Bill Sacks
12-26-14

Here are the commands used to generate the files with date stamp 141219. This
was done from tag

https://svn-ccsm-models.cgd.ucar.edu/clm2/branch_tags/dynlu_crops3_tools_tags/dynlu_crops3_tools_n02_clm4_5_1_r101

----------

STANDARD_RES=360x720cru,48x96,0.9x1.25,1.9x2.5,10x15,ne30np4
mksurfdata.pl -glc_nec 10 -y 1850,2000 -res $STANDARD_RES
mksurfdata.pl -glc_nec 10 -y 1850-2100 -rcp 2.6,4.5,6,8.5 -res $STANDARD_RES

mksurfdata.pl -glc_nec 10 -y 2000 -res 4x5

mksurfdata.pl -y 2000 -res 5x5_amazon,1x1_brazil -hirespft
# Note that the 1x1_brazil 1850 & transient datasets do NOT use -hirespft, in contrast to the year-2000 file (for consistency with the previous version of these datasets)
mksurfdata.pl -y 1850-2100 -rcp 8.5 -res 1x1_brazil

YYMMDD=141219
mv surfdata_1x1_brazil_rcp8.5_simyr1850_c${YYMMDD}.nc surfdata_1x1_brazil_simyr1850_c${YYMMDD}.nc
mv surfdata_1x1_brazil_rcp8.5_simyr1850_c${YYMMDD}.log surfdata_1x1_brazil_simyr1850_c${YYMMDD}.log

mksurfdata.pl -y 1850,2000 -res 1x1_tropicAtl
mksurfdata.pl -y 1850-2000 -res 1x1_tropicAtl

# ---------------------------------------------------------------------
# NOTE: Before making the following crop datasets, move the already-created datasets to somewhere else; otherwise, some of them will be overwritten by the crop versions.
# ---------------------------------------------------------------------

# CROP; note that files need to be renamed to have _mp24 in their names
mksurfdata.pl -crop -hirespft -glc_nec 10 -y 2000 -r 1.9x2.5,10x15
mksurfdata.pl -crop -hirespft -y 2000 -r 1x1_numaIA
mksurfdata.pl -crop -y 2000 -r 1x1_smallvilleIA -pft_idx 17,18,19,20,21,22,23,24 -pft_frc 30,5,24,6,13,7,7,8

YYMMDD=141219
mv surfdata_1.9x2.5_simyr2000_c${YYMMDD}.nc surfdata_1.9x2.5_mp24_simyr2000_c${YYMMDD}.nc 
mv surfdata_10x15_simyr2000_c${YYMMDD}.nc surfdata_10x15_mp24_simyr2000_c${YYMMDD}.nc 
mv surfdata_1x1_numaIA_simyr2000_c${YYMMDD}.nc surfdata_1x1_numaIA_mp24_simyr2000_c${YYMMDD}.nc 
mv surfdata_1.9x2.5_simyr2000_c${YYMMDD}.log surfdata_1.9x2.5_mp24_simyr2000_c${YYMMDD}.log 
mv surfdata_10x15_simyr2000_c${YYMMDD}.log surfdata_10x15_mp24_simyr2000_c${YYMMDD}.log 
mv surfdata_1x1_numaIA_simyr2000_c${YYMMDD}.log surfdata_1x1_numaIA_mp24_simyr2000_c${YYMMDD}.log 
mv surfdata_1x1_smallvilleIA_simyr2000_c${YYMMDD}.nc surfdata_1x1_smallvilleIA_mp24_simyr2000_c${YYMMDD}.nc 
mv surfdata_1x1_smallvilleIA_simyr2000_c${YYMMDD}.log surfdata_1x1_smallvilleIA_mp24_simyr2000_c${YYMMDD}.log 


# URBAN; note that some fields are overwritten via ncks commands in the mksurfdata.pl script
mksurfdata.pl -y 2000 -r 1x1_camdenNJ,1x1_vancouverCAN,1x1_mexicocityMEX


# URBAN; for this one, an error check needs to be commented out in the source code, as shown below:
mksurfdata.pl -y 2000 -r 1x1_urbanc_alpha

Index: mkurbanparMod.F90
===================================================================
--- mkurbanparMod.F90	(revision 47505)
+++ mkurbanparMod.F90	(working copy)
@@ -754,7 +754,7 @@
                   write(6,*) 'n: ', n
                   write(6,*) 'region: ', region_o(n)
                   write(6,*) 'urbn_classes_gcell_o(n,k): ', urbn_classes_gcell_o(n,k)
-                  call abort()
+!                  call abort()
                end if
             end do
          end if

-----------

The following documents the steps used to create the test dataset,
landuse.timeseries_1x1_tropicAtl_TEST_simyr1939-1943_c141219.nc. As the name
implies, this dataset is based off of
landuse.timeseries_1x1_tropicAtl_hist_simyr1850-2005_c141219.nc, but contains just
a subset of years. The purpose is to test some edge cases of reading the
transient PFT dataset.

I started by taking a subset of the years from the original dataset. However, I
then modified the PCT_NAT_PFT array so that each year has a unique breakdown
into PCT_NAT_PFT - to catch any problems with reading in the wrong year of data.

Here were the specific steps:

ncks -d time,89,93 landuse.timeseries_1x1_tropicAtl_hist_simyr1850-2005_c141219.nc landuse.timeseries_1x1_tropicAtl_TEST_simyr1939-1943_c141219.nc

From R:
> dat.subset <- read.ncdf('landuse.timeseries_1x1_tropicAtl_TEST_simyr1939-1943_c141219.nc')
> for (time in 2:5) {
+ dat.subset$PCT_NAT_PFT$data[5,time] <- dat.subset$PCT_NAT_PFT$data[5,time] - time
+ dat.subset$PCT_NAT_PFT$data[16,time] <- dat.subset$PCT_NAT_PFT$data[16,time] + time
+ }
> write.ncdf('new.pct_nat_pft.nc', dat.subset)

ncks -A -v PCT_NAT_PFT new.pct_nat_pft.nc landuse.timeseries_1x1_tropicAtl_TEST_simyr1939-1943_c141219.nc