Create a weather file from measured data

edit

You also need 1) pressure (if you don't have it, use an elevation-adjusted average value), 2) wind speed and direction (especially important if you're doing natural ventilation), and as others have alluded, 3) global horizontal and direct normal solar radiation. I'm very much against mashing together weather data from different sources and time periods, and while that may "work" in the sense it doesn't crash the simulation, I doubt whether you're getting any improvement in the analysis which I would think was the motivation for such data manipulation. A simple illustration - if it's sunny and windy on a particular day in the "typical year" file, how confident are you that would also be true for the same calendar date of your file ?

My suggestion is that you go and look for the missing data elements from the closest weather station for the same time period. An unintended but unfortunate consequence with the proliferation of free "typical year" weather data is the perception that actual historical weather data are rare and hard to obtain, whereas the truth is exactly the opposite - you need 10-25 years of historical data just to create a "typical year" weather file. I recently participated in a webinar organized by the Chicago chapter of IBPSA-USA where I described various sources for historical data. The webinar presentations are available here.

Do you know about elements?

Also see answers to this question. As @Joe Huang stated:

solar radiation is highly correlated to other climatic factors, so if you merge solar from one place in time to temperature, humidity, etc. for another place at a different time, I'm really not sure how much improvement you're getting, or maybe even making it worse.