Observed Data Processing
The steps in sim2seis
are typically executed within an ert
-controlled ensemble run, which includes multiple realisations. Since the observed seismic data remains constant, only a single run is required, and this is performed outside of ert
via the command line:
# Normal run
> sim2seis_observed_data -s <start_dir> -c <config_dir> -f <config_file>
>
# Run with verbose output
> sim2seis_observed_data -s <start_dir> -c <config_dir> -f <config_file> -v true
>
# Run without generating attributes - depth conversion only
> sim2seis_observed_data -s <start_dir> -c <config_dir> -f <config_file> -n true
Figure 1: Command line interface to observed data processing.
Similar to the sim2seis
workflow, the processing of observed data is controlled by YAML configuration files. There is one parameter file, shown in Figure 2, and one interval definition file for generating attribute maps. The interval definition file follows the format described in attribute maps.
########################################################################################################################
#
# Configuration file for observation data in connection with a sim2seis runs
#
# Lines that are commented out are parameters that users very seldom have to change, they will use validated
# default values. However, they can be changed at need, but please verify that the values entered are valid
# before running any of the parts in the sim2seis workflow
#
########################################################################################################################
## For the main class: relative path to the config file
#______________________________________________________________________________________________________________________#
attribute_definition_file: observed_data_intervals_drogon.yml
# pickle_file_output_path: ../../share/observations/pickle_files
# pickle_file_prefix: observed_data
# rel_path_global_config: ../../fmuconfig/output
# observed_data_path: ../../share/observations/cubes
## Section for depth maps - inputs only
#______________________________________________________________________________________________________________________#
observed_depth_surf:
# paths and standard naming. Depth maps are read from horizon folder INPUT
suffix_name: depth
horizon_names: [MSL, TopVolantis, BaseVolantis, BaseVelmodel]
# depth_dir: ../../share/observations/maps
## Section for depth conversion
#______________________________________________________________________________________________________________________#
depth_conversion:
z_inc: 4.0
min_depth: 1500.0
max_depth: 2000.0
# Section for ert and webviz export
#______________________________________________________________________________________________________________________#
webviz_map:
# grid_file: simgrid.roff
# zone_file: simgrid--zone.roff
# region_file: simgrid--region.roff
attribute_error: 0.07
# grid_path: ../../sim2seis/input/pem
output_path: ../../share/observations/tables
## Section for time data
#______________________________________________________________________________________________________________________#
#observed_time_data:
# time_cube_dir: ../../share/observations/cubes
# time_cube_prefix: seismic--
# time_suffix: --time.gri
# horizon_dir: ../../share/observations/maps
## End of config file
#______________________________________________________________________________________________________________________#
Figure 2: Example YAML configuration file for observed data.
When to run Observed Data Processing from ert
If sim2seis
is run with structural uncertainty, the depth conversion of observed seismic data should be repeated for each realisation, as it will vary. If not, cubes and attribute maps in depth domain will not be consistent between observed and modelled data. All required parameters are defined in ert configuration.
-- Optional run of depth conversion of observed data:
FORWARD_MODEL OBSERVED_DATA(<START_DIR>=<JOB_STARTDIR>, <CONFIG_DIR>=<RELPATH_CONFIG_FILES>, <CONFIG_FILE>=<OBS_DATA_CONFIG>, <VERBOSE>=<VERBOSE_OUTPUT>)