how to do a best well tie?

this was a question from brother zhuhuan

i think i may help alittle about this subject by adding some tips.

well to seismic tie is the first step in seismic interpretation , any seismic interpretor should start from this point ... why do we need well to seismic tie ?
- to check the TD relation at each well.
- to understand the seismic responce of the main interesting geological layers ( targets , geological regional markers , seismic regional markers ,.. etc.)
- to understand the relation and difference between geological markers and seismic image (due to seismic defects , multiples, low resolution, noise, difference in velocity ...etc.)
- the regional change of all previously mentioned from well to another and from area to another...

from all of this .... we must understand that seismic to well tie or synthetic seismogram is not just a mathematical operation generate a figure to put in your report ,,,, it's important step which you should care about.

how to do the best well to seismic tie?

it's a process of well to seismic tie not seismic to well tie , so what we need to do is to convert the well to a seismic trace looks like the recorded seismic trace at the place of the well , in other words , we need to generate seismic trace from the well data , and play with them untill reaching the most close semilarity with the real seismic trace represent the well ... as following steps :
1- prepare the needed well logs for synthetic seismogram ( density , Vp, Vs, resistivity )
2- the most important is Vp .... then density log , Vs is important for offset calculation , R is helpful in recognizing gas effict.
3- if Vs is not available, you may claculate from Vp , the same if you don't have density , but this will effect your results.
4- QC all these logs , together with lithology information and GR , check logs defects , the unlogic values , spikes and borehole conditions and solve all problems before starting generation.
5- the frequency of logs is critical point .... seismic frequency is too far from the logs frequency , so it would be realistic to filter your logs, but don't forget that very high contrast thin layers would be resolvable.
6- many softwares has modules for synthetic seismogram ... but i prefere Hampson Russel , i feel it is specialzed in this subject , so i advice.
7- the seismic trace you have is originally multi offsets traces stacked together , so zero offset synthetic method which is the most common will not be realistic .. here is the strong point of hampson russel .
8- you need to know your seismic parameters (cable length , frequncy content , amplitude content ranges )
9- the optimum workflow of generating synthetic seisogram by hampson russel software is fully described in the software manual .. so no need to repeat these steps here , simply read and follow these steps.
10- extracting the wavelt and adjusting the phase are the most important steps you need to care about.


i am not sure if i mentioned every thing, i wrote just this subject from memory , so may be i will edit in the future , if any one will find a mistake or wrong information i will like to be informed.
thank you fore reading and i am ready for any question

note : read carefully the manuals of hampson russel (avo_guide, checkshot , RWtheory and elog) , you will find good details about options and techniques , will be useful even if you will use another program.