PPI

PPI is a weather radar product display running in the browser. Despite it's name referring to a very specific product, PPI, it can (theoretically) be used to visualize any raster products in Cartesian coordinates.

There's a deployment running at https://radar.a3h.net/ that displays FMI's PPI reflectivity products for individual radars.

Features

• reprojection
• coloring
• product animation
• decent base map
• keyboard shortcuts (SPACE, LEFT, RIGHT, S, P, F)
• rudimentary cursor tool (show value under cursor)

Future plans

• show distance to points of interest, e.g. to closest localities / max reflectivity
• range rings
• more products (or moments, e.g. radial velocity)
• more color scales
• product metadata display (PRF etc)
• rendering optimizations
• different animation speeds
• time rather than product based animation
• atmospheric pressure data overlay
• other than FMI data (NEXRAD?)

Known issues

• animation state doesn't currently survive page refresh
• does unnecessary product loads at initialization

Components

There are three main components:

1. product downloader (only for FMI right now),
2. distribution collector (collects downloaded products, saves them for web distribution),
3. client software (JavaScript single page app written in modern JS; React, Redux, OpenLayers)

Create deployment

cd client
make clean && make
cd ..
mkdir -p client/build/radar
python fmi/dist_builder/collect_radar_products.py /path/to/where/you/downloaded/the/stuff | \
    python collect.py fmi/dist_builder/raster_to_json.py client/build/data
cd client
make

Thoughts and ideas

• Line density display: https://twitter.com/archillect/status/938533533810937856
• FMI summer color scale for reflectivity http://wms.fmi.fi/fmi-apikey/<>/geoserver/Radar/ows?service=WMS&request=GetLegendGraphic&format=image%2Fpng&width=500&height=100&layer=anjalankoski_dbzh&style=Radar+dbz+Summer
  • From QGIS coloring, should look at the above legend graphic instead

    FMI reflectivity (summer)
    0 => 255, 255, 255   (nodata)
    21 => 108, 235, 243  (kohtalainen)
    62 => 65, 154, 90    (kohtalainen)
    84 => 241, 243, 90   (sakea)
    168 => 206, 2, 2     (sakea)
    189 => 131, 10, 70   (hyvin sakea)
    251 => 244, 244, 244 ()
    

• Smoothing where every value around is some data value and the value in the point is not scanned or no echo.

Licensing

GPLv3