def test_load_snb_parms():
fname = resource_filename('conductor', 'tests/input/snow_band.txt')
cells = load_snb_parms(fname, 15)
assert len(cells) == 6
assert len(cells['369560']) == 15
expected_zs = [ 2076, 2159, 2264, 2354, 2451, 2550, 2620, 2714, 2802,\
2900, 3000, 3100, 3200, 3300, 3400 ]
zs = [ band.median_elev for band in cells['368470'] ]
assert zs == expected_zs
def test_load_snb_parms():
fname = resource_filename('conductor', 'tests/input/snow_band.txt')
cells = load_snb_parms(fname, 15)
assert len(cells) == 6
assert len(cells['369560']) == 15
expected_zs = [ 2076, 2159, 2264, 2354, 2451, 2550, 2620, 2714, 2802,\
2900, 3000, 3100, 3200, 3300, 3400 ]
zs = [band.median_elev for band in cells['368470']]
assert zs == expected_zs
def large_merge_cells_unit_test_parms():
""" Uses input from a large real-world set of VIC snow band and vegetation
parameter files to test merge_cells()
"""
fname = resource_filename('conductor', 'tests/input/snow_band.txt')
elevation_cells = load_snb_parms(fname, 15)
fname = resource_filename('conductor', 'tests/input/veg.txt')
hru_cells = load_veg_parms(fname)
expected_zs = [ 2076, 2159, 2264, 2354, 2451, 2550, 2620, 2714, 2802, 2900,\
3000, 3100, 3200, 3300, 3400 ]
expected_afs = { 0.000765462339, 0.000873527611, 0.009125511809,\
0.009314626034, 0.004426673711, 0.004558753487, 0.001388838859, 0.000737445417 }
return elevation_cells, hru_cells, expected_zs, expected_afs
def main():
print('\n\nVIC + RGM ... together at last!')
# Parse command line parameters
vic_global_file, rgm_params_file, surf_dem_in_file, bed_dem_file, \
pixel_cell_map_file, init_glacier_mask_file, output_trace_files, \
glacier_root_zone_parms, open_ground_root_zone_parms, band_size = parse_input_parms()
# Get all initial VIC global parameters from the global parameter file
with open(vic_global_file, 'r') as f:
global_parms = Global(f)
assert global_parms.state_format == 'NETCDF', \
"{} only supports NetCDF input statefile input as opposed "\
"to the specified {}. Please change change this in your "\
"global file {}".format(__name__, global_parms.state_format,
vic_global_file)
# Initial VIC output state filename prefix is determined by STATENAME in the global file
state_filename_prefix = global_parms.statename
# Apply custom glacier_id and open_ground_id Band attributes, if provided
if global_parms.glacier_id is None:
print(
'No value for GLACIER_ID was provided in the VIC global file. Assuming default value of {}.'
.format(Band.glacier_id))
else:
Band.glacier_id = global_parms.glacier_id
# FIXME: reinstate the following commented-out code once OPEN_GROUND_ID is supported in VIC
# Numeric code indicating an open ground vegetation tile (HRU)
# if global_parms.open_ground_id is None:
# print('No value for OPEN_GROUND_ID was provided in the VIC global file. Assuming default value of {}.'.format(Band.open_ground_id))
# else:
# Band.open_ground_id = global_parms.open_ground_id
if band_size:
Band.band_size = band_size
# Load parameters from Snow Band Parameters File
num_snow_bands, snb_file = global_parms.snow_band.split()
num_snow_bands = int(num_snow_bands)
elevation_cell_dict = load_snb_parms(snb_file, num_snow_bands)
# Load vegetation parameters from initial Vegetation Parameter File
hru_cell_dict = vegparams.load_veg_parms(global_parms.vegparam)
# Apply custom HRU root_zone_parms attributes, if provided
if glacier_root_zone_parms or open_ground_root_zone_parms:
cells.apply_custom_root_zone_parms(hru_cell_dict,
glacier_root_zone_parms,
open_ground_root_zone_parms)
Band.glacier_root_zone_parms = glacier_root_zone_parms
Band.open_ground_root_zone_parms = open_ground_root_zone_parms
# Merge all VIC cells info gathered from Snow Band and Vegetation Parameter files and custom parameters
cells = merge_cell_input(hru_cell_dict, elevation_cell_dict)
# TODO: Do a sanity check to make sure band area fractions in Snow Band Parameters file add up to sum of HRU
# area fractions in Vegetation Parameter File for each cell?
#assert (area_fracs == [sums of HRU area fracs for all bands])
# The RGM will always output a DEM file of the same name (if running RGM for a single year at a time)
rgm_surf_dem_out_file = temp_files_path + 's_out_00001.grd'
# Open and read VIC-grid-to-RGM-pixel mapping file
cellid_map, elevation_map, cell_areas, num_cols_dem, num_rows_dem = get_rgm_pixel_mapping(
pixel_cell_map_file)
# Get DEM xmin, xmax, ymin, ymax metadata of Bed DEM and check file header validity
dem_xmin, dem_xmax, dem_ymin, dem_ymax, num_rows, num_cols = read_gsa_headers(
bed_dem_file)
# Verify number of columns & rows agree with what's stated in the pixel_to_cell_map_file
assert (num_cols == num_cols_dem) and (
num_rows == num_rows_dem), 'Mismatch of stated dimension(s) \
between Bed DEM in {} (num rows: {}, num columns: {}) and the VIC-grid-to-RGM-pixel map in {} \
(num rows: {}, num columns: {}). Exiting.\n'.format(
bed_dem_file, num_rows, num_cols, pixel_cell_map_file,
num_rows_dem, num_cols_dem)
# Read in the provided Bed Digital Elevation Map (BDEM) file to 2D bed_dem array
bed_dem = np.loadtxt(bed_dem_file, skiprows=5)
# Check header validity of Surface DEM file
_, _, _, _, num_rows, num_cols = read_gsa_headers(surf_dem_in_file)
# Verify number of columns & rows agree with what's stated in the pixel_to_cell_map_file
assert (num_cols == num_cols_dem) and (
num_rows == num_rows_dem), 'Mismatch of stated dimension(s) \
between Surface DEM in {} (num rows: {}, num columns: {}) and the VIC-grid-to-RGM-pixel map in {} \
(num rows: {}, num columns: {}). Exiting.\n'.format(
surf_dem_in_file, num_rows, num_cols, pixel_cell_map_file,
num_rows_dem, num_cols_dem)
# Read in the provided Surface Digital Elevation Map (SDEM) file to 2D surf_dem array
surf_dem_initial = np.loadtxt(surf_dem_in_file, skiprows=5)
# Check agreement between elevation map from VIC-grid-to-RGM-pixel file and the initial Surface DEM
assert (np.equal(elevation_map, surf_dem_initial)), 'Values mismatch between provided initial Surface DEM \
file (num rows:{}, num columns:{}) and VIC-grid-to-RGM-pixel file (num rows:{}, num columns:{}). Exiting.\n'\
.format(num_rows, num_cols, num_rows_dem, num_cols_dem)
# Check header validity of initial Glacier Mask file
_, _, _, _, num_rows, num_cols = read_gsa_headers(init_glacier_mask_file)
# Verify number of columns & rows agree with what's stated in the pixel_to_cell_map_file
assert (num_cols == num_cols_dem) and (
num_rows == num_rows_dem), 'Mismatch of stated dimension(s) \
between Glacier Mask in {} (num rows: {}, num columns: {}) and the VIC-grid-to-RGM-pixel map in {} \
(num rows: {}, num columns: {}). Exiting.\n'.format(
init_glacier_mask_file, num_rows, num_cols, pixel_cell_map_file,
num_rows_dem, num_cols_dem)
# Read in the provided initial glacier mask file to 2D glacier_mask array
glacier_mask = np.loadtxt(init_glacier_mask_file, skiprows=5)
# Apply the initial glacier mask and modify the band and glacier area fractions accordingly
update_area_fracs(cells, cell_areas, cellid_map, num_snow_bands,
surf_dem_initial, num_rows_dem, num_cols_dem,
glacier_mask)
temp_snb = temp_files_path + 'snb_temp_' + global_parms.startdate.isoformat(
) + '.txt'
snbparams.save_snb_parms(cells, temp_snb, band_map)
temp_vpf = temp_files_path + 'vpf_temp_' + global_parms.startdate.isoformat(
) + '.txt'
vegparams.save_veg_parms(cells, temp_vpf)
# Run the coupled VIC-RGM model for the time range specified in the VIC global parameters file
time_iterator = run_ranges(global_parms.startdate, global_parms.enddate,
global_parms.glacier_accum_startdate)
for start, end in time_iterator:
print('\nRunning VIC from {} to {}'.format(start, end))
# 1. Write / Update temporary Global Parameters File, temp_gpf
temp_gpf = temp_files_path + 'gpf_temp_{}.txt'.format(
start.isoformat())
# set global parameters for this VIC run
global_parms.vegparm = temp_vpf
global_parms.snow_band = '{} {}'.format(num_snow_bands, temp_snb)
global_parms.startdate = start
global_parms.enddate = end
global_parms.statedate = end
global_parms.write(temp_gpf)
print('invoking VIC with global parameter file {}'.format(temp_gpf))
# 2. Run VIC for a year. This will save VIC model state at the end of the year, along with a Glacier Mass Balance (GMB) polynomial for each cell
subprocess.check_call([vic_full_path, "-g", temp_gpf],
shell=False,
stderr=subprocess.STDOUT)
# 3. Open VIC NetCDF state file and get the most recent GMB polynomial for each grid cell being modeled
state_file = state_filename_prefix + "_" + start.isoformat() ## FIXME
print('opening VIC state file {}'.format(state_file))
state = h5py.File(state_file, 'r+')
gmb_polys = get_mass_balance_polynomials(state, state_file, cell_ids)
# 4. Translate mass balances using grid cell GMB polynomials and current veg_parm_file into a 2D RGM mass balance grid (MBG)
mass_balance_grid = mass_balances_to_rgm_grid(gmb_polys,
pixel_to_cell_map,
num_rows_dem,
num_cols_dem, cell_ids)
# write Mass Balance Grid to ASCII file to direct the RGM to use as input
mbg_file = temp_files_path + 'mass_balance_grid_' + start.isoformat(
) + '.gsa'
write_grid_to_gsa_file(mass_balance_grid, mbg_file, num_cols_dem,
num_rows_dem, dem_xmin, dem_xmax, dem_ymin,
dem_ymax)
# 5. Run RGM for one year, passing MBG, BDEM, SDEM
subprocess.check_call([
rgm_full_path, "-p", rgm_params_file, "-b", bed_dem_file, "-d",
surf_dem_in_file, "-m", mbg_file, "-o", temp_files_path, "-s", "0",
"-e", "0"
],
shell=False,
stderr=subprocess.STDOUT)
# remove temporary files if not saving for offline inspection
if not output_trace_files:
os.remove(mbg_file)
os.remove(rgm_surf_dem_file)
# 6. Read in new Surface DEM file from RGM output
rgm_surf_dem_out = np.loadtxt(rgm_surf_dem_out_file, skiprows=5)
temp_surf_dem_file = temp_files_path + 'rgm_surf_dem_out_' + start.isoformat(
) + '.gsa'
os.rename(rgm_surf_dem_out_file, temp_surf_dem_file)
# this will be fed back into RGM on next time step
surf_dem_in_file = temp_surf_dem_file
# 7. Update glacier mask
glacier_mask = update_glacier_mask(rgm_surf_dem_out, bed_dem,
num_rows_dem, num_cols_dem)
if output_trace_files:
glacier_mask_file = temp_files_path + 'glacier_mask_' + start.isoformat(
) + '.gsa'
write_grid_to_gsa_file(glacier_mask, glacier_mask_file)
# 8. Update areas of each elevation band in each VIC grid cell, and update snow band and vegetation parameters
update_area_fracs(cells, cell_areas, cellid_map, num_snow_bands, \
rgm_surf_dem_out, num_rows_dem, num_cols_dem, glacier_mask)
temp_snb = temp_files_path + 'snb_temp_' + start.isoformat() + '.txt'
snbparams.save_snb_parms(cells, temp_snb, band_map)
temp_vpf = temp_files_path + 'vpf_temp_' + start.isoformat() + '.txt'
vegparams.save_veg_parms(cells, temp_vpf)
# 11 Update HRUs in VIC state file
# don't forget to close the state file
# Get ready for the next loop
global_parms.init_state = "{}_{}.txt".format(global_parms.statename,
end.strftime("%Y%m%d"))
global_parms.statedate = end
def main():
print('\n\nVIC + RGM Hydro-Conductor starting...')
# Parse command line parameters
vic_path, rgm_path, output_path, vic_global_file, rgm_params_file, \
surf_dem_in_file, bed_dem_file, pixel_cell_map_file, \
init_glacier_mask_file, glacier_thickness_threshold, output_trace_files,\
glacier_root_zone_parms, open_ground_root_zone_parms, band_size,\
loglevel, output_plots\
= parse_input_parms()
# Set up logging
numeric_loglevel = getattr(logging, loglevel.upper())
logging.basicConfig(filename=output_path+'/hydrocon.log.'+\
strftime("%d-%m-%Y_%H:%M"), level=numeric_loglevel,\
format='%(levelname)s %(asctime)s %(message)s')
logging.info('------- VIC-RGM Hydro-Conductor Startup -------')
logging.info('VIC executable at {} '.format(vic_path))
logging.info('RGM executable at {} '.format(rgm_path))
# Get all initial VIC global parameters from the global parameter file
logging.info('Loading initial VIC global parameters from %s', vic_global_file)
with open(vic_global_file, 'r') as f:
global_parms = Global(f)
assert global_parms.state_format == 'NETCDF',\
'VIC only supports NETCDF input statefile input format.'\
'(STATE_FORMAT in global file is currently set as {})'\
.format(global_parms.state_format)
# Apply custom glacier_id and open_ground_id Band attributes, if provided
if global_parms.glacier_id is None:
logging.info('No value for GLACIER_ID was provided in the VIC global file. \
Assuming default value of {}.'.format(Band.glacier_id))
else:
Band.glacier_id = global_parms.glacier_id
# FIXME: reinstate the following commented-out code once OPEN_GROUND_ID
# is supported in VIC
# Numeric code indicating an open ground vegetation tile (HRU)
# if global_parms.open_ground_id is None:
# print('No value for OPEN_GROUND_ID was provided in the VIC global file.
# Assuming default value of {}.'.format(Band.open_ground_id))
# else:
# Band.open_ground_id = global_parms.open_ground_id
# Set custom elevation band size, if provided
if band_size:
Band.band_size = band_size
logging.info('Elevation band size set to {} meters.'.format(Band.band_size))
# Create temp_files_path, if it doesn't already exist
temp_files_path = output_path + '/hydrocon_temp/'
os.makedirs(temp_files_path, exist_ok=True)
logging.info('Temporary output files will be written to {}.'.format(temp_files_path))
# Load parameters from Snow Band Parameters File
num_snow_bands, snb_file = global_parms.snow_band.split()
num_snow_bands = int(num_snow_bands)
logging.info('Loading initial VIC snow band parameters from %s', snb_file)
elevation_cell_dict = load_snb_parms(snb_file, num_snow_bands)
# Load vegetation parameters from initial Vegetation Parameter File
logging.info('Loading initial VIC vegetation parameters from %s',\
global_parms.vegparam)
hru_cell_dict = load_veg_parms(global_parms.vegparam)
# Apply custom HRU root_zone_parms attributes, if provided
if glacier_root_zone_parms or open_ground_root_zone_parms:
cells.apply_custom_root_zone_parms(hru_cell_dict, glacier_root_zone_parms,\
open_ground_root_zone_parms)
Band.glacier_root_zone_parms = glacier_root_zone_parms
Band.open_ground_root_zone_parms = open_ground_root_zone_parms
# TODO: Do a sanity check to make sure band area fractions in Snow Band
# Parameters file add up to sum of HRU area fractions in Vegetation
# Parameter File for each cell?
#assert (area_fracs == [sums of HRU area fracs for all bands])
# Create Ordered dictionary of Cell objects by merging info gathered from
# Snow Band and Vegetation Parameter files and custom parameters
cells = merge_cell_input(hru_cell_dict, elevation_cell_dict)
# Open and read VIC-grid-to-RGM-pixel mapping file.
logging.info('Loading VIC-grid-to-RGM-pixel mapping from %s',\
pixel_cell_map_file)
vic_cell_mask, cell_areas, num_cols_dem, num_rows_dem\
= get_rgm_pixel_mapping(pixel_cell_map_file)
# Get DEM xmin, xmax, ymin, ymax metadata of Bed DEM and check file header
# validity
dem_xmin, dem_xmax, dem_ymin, dem_ymax, num_rows, num_cols\
= read_gsa_headers(bed_dem_file)
# Verify that number of columns & rows agree with what's stated in the
# pixel_cell_map_file
assert (num_cols == num_cols_dem) and (num_rows == num_rows_dem),\
'Mismatch of stated dimension(s) between Bed DEM in {} (num rows: {}, '
'num columns: {}) and the VIC-grid-to-RGM-pixel map in {} (num rows: {}, '
'num columns: {}). Exiting.\n'.format(bed_dem_file, num_rows, num_cols,
pixel_cell_map_file, num_rows_dem, num_cols_dem)
# Read in the provided Bed Digital Elevation Map (BDEM) file to 2D bed_dem
# array
logging.info('Loading Bed Digital Elevation Map (BDEM) from %s', bed_dem_file)
bed_dem = np.loadtxt(bed_dem_file, skiprows=5)
# Check header validity of Surface DEM file
_, _, _, _, num_rows, num_cols = read_gsa_headers(surf_dem_in_file)
# Verify number of columns & rows agree with what's stated in the
# pixel_to_cell_map_file
assert (num_cols == num_cols_dem) and (num_rows == num_rows_dem),\
'Mismatch of stated dimension(s) between Surface DEM in {} (num rows: {}, '
'num columns: {}) and the VIC-grid-to-RGM-pixel map in {} (num rows: {}, '
'num columns: {}). Exiting.\n'.format(surf_dem_in_file, num_rows, num_cols,\
pixel_cell_map_file, num_rows_dem, num_cols_dem)
# Read in the provided Surface Digital Elevation Map (SDEM) file to 2D
# surf_dem array
logging.info('Loading Surface Digital Elevation Map (SDEM) from %s',\
surf_dem_in_file)
current_surf_dem = np.loadtxt(surf_dem_in_file, skiprows=5)
# Check if Bed DEM has any points that are higher than the Surface DEM
# in the same location. If so, set these Bed DEM points to equal the
# Surface DEM values, thus avoiding producing negative values when the
# two are subtracted during glacier mask update. This reconciliation is
# necessary because the two DEMs come from different sources, and could
# have some overlapping elevation points.
dem_diffs = current_surf_dem - bed_dem
neg_val_inds = np.where(dem_diffs < 0)
num_neg_vals = len(neg_val_inds[0])
if num_neg_vals > 0:
bed_dem[neg_val_inds] = current_surf_dem[neg_val_inds]
new_bed_dem_file = bed_dem_file[0:-4] + '_adjusted.gsa'
logging.warning('The provided Bed DEM (%s) has %s elevation points \
(out of a total of %s elevation points in the domain) higher than those \
in the provided Surface DEM (%s), probably because they come from different \
data sources. The Bed DEM has been adjusted to equal the Surface DEM elevation \
at these points and written out to the file %s.',\
bed_dem_file, num_neg_vals, len(bed_dem), surf_dem_in_file, new_bed_dem_file)
bed_dem_file = new_bed_dem_file
write_grid_to_gsa_file(bed_dem, bed_dem_file, num_cols_dem, num_rows_dem,\
dem_xmin, dem_xmax, dem_ymin, dem_ymax)
# Check header validity of initial Glacier Mask file
_, _, _, _, num_rows, num_cols = read_gsa_headers(init_glacier_mask_file)
# Verify number of columns & rows agree with what's stated in the
# pixel_to_cell_map_file
assert (num_cols == num_cols_dem) and (num_rows == num_rows_dem),\
'Mismatch of stated dimension(s) between Glacier Mask in {} (num rows: {}, '
'num columns: {}) and the VIC-grid-to-RGM-pixel map in {} (num rows: {}, num '
'columns: {}). Exiting.\n'.format(init_glacier_mask_file, num_rows, num_cols,\
pixel_cell_map_file, num_rows_dem, num_cols_dem)
# Read in the provided initial glacier mask file to 2D glacier_mask array
logging.info('Loading initial Glacier Mask from %s', init_glacier_mask_file)
glacier_mask = np.loadtxt(init_glacier_mask_file, skiprows=5)
# Apply the initial glacier mask and modify the band and HRU area
# fractions according to their digitized fractions of the DEM
logging.debug('Applying initial band and HRU area fraction digitization.')
band_areas, glacier_areas = bin_bands_and_glaciers(cells, cell_areas,
vic_cell_mask, num_snow_bands, current_surf_dem,
glacier_mask)
digitize_domain(cells, cell_areas, band_areas, glacier_areas)
# Set the VIC output state file name prefix (to be written to STATENAME
# in the global file)
state_filename_prefix = temp_files_path + 'vic_hydrocon_state'
# Set the VIC results output file name prefix to NETCDF_OUTPUT_FILENAME given
# in the original global file.
netcdf_output_filename_prefix = global_parms.netcdf_output_filename
# The RGM will always output a DEM file of the same name (if running RGM for
# a single year at a time)
rgm_surf_dem_out_file = temp_files_path + 's_out_00001.grd'
# (initialisation done)
# Display initial surface DEM and glacier mask
if output_plots:
figure = GlacierPlotter(current_surf_dem, glacier_mask, bed_dem,
global_parms.startdate.isoformat(), output_trace_files, temp_files_path,
glacier_thickness_threshold)
#### Run the coupled VIC-RGM model for the time range specified in the VIC
# global parameters file
time_step = 0
time_iterator = run_ranges(global_parms.startdate,
global_parms.enddate,
global_parms.glacier_accum_startdate)
for start, end in time_iterator:
# Write temporary VIC parameter files
temp_snb = temp_files_path + 'snb_temp_' + start.isoformat() + '.txt'
logging.debug('Writing temporary snow band parameter file %s', temp_snb)
save_snb_parms(cells, temp_snb)
temp_vpf = temp_files_path + 'vpf_temp_' + start.isoformat() + '.txt'
logging.debug('Writing temporary vegetation parameter file %s', temp_vpf)
save_veg_parms(cells, temp_vpf)
temp_gpf = temp_files_path + 'gpf_temp_{}.txt'.format(start.isoformat())
logging.debug('Writing temporary global parameter file %s', temp_gpf)
global_parms.vegparam = temp_vpf
global_parms.snow_band = '{} {}'.format(num_snow_bands, temp_snb)
global_parms.startdate = start
global_parms.enddate = end
global_parms.statedate = end
global_parms.statename = state_filename_prefix
global_parms.netcdf_output_filename = netcdf_output_filename_prefix \
+ start.isoformat() + '-' + end.isoformat() + '.nc'
if time_step > 0:
global_parms.glacier_accum_start_year = start.year
global_parms.glacier_accum_start_month = start.month
global_parms.glacier_accum_start_day = start.day
global_parms.write(temp_gpf)
# Run VIC for a year, saving model state at the end
print('\nRunning VIC from {} to {}'.format(start, end))
logging.info('\nRunning VIC from %s to %s using global parameter file %s',\
start, end, temp_gpf)
try:
subprocess.check_call([vic_path, "-g", temp_gpf], shell=False,\
stderr=subprocess.STDOUT)
except subprocess.CalledProcessError as e:
logging.error('Subprocess invocation of VIC failed with the following \
error: %s', e)
sys.exit(0)
# Open VIC NetCDF state file and load the most recent set of state
# variable values for all grid cells being modeled
state_file = state_filename_prefix + '_' + end.isoformat()
logging.info('Reading saved VIC state file %s', state_file)
# leave the state file open for modification later
state_dataset = netCDF4.Dataset(state_file, 'r+')
state_dataset.set_auto_mask(False)
# these should never change within a run of the Hydro-Conductor:
Cell.Nlayers = state_dataset.state_nlayer
Cell.Nnodes = state_dataset.state_nnode
# drop unused fit error term from glacier mass balance polynomial
Cell.NglacMassBalanceEqnTerms = state_dataset.state_nglac_mass_balance_eqn_terms - 1
state = state_dataset.variables
# read new states of all cells
read_state(state, cells)
# optionally leave the last VIC state file on disk
if not output_trace_files:
os.remove(state_file)
gmb_polys = {}
cell_ids = []
for cell_id in cells:
# Make sure VIC cell IDs in the state file agree with those in the
# vic_cell_mask, which is derived from the pixel_cell_map_file
if int(cell_id) not in vic_cell_mask:
print('Cell ID {} read from the VIC state file {} was not found in '
'the VIC cell mask derived from the given RGM-Pixel-to-VIC-Cell map '
'file (option --pixel_map) {}. Exiting.'
.format(cell_id, state_file, pixel_cell_map_file))
logging.error('Cell ID {} read from the VIC state file {} was not '
'found in the VIC cell mask derived from the given '
'RGM-Pixel-to-VIC-Cell map file (option --pixel_map) {}')
sys.exit(0)
cell_ids.append(cell_id)
# Read Glacier Mass Balance polynomial terms from cell states;
# leave off 4th the "fit error" term at the end of the GMB polynomial.
gmb_polys[cell_id] = cells[cell_id].cell_state.variables\
['GLAC_MASS_BALANCE_EQN_TERMS'][0:Cell.NglacMassBalanceEqnTerms]
# Translate mass balances using grid cell GMB polynomials and current
# surface DEM into a 2D RGM mass balance grid (MBG) and write the
# MBG to an ASCII file to give as input to the RGM
mbg_file = temp_files_path + 'mass_balance_grid_' + end.isoformat()\
+ '.gsa'
logging.debug('Converting glacier mass balance polynomials to 2D grid \
and writing to file %s', mbg_file)
mass_balance_grid = mass_balances_to_rgm_grid(gmb_polys, vic_cell_mask,\
current_surf_dem, bed_dem, num_rows_dem, num_cols_dem)
write_grid_to_gsa_file(mass_balance_grid, mbg_file, num_cols_dem,\
num_rows_dem, dem_xmin, dem_xmax, dem_ymin, dem_ymax)
# Write modified surface DEM with all pixels lying outside of VIC
# domain set equal to the bed DEM
rgm_surf_dem_in_file = temp_files_path + 'rgm_surf_dem_in_'\
+ end.isoformat() + '.gsa'
write_grid_to_gsa_file(current_surf_dem, rgm_surf_dem_in_file, num_cols_dem,\
num_rows_dem, dem_xmin, dem_xmax, dem_ymin, dem_ymax)
# Run RGM for one year, passing it the MBG, BDEM, SDEM
logging.info('Running RGM for current year with parameter file %s, \
Bed DEM file %s, Surface DEM file %s, Mass Balance Grid file %s',\
rgm_params_file, bed_dem_file, rgm_surf_dem_in_file, mbg_file)
try:
subprocess.check_call([rgm_path, "-p", rgm_params_file, "-b",\
bed_dem_file, "-d", rgm_surf_dem_in_file, "-m", mbg_file, "-o",\
temp_files_path, "-s", "0", "-e", "0" ], shell=False,\
stderr=subprocess.STDOUT)
except subprocess.CalledProcessError as e:
logging.error('Subprocess invocation of RGM failed with the following \
error: %s', e)
sys.exit(0)
# Read in new Surface DEM file from RGM output
logging.debug('Reading Surface DEM file from RGM output %s',\
rgm_surf_dem_out_file)
current_surf_dem = np.loadtxt(rgm_surf_dem_out_file, skiprows=5)
temp_surf_dem_file = temp_files_path + 'rgm_surf_dem_out_'\
+ end.isoformat() + '.gsa'
os.rename(rgm_surf_dem_out_file, temp_surf_dem_file)
# this will be fed back into RGM on next time step:
rgm_surf_dem_in_file = temp_surf_dem_file
# remove temporary files if not saving for offline inspection
if not output_trace_files:
os.remove(mbg_file)
os.remove(rgm_surf_dem_in_file)
os.remove(rgm_surf_dem_out_file)
# Update glacier mask
logging.debug('Updating Glacier Mask')
glacier_mask = update_glacier_mask(current_surf_dem, bed_dem,
num_rows_dem, num_cols_dem, glacier_thickness_threshold)
if output_trace_files:
glacier_mask_file = temp_files_path + 'glacier_mask_'\
+ end.isoformat() + '.gsa'
logging.debug('Writing Glacier Mask to file %s', glacier_mask_file)
write_grid_to_gsa_file(glacier_mask, glacier_mask_file, num_cols_dem,
num_rows_dem, dem_xmin, dem_xmax, dem_ymin, dem_ymax)
if output_plots:
figure.update_plots(current_surf_dem, glacier_mask,
glacier_thickness_threshold, bed_dem, end.isoformat())
# Update HRU and band area fractions and state for all VIC grid cells
logging.debug('Updating VIC grid cell area fractions and states')
update_area_fracs(cells, cell_areas, vic_cell_mask, num_snow_bands,
current_surf_dem, glacier_mask)
# Update the VIC state file with new state information
new_state_date = end + one_day
new_state_file = state_filename_prefix + '_' + new_state_date.isoformat()
logging.debug('Writing updated VIC state file %s', new_state_file)
# Set the new state file name VIC will have to read in on next iteration
global_parms.init_state = new_state_file
new_state_dataset = netCDF4.Dataset(new_state_file, 'w')
write_state(cells, state_dataset, new_state_dataset, new_state_date)
logging.debug('Closing old and updated NetCDF state files.')
state_dataset.close()
new_state_dataset.close()
time_step = time_step + 1
def toy_domain_64px_cells():
# NOT USED BUT USEFUL INFO: initial band map of lower elevation bounds for
# all existing (valid) bands
# test_band_map = {
# allows for glacier growth at top:
# '12345': [2000, 2100, 2200, 2300, 0],
# allows for glacier growth at top, and revelation of lower band at bottom:
# '23456': [0, 1900, 2000, 2100, 0]}
fname = resource_filename('conductor', 'tests/input/snb_toy_64px.txt')
elevation_cells = load_snb_parms(fname, 5)
fname = resource_filename('conductor', 'tests/input/vpf_toy_64px.txt')
hru_cells = load_veg_parms(fname)
cells = merge_cell_input(hru_cells, elevation_cells)
cell_ids = list(cells.keys())
# We have a total allowable number of snow bands of 5, with 100m spacing
num_snow_bands = 5
band_size = 100
# Spatial DEM layouts
bed_dem_by_cells = {
cell_ids[0]:
np.array([
[2065, 2055, 2045, 2035, 2025, 2015, 2005, 2000],
[2075, 2085, 2100, 2100, 2100, 2100, 2100, 2005],
[2085, 2100, 2210, 2230, 2220, 2200, 2110, 2010],
[2090, 2100, 2240, 2377, 2310, 2230, 2125, 2015],
[2070, 2110, 2230, 2340, 2320, 2230, 2130, 2020],
[2090, 2105, 2200, 2210, 2220, 2220, 2120, 2015],
[2090, 2100, 2105, 2110, 2140, 2150, 2130, 2010],
[2080, 2075, 2065, 2055, 2045, 2035, 2020, 2000]
]),
# note: bed elev 2085 at position [1,1] above will be used to demonstrate
# glacier receding to reveal more band area fraction for Band 0
cell_ids[1]:
np.array([
[1970, 1975, 1850, 1799, 1975, 1965, 1960, 1960],
[1970, 2000, 2025, 2035, 2005, 2005, 2000, 1965],
[1975, 2000, 2100, 2125, 2130, 2110, 2000, 1970],
[1985, 2005, 2105, 2130, 2150, 2100, 2000, 1975],
[1990, 2010, 2110, 2120, 2110, 2105, 2005, 1980],
[1980, 2005, 2105, 2105, 2110, 2100, 2000, 1980],
[1970, 2000, 2000, 2020, 2035, 2025, 2000, 1970],
[1965, 1965, 1970, 1970, 1975, 1960, 1950, 1960]
])
}
initial_surf_dem_by_cells = {
cell_ids[0]:
np.array([
[2065, 2055, 2045, 2035, 2025, 2015, 2005, 2000],
[2075, 2100, 2120, 2140, 2130, 2120, 2100, 2005],
[2085, 2110, 2250, 2270, 2260, 2240, 2110, 2010],
[2090, 2120, 2260, 2377, 2310, 2250, 2125, 2015],
[2070, 2120, 2250, 2340, 2320, 2250, 2130, 2020],
[2090, 2105, 2200, 2210, 2220, 2220, 2120, 2015],
[2090, 2100, 2105, 2110, 2140, 2150, 2130, 2010],
[2080, 2075, 2065, 2055, 2045, 2035, 2020, 2000]
]),
cell_ids[1]:
np.array([
[1970, 1975, 1995, 1995, 1975, 1965, 1960, 1960],
[1970, 2000, 2045, 2055, 2005, 2005, 2000, 1965],
[1975, 2000, 2100, 2155, 2160, 2140, 2000, 1970],
[1985, 2005, 2105, 2160, 2180, 2130, 2000, 1975],
[1990, 2010, 2110, 2150, 2140, 2105, 2005, 1980],
[1980, 2005, 2105, 2105, 2110, 2100, 2000, 1980],
[1970, 2000, 2000, 2020, 2035, 2025, 2000, 1970],
[1965, 1965, 1970, 1970, 1975, 1960, 1950, 1960]
])
}
initial_glacier_mask_by_cells = {
cell_ids[0]:
np.array([
[ 0, 0, 0, 0, 0, 0, 0, 0 ],
[ 0, 1, 1, 1, 1, 1, 0, 0 ],
[ 0, 1, 1, 1, 1, 1, 0, 0 ],
[ 0, 1, 1, 0, 0, 1, 0, 0 ],
[ 0, 1, 1, 0, 0, 1, 0, 0 ],
[ 0, 0, 0, 0, 0, 0, 0, 0 ],
[ 0, 0, 0, 0, 0, 0, 0, 0 ],
[ 0, 0, 0, 0, 0, 0, 0, 0 ]
]),
cell_ids[1]:
np.array([
[ 0, 0, 1, 1, 0, 0, 0, 0 ],
[ 0, 0, 1, 1, 0, 0, 0, 0 ],
[ 0, 0, 0, 1, 1, 1, 0, 0 ],
[ 0, 0, 0, 1, 1, 1, 0, 0 ],
[ 0, 0, 0, 1, 1, 0, 0, 0 ],
[ 0, 0, 0, 0, 0, 0, 0, 0 ],
[ 0, 0, 0, 0, 0, 0, 0, 0 ],
[ 0, 0, 0, 0, 0, 0, 0, 0 ]
])
}
def build_padded_dem_aligned_map(array_1, array_2, padding_thickness,\
fill_value):
""" Helper function to create a map that is aligned with a DEM made up
of two adjacent rectangular pixel arrays (representing 2 VIC cells),
with NaN padding around the edges (simulating the output from
get_rgm_pixel_mapping(), where some pixels that are read in do not
belong to either VIC cell). This can be used to generate test
instances of surf_dem, bed_dem, cellid_map, glacier_mask
"""
vertical_pad = np.empty((padding_thickness, 2*padding_thickness+len(array_1[0])+len(array_2[0])))
vertical_pad.fill(fill_value)
horizontal_pad = np.empty((len(array_1), padding_thickness))
horizontal_pad.fill(fill_value)
padded_map = np.concatenate((array_1, array_2), axis=1)
padded_map = np.concatenate((horizontal_pad, padded_map), axis=1)
padded_map = np.concatenate((padded_map, horizontal_pad), axis=1)
padded_map = np.concatenate((vertical_pad, padded_map), axis=0)
padded_map = np.concatenate((padded_map, vertical_pad), axis=0)
return padded_map
# Create cellid_map that is padded by 2 rows & columns of np.NaNs (to simulate mapping
# of valid pixels to VIC cells, as read in by get_rgm_pixel_mapping())
cellid_map_cell_0 = np.empty((8,8))
cellid_map_cell_0.fill(cell_ids[0])
cellid_map_cell_1 = np.empty((8,8))
cellid_map_cell_1.fill(cell_ids[1])
cellid_map = build_padded_dem_aligned_map(cellid_map_cell_0,\
cellid_map_cell_1, 2, 9999)
# Create bed_dem with padding of 2
bed_dem = build_padded_dem_aligned_map(bed_dem_by_cells[cell_ids[0]],\
bed_dem_by_cells[cell_ids[1]], 2, 9999)
# Create initial surf_dem with padding of 2
surf_dem = build_padded_dem_aligned_map(initial_surf_dem_by_cells[cell_ids[0]],\
initial_surf_dem_by_cells[cell_ids[1]], 2, 9999)
# Non-padded version of surf_dem:
#surf_dem = np.concatenate((initial_surf_dem_by_cells[cell_ids[0]], \
# initial_surf_dem_by_cells[cell_ids[1]]), axis=1)
# Create initial glacier mask with padding of 2
glacier_mask = build_padded_dem_aligned_map(initial_glacier_mask_by_cells[cell_ids[0]],\
initial_glacier_mask_by_cells[cell_ids[1]], 2, 9999)
# The toy surface DEM above, broken down by elevation bands. Useful for checking median elevations
cell_band_surf_pixel_elevations = {
cell_ids[0]: [
# band 0, median: 2040
[2065, 2055, 2045, 2035, 2025, 2015, 2005, 2000,
2075, 2005,
2085, 2010,
2090, 2015,
2070, 2020,
2090, 2015,
2090, 2010,
2080, 2075, 2065, 2055, 2045, 2035, 2020, 2000],
# band 1, median: 2120
[2100, 2120, 2140, 2130, 2120, 2100,
2110, 2110,
2120, 2125,
2110, 2130,
2105, 2120,
2100, 2105, 2110, 2140, 2150, 2130],
# band 2, median: 2250
[2250, 2270, 2260, 2240,
2260, 2250,
2250, 2250,
2200, 2210, 2220, 2220],
# band 3, median: 2330
[2377, 2310,
2340, 2320]
],
cell_ids[1]: [
# band 0, median: 1970
[1970, 1975, 1995, 1995, 1975, 1965, 1960, 1960,
1970, 1965,
1975, 1970,
1985, 1975,
1990, 1980,
1980, 1980,
1970, 1970,
1965, 1965, 1970, 1970, 1975, 1960, 1950, 1960],
# band 1, median: 2005
[2000, 2045, 2055, 2005, 2005, 2000,
2000, 2000,
2005, 2000,
2010, 2005,
2005, 2000,
2000, 2000, 2020, 2035, 2025, 2000],
# band 2, median: 2120
[2100, 2155, 2160, 2140,
2105, 2160, 2180, 2130,
2110, 2150, 2140, 2105,
2105, 2105, 2110, 2100],
]
}
return cells, cell_ids, num_snow_bands, band_size, cellid_map, bed_dem,\
surf_dem, glacier_mask, cell_band_surf_pixel_elevations