Auxiliary fortran packages

tmtools

The fortran programs use the tmtools package for interpolation and file I/O. This contains the following subroutines:

• writetmdata(kout,iyear,nx,ny,nz,ntr,ntim,fname, xlabelx,zdata) writes the output file. The input arguments are:
  • kout: file unit
  • iyear: base year for data (yyyy)
  • nx,ny,nz: spatial dimension in longitude, latitude, vertical (in general im,jm,lm)
  • ntr: number of species (ntr)
  • ntim: number of timesteps (1=annual,4=seasonal,12=monthly)
  • fname: filenname
  • xlabelx: header
  • zdata: data array with dimensions (nx,ny,nz,ntr,ntim)
• subroutine fractions(fx,fy,dxyp1,dxyp2,mx,my): calculate fractions fx, fy that gridcells nx,ny overlap with im,jm.
• subroutine fractions2(fx,fy,dxyp1,dxyp2,mx,my): calculate fractions fx, fy that gridcells nx,ny overlap with im,jm. For Olsson data.
• subroutine fill(z1,mx,my,fx,fy,dxyp1,dxyp2,z2): fill target grid with overlapping fractions.

• subroutine geomtry(nx2,ny2,dxyp,ytop): calculates the surface area of the grid cells of the target model, TM.
• subroutine linear(x,y,n,x0,y0,m): Linear interpolation from one vector to another. The arguments are:
  • x: the array of x-values, locations where the function is given (input)
  • y: the array of y-values, values of the function (input)
  • n: the no. of x and y values, dimension of x and y (input)
  • x0: the x-locations where values are to be calculated by interpolation (input)
  • y0: the interpolated values (output)
  • m: no of values to interpolate, dimension of x0 and y0 (input)
• subroutine muherm(x,y,n,x0,y0,m): Hermite cubic interpolation from one vector to another. The arguments are:
  • x: the array of x-values, locations where the function is given (input)
  • y: the array of y-values, values of the function (input)
  • n: the no. of x and y values, dimension of x and y (input)
  • x0: the x-locations where values are to be calculated by interpolation (input)
  • y0: the interpolated values (output)
  • m: no of values to interpolate, dimension of x0 and y0 (input)

hdf

The fortran programs use the hdf f90 package for file I/O in HDF format. This contains the following subroutines:

• IO_WRITE4D_32(sd_id,im,labim,jm,labjm,lm,lablm,nt,labnt,data,name)
• IO_WRITE3D_32(sd_id,im,labim,jm,labjm,lm,lablm,data,name)
• IO_WRITE3D_32d(sd_id,im,labim,jm,labjm,lm,lablm,data,name,idate)
• IO_WRITE2D_32(sd_id,im,labim,jm,labjm,data,name)
• IO_WRITE2D_I16(sd_id,im,jm,data,name)
• IO_WRITE2D_32d(sd_id,im,labim,jm,labjm,data,name,idate)
• IO_READ4D_32(sd_id,im,jm,lm,nt,data,name,ifail)
• IO_READ3D_32(sd_id,im,jm,lm,data,name,ifail)
• IO_READ2D_32(sd_id,im,jm,data,name,ifail)
• IO_READ2D_32f(sd_id,im,jm,data,number,name,ifail)
• IO_READ2D_I16(sd_id,im,jm,data,name)

There is also an include file hdf.i which is a Fortran header file for HDF routines