Today is the last day of October so it’s a good time to report on what I’ve been doing in the past month. I am halfway to reaching my goal of displaying the detailed results of our high-resolution Alpine forecasts, and I would still appreciate your support! In October, I made several experiments to find a technical solution that scales to our needs, I migrated soarV2 to the selected technology, and I partially implemented the rendering of our high-resolution forecasts. You can see the result here: https://soarwrf1.soaringmeteo.org/v2-dev In the “layers” menu, select the WRF model, and one of the covered zone. Then, play with the various overlays (e.g. thermal velocity, rain, wind at various altitudes, etc.). However, note that I did not yet implement the meteograms for the WRF model. You can also watch a video that I just recorded where I look at the XC flying potential over the Europe, and I zoom into the South of France, and switch to the WRF model for more precision. It seems that around Col de Bleyne will be flyable tomorrow! https://soarwrf1.soaringmeteo.org/v2-dev/demo.webm Last, note that I have also changed the way we render the GFS model. Instead of dealing with “points”, we now deal with “zones”. In the process, I have significantly expanded the coverage of the South Hemisphere, I hope this will be useful to the pilots there since the season just started! I am super happy about what I’ve done, but as I said above, I didn’t reach my goal yet. My progress was slowed down by unanticipated challenges when I migrated soarV2 to the new engine. I still need to update the implementation of the meteograms and sounding diagrams to handle the WRF forecast data, and then I will improve the user interface to make it more intuitive and easy to use on small screens. Now comes the part where I request again your support to fund my work. I received a lot of money already: 5,500 € in total (some donators preferred to give directly to me instead of going through Leetchi), and I am extremely grateful to you! I could spend almost all of my time on the project (17 days in total, to be precise). However, since I now estimate that I need more weeks to complete the project, I will keep the money pot open, and I would appreciate if you can still support me!

I just published a demo of the new engine here: https://soarwrf1.soaringmeteo.org/v2-dev/ I invite you to try it and let me know if it works well on your setup or not (send an email to [email protected]). I’ve already noticed some rendering artifacts on the map overlays, on Firefox Android… I hope I will find a solution to fix that issue soon. The most visible changes are: - We always use a simple arrow to show the wind on the map instead of wind barbs, - When you zoom out, the content of the overlay (e.g., the XC flying potential) is not anymore down-sampled. That change is precisely what will enable us to render the results of our Alpine forecast on the map! Other than that, I applied many other changes that are not visible: - The underlying map engine is OpenLayers instead of Leaflet, - The backend now stores the forecast data on the disk instead of loading everything in the memory, - The backend now produces PNG images and vector tiles in addition to JSON documents. With those changes, we have an infrastructure able to process around 50 variables (temperature, boundary layer depth, etc.) on more than 100k points and 70 time steps. That’s about 350 million data points in total! As a matter of comparison, our previous coverage of the GFS model had 5 times less points. Getting there was not easy and I faced several challenges along the way. Those changes will allow us to render the results our Alpine forecast. That being said, processing more points takes more time (unless we buy more servers). This is why the link above shows only the Europe instead of the whole world. I still need to figure out some more optimizations! The next steps are the following: - Find a solution to not reduce our coverage of the GFS model. - Clean up the code, test it thoroughly, and update soarV2 with the new engine. This should be done before the end of the month. - Finally, integrate our Alpine forecast in the new engine! That should take a couple more weeks at least.