Skip to content

Model Catalog ​

njord can poll any weather model available through the Open-Meteo API. This page lists common models with their coverage and capabilities.

Choose models based on your location and how far ahead you need forecasts. Regional models offer higher resolution for shorter horizons; global models cover everywhere but at coarser resolution.

Global models ​

Model IDProviderMax HorizonResolutionNotes
icon_globalDWD180h~13 kmGerman Weather Service global
ecmwf_ifs025ECMWF240h~25 kmEuropean Centre high-res
gfs_seamlessNOAA384h~13 kmUS GFS blended with HRRR
ukmo_seamlessUK Met Office168h~10 kmUK Met Office global
arpege_worldMeteo-France96h~40 kmFrench global model
gem_seamlessECCC240h~15 kmCanadian blended model
jma_seamlessJMA264h~20 kmJapan Meteorological Agency
kma_seamlessKMA288h~12 kmKorean Meteorological Administration
bom_access_globalBoM240h~25 kmAustralian Bureau of Meteorology
cma_grapes_globalCMA240h~15 kmChina Meteorological Administration

European models ​

Model IDProviderMax HorizonResolutionNotes
icon_euDWD120h~7 kmDWD European domain
arpege_europeMeteo-France96h~11 kmFrench European domain
knmi_harmonie_arome_europeKNMI60h~5.5 kmDutch Met pan-European
dmi_harmonie_arome_europeDMI60h~5.5 kmDanish Met pan-European

Regional models ​

Model IDProviderRegionMax HorizonResolution
icon_d2DWDGermany, Switzerland, Austria48h~2 km
knmi_harmonie_arome_netherlandsKNMINetherlands, Belgium48h~2.5 km
metno_nordicMET NorwayScandinavia60h~2.5 km
meteoswiss_icon_ch1MeteoSwissSwitzerland33h~1 km
meteoswiss_icon_ch2MeteoSwissSwitzerland45h~2 km
arome_franceMeteo-FranceFrance48h~1.3 km
arome_france_hdMeteo-FranceFrance48h~1.3 km
geosphere_arome_austriaGeoSphereAustria48h~2.5 km
arpae_2iARPAEItaly48h~2 km
ukmo_uk_2kmUK Met OfficeUnited Kingdom, Ireland54h~2 km
hrrr_us_conusNOAAContinental US48h~3 km
nbm_us_conusNOAAContinental US192h~2.5 km
jma_msmJMAJapan, Korea78h~5 km
kma_ldpsKMAKorea48h~1.5 km
gem_hrdps_continentalECCCCanada48h~2.5 km

Tips for model selection ​

  1. Start small. Two or three models are enough to get useful forecasts and enable consensus. Adding more models increases API usage proportionally.

  2. Mix global and regional. A global model provides the long-range outlook; a regional model gives higher-resolution short-term detail.

  3. Use per-location models for regional coverage. Add regional models only to the locations they cover. See per-location model configuration.

  4. Check horizon vs. your needs. If you only display +24h forecasts, a regional model with a 48h horizon is plenty. Long-range planning needs global models.

  5. Consider the budget. Each model-location pair is one API request per poll cycle. See budget calculation for details.

Open-Meteo data is licensed under CC BY 4.0.