Changes between Version 12 and Version 13 of Branches/Driver_Improvements


Ignore:
Timestamp:
2016-05-26T16:50:50+02:00 (8 years ago)
Author:
jpolcher
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Branches/Driver_Improvements

    v12 v13  
    2727=== Jan's understanding of the working of dim2driver (Jan 26/05/2016) === 
    2828 
     29For all model time steps (itauin) between the forcing time steps n-1 (all variables in '*_nm1') and n (variables in '*_n') a linear interpolation will be done with the forcing values to obtain the correct value at itauin. There will be 'split' time steps of ORCHIDEE between n-1 and n. Once these 'split' number of time steps have been execute variables *_n will be copied into *_nm1 and the next forcing time step of the file written into *_n. 
     30 
     31As explained above for all time steps between itauin(n-1) and itauin(n) the interpolation will be correct for instantaneous variables as there is a perfect match between the time step space and the physical time. 
     32 
     33The result is less predictable for fluxes which are averaged over an interval from which we only know that it contains the physical timestep of n. readdim2.f90 has not analysed the time axis of the fluxes and thus does not know where in this interval n is exactly situated. Only when n is at the centre of the time step will the interpolation be correct. 
     34 
    2935The graphic below explains probably better the above description. 
    3036 
    3137[[Image(Interpolation_dim2driver.jpeg, width=50%)]] 
    3238 
    33 The following changes in nomenclature are used in the figure : t has become Tf and thus the forcing considered is between Tf and Tf+1 for the interpolation toward t'. 
    34  
    35 The figure is especially meant to illustrate that the interpolation for fluxes assumes that the interval of validity of the fluxes is assumed to be between Tf-1/2 and Tf+1/2 or, t-dt/2 and t+dt/2. 
     39The figure is especially meant to illustrate that the interpolation for fluxes assumes that the interval of validity of the fluxes is assumed to be between n-3/2 and n-1/2 for variables *_nm1 and n-1/2 and n+1/2 for variables *_n. 
    3640 
    3741== Add an explicit time information within the forcing files ==