Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015...

27
Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

Transcript of Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015...

Page 1: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

Trivial applications of NeoKinemawhich illustrate its algorithm

by Peter BirdUCLA

2002/2009/2015

Support from NSF & USGS are gratefully acknowledged.

Page 2: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

Many grids used for testing overlap the pole and the international date line.Thus, they demonstrate nicely that there are no singularities, and that the spherical algebra was done correctly.

Page 3: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST01: No input data, zero velocity at two nodes → static solution.

Page 4: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST02: No input data, one fixed node, and one boundary node that rotatesaround it.

Result is rigid-plate rotation, at rates on the order of 7×10-16{radian}/s.

Page 5: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

Strain-rates in earlyversions of NK:

Maximumstrain-rate of 210-20 /s =

0.3% in 4 Ga.

Page 6: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

Now, using NKv4.1: Maximum

strain-rate becomes 810-23 /s

[7 orders smaller thanthe rotation rate]

= 0.001% in 4 Ga.

Page 7: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST03: No input data, uniform extension driven by boundary conditions.

Page 8: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

Note: In absence ofdata, lithospherebehaves as a uniformviscous sheet.Therefore, in uniformstress field far from BCs, it undergoes equal vertical andhorizontal shortening.

Page 9: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST04:Data on the azimuth of themost-compressive horizontal principal stress are given:

Page 10: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

and NeoKinema interpolates these directions by the algorithm of Bird & Li [1996]:

N.B. This interpolationwas done with theindependent-data variantof the method ofBird & Li [1996].The alternate clustered-datamethod would infer largeruncertainties in the results.

Page 11: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

When the same velocity boundary conditions are given (as in Test03), NeoKinema attempts to find a velocity solution that will honor these interpolated directions:

Page 12: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

The new solution hasnonuniform strain-rates(large where stress directions are compatible with the velocity BCs; small elsewhere):

Page 13: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

and here is the (mis)match between the principal strain-rate azimuths of the solution and the target azimuths derived from the stress data:(Note that many of the red target azimuths are hidden by the yellow actual azimuths.)

Page 14: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST05:Three faults of unknown slip-rate make up a plate boundary system. Each dip-slip fault has assumed rake of 90°.(Same velocity BCs as in Test03; however, no stress-direction data as in Test04.)

Page 15: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

NeoKinema finds a solution with most of the deformationassigned to slip on the faults:

Page 16: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

However, the requirement of purely dip-slip faulting on the twonormal faults requires some significant continuum strain-rates:

Page 17: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST06:The same 3 faults of unknown slip-rate make up a plate boundary system. But now, each dip-slip fault has assumed rake of 90°20° ().(Same velocity BCs as in Tests03~05.)

Page 18: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

NeoKinema finds a solution with virtually all of the deformationassigned to slip on the faults, and continuum strain-rates are smaller:

Note oblique slip on these faults.

Page 19: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST08:A strike-slip fault whosetrace is a small circleis entered with unknownslip rate.(Note that it is notnecessary to outlinefault zones with slenderelements, althoughone may choose todo so.)

Page 20: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

When the solution is driven indirectly by velocity at one node,the solution is Eulerian plate tectonics, with minimal strain-rates:

Page 21: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST09:Example of a conveniencefeature, the type-4 boundarycondition, which allowsboundary nodes to be assignedto a major plate by simplygiving its abbreviation(e.g., “NA”, “PA”);the necessary velocity iscalculated within NeoKinemaby the Euler formula.

(Note: Lacking any data,such as fault locations,the program finds a uniform-viscous-sheet solution to thisproblem.)

PA

NA

Page 22: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST10:Synthetic GPS velocities,which are consistent withuniform plate rotation,are input at many internalpoints.Model boundaries are free,except at 2 boundarynodes which are fixed:

Page 23: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST11:Same as Test10,except that now thevelocity reference frameof the GPS data istreated as unknown, or free-floating.

The result is that the velocity reference frameis determined by the 2fixed boundary nodes,and all motion is reducedto less than 0.0004 mm/a.

Page 24: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST13: Conversion of input short-term interseismic GPS velocities (left)to long-term corrected velocities (right), along a strike-slip faultwhich is temporarily locked down to 100 km depth.(In this test, velocity BCs “enforce” the right plate-motion solution.) INPUT: OUTPUT:

Page 25: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST14: Conversion of input short-term interseismic GPS velocities (left)to long-term corrected velocities (right), along a strike-slip faultwhich is temporarily locked down to 100 km depth.(In this test, the southwestern plate is free, and GPS data determines its velocity.) INPUT: OUTPUT:

Page 26: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

TEST15: Conversion of input short-term GPS velocities (left)to long-term corrected velocities (right), along the Cascadia subduction thrust,which is temporarily locked from 10 km to 40 km depths.(Note that long-term relative velocities within NA are less than short-term.) INPUT: OUTPUT:

Page 27: Trivial applications of NeoKinema which illustrate its algorithm by Peter Bird UCLA 2002/2009/2015 Support from NSF & USGS are gratefully acknowledged.

(end)