Example 10 — Masked Zonal Statistics (3D)
Demonstrates combining setFillValue (masking) with zonalStats on 3D temporal data:
- Computes cattle density statistics for all states (before masking)
- Applies a mask to keep only Minas Gerais (UF=18)
- Recomputes statistics showing temporal evolution (1990–2012) for the masked region
Source Code
program main
use fpl
implicit none
type(nc2d_byte_lld) :: mask, zones
type(nc3d_double_lld_ti) :: cattle
integer(kind=intgr), parameter :: nzones = 28
integer(kind=intgr), parameter :: uf_minas_gerais = 18
integer(kind=intgr), allocatable, dimension(:,:) :: zcount
real(kind=double), allocatable, dimension(:,:) :: zmean, zmin, zmax, zsum, zvar
character(len=20) :: uf_name(28)
integer(kind=intgr) :: z, k
character(len=100) :: ofile
uf_name = ""
uf_name(18) = "Minas Gerais"
! ... (other states)
! Read mask and zone grids
mask%varname = "UF"
mask%lonname = "lon"
mask%latname = "lat"
mask%varunits = "class"
call readgrid("database/brazil_UF.nc", mask)
zones%varname = "UF"
zones%lonname = "lon"
zones%latname = "lat"
zones%varunits = "class"
call readgrid("database/brazil_UF.nc", zones)
! Read cattle density data (1990-2012)
cattle%varname = "Cattle"
cattle%lonname = "lon"
cattle%latname = "lat"
cattle%timename = "time"
call readgrid("database/CATTLE19902012.nc", cattle)
! Allocate output arrays
allocate(zcount(nzones, cattle%ntimes))
allocate(zmean(nzones, cattle%ntimes))
allocate(zmin(nzones, cattle%ntimes))
allocate(zmax(nzones, cattle%ntimes))
allocate(zsum(nzones, cattle%ntimes))
allocate(zvar(nzones, cattle%ntimes))
! Step 1: Unmasked statistics
call zonalStats(zones, cattle, nzones, zcount, zmean, zmin, zmax, zsum, zvar)
! ... print results for all states ...
! Step 2: Apply mask — keep only Minas Gerais
call setFillValue(mask, cattle, uf_minas_gerais)
! Step 3: Masked statistics — temporal evolution
call zonalStats(zones, cattle, nzones, zcount, zmean, zmin, zmax, zsum, zvar)
! ... only zone 18 (Minas Gerais) has data now ...
! Save to CSV
ofile = "database/zonalstats_cattle_masked_mg.csv"
open(unit=10, file=trim(ofile), status='replace', action='write')
write(10,'(a)') "state,year,pixels,mean,min,max,sum,var"
z = uf_minas_gerais
do k = 1, cattle%ntimes
write(10,'(a,",",i0,",",i0,",",es14.6,",",es14.6,",",es14.6,",",es14.6,",",es14.6)') &
trim(uf_name(z)), int(cattle%times(k)), zcount(z,k), &
zmean(z,k), zmin(z,k), zmax(z,k), zsum(z,k), zvar(z,k)
end do
close(10)
deallocate(zcount, zmean, zmin, zmax, zsum, zvar)
call dealloc(mask)
call dealloc(zones)
call dealloc(cattle)
end program main
Compile & Run
gfortran -fopenmp -o ex10.out ex10_zonalstats_masked_3d.f90 -I/usr/lib64/gfortran/modules/ -lFPL
./ex10.out
Figure
Output
======================================================================
BEFORE masking — Cattle Mean Density (head/ha): 1990 vs 2012
======================================================================
State Mean 1990 Mean 2012
----------------------------------------------------------------------
Tocantins 0.4558 1.2741
Minas Gerais 0.7255 1.6883
Goias 0.8503 1.4730
Mato Grosso do Sul 1.0375 1.7206
======================================================================
Applying mask: keeping only UF=18 (Minas Gerais)
======================================================================
AFTER masking — Cattle in Minas Gerais: Temporal Evolution (1990-2012)
======================================================================
Year Pixels Mean Min Max Total
----------------------------------------------------------------------
1990 722848 0.7255 0.0000 1.4670 524449.24
2000 722848 0.8844 0.0000 1.9470 639275.11
2012 722848 1.6883 0.0000 6.5430 1220352.83
======================================================================
Results saved to: database/zonalstats_cattle_masked_mg.csv