Dear Rexha,

a question similar to yours was already asked in the past, you can look at:

https://www.researchgate.net/post/Can_we_analyse_focal_mechanism_for_borehole_microseismic_data_Is_anyone_ever_try_that

proposed by Massita Cindy

Regards,

Lorenzo

 There must be something I don't understand.  How are you locating the micro-earthquakes if you only have a single vertical array of sensors?  If that is all you have, then strictly speaking you cannot really locate the events nor determine their focal mechanisms.

You may need a basic seismology book.  If you cannot obtain one, you could check    https://sites.google.com/site/themetalibrary/library-genesis

At the sites they point to, there are some basic older books available that would serve you well.

Dear Lorenzo

Oh there is someone who already asked it. Thank you for your information. I will check all of the references. As further discussion I read, it is written that it is possible, but unfortunately the the results will be very poor. Do you have any suggestion about this problem? Installing at least 2 boreholes seimometer isn't really an option for me actually.

Regards,

Dear Allan Lindh

I am sorry, you may be misunderstand "single vertical array of sensors" with "one monitoring seismometer". Actually, I have 8 seismometers which is deployed in single vertical aray. And for locating the events and reducing ambiguity of misfit, I used polarization analyses to calculate the back azimuth, then used Grid Search Optimization in this back azimuth area.

Regards,

Dear Rezha

My only thought is if you have a location for what I suppose is a cluster, is there any chance of installing 3 surface stations in a triangle centered on the cluster, with the dimensions of the triangle approximately the same as the estimated depths of the cluster?  Even for a short period of time, 3 additional stations would clarify your locations, particularly the depths.  And while 4 stations, 3 + borehole, will not strictly speaking produce a focal mech for individual events, if the event locations vary, a composite mechanism might be possible.  By itself, I cannot see that a vertical string alone will produce much constraint on focal mechanisms, unless by luck a nodal plane passes through your vertical array, which you would see as a change in first motion along the string.  If you have very many events, and had many events whose nodal plane cut the array, I suppose one might try to tease a composite mech out of that data, although not sure

If you record good onscale P&S amplitudes, and they vary coherently along the string, there is information there on the focal mech.  Enough events with clear variation, and if that variation varied with event location, a joint inversion for a composite mech might be possible, if one assumed all events had the same mechanism.  Makes my head hurt just trying to think about that however.

Good luck

Dear Allan Lindh

Yes I agree that installing surface seismometer will be very helpful. It will be a suggestion for next study, but unfortunately for now I have to deal with our available data first (which is recorded using single array vertical borehole).

I thing it is a great idea to observe all changes in first motion along the string. I am still working on composite mechanism, but like you said, it is hard.

But anyway, thank you so much for your suggestion, I will try my best.

Regards, 

Dear Rexha

Do you mind if I ask how far the cluster of seismicity is from your borehole array?

And the depth of the seismicity, and the depth of your string?

Dear Rexha,

If the data you have are already acquired by a single array in vertical borehole I agree with Mr Lindh that it is a hard story. You can try what discussed by Nolen-Hoeksema&Ruff, but I don't remeber the details.

Also for localisation you must be careful because the results are reliable only at the depths spanned by the array and not too far fromt it horizontally.

Regards,

Lorenzo

Dear Allan

I am sorry for this late reply. The distance of the event's cluster to the seismometer is about 300 meters. The depth of the borehole seismometer is from 500 until 700 BSL. And the depth of cluster is about 575 km.

If you do not mind, you can find the pictures in attachment below. Thank you for your consideration.

Regards,

Dear Rexha,

thank you for the very interesting info. To me the acquisition geometry in comparison with the posistion of the microseismic cluster seem ok, what do you think Allan?

In your slide 4 (map view) it is clear the influence of the array as you can see the circular like distribution of events araound the wellhead, which is the typical result of a vertical array.

About focal mechanism, what I can suggest at the moment is to try to invert using the double couple assumtpion to reduce the complexity of the problem.

I let you know if I have some other idea.

Regards,

Lorenzo

Yes I agree with Lorenzo.  If you have information about the source mechanism, it will be in the amplitudes of the P & S arrivals.

Being a simple minded old fellow, I would do the simplest things, which you have probably already done.

If you plot a record section of individual events, can you see the moveout of the P wave arrival?  This should tell you the approximate depth of the event.

Can you see the moveout of the SV apprival?  (I assume you have single component siesmometers?)  If so, this should give you a distance from the well estimate.

If you have usable P&S arrivals, then for onscale events you could measure the P and SV amplitudes, take their ratio, and plot as a function of depth in the array.  (Might need to filter the records to reduce the hi-freq noise.)  If you have information about the mechanism, it should show up as a coherent pattern in that plot of amplitude ratios.

It is an interesting problem.  Good luck.

Dear Lorenzo

Thank you for your suggestion. I am still working on it.

Dear Allan

Yes , I can see the moveout of P wave arrival and SV arrival. Thank you for your suggestion. I think it is a really good idea to simplify the problem like Lorenzo and you said, and  look for the ratio of P and SV.

Regards,

Dear Rexha,

the suggestion from Allan made me remember what done in:

Rutledge, J. T. & Phillips, W. S. Hydraulic stimulation of natural fractures as revealed by induced microearthquakes, Carthage Cotton Valley gas field, east Texas. Geophysics 68, 441–452 (2003).

I gave a fast read to the paper and I guess Allan's suggestion is on the right way.

I would add that you could try to group the events in diffferent families on the basis of the waveform/polarity of P and S waves. Then, if you have some sw tool to do it, you could perform repeated modelling of different fault geometries to find out which best approaches your waveforms. That is, you try to find the solution by solving many time the direct problem instead of solving the inversion. It is a brute force approach, not very elegant, but sometimes it works.

Regards,

Lorenzo