From api-users at lists.met.no Fri Nov 1 13:11:25 2019 From: api-users at lists.met.no (api.met.no mailing list) Date: Fri, 1 Nov 2019 14:11:25 +0100 Subject: [api.met.no] Change in CAP alerts for Norway + some FAQs Message-ID: Dear API users, For users in Norway we would like to inform about a change in the CAP output used on MetAlerts regarding icy driving conditions. The eventCode "drivingConditions" will be replaced by the more specific "ice" and "blowingSnow" eventCodes. More information here (in Norwegian): https://www.met.no/nyhetsarkiv/farevarsling-av-is-og-snofokk-erstattar-vanskelege-koyretilhove At the same time we would like to address a few common questions, as well as correct an error in the previous mail. Locationforecast ---------------- The new model output available under /test is nearing completion, and we do not anticipate any further changes before going into production at the end of the month. During the test period we've made some minor changes and fixed some bugs, here's a quick summary: - the model name is now called met_public_forecast instead of FORTI - the global model is now updated every 6 hours instead of 12 (not 3 instead of 6 as stated in the previous email) - "runended" and "nextrun" has had some bugfixes and now work as in the old (current) version Example: https://api.met.no/weatherapi/locationforecast/1.9/test?lat=60.10&lon=9.58&msl=70 Q. Until now I have been looking at the 'runended' parameter to tell me when the forecast model was last updated (i.e. how "fresh" the forecast is). How do I find out when the underlying forecast model was updated for the received data? A. The "runended" parameter is only meaningful when comparing two different forecasts for the same location (the newer one hopefully being more correct). To find out when to download newer versions, use the "nextrun" parameter. Note however that the "nextrun" timestamp is *not* an exact promise, but an approximate guess of when the next model will be ready (variable depending on supercomputer load) plus a random delay to allow for model copying latency, XML caching timeouts etc. Q. Why can't I just add a constant time interval to the "runended" parameter? A. In addition to the reasons above, the model update periods for the various areas is expected to change without notice as we get more data and constantly try to improve the product. When the model update frequency increases, this will be reflected in the "nextrun" parameter. Q. Are historical weather forecasts available via the api? A. Unfortunately no, as the sheer amount of data makes it impossible. Our Nordic model run data are available as NetCDF files as described in the API documentation. Unfortunately we cannot distribute global models due to licensing issues. Radar animations ---------------- Q. Will the GIF animations be discontinued at the start of 2020? What about the Javascript slideshow? A. We are looking into the possibility of open-sourcing the timeslider library from gaf.met.no and other sites, but it is not trivial to decouple from React and other dependencies. For this we need to allocate more resources, but at the moment our hands are full migrating over legacy systems and hardware. As such there is no firm date when we will stop producing GIF radar animations. Best regards, Geir Aalberg API developer From api-users at lists.met.no Wed Nov 27 14:31:06 2019 From: api-users at lists.met.no (api.met.no mailing list) Date: Wed, 27 Nov 2019 15:31:06 +0100 Subject: [api.met.no] Changes to locationforecast/1.9 and textforecast/2.0 on Dec 3 Message-ID: Dear API users, on December 3rd we will be launching the new backend for locationforecast/1.9. As mentioned previously there are some minor differences, e.g. the 2- and 3-hour intervals have been discontinued, and the model names changed. Should you experience problems with the new format, we advice you to temporarily switch to using LocationforecastLTS until you can make the necessary changes. The LocationforecastLTS product is now deprecated, but will continue running for at least 3 more months. https://api.met.no/weatherapi/locationforecastlts/1.3/documentation At the same time we will also be launching the new *sea* text forecast. The existing *seahigh* and *seabank* forecasts will continue until February 1st. At the same time the XML schema will be upgraded to version 0.3: https://schema.api.met.no/schemas/textforecast-0.3.xsd Regards, Geir Aalberg API developer