Aspect Geoid Test

ASPECT
1

In my current work, I am using geoid to evaluate time-dependent mantle convection models. The models are compressible.

I found that the geoid pattern and amplitude are reproduced reasonably well when I use the Simple Compressible material model. However, in this material model, the viscosity does not depend on temperature. When I use the Steinberger material model with a similar viscosity profile, the geoid pattern becomes different (without temperature-dependent viscosity). If I also include temperature-dependent viscosity, the geoid and dynamic topography amplitudes become unrealistically large, and the geoid pattern becomes worse.

I previously tested incompressible mantle models and did not see this problem. It only appears when I use compressible mantle models. I also tested the S40RTS tomography model in ASPECT and found the same problem.

The results are compiled in the pdf file.

Aspect_geoid_test.pdf (997.6 KB)

2

@DebanjanPal1995 I have no experience with these things, but if I understand right, then by “geoid” you mean the gravity anomaly not the dynamic topography, right? If so, gravity is caused by density, and to debug what happens with the gravity you might want to look at the density variations in your model. If you get geoids that are 20x as large as expected, then you must have density variations that are 20x as large as expected.

Best
WB

3

Thanks for the reply @bangerth.

The geoid is the contribution from density anomalies within the mantle and associated dynamic topographies at the surface and CMB, which can give rise to mass excesses or mass deficits. Therefore, dynamic topography is also an integral component of the geoid (Richards and Hager, 1989). I have also plotted the dynamic topography to show how temperature-dependent viscosity in the Steinberger material model strongly amplifies dynamic topography, which appears unrealistic. I wonder what might be causing this unrealistically large topography.

Aspect_forum_DT.pdf (253.3 KB)

4

@DebanjanPal1995 So if I understand you right, then the unrealistic size of the geoid is caused by the unrealistic dynamic topography in your model? If so, then that’s where you should keep poking. Unrealistic dynamic topography is typically caused by unrealistically large velocities, for example.

Best
WB