dwdGribExtractor 0.0.7

dwdGribExtractor: A tiny Python interface to request and extract NWP grib file data from opendata.dwd.de

dwdGribExtractor is a smart package to easely extract data from numerical weather prediction grib files provided by DWD. The difference to other packages is the location flexibility. So every point on the 2.2km x 2.2km grid (ICON-D2) can be retrieved for the next X forecast hours.

At the moment only ICON_D2 is supported.

Supported weather variables

For currently available weather variabels see: ICON User Manual Chapter 6.1.4 table 6.4

Installation

Install with pip:

$ pip install dwdGrilbExtractor

Dependencies

To read the grib2 files xarray with cfgrib engine is used. The easiest way to install cfgrib and all its binary dependencies is via Conda:

$ conda install -c conda-forge cfgrib

alternatively, if you install the binary dependencies yourself, you can install the Python package from PyPI with:

$ pip install cfgrib

Example

locationList = {
    "Vienna": {
        "lat": 48.20,
        "lon": 16.37
    },
    "Graz": {
        "lat": 47.07,
        "lon": 15.43
    }
}

variables = ["aswdir_s", "aswdifd_s", "t_2m"]

forecast = ICON_D2(locations = locationList, forecastHours = 3)
data = forecast.collectData(varList = variables, cores = None) # Disable multiprocessing
#data = forecast.collectData(varList = variables, cores = 4)


#### Indexing one location
loc = "Graz"
result = data.loc[loc]
result = data.loc[loc, "2021-09-06 06:15:00"]

#### Indexing multiple locations
loc = ["Graz", "Vienna"]
result = data.loc[loc]

#### Indexing one location with datetime condition
loc = "Graz"
mask = data.loc[loc].index.get_level_values(0) > np.datetime64('2021-09-06T06:15:00')
mask = data.loc[loc].index.get_level_values(0).hour == 8
result = data.loc[loc][mask]

#### Indexing multiple locations with datetime condition
loc = ["Graz", "Vienna"]
mask = data.loc[loc].index.get_level_values(1) > np.datetime64('2021-09-06 06:15:00')
mask = data.loc[loc].index.get_level_values(1).hour == 8
result = data.loc[loc][mask]

Knwon Issues

Windows

• Multiprocessing on some windows machines may not work. Disable it by setting forecast.collectData(varList = variables, cores = None)

• Spyder IDE does not produce print outputs if multiprocessing is enabled.

• Dont run the code in Spyder with F5 or debug mode. This calls runfile() and sometimes crashes memory.

Author

Manuel Strohmaier

License