I'd push back a bit on the premise - they didn't really "recreate" weather patterns so much as piece together what actually happened using what was documented at the time. The difference matters because we're not talking about reconstruction or modeling; we're talking about finding the historical record that was already there, just scattered across different sources and often buried in archives.
The sources are pretty diverse. Weather stations that were operating during the war kept daily logs - temperature, pressure, wind direction, all that stuff. Pilot reports from bomber missions are gold because pilots had to note weather conditions, cloud cover, visibility; those observations were precise and numerous. Ship captains kept detailed logs too, especially in the Atlantic where weather could make or break naval operations. Beyond official records, there's a surprising amount of detail in letters, diaries, and even photographs - a soldier describing "gray skies and mud" or a photo showing snow depth tells you something concrete about conditions on a specific day and location.
What really made the difference was systematic cross-referencing. If you have a pilot's report saying they encountered heavy cloud cover over a specific area on June 5th, and you've got a weather station 50 miles away recording pressure drops and wind shifts on the same day, and a naval log mentioning rough seas - suddenly you've got a much clearer picture of what the actual weather was. Modern historians and meteorologists went through vast amounts of this material to reconstruct not just what happened, but why certain operations succeeded or failed. The D-Day forecasting question is the famous example - whether meteorologists actually predicted a break in the weather that allowed the invasion to proceed, rather than getting lucky.