The sites table contains the full input sites list. This means that not all sites in the site's table may have corresponding im_data in it.

There will be a copy of each DB for

IMDB (non-parametric)

• h5
• Tables
  • sites  - dataframe(name, lat, lon)
  • simulations - dataframe(id, simulation_name)
  • im_data/[station_name] - dataframe(sim_index, {im_types})
• Attributes
  • imdb_type: non-parametric
  • ims - list of ims contained in db

EMPDB (parametric)

• h5
• Tables
  • sites  - dataframe(name, lat, lon)
  • ruptures - dataframe(index_no, rupture_name)
  • im_data/[station_name] - dataframe(rupture_index, {im_types}, {im_types}_sigma)
• Attributes
  • imdb_type: parametric
  • ims - list of ims contained in db

SSDB (site-source db)

• h5
• Tables
  • Faults - datframe(index_no, name)
  • sites - dataframe(name, lat, lon)
  • distances/[station_name]  - dataframe (fault_index, rrup, rjb, rx, ry, rtvz)

Metadata

Each of the above tables have the below attributes stored:

• erf_version (name of the input file)
• station_version (name of the input file)
• source_type (Currently either fault or distributed-seismicity)
• data_created (time of creation)
• numpy_version
• pandas_version
• tables_version

Source

• NHM file (txt)
  • https://github.com/ucgmsim/Pre-processing/blob/master/SrfGen/NHM/NZ_FLTmodel_2010_v18p6.txt
• Distributed Seismicity (txt)
  • Lon, Lat, Mw, Depth, tect_type, Annual Rate of Recurrance

Every ensemble needs to be compromised of the same station list

Site

• ll file (txt)
  • non_uniform_whole_nz_with_real_stations-hh400_v18p6p2.ll
• vs30 file (txt)
  • non_uniform_whole_nz_with_real_stations-hh400_v18p6p2.vs30