23 Sep 2020 |
@_slack_fatiando_UMFSBQVMG:matrix.org | In reply to@_slack_fatiando_UT0BDHKPE:matrix.org yes computation Ok, but you were just struggling with computation time, right? I mean, no crashes because of lack of memory... | 13:23:13 |
@_slack_fatiando_UT0BDHKPE:matrix.org | In reply to@_slack_fatiando_UMFSBQVMG:matrix.org Ok, but you were just struggling with computation time, right? I mean, no crashes because of lack of memory... I had an OOM erorr once, for computing terrain correction, but the grid was much bigger | 13:24:52 |
@_slack_fatiando_UMFSBQVMG:matrix.org | In reply to@_slack_fatiando_UT0BDHKPE:matrix.org I had an OOM erorr once, for computing terrain correction, but the grid was much bigger Ok! I'm pretty sure that the forward model wouldn't consume too much memory besides the already stored prisms and observation points. But if you find any memory leakage, please feel free to report it | 13:26:15 |
@_slack_fatiando_U013EU24QKH:matrix.org | Richard Scott thanks, looking forward to it... | 13:32:48 |
@_slack_fatiando_U0156QCM6AH:matrix.org | Google use the disturbance | 21:00:05 |
24 Sep 2020 |
@_slack_fatiando_UMSRSPEMA:matrix.org | Thanks for the shoutout Richard Scott đ Unfortunately that paper hasnât been accept and we kind of let it fall through the cracks of covid. Weâll definitely keep trying, though. For now, I highly recommend this book to anyone using gravity data: https://www.elsevier.com/books/understanding-the-bouguer-anomaly/pasteka/978-0-12-812913-5 | 07:29:51 |
@_slack_fatiando_UMSRSPEMA:matrix.org | To answer your question, think of the gravity disturbance as the geophysically correct âfree-air anomalyâ. So after the normal gravity correction you still need to remove the effect of topography. To do that, you have to model the effect of topography. One approximation for that is the Bouguer correction. Another would be to approximate topography with prisms and compute their effects (Lorenzo and santisoler were discussing this earlier). In the literature, the names of these corrections is slightly confusing. Some people use âterrain correctionâ to mean doing the Bouguer correction and then using a prism model to correct the flaws of the Bouguer plate approximation. Youâll also see âtopographic correctionâ used to mean just calculate the topographic effect with prisms and donât do a Bouguer correction at all (which is my preferred option). The Bouguer correction is nice and fast but if you have the software to do topographic correction then I donât see the reason to use it. | 07:34:15 |
@_slack_fatiando_UMSRSPEMA:matrix.org | The same is true for the free-air correction. In geophysics, we only used it because the Somigliana formula for normal gravity only calculates on the surface of the ellipsoid. So itâs a rough first-order approximation from the time when there was no software but excell :-) But with Boule, we have a closed-form expression for normal gravity anywhere outside the ellipsoid. So the approximation (that 0.30...h factor) is completely unnecessary. | 07:37:20 |
@_slack_fatiando_UMSRSPEMA:matrix.org | If youâre interested in our rejected paper (đ), a preprint is still available here: https://www.leouieda.com/papers/use-the-disturbance.html | 07:38:37 |
@mark-wieczorek:matrix.org | Thanks for the preprint. Geophysicists take a bit of a different perspective on this: The "anomaly" is the difference between what you measure and what you expect (at the same point), irregardless of the reference ellipsoid that is used. | 09:31:13 |
@_slack_fatiando_UT0BDHKPE:matrix.org | leouieda many thanks for your explanation. So basically you would interpret your data after (gravity.disturbance - topographic correction ) isnât? For local (small region) you should also remove the trend (applying vd.Trend to (gravity.disturbance - topographic correction ) ? | 11:48:01 |
@_slack_fatiando_UT0BDHKPE:matrix.org | (edited) ... the trend ... => ... the regional trend ... | 12:05:34 |
@_slack_fatiando_UMSRSPEMA:matrix.org | > The "anomaly" is the difference between what you measure and what you expect (at the same point),
That's the major point of confusion actually. The definition of "gravity anomaly" from geodesy is "gravity at the geoid - normal gravity at the ellipsoid". So it's at two different points and it makes no sense talking about "free-air anomaly at 250km height" since there is no "height". It's defined that way because Stoke's equation allows calculation of the geoid height from those gravity anomalies.
But geophysicists took that concept and applied it to modelling density anomalies and that's where the problem started. For modelling, it only makes sense to use gravity disturbances (defined in geodesy as "gravity at a point - normal gravity at the same point"). In fact, modern geodesy is moving away from anomalies and towards disturbances since we can also calculate the geoid from gravity disturbances and they are less complicated conceptually. | 12:38:18 |
@_slack_fatiando_UMSRSPEMA:matrix.org | In practice, it's not a huge difference and a lot of geophysicists say "gravity anomaly" but then go an calculate a gravity disturbance instead. In practice, the only difference is using heights relative to the ellipsoid (for disturbances) instead of relative to the geoid (for anomalies). | 12:39:56 |
@_slack_fatiando_UMSRSPEMA:matrix.org | (edited) ... (for anomalies). => ... (for anomalies). So if you're using GPS/GNSS heights, you are already calculating disturbances. | 12:40:34 |
@mark-wieczorek:matrix.org | Personally, I suspect that the vast majority of my colleagues don't know the difference between an anomaly and a disturbance. They all use disturbances without knowing it. Hope the paper gets accepted! | 13:26:36 |
@_slack_fatiando_UMSRSPEMA:matrix.org | Thanks! We have to work on it to expand some of the points we made.
That's my experience as well. On the teaching side of things, it's so much easier to talk about disturbances. The geoid doesn't even come up, which is often something that confuses students. | 13:30:26 |
@mark-wieczorek:matrix.org | without reading the paper... One thing that you might want to add is a quick discussion of this in spherical harmonics. If you have the potential coefficients, muliplying these by GM(degree+1)/r2 gives you the distubance, but GM(degree-1)/r2 gives you the anomaly. This is often misused, but in practice, the difference is so small that it doesn't matter... | 13:38:56 |
25 Sep 2020 |
@_slack_fatiando_U0156QCM6AH:matrix.org | The gravity books I have looked at weren't models of clarity anyway :) | 02:04:25 |
@_slack_fatiando_U013EU24QKH:matrix.org | In reply to@_slack_fatiando_UMSRSPEMA:matrix.org Thanks for the shoutout Richard Scott đ Unfortunately that paper hasnât been accept and we kind of let it fall through the cracks of covid. Weâll definitely keep trying, though. For now, I highly recommend this book to anyone using gravity data: https://www.elsevier.com/books/understanding-the-bouguer-anomaly/pasteka/978-0-12-812913-5 many thanks for the explanation sir, hope the paper accepted soon... | 09:40:57 |
1 Oct 2020 |
| @_slack_fatiando_U0156QCM6AH:matrix.org changed their profile picture. | 23:48:49 |
8 Oct 2020 |
@_slack_fatiando_U0156QCM6AH:matrix.org | In reply to@mark-wieczorek:matrix.org without reading the paper... One thing that you might want to add is a quick discussion of this in spherical harmonics. If you have the potential coefficients, muliplying these by GM(degree+1)/r2 gives you the distubance, but GM(degree-1)/r2 gives you the anomaly. This is often misused, but in practice, the difference is so small that it doesn't matter... So from things like GRACE data Mark Wieczorek? | 08:55:10 |
@mark-wieczorek:matrix.org | For any spherical harmonic model of the gravitational potential. | 08:57:14 |
9 Oct 2020 |
@_slack_fatiando_U0156QCM6AH:matrix.org | In reply to@_slack_fatiando_U0156QCM6AH:matrix.org So from things like GRACE data Mark Wieczorek? http://www2.csr.utexas.edu/grace/RL06_mascons.html | 03:03:23 |
13 Oct 2020 |
@_slack_fatiando_U0156QCM6AH:matrix.org | Redacted or Malformed Event | 23:21:46 |
@_slack_fatiando_U0156QCM6AH:matrix.org | In reply to@_slack_fatiando_U0156QCM6AH:matrix.org http://www2.csr.utexas.edu/grace/RL06_mascons.html Mark Wieczorek - degree is the degree of the expansion used in the potential coefficient calculation for a particular model? | 23:22:22 |
14 Oct 2020 |
@mark-wieczorek:matrix.org | Hi Richard Scott. Here is a summary of spherical harmonics that might be useful: https://shtools.oca.eu/shtools/public/real-spherical-harmonics.html | 08:15:39 |
@_slack_fatiando_U0156QCM6AH:matrix.org | With a library - thanks Mark. | 08:42:26 |
15 Oct 2020 |
@_slack_fatiando_U0156QCM6AH:matrix.org | Which looks really impressive! | 03:46:34 |
@_slack_fatiando_UT0BDHKPE:matrix.org | Redacted or Malformed Event | 06:30:53 |