choosing the correct increment while gridding data
I am currently gridding data using the command "surface" and pre-processed the data using the command "blockmean".
While going through these steps, I was wondering what was the best strategy to adopt while choosing the increment (-I) of the "blockmean" and "surface" commands?
Indeed, I am working on seismic reflection profiles and the data I want to grid correspond to horizons that have been picked on seismic lines that are 400 to 800 m apart.
Should the increment of the "blockmean" and "surface" commands be the same and, at least, of 800 m? Since the larger distance between the profile is 800 m ? So an increment of about -I0.007 (degree) since the data are in degree decimal? Or can it be independent and in that case, which value to use that will make sense too? I am a bit confused.
The study area is also relatively small (0.1 degree of difference in latitude and 0.3 degree of difference in longitude), so the increment should not be too big, but how small can it be so that the interpolation will still mean something?
Also, could you please confirm me that if I use an increment of -I0.01 for another (larger) study area this corresponds well to an increment of about 1 km? And in that case, would it make sense to "blockmean" and "surface" the data with such an increment knowing that the profiles are 8 to 10km apart?
Thanks for your help!
We typically use the same increment for blockmean and surface. While you could use a finer interval for surface you would never do the opposite (aliasing). You say the largest distance between profile is 800 m but what is important is the shortest distances as well as the sampling interval along your seismic lines. Also, things like 0.007 as an interval is a bit jarring and is unlikely to work well with your boundary limits. I would choose an interval in arc seconds, e.g., -I15s or -I30s instead. You may have to experiment with the spacing. one arc second is ~30 meters depending on your latitude.