Nobody warns you that the hardest skill in remote fieldwork is sitting still. I arrived at [STATION NAME] with a sampling plan built in a warm office, laminated, colour-coded, and entirely dependent on the assumption that we could get where we wanted to go on the days we wanted to go there.
We executed roughly forty per cent of that plan. It was, by any reasonable measure, a successful season.
The arithmetic of a field season
One hundred and eighteen days on station. Of those, thirty-one had weather good enough to move. Of those thirty-one, nineteen coincided with the equipment being serviceable and the ice being where we needed it. Nineteen working days from four months — and that ratio is not unusual.
- Two hundred and forty kilograms of gear moved by hand across sea ice.
- Sixty-one under-ice samples, against a target of one hundred and forty.
- One generator rebuilt twice, in a tent, at minus twenty-six.
- Zero injuries, which is the number that actually matters.
“You do not beat the weather. You arrange your life so that when the weather relents, you are already dressed.”
Designing for a bad season
The lesson I took home was structural. Our sampling design assumed a steady rate of data collection and degraded badly when that rate collapsed. The redesign — now used across the group's field programme — front-loads the irreplaceable measurements and treats everything else as optional.
It sounds obvious written down. It took a bad season to learn it properly.
I flew out in the last week of the season with less data than I wanted and a much better field programme than I arrived with. On balance, I would take that trade again.
