Feature request: allow selecting Open-Meteo weather model (e.g. Canadian GEM) for more accurate regional forecasts #140

New Issue

2026-10-05T01:25:35+02:00

adam commented

2026-10-05 01:25:35 +02:00

Is your feature request related to a problem? Please describe.

get_wx_meteo() in mesh_bot.py currently calls Open-Meteo's default "best_match" blended model. For users near large lakes or other locally-complex terrain, this can diverge noticeably from what national weather services (e.g. Environment Canada) forecast for the same area, since best_match blends several global models rather than prioritizing the highest-resolution regional model for that location.

For example, users near the Great Lakes in Ontario often see meaningfully different fog/precipitation forecasts from mesh_bot (Open-Meteo best_match) versus Environment Canada, likely due to lake-effect weather being harder for coarser global models to resolve.

Describe the solution you would like

Open-Meteo's existing /v1/forecast endpoint already supports a models query parameter, so no new endpoint integration is needed. For example, Canadian users could request Environment Canada's own GEM model directly:

https://api.open-meteo.com/v1/forecast?...&models=gem_seamless

Suggest adding an optional config.ini setting under [location], e.g.:

# Optional: specify an Open-Meteo model for more locally-accurate forecasts.
# Leave blank for Open-Meteo's automatic best_match selection.
# Examples: gem_seamless (Canada), icon_seamless (Europe), jma_seamless (Japan)
meteoWxModel =

Then in get_wx_meteo(), conditionally add params["models"] = my_settings.meteo_wx_model if the setting is non-empty.

Describe alternatives you have considered

Manually patching get_wx_meteo() locally, but this would need to be redone on every update, and other users in Canada (or other regions with a strong local model) would benefit from this being a supported option rather than everyone patching it themselves.

Additional context

Full list of available Open-Meteo models: https://open-meteo.com/en/docs (see "Weather models" dropdown). This would be a small, backward-compatible change - leaving the setting blank preserves current default behavior exactly.

**Is your feature request related to a problem? Please describe.** get_wx_meteo() in mesh_bot.py currently calls Open-Meteo's default "best_match" blended model. For users near large lakes or other locally-complex terrain, this can diverge noticeably from what national weather services (e.g. Environment Canada) forecast for the same area, since best_match blends several global models rather than prioritizing the highest-resolution regional model for that location. For example, users near the Great Lakes in Ontario often see meaningfully different fog/precipitation forecasts from mesh_bot (Open-Meteo best_match) versus Environment Canada, likely due to lake-effect weather being harder for coarser global models to resolve. **Describe the solution you would like** Open-Meteo's existing /v1/forecast endpoint already supports a `models` query parameter, so no new endpoint integration is needed. For example, Canadian users could request Environment Canada's own GEM model directly: `https://api.open-meteo.com/v1/forecast?...&models=gem_seamless` Suggest adding an optional config.ini setting under [location], e.g.: ``` # Optional: specify an Open-Meteo model for more locally-accurate forecasts. # Leave blank for Open-Meteo's automatic best_match selection. # Examples: gem_seamless (Canada), icon_seamless (Europe), jma_seamless (Japan) meteoWxModel = ``` Then in get_wx_meteo(), conditionally add `params["models"] = my_settings.meteo_wx_model` if the setting is non-empty. **Describe alternatives you have considered** Manually patching get_wx_meteo() locally, but this would need to be redone on every update, and other users in Canada (or other regions with a strong local model) would benefit from this being a supported option rather than everyone patching it themselves. **Additional context** Full list of available Open-Meteo models: https://open-meteo.com/en/docs (see "Weather models" dropdown). This would be a small, backward-compatible change - leaving the setting blank preserves current default behavior exactly.

adam closed this issue

2026-10-05 01:25:35 +02:00

adam commented

2026-10-05 01:25:36 +02:00

let me know if that didn't handle it :)

let me know if that didn't handle it :)

adam commented

2026-10-05 01:25:37 +02:00

It worked! Thanks!

It worked! Thanks!

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: adam/meshing-around#140