Version 6.2.6 New method to handle Multiple Treatment or Replicates

The new method allow simple shift between various treatment/replicates by using a Filter Variable in the input files for validation. The Filter Variables already existed in previous version 5.4.1. The new methods to facilitate the use of the filter variable is to reshape normal PG-input files from containing many variables representing various position in a field based on the corresponding information available in the ID-field of the PG files. Instead of having many variables the new reshaped files will have fewer variables and instead replicated dates with addition variables to specify a number to represent the position.

A simple example may be that we have 4 treatment and each treatment is represented by one variable like in the examaple below using a file called FlakaLAI.bin

In that particular example we simple have Id specied by C, I, F and IL.

By selectiing the tab Create New File we can select the new option Reshape current….

And a new Window will be shown that display information about how the file can be converted to a new file:

Translation codes will be suggested based on default number. In the example above 9 translation codes are suggested with corresponding pairs of Treatment id and Treatment No. Note that hits are for each of the 4 first treatments and no hits for the remaining Treatment ids. Providing this is a suitable representation you can simple click on the button Create New File. Otherwise if you change the translation codes you have to click on the button Update Translation prior you click on Create New File.

Note that the new file will have a suggested name with _Reshaped added to the previous file and will contain in our case 4 time as many records (rows) but now with only 2 variables instead of the previous 4.

Note that the LAI variable is missing for many time points but continously represented by the Treatment variable for each record.

To make a common calibration effort representing different inputs of both water and nutrients we can select different representation for in our case both meteorological input file and External N Inputs.

Those inputs files have in the case about 5 Repetitions to represent 5 different possible inputs that are indentified by parameters SeqNo Met File and Filter Var Value in a possible multirun.

Those parameters are selected for MultiRuns after appropriate switches for Dynamic indexed input file

and UseFilterParameters are put to On (Note this is a new switch for version 6.2.6)

Note that the switch UseFilterParameters will overwrite the values of the filter variable in the Validation list shown in the validation tab.

by using the values from the parameter table.

The values for the parameters are defined as a Table in the first Dimension.

Note that conventional parameters for random values sampling are defined for Dim 2.

This will make a pattern where the 1st dimension varies with higher frequence compared to the 2nd dimension.

A typical performance will be with variability for each of the treatments and sequence of shift in the parameters.

Leave a Reply

Your email address will not be published. Required fields are marked *