Exact advection with velocity flow field

In this release we still use the exact velocity to propagate the parcels. In later releases the velocity field is computed from the vorticity.

What's Changed

• RK4 and README by @sjboeing in #1
• Added optimised merging code by @sjboeing in #2
• Changing order of compilation (needed for compilation on my laptop) by @sjboeing in #3
• Fix 2d elliptic model by @matt-frey in #8
• Default choices for parameters by @matt-frey in #9
• Add python plotting tools by @matt-frey in #10
• write diagnostics to h5; fix bug in main model loop by @matt-frey in #12
• Animation with matplotlib by @matt-frey in #14
• write lambda parameter to h5; update python scripts by @matt-frey in #16
• Add halo cells by @matt-frey in #20
• fix step number of h5 writes by @matt-frey in #22
• Improve python scripts by @matt-frey in #23
• Add combine python script by @matt-frey in #24
• change signature of grid2par by @matt-frey in #19
• fix division by zero warning by @matt-frey in #30
• add disable-python-check flag by @matt-frey in #31
• Add merger module by @matt-frey in #32
• Alternative time stepping methods by @matt-frey in #27
• allow a single parcel by @matt-frey in #42
• Use periodic BCs when merging, apply parcel BCs after merge by @sjboeing in #36
• Small optimisations in merging procedure by @sjboeing in #39
• fix ellipse orientation if evec(1) = 0 and evec(2) = epsilon by @matt-frey in #41
• write integer step attribute; write number of parcels per step by @matt-frey in #46
• fix evec update by @matt-frey in #47
• Issue 5 by @sjboeing in #37
• Update number of time steps in h5 by @matt-frey in #49
• Towards serial MPIC by @matt-frey in #48
• Adjusting LS-RK4 setup to be in line with classic RK4 by @sjboeing in #56
• Convergence checks by @sjboeing in #57
• Time stepping optimisations by @sjboeing in #59
• remove nearest grid point interpolation by @matt-frey in #64
• Error in horizontal parcel distance by @matt-frey in #68
• Getting scalasca to work with epic by @matt-frey in #66
• Error in B matrix scaling and unit test for parcel split by @matt-frey in #71
• Revert to unscaled ellipses/ellipsoids by @matt-frey in #69
• Update unit test with unscaled B matrix version by @matt-frey in #72
• remove state instance in classic rk4 by @matt-frey in #73
• Add diverge module by @matt-frey in #53
• Add unit test for parcel merge by @matt-frey in #76
• Adding comments in nearest.f90, changing variable names for clarity by @sjboeing in #78
• Integrate tri-diverge.f90 by @matt-frey in #81
• Make sure we can completely disable procedures by @matt-frey in #83
• Split index vector ij into is and js by @matt-frey in #86
• Need to divide by vcell in phi? by @matt-frey in #89
• unit tests: interpolation module --> parcel_interpl by @matt-frey in #94
• Add maximum absolute volume error to the H5 file by @matt-frey in #93
• Multi-parcel merge? by @matt-frey in #88
• Add more plotting functions by @matt-frey in #96
• Adaptive time stepping by @matt-frey in #98
• Update readme by @matt-frey in #95
• Add vorticity inversion by @matt-frey in #101
• Restructuring by @matt-frey in #102
• Ensure double precision by @matt-frey in #103
• Review/add to namelist parameters and double precision by @matt-frey in #99
• Enable buoyancy and humidity in parcel splitting and merging by @matt-frey in #106
• Gradient correction by @sjboeing in #91
• Maximum volume fraction by @sjboeing in #108
• Add plot_parcel_number plot by @matt-frey in #109

Full Changelog: https://github.com/matt-frey/epic/commits/0.2.0

Scaled ellipses

Up to this point we use scaled ellipses.

What's Changed

• RK4 and README by @sjboeing in #1
• Added optimised merging code by @sjboeing in #2
• Changing order of compilation (needed for compilation on my laptop) by @sjboeing in #3
• Fix 2d elliptic model by @matt-frey in #8
• Default choices for parameters by @matt-frey in #9
• Add python plotting tools by @matt-frey in #10
• write diagnostics to h5; fix bug in main model loop by @matt-frey in #12
• Animation with matplotlib by @matt-frey in #14
• write lambda parameter to h5; update python scripts by @matt-frey in #16
• Add halo cells by @matt-frey in #20
• fix step number of h5 writes by @matt-frey in #22
• Improve python scripts by @matt-frey in #23
• Add combine python script by @matt-frey in #24
• change signature of grid2par by @matt-frey in #19
• fix division by zero warning by @matt-frey in #30
• add disable-python-check flag by @matt-frey in #31
• Add merger module by @matt-frey in #32
• Alternative time stepping methods by @matt-frey in #27
• allow a single parcel by @matt-frey in #42
• Use periodic BCs when merging, apply parcel BCs after merge by @sjboeing in #36
• Small optimisations in merging procedure by @sjboeing in #39
• fix ellipse orientation if evec(1) = 0 and evec(2) = epsilon by @matt-frey in #41
• write integer step attribute; write number of parcels per step by @matt-frey in #46
• fix evec update by @matt-frey in #47
• Issue 5 by @sjboeing in #37
• Update number of time steps in h5 by @matt-frey in #49
• Towards serial MPIC by @matt-frey in #48
• Adjusting LS-RK4 setup to be in line with classic RK4 by @sjboeing in #56
• Convergence checks by @sjboeing in #57
• Time stepping optimisations by @sjboeing in #59
• remove nearest grid point interpolation by @matt-frey in #64
• Error in horizontal parcel distance by @matt-frey in #68
• Getting scalasca to work with epic by @matt-frey in #66
• Error in B matrix scaling and unit test for parcel split by @matt-frey in #71
• Revert to unscaled ellipses/ellipsoids by @matt-frey in #69

Full Changelog: https://github.com/matt-frey/epic/commits/0.1.0