First under the unix terminal:
-> gunzip das.tar.zip
-> tar -xvf das.tar

you will get a folder 'DAS'

Directorey information

truth:: the truth from nature run
obs  :: the observations to be assimilated, which is based on the nature run. It is the nature run plus random perturbation 
      with the standard deviation defined in the file 'common_speedy.f90'.


letkf ::  the code for letkf 

         script to run the code in the directory '/DAS/letkf'::
	 letkf.sh

         maile code:
	 letkf.f90 

	 subroutine module includes:
	 common_letkf.f90  (followed the equation of letkf)   
	 letkf_tools.f90   (the interface between model and data assimilation)
         
	 utilities includes::
         common_speedy.f90
	 common_mtx.f90
	 netlib.f
	 mt19937ar.f90

	 Calculate the mean state
	 ensmean.sh

tdvar :: the code for 3D-Var (by Takemasa Miyoshi)

         script to run the code '/DAS/tdvar'::
	 tdvar.sh

         main code:: 
	 tdvar.f90

	 subroutine modeul includs::
	 tdvar_tools.f90, which uses the minimizer in the file 'minimizelib.f90'.

	 
The results will be at '/DAS/DAS_result'
The results for 3D_Var is in the folder '/DAS/DAS_result/3dvar
'anal' is the folder of analysis result
'analf' is the folder containing filtered value
'gues' is the forecast

The results for LETKF is in the folder '/DAS/DAS_result/letkf
'anal' is the folder of analysis result, each folder inside represents the results from each ensemble member.
'analf' is the folder containing filtered value
'gues' is the forecast


RMS error can be calculated based on the data in the 'truth' folder
test_3dvar.gs is an example to plot the RMS error for 3dvar