The valuable chart archives on the Wetterzentrale site provides a number of reanalysis series, from CFSR, ERA and NOAA that you can download daily charts from. The CFSR series is at 0.5° resolution, whilst data from ERA is 1° and NOAA 2° resolution. Here’s an example of all the daily charts (00 UTC) for last winter [DJF 2018-19] using CFSR reanalysis data of MSLP and geopotential height at 500 hPa.
Trying to find the start and end of any cold or warm spell in a series of weather charts by quantifying just how cold or warm it’s been in any particular month or season is not easy. Likewise comparing one period to another can also be difficult to do. So I came up with what I think is a novel way of quantifying values of either heights or temperature for any of these reanalysis charts by careful examination of the pixels that go to make up the colour filled contours of each chart. In software it’s not difficult to add up the total number of each coloured pixel that make up the colour filled contours. The data grid below is the analysis of all the charts in the grid of images for last winter, the percentage values are for each colour representing a different geopotential height.
With this table it’s now quite easy to see how the cold spell of early December 2018 set in, with all the talk of an early SSW event, giving way to an anticyclonic Christmas, and how January turned gradually colder as the month went on, and the surprisingly record breaking mild spell at the end of February 2019. Counting pixels is quite simple, although many of the colour filled contours are lost to black isobars and values (black text on white labels), so at best it’s a not too perfect approximation. But once you have defined your area it will work for anywhere in the world as long as they stick to the colour scheme as laid out in the legend.