Model Outputs

Model Outputs

Every output folder and file name comes from one function, Setup/chess_paths.m, so the library and the GUI write the same layout.

Folder layout

<outfolder>/<project_name>/
    Event_Database/
        <db>.mat                              storm + prob_mass + db_info (shared by every PSE)
        <db>_raw_files.mat                    raw converted CHS files (CHS_Data)
    Tides/                                    GUI only
    CHESS_project_<project_name>.mat       GUI only
    PSE/<struc_id>/<case_name>/<WF>/
        <db>_<wflow>_config_file.mat          config of the run
        <db>_<wflow>_project_forcing.mat      project forcing (cache for hotstart)
        <db>_<wflow>_project_responses.mat    responses (Resp)
        <db>_<wflow>_sampling_seed.mat        LCS only: sampled life cycles
        RB1_<dataset>/  RB3/  FB1_<dataset>/  FB3/  LCS_DPA/     plots and per-dataset outputs

Level	Meaning
project_name	The project. One storm database per savepoint, shared by everything below
struc_id	The PSE (structure transect)
case_name	An alternative of that PSE (e.g. existing vs. raised crest)
<WF>	The workflow folder

workflow	<WF>	<wflow>
1	PROS	PROS-RB
2	DET	DET
3	LCS	LCS
4	PROS	PROS-FB

<db> is the storm database name: config.storm_db when set, else <region>_SP<id> for CHS input (e.g. CHS-NA_SP05572), or the file name of a storm + prob_mass .mat.

Files

File	Written by	Holds
<db>_raw_files.mat	call_chs_data_formater.m	CHS_Data: every converted CHS storm file (data, attributes, headers, units)
<db>.mat	call_chs_data_formater.m	storm (by storm type → data type → dataset), prob_mass, db_info
*_config_file.mat	setup, updated through the run	config and the PSE geometry
*_project_forcing.mat	call_project_forcing_formater.m	project_forcing: replicates (PROS), life cycles (LCS) or storms (DET)
*_project_responses.mat	call_project_response.m	Resp: responses, hazard curves or life-cycle summaries
*_sampling_seed.mat	chess_lcs.m	Sampling mode, sampled storm IDs per life cycle, raw XC copula samples

Storm database details (db_info)

Set when the database is built and applied to config on every run:

• Bias and uncertainty for SWL and Hm0, and where they came from (chs_dependencies or manual).
• Offshore elevation of the savepoint (study grid, positive up), its source and the distance to the grid point.
• Savepoint coordinates, Nyrs_XC / Nstm_XC, the source file and the build date. A rebuilt database has a new build date, which makes later runs resample (Hotstart Levels).

Sharing and reuse

• The storm database is built once and always holds the whole CHS input: every storm type and data type present, whatever the first run asked for. Each run selects what it needs.
• A run asking for something the database lacks (e.g. WLP) rebuilds <db>.mat from <db>_raw_files.mat without converting the HDF5 files again.
• A different savepoint or region writes a different <db>; nothing is overwritten.
• An LCS reference case (ref_case_name) is looked up in the same project and PSE, with the same <db>.
• Paths stored inside saved files are not used to find them, so a moved or renamed project folder still works.