User Q & A

An unordered list of user questions received after the release of the 2014 model and maps. In many cases, the questions were prompted by inconsistencies or errors in the hazard code or input files. These have been addressed as issues with associated corrections.

Q: In the grid source input files, there is a reference to the file meanrjb.bin located in the bin directory. I do not see this file or directory in the GitHub repository?

A: The file getmeanrjf.v2.f should be compiled and used to create the various 'meanrjb' distance correction tables that are required when processing grid sources. They are not stored in the GitHub repository. (Answered by @pmpowers-usgs on 6/10/15)

Q: In the USGS source code for the 2014 NSHM, there is a bin2ascii utility program. Is there a program to go the other way (ascii2bin)?

A: There is no ascii2bin program. Essentially, all you need to do is store the relevant activity matrix in a 1D vector and write it using "putbuf2" from the iosubs library. (Answered by Steve Harmsen on 8/3/15)

Q: Why are the number of records in the Charleston agrid files (e.g. agrid_char_l.2013.ssc) higher than the box limit listed in the Charleston input files?

A: The Charleston agrids have longer headers than other agrid files. The header is 896 bytes long and basically replicates the information in the input files. The long header is indicated by the 20 in line 11 of the Charleston input files (e.g. CEUSchar_2014_l.ssc67.in). (Answered by @pmpowers-usgs on 8/27/15)

Q: Can users interpolate the amplification factors for any desired period/frequency? If yes, is it a simple linear interpolation?

A: The USGS does not have any guidance on how to extend the GMMs we use and you should contact the author/developer of the GMM directly. (Answered by @pmpowers-usgs on 9/9/15)

Q: Frequency-dependent factors are provided in Table 14 of the OFR to convert from hard rock (NEHRP site class A) to firm rock (NEHRP site class B/C). When comparing the values in file P11A_Rcd.rev to the hard rock values, the amplification factors are different from Table 14 values. How are the GMMs corrected for the B/C boundary?

A: The table files are all for hard rock (NEHRP site class A). In the fortran GMM implementation of the above models (Atkinson and Boore (2006), Pezeshk et al. (2011), and Atkinson (2008)), there is a coefficient called 'bcfac' that is added if the site is soft-rock (search for the amean11 function in hazFXnga13l.f and hazgrdXnga13l.f). These models all work in log10 ground-motion space (not natural log space, which is more common) and in units of cm/s² (not m/s²). 'bcfac' is also in this space and in these units. Table 14 was created to give a scaling that is comparable to other GMM implementations so some conversion has been applied and it is inappropriate to use the values from the table to scale values coming out of the GMMs. (Answered by @pmpowers-usgs on 9/9/15)

Q: There is a comment in the fortran code hazgridXnga13l.f which says 'Jan 3 2014: Apply same high-limit (40 hz) median clamp in AB06',A08',Pez11.' Would you explain more about this modification?

A: The comment refers to the soft rock period <=0.024 (40 Hz) distance decay (clamping) that is applied per Gail Atkinson's recommendation. The relevant lines from function amean11 in the fortran code are: if(jf.eq.12.or.jf.eq.11)then, amean = amean -0.3 + 0.15*rl (jf is a period index) (Answered by @pmpowers-usgs on 9/9/15)

Q: For the fault input files in the WUS, why are some faults missing from or inconsistently represented across the geodetic and/or geologic models?

A: See issues #28, #29, #30, #31, #32, #34, and #40.

Q: Apart from normal faulting in which three different branches are considered based on dip value to address epistemic uncertainty, why are there a variety of weights (1.0, 0.5, 0.75, 0.0) for the other senses of slip in WUS fault input files (e.g. 2014WUSbird.char.in)?

A: In 2008 sources were separated by region and GR vs CH weights were 50/50 in the PacNW and 33/67 in the rest of the WUS. In combining files in the 2014 model (geo vs. geodetic), corrections were made to the CH/GR weights such that after combining results, the original 50/50 or 33/67 weights were preserved; that is where the 0.75, for example, comes from. For some of the faults (e.g. Yaquina), the faults were not supplied to the geodetic modelers (so you see some weights of 0.0). FOr clarity, faults with a weight of 0.0 in the goedetic models have been removed in issue #31. The weights in these instances do not reflect the sense of slip or dip. (Answered by @pmpowers-usgs on 8/24/15)

Q: How is the fault-rupture model considered for IMW faults? There are different weights in the logic tree based on either partial rupture or full rupture. How are these weights considered in the input and/or combine files?

A: See answer above about weight correction from the 2008 to 2014 model. The fault-rupture model(s) are embedded in the weights within the geologic-bird-zeng model files. (Answered by @pmpowers-usgs on 8/24/15)

Q: What is the sense of slip for the 1349 Eastern Monitor Range fault zone? In the Geo and Zeng input files (e.g. 2014WUSgeo.char.in and e.g. 2014WUSzeng.char.in), it is considered a normal fault with three dips, but it is modeled as strike-slip with a dip of 60 degrees in the Bird model (e.g. 2014WUSbird.char.in). Same with the 871 Turner and Mill Creek faults, which are modeled as reverse, except in the Bird model where it is strike-slip?

A: Geodetic modeling generally provides a rake. There are cases where Bird's rake is below the 45 degree cutoff and therefore it is assigned strike-slip, despite the fact that the fault dips significantly. Please follow what the model prescribes. (Answered by @pmpowers-usgs on 8/24/15)

Q: Are grid files available in ascii format for both the CEUS and the WUS?

A: There are XML forms of the grid sources in the nshmp-model-cous-2014 directory, but they have not yet been finalized. (Answered by @pmpowers-usgs on 8/24/15)

Q: Are the UCERF3 grid files available for download?

A: Yes, the UCERF3 solutions are available here. There are also reformatted and reduced versions, the likes of which will be used in future NSHM products and hazard analyses here. (Answered by @pmpowers-usgs on 9/22/15)