4-1-05: check the min depth (h+eta) called in stability_report() why in fCANH96.init is it 0.42 m?? (DONE) 4-1-05: make a separate initial_condition.[ch] file that loads up and sets U, V, and ETA - U & V initial conditions (DONE) 4-1-05: check the ouput of total_potential_energy() called in verbose output * in some wave runs it's almost as if potential_energy is 2X kinetic energy! => actually I forgot the 1/2 in from of the g*\eta^2 so that is why (FIXED) 4-4-05: fix the GUI window so that the Run Time is right (DONE) 4-4-05: sponge.[ch] : write sponge_layer_stability_report() (DONE) 4-5-05: bousinessq_dynamics.c : speedup b_calc_eta_EP3(...), the diff_y term doesn't have to be calculated each time! (DONE) 4-5-05: field.c : in field2D_barx(FB, F) - when FB has N+2 x-points realtive to F, the i=0 & i=N+1 points are made to be the same as the i=1 and i=N FB points. Maybe it shuld just be the F(0,:) and F(N-1,:) row instead? (DONE) 4-5-05: WEI_KIRBY dynamics: w/ Fx!=0, reducing dt does not help stability. It seems that the model blows up regardless and it depends only on the number of iterations going on. This indicates an algorithm problem not a stability problem. -> I went through most of functions preliminary_calculations() and rhs_eta_equation() and found lots of small bugs. -> need to do rhs_uvmom stuff (DONE) 4-6-05: there is suggestion that biharmonic friction is actually leading to instabilites in the bousinessq wave modes. Every time I reduce the biharmonic friction it results in long run times and less global type instabilities! Wierd. -> Ok, this is because the explicit wave model was unstable. This is now fixed and biharmonic stuff is ok. (DONE) 4-8-05: * added the tridiagonal solver for U (DONE) - seems to work for a simple wave maker case * need to add the periodic solver for V (DONE) 4-1-05: make a separate initial_condition.[ch] file that loads up and sets U, V, and ETA - U & V initial conditions (DONE) - add a separate ETA line in the .init file (DONE) 4-11-05: WORK on the wave breaking stuff! - Zelt (DONE) - output breaking eddy viscosity (DONE) 4-11-05: BUG!!! w/ the new V tridiagonal solver the constant alongshore forcing fCtest.init doesn't work!!! it blows up slowly but at the first time step things are really wierd. * try with NSWE dynamics first - this is buggy too! => OK I fixed the bug - I wasn't doing the iAA22 term correctly in tridiagonal.c: - tridiagonal solve for v --- (DONE) 5-11-05: I changed the model code so that instead of doing the wierd integration of F1 & G1 (Bousinessq terms) that funwave does, I used dU/dt and dV/dt estimates that are used in the Ft and Gt terms instead. This now runs without blowing up in the fChump.init case example. Previously, with F1 and G1 it blew up pretty fast. But strangely, with only one or the other of F1 or G1 active it worked fine. Maybe there is a instability type thing with these terms on a Cgrid??? Who knows. 5-17-05: BUG!!! the Ft and Gt terms are zeroed out!!!!!! Holy shit. Everything could be wrong. I have to check this and rerun the solitary wave cases. (FIXED) * 5-18-05: this seems to be ok with solitary wave cases (OK) 5-18-05: BUG! write a function to calculate U = u + hb1( h u_xx + (hu)_xx) etc and V = v + h b1( h v_yy + (hv)_yy) etc. - and apply to initial conditions (DONE) 5-22-05: if wave breaking is turned on then the breaking eddy viscosity is added on to the background kappa value for the tracer. (DONE) 5-12-05: * make bousinessq wave dynamics not a compile time choice (DONE) * in field.[ch], make a general _self_add_[two,three] (w/ multiple args A+= B+C+D) etc. & apply (DONE) 5-12-05: Run bousinessq model tests (DONE) * flat bottom oblique waves * planar slope, oblique non-breaking waves (fCangle) * do a solitary wave test (fCsolitary) => one in x direction => one in y direction 5-26-05: Add the memory report thing to level 2 and 3 output as well. (DONE) 5-23-05: * in function 'b_calc_eta_EP3' - I think there was a bug with the hm at j=0. I changed it to 'hm = *(hbxy+MN)' to coincide with v_yy etc. Check this with nonuniform bathymetry. I think this didn't show up because the bathymetry near the eges of the Y-domain were always uniform (DONE) 5-24-05: remove dissipation.[ch] from the CVS repository (DONE) 5-26-05: Allow for an option to set the start time of the output a la kepsilon (resuse that code! - already DONE) 5-26-05: Finish coding the newtonian0 friction component, this will allow greater dx/dy ratio differences (DONE) 5-26-05: For snapshot output there is always a file of 0 bytes length written without the _????.dat format. (DONE) 6-24-05: The linear dynamics d(eta)/dt equation was using div( (h+eta)U ) not div( hU ) - which is a bug! Changed the eta dynamics to use a switch and did some simple tests (DONE) 6-24-05: did another VALGRIND pass: I found another bug in AB2 time stepping w/ dv/dt!!! simple typo that shouldn't affect longer time integrations too much. (DONE)
