Made more winter graphs. Counted all the January and February days from 1961-2010 (except for Paris, which was 1951-2000); rounded to the nearest °F. The y-axis is in a decimal, because I divided the count per °F by the total number of days. Anyway, higher value means a particular temperature is more frequent. Red line is the mean for the period (both January & February); yellow is the coldest temperature on an average year. For some places, might be very different from the record low if only some years have extreme cold. Helsinki has a strange shape:
Extremely skewed, and the mode (most frequent value) is the same for both the max and min! Don't think I've graphed a location with a pattern like that. The mode seems to correspond to just above freezing. Only way to have the same mode for both the min and max is by warmer than average days having a very low diurnal range, which makes from its geography.
Bucharest has a skew in the opposite direction for the maxes, though nowhere as extreme as Helsinki
Mins have the same long tail towards cold, and getting much above freezing is infrequent.
I graphed Bridgeport, Connecticut rather than New York City. Similar climates, Bridgeport gets some milder effect from the coast like New York City but not as much urban heat island. Maxes don't have much skew.
Mins much above freezing are unlikely, creating a skew. Bridgeport has a similar range and extremes as Bucharest, maybe similar slightly more extreme; but its temperatures are less peaked around a certain value; within 7-10°F of the mean the frequency doesn't change much.
Now, 100 miles north. Max graph is roughly similar
mins much colder, and rather variable. Helsinki looks just as variable in that its capable of as anomalous temperatures but doesn't get them as often, the "cold tail" has low values
Now for some place completely different. Unlike the temperate climates, Churchill has a sharp cut-off for much colder than average temperatures, while much warmer than average are possible but don't happen often. This makes sense; it's at the source of cold rather than getting cold incursions. But it can get occasional warm incursions from the US Midwest ?
Still on average year, it reaches 22°F below the average min; about the same as Amherst. But the records aren't much lower than that.
Brr.
And reposting Paris, looks refreshing after Churchill's numbers
another way to compare: the 1st, 5th, 50th and 95th percentile of Jan & Feb minimums and maximums. 5th percentile = 5% of winter temperature are colder than that temperature, 95% are warmer. 50th percentile is the median
Amherst's minimums distribution is similar to Helsinki, maximums a little cooler than Bucharest. Bridgeport matches Bucharest well, but there's no equivalent for Amherst in Europe. Its diurnal range is too big. Amherst and Helsinki have the same 90% range (5th to 95th percentile), Bridgeport and Bucharest a bit less, more so for mins. Churchill's maxes have a very wide spread; wider than the other locations. Paris has a relatively narrow range.