More flexible weather data handling

The simple and none weather models are gone

The new const model (see later) can sometimes be used as replacement.

The default weather model is rewritten

There are probably many new bugs, and the error messages are likely less informative. However, it is more flexible as described below.

More fine grained weather data available weather files

You can now specify Minute, Second, and even Microsecond in the weather files. If you have set minimal_timestep, Daisy will try to make the timesteps match the weather data.

You can also specify time with a Date (or time) column. The format is one of

  yyyy-mm-dd
  yyyy-mm-ddThh
  yyyy-mm-ddThh:mm
  yyyy-mm-ddThh:mm:ss
  yyyy-mm-ddThh:mm:ss.uuuuuu

depending on the desired resolution.

All weather data describe the preceding time step, except for daily data which describe the specified date. So

  1999-02-05

will specify the weather for the fifth of February, 1999, while

  1999-02-05T04

will specify the weather for the the hour between 3 and 4 the same day (assuming hourly data).

Keywords, columns, and parameters interchangeable

All numeric data about the weather files can be specified as keywords, columns, or parameters. For example, to run with a constant vapour pressure of 1000 Pascal, you can specify

  VapPres: 1000 Pa

in the weather file header. Or alternatively, as a parameter:

  (weather default "dk-taastrup.dwf"
    (VapPres 1000 [Pa]))

Doing the later will overwrite any value specified in dk-taastrup.dwf.

Data that are not simple numbers (like Surface or PrecipCorrect) can be specified as keywords or parameters.

Single number allowed for PrecipCorrect and PrecipScale

If you specify a single number, it is assumed to represent the whole year.

New table model

The new table model differs from the default model on three points.

1. The file name is specified with an explicit file parameter. E.g.

  (weather table (file "dk-taastrup.dwf"))

2. The `missing_years' parameter is not supported.
3. The table model can be used with the combine and time models, explained below.

New const model

This is identical to the table model, except that the file parameters is not accepted. This means weather data must be specified as parameters. The value will therefore be constant, except the radiation which will still follow a day cycle based on relative extraterrestrial radiation.

New combine model

The combine model will allow you to combine multiple sources of weather data. It has a single parameter entry, which is a list of weather sources entries. Each entry have begin, end, use, and source parameters. The begin and end parameters specify the time interval where we should use this particular source. The use parameter specified the data we should extract from this particular source. The particular value Any specifies that we should use all the data provided by the source. Finally, source specifies the source. This can be any weather model but default. For each type of weather data, Daisy will examine each entry in sequence, and use the first match. If the begin and end parameters are unspecified, the entry will cover from the start to the end of the simulation. The default value for use is Any.

Fo example, we might want to use weather data from the file dk-taastrup.dwf, except for precipitation during the summer, where we instead use data from file dk-taastrup-hourly.dwf.

  (weather combine
           (entry ((begin 2006 6 1 0)
                   (end   2006 9 1 0)
                   (use Precip)
                   (source table (file "dk-taastrup-hourly.dwf")))
                  ((source table (file "dk-taastrup.dwf")))))

New time model

The time model allows you to use weather data from a different period. E.g.

   (weather time (from 2002) (to 2006)
            (source table (file "dk-taastrup.dwf")))

will use four years old weather data in the simulation. Note that all weather data is mapped, so 2001 data will be used for 2005, and 2003 data for 2007. It can be combined with the combine model to map only a selected period. For example

  (weather combine
           (entry ((begin 2006 1 1 0)
                   (end   2007 1 1 0)
                   (source time (from 2002) (to 2006)
                           (source table (file "dk-taastrup.dwf"))))
                  ((source table (file "dk-taastrup.dwf")))))

will use data from dk-taastrup.dwf, with the exception that 2002 data will be used for 2006.

Note that while dates until the 28 of February will be an exact match for the two specified years, other dates may be one day off due to leap years.

If you don't want to map year to year, you can specify offset instead of from and to.

  (weather time (source table (file "dk-taastrup.dwf"))
           (offset (days 1) (hours 0) (minutes 0) (seconds 0) (microseconds 0)))

This will use yesterdays weather data. So will this

  (weather time (source table (file "dk-taastrup.dwf"))
           (offset (days 1)))

as all the offset parameters default to zero. There is no year parameter to offset, since that is not a well defined length of time.