search

jdemaeyer / brightsky

JSON API for DWD's open weather data.
https://brightsky.dev/
MIT License
290 stars 18 forks source link

Add "apparent temperature" parameter #112

Open kaerbr opened 3 years ago

kaerbr commented 3 years ago

Is it possibe to add a "feels like temperatur"? (:

jdemaeyer commented 3 years ago

Hi @Kaerbr

I am a little hesitant to add this as we're trying to stay close to the DWD raw data and they do not provide an apparent temperature. On the other hand, we have already added an icon field (without basis in the DWD data) on the grounds that it's useful for many use cases and helps with replacing Dark Sky, so there's definitely a good case to be made for adding an apparent_temperature field.

I found a simple wind chill index formula used in North America. The DWD apparently uses a more complex "Klima-Michel-Modell" for which I couldn't find any formula(s) and which is not published on the open data server. Do you have any experience with these?

kaerbr commented 3 years ago

Hi @jdemaeyer ,

I do see your point, in staying as close as possible to the DWD API / raw data. Furthermore I must admit, that I havent looked up what data is provided from DWD. 🙈

I'm very much no expert with climate models. But on a short research on the Klima-Michel-Modell it seems like there is too much variation that one has to take in account that its not worth the effort.

The wind chill index formular seems more straight forward but I guess it's more for really cold environments?!

But thanks for the detailed comment! (:

jdemaeyer commented 3 years ago

Apparently people are pretty upset with the wind chill index and, like you say, it's geared towards frostbite warning, and not meant as a more general "feels like" temperature.

There is an apparent temperature formula developed by Robert Steadman which could match what we're looking for. However, it requires knowing the solar radiation. We're currently not parsing this parameter but it is available in both real-world observations (albeit at only very few stations) and MOSMIX (Rad1h), so if we add parsing for these we could also spit out an apparent_temperature according to Steadman's formula.

I find this intriguing and it's definitely a useful feature for Bright Sky, however I cannot make any promises when I'll get to it :/ (I am open for PRs of course ;))