I often need to calculate heat balances for large biogas plants, an important factor is outside temperature. I normally use the weather data from Klimadiagramme.de where I can find the mean monthly temperature and precipitation. This allows me to calculate the heat demand of my plant for a given month reasonably well. However, I also need to know extreme weather. I'm not sure what the metric here is, but I'm not looking for "the coldest night in the last 100 years" but "this was the coldest average over 2-3 days in the last x years" or something similiar.

So, what is a good, free source of climate data that gives me information about extreme weather events as described?

  • $\begingroup$ Worldwide, or for a specific country? These things are likely country-dependent. $\endgroup$ – gerrit Jun 17 '14 at 14:23
  • $\begingroup$ Ideally Europe and Asia,if no worldwide data-pool is available I might ask country by country. $\endgroup$ – mart Jun 17 '14 at 14:36
  • $\begingroup$ I think this is going to be difficult. It seems weatherspark.com has some information on extreme temperatures, but essential information is missing, so it's hard to use for any serious purpose. $\endgroup$ – gerrit Jun 17 '14 at 14:38
  • $\begingroup$ would a gridded daily reanalysis dataset do? have you looked at fetchclimate2.cloudapp.net or daymet.ornl.gov? $\endgroup$ – David LeBauer Jun 18 '14 at 1:01
  • 1
    $\begingroup$ @mart It depends on your purpose. I think weatherspark aggregates whatever data are available. This means that for some stations, the extreme is taken for a 5-year period, whereas for others, it is for a 100-year period. This means one should be careful in comparing extremes. $\endgroup$ – gerrit Jun 18 '14 at 13:54

CLIMDEX, the "Datasets for Indices of Climate Extremes" is probably the closest you are going to get, for any pre-computed indices. The 27 indices it includes are listed on the indices page of that site.

Unfortunately, it doesn't include anything like "coldest 3-day average per year", let alone "coldest 3-day average over the last 5 years". I doubt there would be any pre-built dataset with exactly that index, and there is unlikely to be online tool that would calculate unusual indices, because the data sets are huge (HadCRUt4 is 20MB for monthly data over the last 150 years), and unusual indices would have to be calculated on the fly - something that's just not feasible on the 'web.

I would guess that you'd have to calculate that yourself, from a raw daily temperature dataset. Unfortunately, I don't think such a thing exists on a global scale. However, a reanalysis product like the NCEP/NCAR Reanalysis is probably good enough for your purposes. However, you'll need to be familiar with large data analysis - the daily surface temperature dataset is over 450MB.

Depending on your exact usage requirements, you might be better off just looking up some of the historical records for neighbouring meteorological stations and using those.


For Europe, I'd look into this dataset : http://onlinelibrary.wiley.com/doi/10.1029/2008JD010201/abstract

This is a high-resolution gridded daily dataset derived from station data. However, this is a gridded dataset (25 km resolution) aimed at validation of Global Climate Model. So, depending on your need, this might be of help or not. if this kind of dataset suits your needs, you might also consider reanalysis such as GLDAS (0.25 degree resolution).


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.