Undervalued Canadian Radar Reflectivities

WeatherInThePeg

Just following up on this as I'm not sure if the RADAR processing techniques have changed since this adjustment or not.

Over the past couple months, I've noticed RADAR returns on the Canadian Prairies looking quite weak...far weaker than they should be. When comparing to other RADAR applications, such as Radarscope, the returns in Windy can often be 10-15 dBz lower than the raw data is suggesting.

Which makes me wonder...if you're ingesting the pre-made imagery, if it's CAPPI imagery then it really needs to be altered between winter and summer (1.0 vs 1.5 km CAPPI). I believe on their GeoMet API, the RADAR imagery has been standardized to 1km, but I'm not familiar with the ins and outs of that.

Anyways, all that to say that I've noticed that the RADAR returns are very cool on the Canadian Prairies since we moved into winter (to the point that sometimes I can barely see precipitation moving through the area) and my gut feeling is that it's due to a 1.5 km CAPPI which isn't particularly applicable in winter time.

I do wish Canada just put out a PPI image and called it a day.

WeatherInThePeg

@WeatherInThePeg Just following up to say that I realized the GeoMet products are on a 1km grid, not a 1km altitude.

Filip_K

Hi @WeatherInThePeg, we were using per-site RAIN part of PRECIP-ET product (docs here, data here) from MSC Datamart.

Unfortunately they are separate for rain and snow. Rain product seems to be okay for warm half the year, snow for the cold one. I have switched our ingress processing to start using the SNOW product for now, which should help temporarily. This is the difference

In the long run, we'll have to find a way to do this automatically, probably combining both products together

idefix37

@Filip_K
Would it be possible, in this particular case where snow precipitation can be separated from rain precipitation, to mix them with a different colour scale. As is done by Meteociel here:
https://www.meteociel.fr/observations-meteo/radar2.php?mode=1
Note this is just a guess of the portion of precipitation that could fall as snow by filtering it with temperature forecast by AROME.

Snow colour scale could be brown -> red -> white.
Just a thought …

Filip_K

Hi @idefix37, yes, my thoughts exactly. We definitely want to use rain/snow information wherever possible, we're just not there yet. Some providers already give this information (in form of hydrometeor clasification, separate rain/snow products, ...) and some don't.

We're in very early stages but this is something we want to upgrade down the road

WeatherInThePeg

@Filip_K Thanks so much for the quick fix to another Canadian RADAR issue.

It is unfortunate; I do wish that either raw data was freely available or they just also produced a PPI image so there wasn't seasonality to the applicable imagery.

They are producing a precipitation type product on their GeoMet API under the Radar_1km_SfcPrecipType product, but I don't think that would be particularly helpful in an automated system.

Thanks again for looking into it and the quick adjustment!