Using own dataset for tomography

[pratik-2227]

while using other dataset for tomography, how to decide the value of n_rpt parameters.
also getting some error which i am unable to understand, attaching the image blow,
Please if anybody has run their dataset guide me

[JingChen-Thu]

Hello @pratik-2227 ,

n_rpt is the number of grid node in depth, longitude, and latitude directions. In local earthquake tomography, typically we can set grid spacing to be 1 km/0.05 degree/0.05 degree for preliminary tests and 0.5 km/0.02 degree/0.02 degree for the final imaging task. In addition, the setting also depends on your computational resources.

In you case, the study region in horizontal direction spans 1 degree by 1.55 degree. Since the region is not large, we can set n_rpt to be [N, 51, 79]. I do not get the depth range in the attached figure. I guess (and recommend) the depth range is [-10, 35], so N can be set as 46.

When you set n_rtp = [46, 51, 79], then the grid spacing is:

1. depth: (35 - (-10)) / (46 - 1) = 1 km
2. latitude: (40 - 39) / (51 - 1) = 0.02 degree
3. longitude (73 - 71.45) / (79 - 1) approximately = 0.02 degree.

The above logic for n_rtp is just for your reference.

As for the problem you met, the error possibly arises from the dismatch of n_rpt and the input model file (e.g. "model.h5"). Be sure the shape of the array in the model file is the same as n_rpt, so that TomoATT can correctly read the model file.

You can easily check the model file using the command "h5ls model_name":

If it is not the reason, maybe we can further check the input files in detail.

If you encounter any problem about TomoATT, please feel free to @JingChen-Thu for further guidance.