precipitation (24 h) and look ahead (for less than 24 h) deliver predictions for different points in time

[ekkards]

Use case: Show weather forecast on a pixel clock

• user may look at clock several times a day and information should update frequently
• a short look ahead time seems desirable, e.g. 3, 6 or 12 hours.
• due to limited space, no context info can be delivered and the assumed context is "for the same time"

Expectation: By choosing a suitable look ahead time a consistent data set of temperature and precipitation can be obtained from the node-red module.

Observation: The time intervals for temperature and precipation do not match

The value "precipitationNext24h" is generated as follows, apparently containing 24 h as a fixed value. "precipitationNext24h" : Math.round(sumFutureValue(node.mosmixStation, "RR1c", 24, forecastDate) * 10) / 10,

Workarounds: Attempted workaround with using "additional fields" also has its short coming

• values are rather course grained, i.e. full day or half day.

  • wwPd % (0..100) Probability: Occurrence of any precipitation within the last 24 hours

  • wwPh % (0..100) Probability: Occurrence of precipitation within the last 12 hours

  • Rd00 % (0..100) Probability of precipitation > 0.0 mm during the last 24 hours

  • Rh00 % (0..100) Probability of precipitation > 0.0mm during the last 12 hours

  • RRdc kg / m2 Total precipitation during the last 24 hours consistent with significant weather

  • RRhc kg / m2 Total precipitation during the last 12 hours consistent with significant weather

• and apparently the value changes only twice daily (probably due to usage of the MOSMIX L dataset).

Suggested change: For look ahead time < 24 h, deliver a value for precipitation, wich is only summed up until the look ahead time. "precipitationNext24h" : Math.round(sumFutureValue(node.mosmixStation, "RR1c", lookAheadHours, forecastDate) * 10) / 10,

[SmartyKNX]

Unfortunately you will not get reliable values anyway for the next hours anyway, because the MOSMIX L dataset is only updated 4 times a day.

I would prefer using the MOSMIX S dataset for this usecase, but therefor the code needs to modified to extract a single station from the KMZ-File containing all stations. (I have no idea, why DWD has no single station files for the MOSMIX S dataset.)