The AerisWeather forecasts include information from a blend of models and other data sources, including radar and satellites. Each model and dataset update at different time intervals throughout the day. Global models typically update every 6 hours, while regional models, such as the US NDFD, update as often as every 30-60 minutes. Other data sets, such as radar and satellite information, update as frequently as every 2-5 minutes.
Additionally, individual models may run at set intervals, such as the global GFS and ECMWF at 00Z, 06Z, 12Z, and 18Z. The completion time of each model run varies, as does the model fetching and processing within the forecast platform. Due to the individual model timing variability and the utilizing of multiple forecast models, we cannot guarantee a specific update availability time.
Since the forecast models and data sets update at different frequencies and a specific availability time for each model cannot be guaranteed, AerisWeather recommends basing your forecast update frequency on your specific use case. For example, if you are querying daily forecasts, updating every 4 to 6 hours is often sufficient. If you are utilizing hourly forecasts, then we recommend updating at least hourly. Or, if utilizing our minute forecasts, we suggest updating forecasts as frequently as every 1 to 2 minutes.
Alternatively, the AerisWeather API includes cache headers with an approximate time to cache the forecast, starting at 20 minutes for most forecasts and 60 seconds for minute forecasts.
Contact the AerisWeather Sales team to learn more about optimally leveraging forecasts for your particular application and regions of interest.