First off, the amount of documentation for MTfit is really excellent. Thank you for putting in the time. It makes the package accessible to a wider audience.
There are a couple of places where the input parameters could benefit from additional details.
On the following page: https://djpugh.github.io/MTfit/inversion.html
Azimuth: numpy matrix of azimuth values in degrees
Is this from station to earthquake or earthquake to station?
Measured: numpy matrix of corrected numerator and denominator amplitude ratio observations, needs to have two columns, one for the numerator and one for the denominator.
In the Pugh et al., 2016 paper, there is a correction (Z) based on the Vp/Vs ratio that is applied to the numerator of the ratio (equation 36). Is this what you mean by the "corrected numerator"? Or are you referring to a Q correction? Note: we are assuming a constant Vp/Vs.
Error: numpy matrix of uncertainty (standard deviation) in the amplitude ratio observations, needs to have two columns, one for the numerator and one for the denominator.
Back to equation 37 of the Pugh et al., 2016 paper, the standard deviation of the individual waves is normalized by the amplitude. Should this correction be applied for the standard deviations for the "error" input, or is this taken care of with theoretical amplitudes inside the code?
First off, the amount of documentation for MTfit is really excellent. Thank you for putting in the time. It makes the package accessible to a wider audience.
There are a couple of places where the input parameters could benefit from additional details.
On the following page: https://djpugh.github.io/MTfit/inversion.html
Is this from station to earthquake or earthquake to station?
In the Pugh et al., 2016 paper, there is a correction (Z) based on the Vp/Vs ratio that is applied to the numerator of the ratio (equation 36). Is this what you mean by the "corrected numerator"? Or are you referring to a Q correction? Note: we are assuming a constant Vp/Vs.
Back to equation 37 of the Pugh et al., 2016 paper, the standard deviation of the individual waves is normalized by the amplitude. Should this correction be applied for the standard deviations for the "error" input, or is this taken care of with theoretical amplitudes inside the code?