Event-Based PSHA

The Event Based PSHA calculator computes and stores stochastic event sets and the corresponding ground motion fields. This calculator can also produce hazard curves and hazard maps, similar to the Classical PSHA calculator. The inset below shows an example of the list of results provided by the OpenQuake engine at the end of an event-based PSHA calculation:

user@ubuntu:~$ oq engine --lo <calc_id>
id | name
10 | Ground Motion Fields
11 | Hazard Curves
12 | Hazard Maps
13 | Realizations
14 | Earthquake Ruptures
15 | Events
16 | Uniform Hazard Spectra

Exporting the outputs from the ruptures will produce a CSV file with the following columns:

1. rup_id: incremental number identifying the rupture

2. multiplicity: how many times the rupture occurs in the effective investigation time

3. mag: float specifying the magnitude of the rupture

4. centroid_lon: longitude of the centroid of the rupture

5. centroid_lat: latitude of the centroid of the rupture

6. centroid_depth: depth (in km) of the centroid of the rupture

7. trt: string specifying the tectonic region type

8. strike: strike angle of the rupture surface

9. dip: dip angle of the rupture surface

10. rake: rake angle of the rupture surface

The outputs from the Ground Motion Fields can be exported in the csv format. Exporting the outputs from the Ground Motion Fields in the csv format results in three csv files, where XXX is a placeholder for the calculation ID:

1. gmf_data_XXX.csv

2. sigma_epsilon_XXX.csv

3. sitemesh_XXX.csv

Here is an example of the gmf_data_XXX.csv file for the EventBasedPSHA demo:

site_id	event_id	gmv_PGA
0	99	0.06888
0	125	0.06498
0	142	0.05881
0	151	0.05569
0	156	0.06918
0	161	0.06629
0	179	0.10528
0	184	0.05503
0	200	0.06455
…	…	…

The sitemesh_XXX.csv file provides the association between the site ids in the Ground Motion Fields csv file with their latitude and longitude coordinates, as in the following example:

site_id	lon	lat
0	0.50000	-0.5

The sigma_epsilon_XXX.csv file provides the inter event residuals and inter event standard deviations for each event, as in the following example:

event_id	rlz_id	sig_inter_PGA	eps_inter_PGA
0	1	2.600000E-01	0.84923
1	1	2.600000E-01	0.25794
2	2	NAN	0.84894
3	2	NAN	0.00495
4	2	NAN	-0.8094396
5	2	NAN	1.30782
6	3	NAN	0.89407
7	0	2.600000E-01	-1.723876
8	2	NAN	-0.4926568
…	…	…	…

Users wanting to know exactly how sigma and epsilon are computed should look at the implementation in the file openquake/hazardlib/calc/gmf.py. For instance the event based demo uses the GMPEs BooreAtkinson2008 and ToroEtAl2002, with the latter being defined only for the total standard deviation, and that explains the NAN in the sigmas for some events.

The Events output produces a csv file with fields event_id, rup_id, rlz_id, year and ses_id. The event_id is a 32 bit integer that identifies uniquely the event; the rup_id is a 64 bit integer that identifies uniquely the rupture; the rlz_id is a 16 bit integer that identifies uniquely the realization. The year field is a random number from 0 to the effective investigation time while the ses_id field is a random number from 0 to ses_per_logic_tree_path. Notice that in previous versions of the engine (before engine-3.11) the semantic of those fields was different. Here is an example:

event_id	rup_id	rlz_id	year	ses_id
0	0	1	14_337	35
1	0	1	14_729	99
2	0	2	26_309	53
3	0	2	34_672	219
4	0	2	27_925	180
5	0	2	31_655	152
6	0	3	37_509	170
7	0	0	4_784	225
8	0	2	25_462	84
…	…	…	…	…

The Realizations output produces a csv file listing the source model and the combination of ground shaking intensity models for each path sampled from the logic tree. An example of such a file is shown below in the table below:

rlz_id	branch_path	weight
0	A~AA	0.36000
1	A~AB	0.24000
2	A~BA	0.24000
3	A~BB	0.16000

The branch_path field is used to determine which GMPEs are being used for each tectonic region type, as explained in the logic tree section of the documentation.