retrieval.c

Go to the documentation of this file.

/*
  This file is part of the AIRS Code Collection.
  
  the AIRS Code Collections is free software: you can redistribute it
  and/or modify it under the terms of the GNU General Public License
  as published by the Free Software Foundation, either version 3 of
  the License, or (at your option) any later version.
  
  The AIRS Code Collection is distributed in the hope that it will be
  useful, but WITHOUT ANY WARRANTY; without even the implied warranty
  of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  General Public License for more details.
  
  You should have received a copy of the GNU General Public License
  along with the AIRS Code Collection. If not, see
  <http://www.gnu.org/licenses/>.
  
  Copyright (C) 2019-2025 Forschungszentrum Juelich GmbH
*/
 
#include <mpi.h>
#include <omp.h>
#include <netcdf.h>
#include "jurassic.h"
 
/* ------------------------------------------------------------
   Macros...
   ------------------------------------------------------------ */
 
#define NC(cmd) {                                    \
  int nc_result=(cmd);                               \
  if(nc_result!=NC_NOERR)                            \
    ERRMSG("%s", nc_strerror(nc_result));            \
}
 
/* ------------------------------------------------------------
   Dimensions...
   ------------------------------------------------------------ */
 
#define L1_NCHAN 34
 
#define L1_NTRACK 135
 
#define L1_NXTRACK 90
 
#define L2_NLAY 27
 
#define L2_NTRACK 45
 
#define L2_NXTRACK 30
 
/* ------------------------------------------------------------
   Structs...
   ------------------------------------------------------------ */
 
typedef struct {
 
  int ncid;
 
  int np;
 
  double l1_time[L1_NTRACK][L1_NXTRACK];
 
  double l1_lon[L1_NTRACK][L1_NXTRACK];
 
  double l1_lat[L1_NTRACK][L1_NXTRACK];
 
  double l1_sat_z[L1_NTRACK];
 
  double l1_sat_lon[L1_NTRACK];
 
  double l1_sat_lat[L1_NTRACK];
 
  double l1_nu[L1_NCHAN];
 
  float l1_rad[L1_NTRACK][L1_NXTRACK][L1_NCHAN];
 
  double l2_z[L2_NTRACK][L2_NXTRACK][L2_NLAY];
 
  double l2_p[L2_NLAY];
 
  double l2_t[L2_NTRACK][L2_NXTRACK][L2_NLAY];
 
  float ret_z[NP];
 
  float ret_p[L1_NTRACK * L1_NXTRACK];
 
  float ret_t[L1_NTRACK * L1_NXTRACK * NP];
 
} ncd_t;
 
/* ------------------------------------------------------------
   Functions...
   ------------------------------------------------------------ */
 
void add_var(
  int ncid,
  const char *varname,
  const char *unit,
  const char *longname,
  int type,
  int dimid[],
  int *varid,
  int ndims);
 
void buffer_nc(
  atm_t * atm,
  double chisq,
  ncd_t * ncd,
  int track,
  int xtrack,
  int np0,
  int np1);
 
void fill_gaps(
  double x[L2_NTRACK][L2_NXTRACK][L2_NLAY],
  double cx,
  double cy);
 
void init_l2(
  ncd_t * ncd,
  int track,
  int xtrack,
  ctl_t * ctl,
  atm_t * atm);
 
void read_nc(
  char *filename,
  ncd_t * ncd);
 
void write_nc(
  char *filename,
  ncd_t * ncd);
 
/* ------------------------------------------------------------
   Main...
   ------------------------------------------------------------ */
 
int main(
  int argc,
  char *argv[]) {
 
  static ctl_t ctl;
  static atm_t atm_apr, atm_clim, atm_i;
  static obs_t obs_i, obs_meas;
  static ncd_t ncd;
  static ret_t ret;
 
  FILE *in;
 
  char filename[LEN];
 
  double chisq, chisq_min, chisq_max, chisq_mean, z[NP];
 
  int channel[ND], m, ntask = -1, rank, size;
 
  /* ------------------------------------------------------------
     Init...
     ------------------------------------------------------------ */
 
  /* MPI... */
  MPI_Init(&argc, &argv);
  MPI_Comm_rank(MPI_COMM_WORLD, &rank);
  MPI_Comm_size(MPI_COMM_WORLD, &size);
 
  /* Measure CPU time... */
  TIMER("total", 1);
 
  /* Check arguments... */
  if (argc < 3)
    ERRMSG("Give parameters: <ctl> <filelist>");
 
  /* Read control parameters... */
  read_ctl(argc, argv, &ctl);
  read_ret(argc, argv, &ctl, &ret);
 
  /* Initialize look-up tables... */
  tbl_t *tbl = read_tbl(&ctl);
 
  /* Read retrieval grid... */
  const int nz = (int) scan_ctl(argc, argv, "NZ", -1, "", NULL);
  if (nz > NP)
    ERRMSG("Too many altitudes!");
  for (int iz = 0; iz < nz; iz++)
    z[iz] = scan_ctl(argc, argv, "Z", iz, "", NULL);
 
  /* Read track range... */
  const int track0 = (int) scan_ctl(argc, argv, "TRACK_MIN", -1, "0", NULL);
  const int track1 = (int) scan_ctl(argc, argv, "TRACK_MAX", -1, "134", NULL);
 
  /* Read xtrack range... */
  const int xtrack0 = (int) scan_ctl(argc, argv, "XTRACK_MIN", -1, "0", NULL);
  const int xtrack1 =
    (int) scan_ctl(argc, argv, "XTRACK_MAX", -1, "89", NULL);
 
  /* Read height range... */
  int np0 = (int) scan_ctl(argc, argv, "NP_MIN", -1, "0", NULL);
  int np1 = (int) scan_ctl(argc, argv, "NP_MAX", -1, "100", NULL);
  np1 = GSL_MIN(np1, nz - 1);
 
  /* Background smoothing... */
  const double sx = scan_ctl(argc, argv, "SX", -1, "8", NULL);
  const double sy = scan_ctl(argc, argv, "SY", -1, "2", NULL);
 
  /* SZA threshold... */
  const double sza_thresh = scan_ctl(argc, argv, "SZA", -1, "96", NULL);
 
  /* ------------------------------------------------------------
     Distribute granules...
     ------------------------------------------------------------ */
 
  /* Open filelist... */
  printf("Read filelist: %s\n", argv[2]);
  if (!(in = fopen(argv[2], "r")))
    ERRMSG("Cannot open filelist!");
 
  /* Loop over netCDF files... */
  while (fscanf(in, "%s", filename) != EOF) {
 
    /* Distribute files with MPI... */
    if ((++ntask) % size != rank)
      continue;
 
    /* Write info... */
    printf("Retrieve file %s on rank %d of %d (with %d threads)...\n",
           filename, rank + 1, size, omp_get_max_threads());
 
    /* ------------------------------------------------------------
       Initialize retrieval...
       ------------------------------------------------------------ */
 
    /* Read netCDF file... */
    read_nc(filename, &ncd);
 
    /* Identify radiance channels... */
    for (int id = 0; id < ctl.nd; id++) {
      channel[id] = -999;
      for (int i = 0; i < L1_NCHAN; i++)
        if (fabs(ctl.nu[id] - ncd.l1_nu[i]) < 0.1)
          channel[id] = i;
      if (channel[id] < 0)
        ERRMSG("Cannot identify radiance channel!");
    }
 
    /* Fill data gaps... */
    fill_gaps(ncd.l2_t, sx, sy);
    fill_gaps(ncd.l2_z, sx, sy);
 
    /* Set climatological data for center of granule... */
    atm_clim.np = nz;
    for (int iz = 0; iz < nz; iz++)
      atm_clim.z[iz] = z[iz];
    climatology(&ctl, &atm_clim);
 
    /* ------------------------------------------------------------
       Retrieval...
       ------------------------------------------------------------ */
 
    /* Get chi^2 statistics... */
    chisq_min = 1e100;
    chisq_max = -1e100;
    chisq_mean = 0;
    m = 0;
 
    /* Loop over swaths... */
    for (int track = track0; track <= track1; track++) {
 
      /* Measure CPU time... */
      TIMER("retrieval", 1);
 
      /* Loop over scan... */
      for (int xtrack = xtrack0; xtrack <= xtrack1; xtrack++) {
 
        /* Store observation data... */
        obs_meas.nr = 1;
        obs_meas.time[0] = ncd.l1_time[track][xtrack];
        obs_meas.obsz[0] = ncd.l1_sat_z[track];
        obs_meas.obslon[0] = ncd.l1_sat_lon[track];
        obs_meas.obslat[0] = ncd.l1_sat_lat[track];
        obs_meas.vplon[0] = ncd.l1_lon[track][xtrack];
        obs_meas.vplat[0] = ncd.l1_lat[track][xtrack];
        for (int id = 0; id < ctl.nd; id++)
          obs_meas.rad[id][0] = ncd.l1_rad[track][xtrack][channel[id]];
 
        /* Flag out 4 micron channels for daytime measurements... */
        if (RAD2DEG(acos(cos_sza(obs_meas.time[0], obs_meas.obslon[0],
                                 obs_meas.obslat[0]))) < sza_thresh)
          for (int id = 0; id < ctl.nd; id++)
            if (ctl.nu[id] >= 2000)
              obs_meas.rad[id][0] = GSL_NAN;
 
        /* Prepare atmospheric data... */
        copy_atm(&ctl, &atm_apr, &atm_clim, 0);
        for (int ip = 0; ip < atm_apr.np; ip++) {
          atm_apr.time[ip] = obs_meas.time[0];
          atm_apr.lon[ip] = obs_meas.vplon[0];
          atm_apr.lat[ip] = obs_meas.vplat[0];
        }
 
        /* Merge Level-2 data... */
        init_l2(&ncd, track, xtrack, &ctl, &atm_apr);
 
        /* Retrieval... */
        optimal_estimation(&ret, &ctl, tbl, &obs_meas, &obs_i,
                           &atm_apr, &atm_i, &chisq);
 
        /* Get chi^2 statistics... */
        if (gsl_finite(chisq)) {
          chisq_min = GSL_MIN(chisq_min, chisq);
          chisq_max = GSL_MAX(chisq_max, chisq);
          chisq_mean += chisq;
          m++;
        }
 
        /* Buffer results... */
        buffer_nc(&atm_i, chisq, &ncd, track, xtrack, np0, np1);
      }
 
      /* Measure CPU time... */
      TIMER("retrieval", 3);
    }
 
    /* ------------------------------------------------------------
       Finalize...
       ------------------------------------------------------------ */
 
    /* Write netCDF file... */
    write_nc(filename, &ncd);
 
    /* Write info... */
    printf("chi^2: min= %g / mean= %g / max= %g / m= %d\n",
           chisq_min, chisq_mean / m, chisq_max, m);
    printf("Retrieval finished on rank %d of %d!\n", rank, size);
  }
 
  /* Close file list... */
  fclose(in);
 
  /* Measure CPU time... */
  TIMER("total", 3);
 
  /* Report memory usage... */
  printf("MEMORY_ATM = %g MByte\n", 4. * sizeof(atm_t) / 1024. / 1024.);
  printf("MEMORY_CTL = %g MByte\n", 1. * sizeof(ctl_t) / 1024. / 1024.);
  printf("MEMORY_NCD = %g MByte\n", 1. * sizeof(ncd_t) / 1024. / 1024.);
  printf("MEMORY_OBS = %g MByte\n", 3. * sizeof(atm_t) / 1024. / 1024.);
  printf("MEMORY_RET = %g MByte\n", 1. * sizeof(ret_t) / 1024. / 1024.);
  printf("MEMORY_TBL = %g MByte\n", 1. * sizeof(tbl_t) / 1024. / 1024.);
 
  /* Report problem size... */
  printf("SIZE_TASKS = %d\n", size);
  printf("SIZE_THREADS = %d\n", omp_get_max_threads());
 
  /* MPI... */
  MPI_Finalize();
 
  return EXIT_SUCCESS;
}
 
/*****************************************************************************/
 
void add_var(
  int ncid,
  const char *varname,
  const char *unit,
  const char *longname,
  int type,
  int dimid[],
  int *varid,
  int ndims) {
 
  /* Check if variable exists... */
  if (nc_inq_varid(ncid, varname, varid) != NC_NOERR) {
 
    /* Define variable... */
    NC(nc_def_var(ncid, varname, type, ndims, dimid, varid));
 
    /* Set long name... */
    NC(nc_put_att_text
       (ncid, *varid, "long_name", strlen(longname), longname));
 
    /* Set units... */
    NC(nc_put_att_text(ncid, *varid, "units", strlen(unit), unit));
  }
}
 
/*****************************************************************************/
 
void buffer_nc(
  atm_t *atm,
  double chisq,
  ncd_t *ncd,
  int track,
  int xtrack,
  int np0,
  int np1) {
 
  /* Set number of data points... */
  ncd->np = np1 - np0 + 1;
 
  /* Save retrieval data... */
  for (int ip = np0; ip <= np1; ip++) {
    ncd->ret_z[ip - np0] = (float) atm->z[ip];
    ncd->ret_p[track * L1_NXTRACK + xtrack] = (float) atm->p[np0];
    ncd->ret_t[(track * L1_NXTRACK + xtrack) * ncd->np + ip - np0] =
      (gsl_finite(chisq) ? (float) atm->t[ip] : GSL_NAN);
  }
}
 
/************************************************************************/
 
void fill_gaps(
  double x[L2_NTRACK][L2_NXTRACK][L2_NLAY],
  double cx,
  double cy) {
 
  double help[L2_NTRACK][L2_NXTRACK];
 
  /* Loop over layers... */
  for (int lay = 0; lay < L2_NLAY; lay++) {
 
    /* Loop over grid points... */
    for (int track = 0; track < L2_NTRACK; track++)
      for (int xtrack = 0; xtrack < L2_NXTRACK; xtrack++) {
 
        /* Init... */
        help[track][xtrack] = 0;
        double wsum = 0;
 
        /* Averrage data points... */
        for (int track2 = 0; track2 < L2_NTRACK; track2++)
          for (int xtrack2 = 0; xtrack2 < L2_NXTRACK; xtrack2++)
            if (gsl_finite(x[track2][xtrack2][lay])
                && x[track2][xtrack2][lay] > 0) {
              const double w = exp(-gsl_pow_2((xtrack - xtrack2) / cx)
                                   - gsl_pow_2((track - track2) / cy));
              help[track][xtrack] += w * x[track2][xtrack2][lay];
              wsum += w;
            }
 
        /* Normalize... */
        if (wsum > 0)
          help[track][xtrack] /= wsum;
        else
          help[track][xtrack] = GSL_NAN;
      }
 
    /* Copy grid points... */
    for (int track = 0; track < L2_NTRACK; track++)
      for (int xtrack = 0; xtrack < L2_NXTRACK; xtrack++)
        x[track][xtrack][lay] = help[track][xtrack];
  }
}
 
/************************************************************************/
 
void init_l2(
  ncd_t *ncd,
  int track,
  int xtrack,
  ctl_t *ctl,
  atm_t *atm) {
 
  static atm_t atm_airs;
 
  double k[NW], p, q[NG], t, zmax = 0, zmin = 1000;
 
  /* Reset track- and xtrack-index to match Level-2 data... */
  track /= 3;
  xtrack /= 3;
 
  /* Store AIRS data in atmospheric data struct... */
  atm_airs.np = 0;
  for (int lay = 0; lay < L2_NLAY; lay++)
    if (gsl_finite(ncd->l2_z[track][xtrack][lay])) {
      atm_airs.z[atm_airs.np] = ncd->l2_z[track][xtrack][lay];
      atm_airs.p[atm_airs.np] = ncd->l2_p[lay];
      atm_airs.t[atm_airs.np] = ncd->l2_t[track][xtrack][lay];
      if ((++atm_airs.np) > NP)
        ERRMSG("Too many layers!");
    }
 
  /* Check number of levels... */
  if (atm_airs.np <= 0)
    return;
 
  /* Get height range of AIRS data... */
  for (int ip = 0; ip < atm_airs.np; ip++) {
    zmax = GSL_MAX(zmax, atm_airs.z[ip]);
    zmin = GSL_MIN(zmin, atm_airs.z[ip]);
  }
 
  /* Merge AIRS data... */
  for (int ip = 0; ip < atm->np; ip++) {
 
    /* Interpolate AIRS data... */
    intpol_atm(ctl, &atm_airs, atm->z[ip], &p, &t, q, k);
 
    /* Weighting factor... */
    double w = 1;
    if (atm->z[ip] > zmax)
      w = GSL_MAX(1 - (atm->z[ip] - zmax) / 50, 0);
    if (atm->z[ip] < zmin)
      w = GSL_MAX(1 - (zmin - atm->z[ip]) / 50, 0);
 
    /* Merge... */
    atm->t[ip] = w * t + (1 - w) * atm->t[ip];
    atm->p[ip] = w * p + (1 - w) * atm->p[ip];
  }
}
 
/*****************************************************************************/
 
void read_nc(
  char *filename,
  ncd_t *ncd) {
 
  int varid;
 
  /* Open netCDF file... */
  printf("Read netCDF file: %s\n", filename);
  NC(nc_open(filename, NC_WRITE, &ncd->ncid));
 
  /* Read Level-1 data... */
  NC(nc_inq_varid(ncd->ncid, "l1_time", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l1_time[0]));
  NC(nc_inq_varid(ncd->ncid, "l1_lon", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l1_lon[0]));
  NC(nc_inq_varid(ncd->ncid, "l1_lat", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l1_lat[0]));
  NC(nc_inq_varid(ncd->ncid, "l1_sat_z", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l1_sat_z));
  NC(nc_inq_varid(ncd->ncid, "l1_sat_lon", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l1_sat_lon));
  NC(nc_inq_varid(ncd->ncid, "l1_sat_lat", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l1_sat_lat));
  NC(nc_inq_varid(ncd->ncid, "l1_nu", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l1_nu));
  NC(nc_inq_varid(ncd->ncid, "l1_rad", &varid));
  NC(nc_get_var_float(ncd->ncid, varid, ncd->l1_rad[0][0]));
 
  /* Read Level-2 data... */
  NC(nc_inq_varid(ncd->ncid, "l2_z", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l2_z[0][0]));
  NC(nc_inq_varid(ncd->ncid, "l2_press", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l2_p));
  NC(nc_inq_varid(ncd->ncid, "l2_temp", &varid));
  NC(nc_get_var_double(ncd->ncid, varid, ncd->l2_t[0][0]));
}
 
/*****************************************************************************/
 
void write_nc(
  char *filename,
  ncd_t *ncd) {
 
  int dimid[10], p_id, t_id, z_id;
 
  /* Create netCDF file... */
  printf("Write netCDF file: %s\n", filename);
 
  /* Read existing dimensions... */
  NC(nc_inq_dimid(ncd->ncid, "L1_NTRACK", &dimid[0]));
  NC(nc_inq_dimid(ncd->ncid, "L1_NXTRACK", &dimid[1]));
 
  /* Set define mode... */
  NC(nc_redef(ncd->ncid));
 
  /* Set new dimensions... */
  if (nc_inq_dimid(ncd->ncid, "RET_NP", &dimid[2]) != NC_NOERR)
    NC(nc_def_dim(ncd->ncid, "RET_NP", (size_t) ncd->np, &dimid[2]));
 
  /* Set new variables... */
  add_var(ncd->ncid, "ret_z", "km", "altitude", NC_FLOAT, &dimid[2], &z_id,
          1);
  add_var(ncd->ncid, "ret_press", "hPa", "pressure", NC_FLOAT, dimid, &p_id,
          2);
  add_var(ncd->ncid, "ret_temp", "K", "temperature", NC_FLOAT, dimid, &t_id,
          3);
 
  /* Leave define mode... */
  NC(nc_enddef(ncd->ncid));
 
  /* Write data... */
  NC(nc_put_var_float(ncd->ncid, z_id, ncd->ret_z));
  NC(nc_put_var_float(ncd->ncid, p_id, ncd->ret_p));
  NC(nc_put_var_float(ncd->ncid, t_id, ncd->ret_t));
 
  /* Close netCDF file... */
  NC(nc_close(ncd->ncid));
}