Not deleting NAs in the data frame

[JohannesOberpriller]

I would prefer to not delete the NAs in the data frame, because when comparing different species in mixed runs this could be difficult to understand when for a species no information is present in the output, although it was initalized and useres may want to have this information.

[trotsiuk]

Motivation

The reasons for removing NAs are the following:

• the output array is of dimension m_month, n_spesies, 10, 15. Some of the output columns are empty. The fortran output had to be in the array, thus we grouped the output to the groups, but some groups have more variables than others. To see all the groups and variables see data("i_output"). For me doesn't make sense to return those empty columns to the long format transf.out.
• Some of the variables are not possible to measure if the forcing or scaling data are not provided. E.g. you can't provide the variable group = weibull if you the settings are correct_bias = 0. Potentially we can output those as NAs, but again not much sense for me.

Suggestion

For the comparing with the observational data I would suggest to use the output array instead of long format output. In the model run set the df_out = FALSE.

• this will eliminate the above mentioned issue
• speed up a lot your simulation. df_out = TRUE is very time consuming.

Regues

For the merge request please suggest the solution to resolve two points mentioned under Motivation