| | 8 | == Planned work == |
| | 9 | 1. XIOS2 `InterpWeight' |
| | 10 | * Release of XIOS2 |
| | 11 | * Grouping of current interpolation (read, `aggregate_p' , apply weights and bring back fractions) in ORCHIDEE into a new module (`interpweight') |
| | 12 | 1. Commit and test both steps on the trunk [done October 2016] |
| | 13 | 1. Open a new branch for the interpolation with XIOS2 |
| | 14 | 1. Work on the interpolation (doing nothing for the calendar) with a new `interpweight_XIOS' but with XIOS2 |
| | 15 | 1. Validation off-line |
| | 16 | 1. Implement it on the current version on the trunk |
| | 17 | 1. Coupling to DYNAMICO |
| | 18 | |
| | 19 | === interpweight === |
| | 20 | This module is used in the re-organization done in the section of the code responsible of the interpolation of the morphology datat (PFT, albedo, LAI, ...) from different files done by L. Fita in the mark of the HEAT project. Previous version of the interpolation of these files, proceed by three basic steps common for all the files: |
| | 21 | 1. Modification of values from file |
| | 22 | 1. Creation of land/sea mask using values from file |
| | 23 | 1. Interpolate values |
| | 24 | This module serves to generalize all these steps and construct a common interface independently of the file. Each step is generalize and the different processes almost specific for each file in the previous form are introduced as different methodologies. Thus now it provides: |
| | 25 | 1. Modification of initial values: Removing that wrong values from the file |
| | 26 | * `interpweight_modifying_input[1/2/3/4]D': subroutines to modify 1D, 2D, 3D or 4D input data from file, using different methods: none |
| | 27 | 1. Masking input: Compute ’on fly’ the values of the land/sea mask taking the values in the file |
| | 28 | * `interpweight_masking_input[1/2/3/4]D': subroutines to compute the mask using data from input file using different methods: nomask, mbelow, mabove, msumrange, var |
| | 29 | 1. Area weights: Get coincident areas from input file to the target projection using `aggregate_p' |
| | 30 | * No changes on it |
| | 31 | 1. Interpolate: Use obtained areas and interpolate values at each grid point at the same format as it will be provided by XIOS |
| | 32 | * `interpweight_provide_fractions[1/2/3/4]D': Perform interpolation for fraction/category data |
| | 33 | * `interpweight_provide_interpolation[2/4]D': Perform interpolation for continuous data |
| | 34 | * `variableusetypes': Variable to provide the values along the additional dimension to perform the interpolation. (e.g.: in case the 13 pfts are non consecutive: 1,3,6,18,23,34,35,39,48,...) |
| | 35 | 1. A new variable is added which explains the ‘availability’ of data to perform the interpolation (0, 1). When it is negative it means that there was no data for that grid point |
| | 36 | {{{ |
| | 37 | availability = SUM(area_source)/Area_target_grid_point |
| | 38 | }}} |
| | 39 | A document with more detail and the working notes is available from here [attachment:interpweight_working_notes.pdf interpweight working notes] |
| | 40 | A short presentation with the general aims of the module is also available here [attachment:interpweight.pdf interpweight: Re-organization of ORCHIDEE’s interpolation] |
| | 41 | |
| | 42 | This module will disappear once interpolation will be done throughout XIOS |
| | 43 | |
