Hi @yosafatbudi, thank You for reporting the issue. Semarang radar site was not delivering new data until not long ago, so it was disabled on our side. It is enabled now:

Hi @Dhanur-Pramono, I believe that Juanda is covered by Surabaya radar site, which has been unavailable since 26th of September. This is last data available, as available directly on BMKG web site:

Hi @Helipaddy, when out of bounds, You get nothing but null values for all the parameters.

Hi @lexolutha, ACCESS model on windy is ACCESS-G (about 12km resolution):

ACCESS-C (1.5km resolution) is separated by region (VT is Victoria/Tasmania, SY is Sydney and so on):

Hi @przemo1986, as mentioned in ECMWF release notes, 00 and 12 runs are 360 hours long (15 days) and 06 and 18 runs are 144 hours (6 days) long.

Previously it was 240 hours (10 days) and 90 hours (3.75 days):

Hi @KGP67, thanks for the bug report. There was indeed an issue that's fixed now.

Hi @stratimi, can You, please, point me to the data? I was only able to find this DMI Open Data page that unfortunately doesn't contain any mention of precipitation type.

Hi @Ken, You are right, we're currently using only 36 hours from the forecast because all runs are at least 36 hours long. We might use full 42 hours from 18z run in the future, but I unfortunately can't give You any estimate.

Hi @Alan-Cameron, thank You for reporting the issue, it is fixed now:

Hi @illili, we currently have no intention of adding AIFS back. It's resolution and parameter set is very inferior to deterministic version of ECMWF forecast (HRES).