This is a major revision compared to previous 6.2.6 version from November Further work on to handle Multiple Treatments – With Predefined Criteria for many ensembles and also a number of bugs have been deleted.
One major bug related to the model behaviour has been deleted related to phenological development during the year and multiple harvests. The two major harvest within a year as identified as Final Type Day with 0 or 1 was creating an error. The 0 should be linked with the Harvest of Plots Prior final Harvest whereas 1 should be linked with Harvest of Plants table. Previous version missed the Final Harvest of Plants table.
Another bug was related to the development of GSI after harvest. The GSI was not considering any delay between GSI 1 to 2 (i.e. The step from emergence to start of grain filling) instead grain filling started directly after harvest). To avoid additional development or new flowering of a plant within a current year the model now is restricted to one single period of grain development per year and plant.
All other major developments have been made to develop the procedures with many ensembles representing different treatments or sites. The previous new feature allowing fewer variables as columns in typical PG-file for validation to new lines has been remade to instead use the previous developed feature with repetitions of the entire structure of a PG-file. The major advantage is that variables in a reshaped file has a double identification by both Filter Variables and by the complete Description of each variable by: name, id, unit, pos, country, stations, lat, long and altitude.
The new version is still experimental – all comments are welcome. A more detailed description will follow. Now I wish both new and old users a Happy New Year. The new calibration methods will be useful to further develop the machine learning methods of the model and facilitate application with also automatic continous updates. Thanks a lot for collaboration during this year.