I am conducting an air pollution case-crossover analysis and I'm in the midst of adding in pertinent weather variables. I cast a wide net and downloaded data from NOAA, National weather service and a few other sources. The problem is that for each variable, there are multiple versions of it. For example, we have maximum temperature, minimum temperature, apparent temperature, etc.

Are there any diagnostics I can employee to work out which variables I can use in my time series? Is there also any agreed upon variable? For example, is there literature out there confirming that perhaps Tmin should generally be used for such an analysis? Any resources or studies you may able to direct me to would be really helpful!

  • 1
    Can you elaborate on how you're doing your analysis? I found this article, but it's still not obvious to me what an "air pollution case-crossover analysis" is, or how you plan to use the temperature variable.
    – csk
    Jul 11 '19 at 21:41
  • None of the variables you specify is a good representation of climate or even temperature, because there can be 2 very different climates that happen to share the same tmin, tmax, or even tavg. For example, the variation between tmax and tmin (for example) may be much more important than either tmax or tmin itself. I suggest using all 3 series and see which one works best (you run into statistical problems doing this, but, it's at least one approach)
    – user3856
    Jul 12 '19 at 17:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.