param_definitions.md

Explanation of config parameters for Strain_2D calculators

[general]

• method: string, one of the methods below
• output_dir: string, path to output parent-directory
• input_vel_file: string, path to text file wtih input velocities
• write_metrics: bool, optional, default 0. Writes a text file with a Kostrov moment calculation and a chi-2 misfit to the data

[strain]

• range_strain: float/float/float/float, representing the target region for strain rate to be calculated upon, in W/E/S/N degrees longitude and latitude
• range_data: float/float/float/float, representing the region of velocity data that will be used for the calculation, in W/E/S/N degrees longitude and latitude
• inc: float/float, representing x/y grid spacing of resulting strain grid, in degrees longitude and latitude

[delaunay]/ [delaunay_flat]

• no parameters

[visr]

• See Native Documentation
• distance_weighting: string, either 'gaussian' or 'quadratic'.
  • If Gaussian: Distance weights (Shen et al., 2015)

$$ L_i = e^{- \Delta R_i^2 / D^2} $$

• If Quadratic: Distance weights (Shen et al., 2015)

$$ L_i = \frac {1} {1 + \Delta R_i^2 / D^2} $$

• spatial_weighting: string, either 'voronoi' or 'azimuth'. Corresponds to choosing a Z_i function in Shen et al. (2015)

• min_max_inc_smooth: float/float/float, representing minimum, maximum, and incremental spatial smoothing constants (I think that's D) (km). the 'Dist' in the output will be calculated from one of these choices (min + n*inc). The result is less block-like if you use finer increments.

• weighting_threshold: float, weighting threshold, Wt, from Shen et al., 2015; it ends up determining the optimal smoothing distance D. It's very important.

• uncertainty_threshold: float, cutoff uncertainty for velocity data input. The VISR examples have 0.5. I've been putting it too low (0.05) and I didn't have bad effects... not sure what units it has.

• num_creep_faults: int, number of creeping fault segments (up to 10 creeping faults)

• creep_file: string, path to txt file containing creeping fault endpoints

• executable: string, path to location of compiled fortran executable, visr.exe or similar

• Based on the VISR documentation, I'm not sure where Shen et al. (2015)'s L_0 (distance weighting threshold for ignoring the weight) gets defined. It might not get defined in the config file.

[gpsgridder]

• See Native Documentation
• poisson: float, poisson's ratio used in gpsgridder -S argument
• fd: float, fudge factor used in gpsgridder -Fd argument. The GMT default value is 0.01
• eigenvalue: float, ratio of the smallest eigenvalue used in the fit to the largest eigenvalue, placed in gpsgridder -C argument

[loc_avg_grad]

• EstimateRadiusKm: float, radius of searching neighborhood in km
• nstations: integer, minimum number of stations within search radius

[wavelets]

• See Native Documentation
• code_dir: string, path to location where surfacevel2strain on your computer system, such as '/Users/usrname/Documents/Software/surfacevel2strain'
• qmin: integer, minimum scale wavelength for the computation
• qmax: integer, maximum scale wavelength for the computation
• qsec: integer, scale wavelength for the secular velocity field. Must be between qmin and qmin.
• output_tag: string, path to the directory with preferred Matlab strain results from compearth. This directory is usually created by the Matlab run.

[geostats]

• model_type: string, one of [Gaussian, Exponential, Nugget].
• sill_east: float, sill value (variance) of Veast in mm^2/yr^2
• range_east: float, range (correlation length scale) of Veast in degrees
• nugget_east: float, point-wise variance of Veast (i.e. data noise level) in mm/yr
• sill_north: float, sill value of northing velocity in mm^2/yr^2
• range_north: float, range of Veast in degrees
• nugget_north: float, point-wise variance for Vnorth in mm/yr
• trend: boolean (TODO: Not Yet Implemented)