The sailing season can be used as a proxy for winter temperature. The start of the sailing season is a proxy of ice break-up, which in turn is a very good proxy for mean winter air temperature. (Tarand A & Nordli P.O, 2001) Custom books from the Stockholm harbour exist from 1533 onwards.
Do the preserved custom books reflect the length of the sailing season? Or rather, can we assume that the sailing season reflect ice-coverage of the western Baltic region? E.g. during the sixteenth century, the tradition of celebrating Christmas seems to have been deeply rooted. The sailing season practically always ended well before Christmas, which may reflect the “little ice age”, but probably more reflects a traditional society.
Another, general, problem with the chosen sources, is that the larger the shipping traffic was, the better the sources reflect ice coverage.
The requirements on “open water” may also differ between a large vessel, to one that is smaller. E.g. in late winter/early spring there is a time when the sea is a mixture of ice-floes, thin ice, open water, as well as more solid ice. In such conditions it might be possible to force passage through slush and ice-floes, which might not be dared in a smaller boat. On the other hand, this problem of the development of vessel-sizes, might be superfluous, since the smaller boats might never have occurred in any sources ever, since such boats never paid anything.
Preliminary result of spring arrivals and departures at Stockholm harbour 1533-1800. Correlation between arrival/departure dates and winter-temperature (for the period 1722-1800, using the Uppsala temperature by Moberg) is c. r**2 = 0.8.