Fault model inconsistencies

[pmpowers-usgs]

During the development of the 2014 NSHM:

• Some faults had been added to the fault dB for the purposes of display but should have been removed prior to hand off to the deformation modelers (e.g. Cheraw).
• Some faults were not provided to deformation modelers (e.g. those in TX).

Issues:

1. Remove Cheraw from WUS geodetic fault files. It's inclusion gave it an additional 0.2 weight with ground motions computed using NGAW2.
2. Adjust weights of TX faults in geologic model to compensate for their absence in geodetic models:
   • East Franklin Mountains fault (900)
   • Campo Grande (902)
   • Acala fault (903)
   • Arroyo Diablo fault (904)
   • Caballo fault (906ab)
   • East Sierra Diablo fault (910)
   • West Delaware Mountains fault zone (911)
   • East Baylor Mountain - Carizzo Mountain fault (912)
   • West Eagle Mountains-Red Hills fault (913)
   • West Indio Mountains fault (915)
   • West Lobo Valley fault zone (918abcd)
3. The Carson Range – Kings Canyon system is inconsistently represented in the 2014 model. The system as a whole gets 0.75 weight, with individual component faults getting 0.25 weight for rupturing independently. These component faults should not have been included in the geodetic models and their weights should have been adjusted in the input files such that they still receive total weights of 0.25. Current inconsistent representation:
   • Carson Range missing from both geodetic models
   • Kings Canyon included in Zeng
   • Indian Hill included in both geodetic models
   • Carson City included in both geodetic models
   • Little Valley – although not strictly part of the Carson–Kings system, it is close by and also considered to have a low probability of occurrence (0.25) and should not be included in geodetic inputs.
4. Adjust Southern Whidbey Island fault weights; alternate model is missing from geodetic models; north and south variants missing just from Bird.
5. Add dip variants to Lost River fault in geodetic models.
6. Saddle Mountain fault variants are identical in geologic model and absent from geodetic. The alternate model developed in 2014 was a faster 'Holocene' slip-rate model. Need to update the 0.25 weighted 'alt2' version event rate to reflect 0.5 mm/yr slip rate (geologic model only).

Noted by @AHajiSoltani and @JMAltekruse

[pmpowers-usgs]

Resolutions

1. Removed Cheraw in geodetic models (9c8c984).
2. Adjusted weights of TX faults. To compensate for missing geodetic branches, the weights of each of these normal faults needs to be scaled by 1.25, resulting in final 35°, 50°, and 65° dip variant weights of 0.25, 0.75, and 0.25 (dca9839).
3. Adjusted weights of Carson Range fault in geologic model to compensate for its absence in the geodetic model. This model already gets a low weight relative to the Carson–Kings Range combined model; 35°, 50°, and 65° dip variant weights increased to 0.0625, 0.1875, and 0.0625, respectively, from 0.05, 0.15, and 0.05.
   • Updated weights of Carson Range in geologic model (8107466).
   • Removed Kings Canyon from Zeng and updated geologic weights (4bdce6d).
   • Removed Indian Hill and Carson City from Zeng and Bird and updated geologic weights (5636813).
   • Removed Little Valley from Zeng and Bird and updated geologic weights (d12bf7a).
4. S. Whidbey Isl. is ok, more or less; there are no weighting errors that result from the geodetic models not considering the extended SWI fault ('alt2', which gets 0.5 weight in the geologic model) as the original source model gets full weight. Moreover, Zeng modeled all three strands and recovered values close to the geologic rate for each strand (6 mm/yr / 3 = 2 mm/yr). Bird's rates are about 3x this so he only modeled the single middle strand. Zeng slip rates have been consolidated onto middle strand and rupture rates updated accordingly ( aab954b ).
5. Added Lost River 35°, 50°, and 65° dip variants to geodetic models/inputs with 0.2, 0.6, and 0.2 weights (78fccbd). Although geodetic slip rates were computed for a 45° dipping fault, the change to a 50° middle branch is nominal so no correction has been made to the geodetic slip rates in the fault database. These slip rates have been used, however, to update the event rates based on a 50° dipping fault, consistent with the geologic model ( bba7882 ).
6. The 'alt2' geologic model was updated to reflect a faster 0.5 mm/yr Holocene slip-rate ( 6b98aa1 ).

See attached worksheet for rupture rate recalculations that were performed using OpenSHA MFD utilities.
2014-WUS-RateCorrections.xlsx