NEXRAD quick plots

Easy Python download and plot NEXRAD N0Q compositive reflectivity. Uses RGB high resolution PNG images of North America.

Tested with pytest, flake8 and mypy type checking.

python -m pip install -e .

Usage

RGB data scaling: NEXRAD N0Q base reflectivity maps.

• Black: No Data
• -32 dBZ .. 90 dBZ
• 0.5 dBZ increment

These data are reduced fidelity RGB images. We use xarray.DataArray and plot image by image. For high-fidelity science data, the lower level data are needed--contact us if interested.

import nexradutils as nq

dat = nq.load('~/data/2015-01-19/nexrad/nexrad2015-01-19T01:15:00.png')

>>> dat
<xarray.DataArray (lat: 540, lon: 1220, color: 3)>
array([[[255, 255, 255],
        [255, 255, 255],
        ...,
        [255, 255, 255],
        [255, 255, 255]]], dtype=uint8)
Coordinates:
  * lat      (lat) float64 23.0 23.05 23.1 23.15 23.2 23.25 23.3 23.35 23.4 ...
  * lon      (lon) float64 -126.0 -125.9 -125.9 -125.8 -125.8 -125.7 -125.7 ...
  * color    (color)  'R' 'G' 'B'
Attributes:
    filename:  ~/data/nexrad2015-01-19T01:15:00.png
    wldfn:     None
    time:      2015-01-19 01:15:00

.lat and .lon are vectors of geodetic latitude and longitude respectively, computed based on the .wld file corresponding to the images.

Download NEXRAD data

Get NEXRAD reflectivity data with parallel download:

download-nexrad start stop outdir

example: download from 2018-01-01 to 2018-01-02 to ~/data/nexrad:

download-nexrad 2018-01-01T00 2018-01-03T00 ~/data/nexrad

Plot NEXRAD reflectivity data

NEXRAD plots are georegistered via Cartopy, which is the replacement for deprecated Basemap

• Plot all data in directory:

  plot-nexrad ~/data/nexrad/

• Plot a specific file (subplots if multiple files specified):

  plot-nexrad ~/data/nexrad/2018-01-01T12:35:00.png

• Plot via file glob match:

  plot-nexrad ~/data/nexrad/ -pat 2018-01-01T12*.png

• Keogram (specify lat or lon and value):

  plot-nexrad ~/data/2018-01-01/nexrad/ -keo lat 40

Notes

Coordinates

EPSG:4326 coordinates (WGS84) are in .wld files, which are generally the same for wide time spans of data. The .wld format is like:

0.005 (size of pixel in x direction)
0.0 (rotation of row) (Typically zero)
0.0 (rotation of column) (Typically zero)
-0.005 (size of pixel in y direction)
-126.0 (x coordinate of centre of upper left pixel in map units--here it's WGS84 longitude)
50.0 (y coordinate of centre of upper left pixel in map units--here it's WGS84 latitude)

Mass image downscaling

For initial analysis, the original Nexrad image size of 12200 x 5400 pixels may be too high to complete in a reasonable time. I choose to downsize by a factor of 10, which takes a long time, but is a one-time process.

mkdir orig
cp *.png orig

nice mogrify -scale 10% "*.png"

If you have trouble with this being very slow, try:

MAGICK_TEMPORARY_PATH=/run/shm nice mogrify -scale 10% "*.png"