Changes after the latest release (v2.5.0)

This is the list of changes made after the release of ASPECT version 2.5.0. All entries are signed with the names of the author.

1. Fixed: The function to determine whether a boundary is an inflow or outflow boundary (for the purpose of deciding whether to apply boundary conditions for compositional fields) did not take into account the mesh deformation. This caused incorrectly applied surface boundary conditions in models with free surface and is fixed now.
   (Anne Glerum, 2024/11/11)

2. Fixed: Several fixes for the Newton solver that would slow down convergence or rarely lead to wrong results. The Newton solver and defect correction Picard solver computed a wrong pressure scaling for the linear solver. This has been fixed, leading to a reduction in linear solver iterations. The solution variable temporarily contained the update to the solution when using the Newton solver or defect correction Picard solver, leading to bugs in models that assumed the vector contains the values. This has been fixed by always storing the values of the solution in the variable and never the update. The Eisenstat-Walker method to compute linear solver tolerances for the Newton solver would compute too coarse linear residuals leading to poor nonlinear convergence. This has been fixed.
   (Rene Gassmoeller, Menno Fraters, 2024/11/08)

3. Added: When using the Geodynamic World Builder, ASPECT will create a file called 'original.wb' in the output directory which contains the exact worldbuilder file used to run the ASPECT model.
   (Daniel Douglas, 2024/11/06)

4. New: The new input parameter "Output directory LFS stripe count" allows for configuring the ASPECT output directory for better performance on Lustre file systems.
   (Wolfgang Bangerth, 2024/10/22)

5. Changed: The time dimension of the surface condition function of the 'compute profile' adiabatic conditions plugin now respects the value of the parameter 'Use years in output instead of seconds'.
   (Ranpeng Li, 2024/10/21) *

6. Changed: The 'random uniform' particle generator no longer uses the input parameters in the subsection 'Probability density function'. Instead it uses its own subsection 'Random uniform'. Additionally, the default parameters of the 'probability density function' generator have been fixed to allow not specifying a probability density function. If no function is specified the 'probability density function' generator now behaves identically to the 'random uniform' generator.
   (Rene Gassmoeller, Qianyi Lu, 2024/10/21)

7. Removed: A number of deprecated source code functions and input parameters have been removed. In particular the deprecated option to specify individual material properties in the parameter 'Postprocess/Visualization/List of output variables' has been removed. Input files will be automatically fixed and updated by the update scripts in the directory contrib/utilities.
   (Rene Gassmoeller, 2024/10/16)

8. New: ASPECT now has a Visual Studio Code extension named ASPECT (https://marketplace.visualstudio.com/items?itemName=zhikui.vscode-aspect). This extension provides syntax highlighting and auto-completion for input parameter files, and can display the documentation of input parameters while writing parameter files in VS Code. It is now one of the recommended extensions for VS Code when using ASPECT.
   (Zhikui Guo, Timo Heister, Rene Gassmoeller, 2024/10/09)

9. Changed: The class Particle::World has been renamed to Particle::Manager. All references to world have been replaced by references to manager in the source code. The existing header file "world.h" is now deprecated and user plugins should instead include "manager.h". In input and output options we now refer to different groups of particles as "particle systems".
   (Rene Gassmoeller, 2024/10/08)

10. New: The interface for particle property plugins has been updated. The new interface can update particles cell-wise and is more efficient. In addition, the interface is more extensible for future improvements. All particle property plugins in ASPECT have been updated, user plugins will have to be updated accordingly before the support for the now deprecated old interface is dropped.
    (Rene Gassmoeller, 2024/09/25)

11. Added: A python script which uses Paraview's pvpython to extract velocities from a global model along user-specified boundaries and saves them to ASCII files which can then be applied as velocity boundary conditions in regional chunk models.
    (Daniel Douglas, 2024/09/05)

12. New: The grain size evolution functionality has been extracted from the grain size material model into the new grain size evolution reaction model. This simplifies using the grain size evolution equations in other material models.
    (Rene Gassmoeller, 2024/07/26)

13. New: The CompositeViscoPlastic rheology in ASPECT now allows the user to select isostress (Reuss) or isostrain (Voigt) averaging of viscosities for multicomponent materials via the parameter "Viscosity averaging scheme".
    (Bob Myhill, 2024/07/25)

14. Added: The functions in the CompositeViscoPlastic rheology model now have grain_size as an argument.
    (Bob Myhill, 2024/07/17)

15. Changed: The CompositeViscoPlastic rheology in ASPECT now uses a log-stress Newton scheme, which requires fewer iterations to converge.
    (Bob Myhill, 2024/07/17)

16. Changed: The CompositeViscoPlastic rheology in ASPECT is now required to use the approximate Peierls flow rheology if Peierls flow is switched on.
    (Bob Myhill, 2024/07/17)

17. Changed: The CompositeViscoPlastic rheology in ASPECT now uses the DruckerPragerPower rheology.
    (Bob Myhill, 2024/07/17)

18. New: ASPECT now has a DruckerPragerPower rheology. This rheology is based on the DruckerPrager rheology, but the strain-rate-independent yield stress is replaced with a power-law rheology. The parameters "Reference plastic strain rate" and "Plastic stress exponent" control how the yield stress varies with strain rate.
    (Bob Myhill, 2024/07/17)

19. Added: The DiffusionCreep rheology module can now calculate viscosities and strain rates with grain size as an variable argument.
    (Bob Myhill, 2024/07/04)

20. Changed: The particle interpolator plugin 'bilinear least squares' now activates a limiter by default, which prevents unwanted over- and undershoots by limiting the returned properties to the bounds of the particle properties in the cell. The limiter can be deactivated by the input parameter 'Use linear least squares limiter'.
    (Rene Gassmoeller, 2024/06/25)

21. Removed: The input parameter 'Update ghost particles' has been deprecated. Ghost particles are now always updated. Existing parameter files can be updated by the script in 'contrib/utilities/update_prm_files.sh'.
    (Rene Gassmoeller, 2024/06/25)

22. Added: The Frank Kamenetskii module now has user inputs for reference temperature and reference pressure. Adiabatic surface temperature and pressure are used as the defaults for those values.
    (Grant Block, 2024/06/24)

23. New: ASPECT now supports compositional fields with different discretizations (continuous or discontinuous) and different polynomial degrees at the same time.
    (Timo Heister, 2024/06/18)

24. Changed: The particle plugin functions declare_parameters and parse_parameters now assume that they are already in the particles subsection. User plugins will have to change accordingly.
    (Menno Fraters, 2024/06/15)

25. Changed: The operator splitting scheme can now use SUNDIALs ARKode to compute the reactions (rather than a hand-written forward Euler scheme). ARKode is the new default, but the original scheme can still be used by setting the input parameter Reaction solver type to fixed step.
    (Juliane Dannberg, 2024/06/14)

26. Added: A postprocessor 'darcy velocity' that visualizes the fluid velocity in models with a compositional field named 'porosity'. This postprocessor is independent of the method the porosity is actually advected with, i.e. it could be static, advected with the Darcy velocity, or advected according to the two-phase flow equations.
    (Daniel Douglas, 2024/06/13)

27. New: There is now a postprocessor that computes statistics about how many iterations ARKode needs to solve ODEs.
    (Juliane Dannberg, 2024/06/12)

28. Fixed: The grain size material model now computes the stress that is used to calculate the grain size reduction rate with the viscosity that has been limited between its minimum and maximum given in the input file, rather than with the unlimited viscosity. This is more consistent, because the limited viscosity is the one we use to compute the strain rate (and stress is 2*viscosity_strain_rate). This updated implementation prevents artificial grain size reduction in regions with large viscosity.
    (Juliane Dannberg, 2024/06/10)

29. New: The CPO particle property plugin can now use the Geodynamic World Builder to set initial grain-sizes and grain orientations
    (Menno Fraters, 2024/06/09)

30. Added: A new particle interpolator plugin 'distance weighted average' that computes the interpolated properties as an average of particle properties weighted by a distance function. This plugin is similarly accurate as the 'cell average' interpolator, but provides smoother results and is less susceptible to jumps in interpolated properties if particles move small distances.
    (Rene Gassmoeller, 2024/06/09)

31. Added: Functionality to limit the timestep based on the Darcy velocity.
    (Daniel Douglas, 2024/06/09)

32. New: There is now a new cookbook which models compressible mantle convection in an annulus.
    (Cedric Thieulot, 2024/06/06)

33. Changed: The grain size material model no longer has the option to advect the logarithm of the grain size. This option used to be helpful for models with large gradients in grain size, but now that ASPECT has particles it is more accurate to advect the grain size on particles instead.
    (Juliane Dannberg, 2024/06/07)

34. Changed: The grain size material model no longer has the option to scale the grain size in the lower mantle. This option used to be helpful in models where the grain size is very different between upper and lower mantle, but now that ASPECT has particles, it is more accurate to advect the grain size on particles instead.
    (Juliane Dannberg, 2024/06/07)

35. New: The CPO particle property plugin set now contains a plugin which can compute the symmetry decompositions of the elastic tensor. There are now also tests which test the whole CPO particle property plugin set.
    (Menno Fraters, 2024/06/06)

36. New: We implement an option to use the Weighted BFBT preconditioner introduced by Rudi et al (2017). This preconditioner was designed for problems involving highly heterogeneous viscosities.
    (Quang Hoang, Timo Heister, 2024/06/06)

37. New: Added a Rayleigh Taylor instability benchmark with a free surface. The benchmark was described by Kaus et al. (2010) and reproduced with ASPECT with different stabilization schemes for the free surface by Rose et al. (2017). It can be used to show that the stabilization scheme implemented by I. Rose allows for larger timesteps, and that the projection of velocities onto the surface leads to significantly different topography.
    (Anne Glerum, 2024/06/06)

38. Changed: The grain size material model now uses SUNDIALs to compute the grain size change (rather than a hand- written forward Euler scheme).
    (Juliane Dannberg, 2024/06/06)

39. New: Added fluid reaction and properties calculations to the Katz2003MantleMelting model in the ReactionModels module. These calculations were taken from MeltSimple and the model now calls Katz2003MantleMelting. Additionally, the katz 2003 option in ReactiveFluidTransport calls these calculations, allowing for melting and melt transport to be composited with any material model via ReactiveFluidTransport.
    (Grant Block, 2024/06/06)

40. New: There is now a python script in the contrib folder that creates netcdf files from an input ascii file formatted following structured ascii data file convention.
    (Arushi Saxena, 2024/06/06)

41. Changed: We no longer support "make test" to run the testsuite. Please use "ctest" directly.
    (Timo Heister, 2024/06/05)

42. New: There is now a new cookbook based on the setup of Allken et al, G3, 2012 which models rift interaction in 3D brittle-ductile coupled systems.
    (Cedric Thieulot, 2024/06/05)

43. Added: A cookbook using a model with simple 2D annulus convection to showcase the capabilities of the pyvista Python package in plotting and conducting simple mesh operations using numpy arrays. The cookbook contains a parameter file, documentation, and figures produces by manipulating data on pyvista and plotting new fields. The relevant script is located in [cookbooks/twoD_annulus_visualization/2D_annulus_example.prm].
    (Madeleine Kerr, 2024/06/05)

44. Added: Cookbook which showcases the use of the tian approximation in the Reactive Fluid Transport material model. The cookbook features a kinematically subducting slab with an initial hydration state dehydrating as it advects through a hot mantle wedge.
    (Daniel Douglas, 2024/06/05)

45. Fixed: There was a bug in spherical harmonic data reading of S40RTS and SAVANI initial temperature models when a maximum degree lower than the maximum degree of the spherical harmonic coefficients of the data file is specified by users to set up temperature field from spherical harmonic summation. This bug is now fixed.
    (Shangxin Liu, 2024/06/04)

46. Changed: In S40RTS and SAVANI initial temperature models, the input parameter, 'Specify a lower maximum order' is changed to 'Specify a lower maximum degree', and the input parameter, 'Maximum order' is changed to 'Maximum degree'. This avoids confusion because the function of these two parameters is to allow users to specify a maximum degree lower than the maximum degree of the spherical harmonic data file to calculate the temperature field.
    (Shangxin Liu, 2024/06/04)

47. Changed: The QT based Parameter GUI has been removed from the ASPECT repository and the documentation as it was no longer maintained and no longer working correctly. Instead, use the parameter list linked from the website or edit directly in Visual Studio Code with the ASPECT plugin.
    (Timo Heister, 2024/06/02)

48. Added: The Function expression plug-in for the Gravity model can now evaluate functions in multiple coordinate systems. The plug-in was amended to read in entries to the "Coordinate system" parameter in the Function subsection and Gravity Model section of the parameter file. There are three added tests for the Gravity model function plug-in to ensure the cartesian, spherical, and depth coordinate systems evaluate functions as expected in a 2D cylindrical annulus.
    (Madeleine Kerr, 2024/06/02)

49. Added: The heating function plugin now supports input in different coordinate systems (cartesian, spherical, depth).
    (Rene Gassmoeller, 2024/06/02)

50. New: Benchmark from Gassmoller et al. 2018 for particle integration schemes.
    (Gabriel Johnston, 2024/06/01)

51. Improved: The limiter for the DG solution can now be selected for each compositional field individually, instead of using a single setting that applies to all compositional fields. If only a single setting is found it will still apply to all fields.
    (Rene Gassmoeller, 2024/06/01)

52. Added: Functionality for multiplying the dislocation and/or diffusion creep viscosities by prefactors which depend on the model composition. Created a specific use-case for multiplying the viscosity by a factor that accounts for bound H2O from Hirth & Kohlstaedt 2004 10.1029/138GM06
    (Daniel Douglas, 2024/06/01)

53. New: The grain size material model now has the option to apply plastic yielding (Drucker-Prager yield criterion) to limit the viscosity. This can be switched on through a new input parameter. At the moment, grain size evolution is not affected by plastic yielding.
    (Juliane Dannberg, 2024/06/01)

54. Changed: Melt fraction calculated by the Katz2003 model in the ReactionModels module was added to the reactive fluid transport model. A test was added compositing the reactive fluid transport and visco plastic models, benchmarking the melt fraction results with Katz et. al. 2003.
    (Grant Block, 2024/06/01)

55. New: ASPECT now outputs the physical units of quantities into .pvtu files.
    (Wolfgang Bangerth, 2024/06/01)

56. Changed: ASPECT's default Stokes preconditioner has been changed from a block AMG preconditioner to the geometric multigrid (GMG) preconditioner described in Clevenger and Heister, 2021 (https://doi.org/10.1002/nla.2375). The GMG preconditioner generally performs better, but requires that the viscosity is averaged using one of the material model averaging functions. If models use features that are not supported by the GMG preconditioner and the GMG preconditioner has not been explicitly set ASPECT will fall back to the AMG preconditioner.
    (Rene Gassmoeller, 2024/06/09)

57. Changed: The grain size material model now uses the phase function class to compute where phase transitions occur. This means that phase transition temperature, depth and Clapeyron slope now have to be specified in a different format in the input file (rather than a comma-separated list, they can now be specified as key:value1|value2|...). This is in line with other material models. Note that the width of phase transitions is still zero in the grain size model.
    (Juliane Dannberg, 2024/05/31)

58. Changed: The reset of the grain size at phase transitions is now determined only by the velocity of the material moving through a transition rather than by a somewhat arbitrary 'transition width' parameter.
    (Juliane Dannberg, 2024/05/30)

59. New: ASPECT now has a DiffusionDislocation rheology. This rheology is identical to the one that was originally in the Material Model of the same name.
    (Bob Myhill, 2024/05/29)

60. Changed: A pressure term was added to the Frank Kamenetskii viscous flow law in the Visco Plastic material model. The new FK function is given by viscosity = A * exp(E * 0.5 * (1.0-(T/ref_T)) + F * (P-ref_P)/(ref_rho*g*h)), where F * (P-ref_P)/(ref_rho*g*h) are the new terms added here.
    (Grant Block, 2024/05/29)

61. Changed: The A block solver in the expensive Stokes iterations has been changed for free surface models from a CG solver to a BiCGStab solver. This should improve the stability of the inner (top left) preconditioner in free surface models. The new solver can also be forced by an input parameter, which can be a useful option for models that crash when solving the top left (A) block of the Stokes system.
    (Rene Gassmoeller, 2024/03/31)

62. Changed: The solver used during the expensive iterations of the AMG preconditioner has been changed from the trilinos implementation to the one provided by deal.II. The new solver should be faster and more stable, but can produce marginally different results.
    (Rene Gassmoeller, 2024/03/26)

63. New: It is now possible to use the "initial topography" plugin system for the spherical shell geometry, including using the ascii data system that can then be used to import a text description of the initial topography. As an example, we also include the file data/geometry-model/initial-topography-model/ascii-data/global_1deg.txt.gz that describes the topography of Earth on a (relatively coarse) 1 degree mesh.
    (Wolfgang Bangerth, 2024/03/21)

64. Changed: The postprocessor 'heat flux map' now outputs the heat flux values in the same format as the 'dynamic topography' postprocessor does for dynamic topography. This in particular means instead of writing all values into a single file 'heat flux map' now writes one file per boundary. Both postprocessors now take care to not include spaces in the data file headers so that separating the header columns by spaces works correctly.
    (Rene Gassmoeller, 2024/03/15)

65. Removed: The outdated input option for the geoid postprocessor named 'Include the contributon from dynamic topography' has been removed. Use 'Include surface topography contribution' or 'Include CMB topography contribution' instead.
    (Rene Gassmoeller, 2024/03/15)

66. Implemented a new fluid-rock interaction scheme for the reactive fluid transport material model based on a parametrization by Tian et al., 2019 G3, https://doi.org/10.1029/2019GC008488.
    (Daniel Douglas, 2024/02/24)

67. Incompatible: When using the "AsciiDataBoundary" class to read initial topographies for the chunk geometry, the input file needed to provide topographies spaced on a grid in latitude and longitude where the longitude was to be provided from -pi (180 degrees west) to +pi (180 degrees east). This is contrary to the rest of ASPECT, which generally uses 0 to 2*pi (0 to 360 degrees east) as a convention. This has now been changed: The initial topography plugin will be asked about topographies in lat-long with longitudes between 0 and 2*pi. This also means that input files have to be changed correspondingly.
    (Wolfgang Bangerth, 2024/02/16)

68. New: There is now a new cookbook that demonstrates how to use the Geodynamic World Builder to set up initial conditions for a mid-ocean ridge model with a transform fault following the setup of Behn et al., 2007.
    (Juliane Dannberg, 2024/02/13)

69. New: There is now a new visualization postprocessor called "surface elevation" that can be used to plot the initial topography or the effects of a free surface.
    (Wolfgang Bangerth, 2024/02/04)

70. New: Solution variables can now also be outputted on the surface mesh of the model domain.
    (Anne Glerum, 2024/05/30)

71. Improved: The geometry function point_is_in_domain now also works for meshes that include initial topography and/or mesh deformation for box and chunk geometries.
    (Anne Glerum, 2023/12/14)

72. Removed: The GeometryModel::Chunk and GeometryModel::TwoMergedChunks classes had member functions depth_wrt_topo(), but these were not used anywhere and did not work with the mesh deformation framework. They have consequently been removed.
    (Wolfgang Bangerth, 2023/12/07)

73. Changed: The interface Postprocess::VisualizationPostprocessors::CellDataVectorCreator::execute() used to return an object of type std::pair<std::string, Vector<float>*> where the second part is just a raw pointer. This has been changed (incompatibly) to std::pair<std::string, std::unique_ptr<Vector<float>>> to avoid the use of raw pointers and ensure that memory de-allocation happens automatically.
    (Wolfgang Bangerth, 2023/11/28)

74. Fixed: The stress and shear stress visualization postprocessors now output the correct stresses when elasticity is enabled.
    (Bob Myhill, Rebecca Fildes, 2023/11/10)

75. Fixed: The 'principal stress' postprocessor used the wrong sign for stresses in models with elastic deformation, leading to principal stress directions that were rotated by 90 degrees. This is fixed now.
    (Rene Gassmoeller, Rebecca Fildes, 2023/11/09)

76. Changed: ASPECT now considers boundaries with no normal flow as boundaries with inflow for the purposes of the parameter 'Allow fixed temperature on outflow boundaries' and the corresponding parameter for composition. This was the default behavior up to ASPECT 2.4.0. This behavior was changed in ASPECT 2.5.0, in which boundaries with no normal flow are treated like outflow boundaries. The new behavior caused unintended side effects, therefore it is reverted back to the original behavior. The reason for the initial change was a bugfix for boundary conditions in the first timestep that is now implemented differently.
    (Rene Gassmoeller, 2023/11/06)

77. Fixed: The RK4 particle interpolation scheme computed wrong particle locations for particles that crossed periodic boundaries in a box geometry. This is fixed now.
    (Rene Gassmoeller, 2023/11/03)

78. New: There is now a new cookbook that demonstrates how to use the grain size material model and how to choose the particle parameters in a model with a complex nonlinear rheology.
    (Juliane Dannberg, 2023/10/07)

79. New: The amount of shear heating in a computation can now be limited by setting a maximum stress to be used, based on a Drucker-Prager yield stress with a user- specified cohesion and friction angle. This is useful in models with unrealistically high stresses.
    (Juliane Dannberg, 2023/10/01)

80. Added: There is now a prescribed viscosity material model plugin, which can overwrite the viscosity of a specified location with a viscosity prescribed by a function.
    (Menno Fraters, 2023/09/27)

81. Changed: ASPECT has been renamed from "Advanced Solver for Problems in Earth's ConvecTion" to "Advanced Solver for Planetary Evolution, Convection, and Tectonics" to reflect that the scope of ASPECT has grown beyond mantle convection.
    (Timo Heister, 2023/09/17)

82. New: There is now an entropy table lookup model in the initial composition model. This model takes an initial temperature field and converts it to an initial entropy field using a lookup table.
    (Haoyuan Li, 2023/08/11)

83. Changed: The input subsection 'Particles' has been moved from the 'Postprocess' subsection into its own subsection at the top level. This is to make it more clear that particles are not simply a postprocessing feature. All included input files have been adjusted and the reformatting script has been adjusted. Users will have to apply the reformatting script to their input files manually.
    (Rene Gassmoeller, 2023/09/10)

84. Changed: ASPECT now requires CMake 3.13.4 (just like deal.II v9.5.0) and no longer supports GeodynamicWorldBuilder older than 0.5.0.
    (Rene Gassmoeller, 2023/09/09)

85. Changed: The benchmarks 'blankenbach' and 'king2dcompressible' have been reworked and extended. Changes include: (1) higher resolution, (2) more reproducible convergence results, (3) additional figures, (4) a comparison between gradient based and CBF heat flux calculation, (5) reproducible Richardson extrapolation of ASPECT reference results.
    (Rene Gassmoeller, 2023/08/29)

86. New: There is now a new viscosity profile data file that is more consistent with how the profile is computed in the original Steinberger and Calderwood (2006) paper.
    (Juliane Dannberg, Bernhard Steinberger, 2023/08/22)

87. New: The steinberger material model now allows the viscosities of different materials to differ from the reference viscosity by a constant factor via the parameter Composition viscosity prefactors.
    (Poulami Roy and Bob Myhill, 2023/08/14)

88. Improved: The 'spherical constant' boundary composition and boundary temperature plugins have been updated. In particular the boundary composition plugin can now prescribe different values for different compositional fields.
    (Rene Gassmoeller, 2023/08/02)

89. Improved: The iteration scheme in the diffusion dislocation material model now uses the derivative of the logarithm of the strain rate to the logarithm of the stress. This improves efficiency and stability of the iterative scheme. The parameter "Strain rate residual tolerance" now corresponds to the residual for the logarithm of the strain rate (which close to the solution corresponds to the relative residual).
    (Haoyuan Li and Bob Myhill, 2023/07/18)

90. New: There is now a strain rate tensor postprocessor that is only outputted on the surface of the model domain.
    (Anne Glerum, 2023/07/17)

91. Added: A sea level postprocessor that computes the sea level for glacial isostatic adjustment modeling. It computes the sea level based on the free surface topography, ocean basin, ice melt from boundary traction, and perturbed gravitational potential of the Earth model from the geoid postprocessor.
    (Maaike Weerdesteijn, 2023/07/14)

92. New: There is now a new material model ('reactive fluid transport') that is designed to advect fluids and compute fluid release and absorption based on different models for fluid-rock interaction. The properties of the solid are taken from a base material model.
    (John Naliboff, 2023/07/14)

93. Fixed: The strain rheology now uses the correct 'old strain' when computing the reaction term updates.
    (John Naliboff 2023/07/14)

94. New: There is now a new cookbook which models the (Poiseuille) flow of the lower crust around a rigid obstacle.
    (Cedric Thieulot, 2023/07/13)

95. Improved: The initial lithostatic pressure plugin for boundary tractions now includes the initial topography of the reference point into its pressure profile instead of the maximum topography within the domain.
    (Anne Glerum, 2023/07/13)

96. Changed: Boundary traction models are now organized in a manager class that allows to assign multiple boundary traction plugins to each boundary. Existing user plugins that use the function this->get_boundary_traction() will need to be modified to use this->get_boundary_traction_manager().
    (Chameera Silva, Rene Gassmoeller, 2023/07/13)

97. New: Added a cookbook on how to set up global instantaneous models based on recent geophysical constraints with a heterogeneous density and viscosity distribution and weak plate boundaries prescribed using different plate boundary configurations.
    (Arushi Saxena, Juliane Dannberg, and Rene Gassmoeller 2023/07/13)

98. New: There is now a new cookbook which is a very simple subduction initiation model as published by Matsumoto and Tomoda (1983).
    (Cedric Thieulot, 2023/07/12)

99. Changed: The entropy plugin now includes a plasticity rheology, which can be set by cohesion and frictional angle. It can also use a viscosity prefactor profile, set the lateral viscosity variation limit, and choose p-T dependent conductivity

    
    (Ranpeng Li & Juliane Dannberg, 2023/07/13)

100. Changed: Material models are now always provided with the current strain rate. Checks if the strain rate is provided are not longer necessary. MaterialModelInputs can no longer be initialized without a strain rate.
     (Rene Gassmoeller, 2023/07/12)

101. New: A cookbook which replicates the model setup of the van Keken et al., 2008 2D corner flow subduction benchmark.
     (Daniel Douglas, Cedric Thieulot, Wolfgang Bangerth, Max Rudolph, 2023/07/12)

102. New: There is now a new cookbook which models the deformation of elliptical and rectangular inclusions in simple shear and pure shear.
     (Cedric Thieulot, 2023/07/11)

103. Fixed: When using the 'artificial viscosity' visualization postprocessor in parallel, we used to put NaNs into the output vector (for ghost and artificial cells) that then triggered floating point exceptions. This is now fixed.
     (Wolfgang Bangerth, 2023/07/10)

104. New: The base code for the computation and output of Crystal Preferred Orientation (CPO) has been added to ASPECT.
     (Menno Fraters, 2023/07/09)

105. Changed: The input parameter 'Include viscoelasticity' has been removed from the visco_plastic material model. When 'Enable elasticity' in the Formulation subsection is switched on, elasticity will automatically be included in the visco_plastic material model.
     (Anne Glerum, 2023/07/09)

106. New: ASPECT now allows it to select a different list of assemblers for each advection field.
     (Juliane Dannberg, 2023/07/09)

107. Improved: The iteration scheme in the Peierls creep rheology now uses the derivative of the logarithm of the strain rate to the logarithm of the stress.
     (Haoyuan Li and Bob Myhill, 2023/07/08)

108. New: The cutoff stress for the Peierls stress could now be applied strictly as the lower limit of the stress in the Peierls rheology
     (Haoyuan Li, Bob Myhill, 2023/07/08)

109. Fixed: The computation of the upper box origin in the 'box with lithosphere boundary indicators' geometry model did not take into account a nonzero origin. This is fixed now.
     (Juliane Dannberg, 2023/07/08)

110. Changed: Default field types are now defined by ASPECT based on Field Names if the types are not defined by the user. The logic is as follows: if the name contains the substring stress, it has type stress. If the name is equal to s11, s12, s22, s13, s23, s33, it has type strain. If the name is equal to grain_size, porosity, density_field, or entropy it has the types grain size, porosity, density or entropy respectively. Otherwise it has the type chemical composition.
     (Bob Myhill, 2023/07/08)

111. New: There is now a new additional benchmark called solubility that demonstrates the mass of water is conserved as water is released, migrates and is being reabsorbed.
     (Juliane Dannberg, 2023/07/08)

112. Improved: ASPECT is now by default compiled in DebugRelease mode which compiles both a debug and a release (optimized) version of the executable. Individual build types can still be selected using 'make debug' or 'make release' or by setting the cmake variable CMAKE_BUILD_TYPE to Debug or Release.
     (Rene Gassmoeller, Timo Heister, 2023/07/07)

113. New: The maximum relative increase in the time step length is now bounded by a factor of 1.91, where before it was unbounded. This value is obtained from theoretical considerations about parabolic problems.
     (Laila Busaleh, 2022/09/15)

114. New: Mesh deformation plugin that uses the landscape evolution code FastScape to deform the surface through erosion and sediment deposition.

     Citations: Neuharth, D., Brune, S., Wrona, T., Glerum, A., Braun, J., & Yuan, X. (2022). Evolution of rift systems and their fault networks in response to surface processes. Tectonics, 41(3), e2021TC007166.

     Neuharth, D., Brune, S., Glerum, A., Morley, C. K., Yuan, X., & Braun, J. (2022). Flexural strike-slip basins. Geology, 50(3), 361-365.

     
     (Derek Neuharth, Anne Glerum, Sascha Brune, Esther Heckenbach 2023/03/01)

115. New: There is now a new initial temperature plugin that adds a fixed number of Gaussian perturbations placed at random locations to the initial temperature.
     (Juliane Dannberg, 2023/02/10)

116. New: The 'grain size' material model now supports the grain size evolution equations of Mulyukova & Bercovici 2018 in addition to the existing formulations. For this purpose the old input parameter 'Use paleowattmeter' has been deprecated and replaced with 'Grain size evolution formulation'. The default behavior of the material model has not changed. Benchmarks and tests for the new feature have been added.
     (Arushi Saxena, Ranpeng Li, Juliane Dannberg, Rene Gassmoeller, 2023/08/23)