biology. At one level, phenology is just something that’s screwing up the quality of the observations. But this leads to interesting statistical problems and, more importantly, means I learn more about the biology. And a good statistical analysis needs to include all these nasty bits of biological reality. So I have no alternative but to enjoy learning new stuff.
I wish spring would come here soon, though. The snow’s not so bad, but the ice is a pain.
1 Eiche V. 1966 Cold damage and plant mortality in experimental provenance plantations with Scots pine in northern Sweden. Stud. For. Suec. 36: 1–218. Outi Savolainen’s group in Oulu have been working, e.g. here’s a pdf from a recent talk
2 i.e. we can get funding for it.
3 “We” mainly means other people, but I’m involved a bit too.