APIs for recent historical weather data

51 views
Skip to first unread message

Cord Kaldemeyer

unread,
May 19, 2021, 10:10:49 AM5/19/21
to openmod-i...@googlegroups.com

Hi everybody,

 

does anybody know services that provide recent historical weather data?

Practical example: I want to obtain the wind speed for a certain point in Germany in 100 m altitude for the day/week before in (quarter-)hourly resolution.

 

The service windy.com provides something like this but only for forecasts: https://api.windy.com/point-forecast

An alternative could be openweathermap: https://openweathermap.org/price

 

But both are commercial services and I am looking for alternatives. Any hints?

 

Best wishes and thanks in advance,
Cord

 

Cord Kaldemeyer

Data Scientist

 

ARGE Netz GmbH & Co. KG

Haus der Zukunftsenergien, Otto-Hahn-Straße 12-16, 25813 Husum                                         

Beisheim Center, Ebertstraße 2, 10117 Berlin

Mobil: +49 (0)176 - 457 013 43

Telefon: +49 (0)4841 - 89 44 - 667

Telefax: +49 (0)4841 - 89 44 - 679

kaldemeyer|at|arge-netz.de

Geschäftsführung: Stephan Frense (CEO), Henning Holst

Sitz der Gesellschaft: Husum

Amtsgericht Flensburg: HRA 6501 FL

USt.-ID Nr.: DE 26 727 1472

 

Madalsa Singh

unread,
May 19, 2021, 12:58:28 PM5/19/21
to openmod initiative
Hi, 

You can use NOAA's Integrated Surface Database for hourly data. It has data for almost 507 stations in Germany and 73 of them have historical timeseries until 2019. 
More info about ISD and variables it documents : https://www.ncei.noaa.gov/data/global-hourly/doc/isd-format-document.pdf
Station level information can be extracted using : https://www.ncdc.noaa.gov/homr/
If there's interest in the community for easy station level data parsing and imputation (from the closest station and/or previous day time of day value), I can push my code for scraping it. Let me know.

Madalsa

Jack Kelly

unread,
May 19, 2021, 2:53:18 PM5/19/21
to Madalsa Singh, openmod initiative
If you want data for a specific geographical location, and if that geographical location doesn't happen to have a weather station nearby (which will be true for all locations 100 m above the surface!) then you probably want either a "reanalysis dataset" (like ERA5); or take time=0 from numerical weather predictions (t=0 NWPs and reanalysis datasets aren't exactly equivalent). NOAA's Global Forecast System NWP is freely available.  Oikolab provides API access to NWPs, from 1950.

--
You received this message because you are subscribed to the Google Groups "openmod initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openmod-initiat...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openmod-initiative/2f0dd580-591d-47cc-b37c-06c17397927cn%40googlegroups.com.

nico....@gmail.com

unread,
May 20, 2021, 3:22:18 AM5/20/21
to openmod initiative
In the Netherlands the Royal Netherlands Meteorological Institute (KNMI) provides a multitude of weather information: https://dataplatform.knmi.nl/dataset/?groups=wind
Maybe Deutscher Wetterdienst can help you out?

Shameless plug: at my employer Alliander (DSO) we recently open-sourced a project to expose various weather information in a uniform format for application to use it:
If you're looking to get the data for your use-case in a production-ready solution, this might be interesting to you. Feel free to reach out to the team of the Weather Provider API with questions.

Best,
Nico Rikken

David Brayshaw

unread,
May 20, 2021, 5:20:39 AM5/20/21
to nico....@gmail.com, openmod initiative

Dear Cord et al,

 

Is worth a look at the COPERNICUS data store (+its associated API):

 

https://cds.climate.copernicus.eu/#!/home

 

This should help access many different forms of climate data (inc models, forecasts, reanalyses etc).  In terms of your specific request, it may not help much with sub-hourly data but be able to assist with hourly (e.g., via ERA5).

 

Best wishes,

 

David

Image removed by sender.

Geschäftsführung: Stephan Frense (CEO), Henning Holst

Sitz der Gesellschaft: Husum

Amtsgericht Flensburg: HRA 6501 FL

Image removed by sender.USt.-ID Nr.: DE 26 727 1472

 

--
You received this message because you are subscribed to the Google Groups "openmod initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openmod-initiat...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openmod-initiative/2f0dd580-591d-47cc-b37c-06c17397927cn%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "openmod initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openmod-initiat...@googlegroups.com.

Robbie Morrison

unread,
May 20, 2021, 6:30:10 AM5/20/21
to openmod-i...@googlegroups.com, Carsten Hoyer-Klick (DLR)

Hello all, hello Carsten (in direct cc)

To add that the Open Energy Ontology covers weather data:

In particular:

  • the use‑case at § 8.3 (p 11,12): data annotation of an energy meteorological time series data set
  • figure 4 (below also) covers time series annotation: is the value instantaneous, averaged, or integrated? if it is a point‑in‑time reading, is the temporal alignment start, mid, or end of span? which timezone is used and if local, is daylight saving applied or not?
[timeseries-attributes]

On re‑reading, it is not clear, to me at least, what "average" means precisely in the articulated use‑case: the arithmetic mean of the the start and end instantaneous values or a synonym for "integrated" or with some nonlinear weighting function applied or some kind of rolling average or some other kind of mean or something else?  (I guess I could look on GitHub but the text needs to be clear too.)

Note that the OEO does not import a specialist parent ontology to cover weather information (as it does for financial concepts and for scientific units, for instance).

with best wishes, Robbie

To view this discussion on the web, visit https://groups.google.com/d/msgid/openmod-initiative/6B8D2DF7-0C14-4A67-9B53-EE833F1B7F2F%40reading.ac.uk.
-- 
Robbie Morrison
Address: Schillerstrasse 85, 10627 Berlin, Germany
Phone: +49.30.612-87617

Robbie Morrison

unread,
May 20, 2021, 6:52:18 AM5/20/21
to openmod-i...@googlegroups.com

To continue, could also be related to intensive and extensive quantities, the former averaged and the latter integrated .. desperately confused, R

Robbie Morrison

unread,
Aug 30, 2021, 7:01:58 AM8/30/21
to Carsten.H...@dlr.de, openmod-i...@googlegroups.com

Hi Carsten, all

Thanks!  On reflection, there are two distinct types of variables alluded to on the y‑axis: intensive and extensive:

The "integrated" line applies to extensive variables and the "averaged" line to intensive variables.

So the problem lies mostly in the diagram and not with the underlying definitions themselves.  I suggest therefore you make two diagrams or otherwise indicate this feature.

By the way, the concept of intensities and extensities applies to domains outside of thermodynamics, including price/quantity (p,Q) pairs in economics.

Sorry to keep flogging the horse, but I was very confused when I first encountered this depiction.

with best wishes, Robbie

On 30/08/2021 12.45, Carsten.H...@dlr.de wrote:

Hi everyone,

 

sorry for the late answer…..

 

Von: Robbie Morrison <robbie....@posteo.de>
Gesendet: Donnerstag, 20. Mai 2021 12:30
An: openmod-i...@googlegroups.com
Cc: Hoyer-Klick, Carsten <Carsten.H...@dlr.de>
Betreff: Re: [openmod-initiative] Re: APIs for recent historical weather data

 

Hello all, hello Carsten (in direct cc)

To add that the Open Energy Ontology covers weather data:

In particular:

  • the use‑case at § 8.3 (p 11,12): data annotation of an energy meteorological time series data set
  • figure 4 (below also) covers time series annotation: is the value instantaneous, averaged, or integrated? if it is a point‑in‑time reading, is the temporal alignment start, mid, or end of span? which timezone is used and if local, is daylight saving applied or not?

[timeseries-attributes]

On re‑reading, it is not clear, to me at least, what "average" means precisely in the articulated use‑case: the arithmetic mean of the the start and end instantaneous values or a synonym for "integrated" or with some nonlinear weighting function applied or some kind of rolling average or some other kind of mean or something else?  (I guess I could look on GitHub but the text needs to be clear too.)

Note that the OEO does not import a specialist parent ontology to cover weather information (as it does for financial concepts and for scientific units, for instance).

 

[CHK] Maybe we should get more precise. Average usually means some integration divided by the time step duration or number of samples taken within the time step. It maybe needs a more precise formulation in the ontology.

 

Carsten

Reply all
Reply to author
Forward
0 new messages