###########################################################################
###
### SEIB-PF02: 
### SEIB-DGVM (ver2.71) with SIR Particle Filter, ver.02
### 20191025, RIKEN R-CCS
###
############################################################################


###Citation:

TBA


###Copy right:

Check the license at:
http://data-assimilation.riken.jp/~opendata/SEIB-PF02/LICENSE


###Developers:

- Data assimilation code for the SEIB-DGVM (SEIB-PF02)
  Hazuki Arakida (RIKEN R-CCS):
   Main developer 
  Shin-ichiro Shima (University of Hyogo): 
   Modified the random number generator in the SEIB-DGVM
   Helped to develop the resampling code for SIR

- Observation data (MODIS_LAI)
  Shunji Kotsuki (RIKEN R-CCS) 


###SEIB-DGVM (ver2.71)
Original SEIB-DGVM code is available at http://seib-dgvm.com/
We thank Dr. Hisashi Sato for providing the code to generate 
the climate forcing data (this code is not included in SEIB-PF02).


###SEIB-PF02#################################################################

- Job scripts
  - spin.sh #Spin-up
  - da_nat.sh #DA and NODA (nat)

- Main code for SEIB-PF
  - seib271_pf200/scr/

    1) Place seib-pf02.diff and the following files of the original SEIB-DGVM (ver. 2.71) in the same directory.
       etc.f90
       initialize.f90
       main.f90
       metabolic.f90
       modules.f90
       output.f90
       physics.f90
       population_regu.f90
       spatial_calc.f90
       start_point.f90 #MPI and DA settings. Compile this file with seib271_pf200/MPI_compile.sh 

    2) Then run the following code to generate SEIB-PF02 code:
       patch -u < seib-pf02.diff

  - seib271_pf200/R_para.R
    #Initial parameter settings for Pmax and Dormancy. Foliage is not used.
    *** Generate "Parameterin.txt" before DA.***

  - seib271_pf200/R_resample.sh
    #Called from start_point.f90, this code runs seib271_pf200/R_pf.R

  - seib271_pf200/R_pf.R
    #Resampling: Change "inf"(Noise to the parameters after resampling) if needed.  

- Inputs (need to be prepared)
  seib271_pf200/data/co2_1750-2100_a1b.txt (from original SEIB-DGVM2.71)
  seib271_pf200/data/land_prop.txt (from original SEIB-DGVM2.71) 
  seib271_pf200/scr/studysites.csv (This table was used for regional experiment at Siberia)
  seib271_pf200/lai/sd_lai000001.txt (e.g. site no = 1)
  seib271_pr200/lai/obs_lai000001.txt (e.g. site no = 1)
  -> r_obs_prep.R is run from spin.sh
  seib271_pf200/forcing/cli000001.txt (e.g. site no = 1. N= 365 days, nine column data (same as the generated data on SEIB-DGVM web-site, .csv, without header)
  -> r_forcing_prep.R is run from spin.sh
  seib271_pf200/omitsites.csv (site number for skipping simulation, if needed

- Outputs (Spin-up)
  at out/SITENUMBER
  
- Outputs (DA)
  at dat
  output.bin: variables
  parameterlog.bin: model parameters

- Outputs (NODA)
  at nat
  output.bin: variables
 
- Make figs from the outputs
  ADD_ALL.sh -> fig...files are used. 
  We separated output to dat1/dat2 and nat1/nat2,
  therefore, two job scripts are prepared.

===cf. These files are generated when job script is run===
  seib271_pf200/scr/ptl.txt (particle numbers)
  seib271_pf200/scr/da.txt (flag for data assimilation, spin-up:1, noDA:2, DA:3)
  seib271_pf200/scr/daloop.txt (91 * (years for DA))
  seib271_pf200/scr/site_start.txt (start site)
  seib271_pf200/scr/site_end.txt (end site)
  seib271_pf200/SITENUMBER/forcing_flag.txt (forcing climate data with missing data: 0, without missing data: 1)
==========================================================


###MODIS observations codes#########################################

- MODIS_LAI/step1: make the centers of the study sites
  studysite_maker.R #R script. 

- MODIS_LAI/step2/003_medianmake-climate: LAI generator
  1)Download MODIS DATA and change the format.
  2)Copy http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/SFMT/VERSIONS/ARCHIVES/SFMT.f90
  to MODIS_LAI/step2/003_medianmake-climate/source
  3)Place the station file in "station". 
  4)Run 001_medianmake.sh. 
  5)Outputs are found in "out".
    Use "modis-median_station..."

- LAI/step3: make LAI for DA / validation of data number after QC
  lai_for_da_20161013.txt #R script. 

- LAI/step4: make figs for LAI
  lai_for_da.txt #R script.