237 |
gradient, $\nabla_h \Phi'_{hyd}$. Note that presently, this term is in |
gradient, $\nabla_h \Phi'_{hyd}$. Note that presently, this term is in |
238 |
fact incorporated to the $\vec{\bf G}_{\vec{\bf v}}$ arrays ({\bf |
fact incorporated to the $\vec{\bf G}_{\vec{\bf v}}$ arrays ({\bf |
239 |
gU,gV}). |
gU,gV}). |
240 |
\marginpar{JMC: Clarify this term?} |
\marginpar{JMC: Clarify ``this term''?} |
241 |
|
|
242 |
|
\fbox{ \begin{minipage}{4.75in} |
243 |
|
{\em S/R TIMESTEP} ({\em timestep.F}) |
244 |
|
|
245 |
|
$G_u^n$: {\bf Gu} ({\em DYNVARS.h}) |
246 |
|
|
247 |
|
$G_u^{n-1}, u^*$: {\bf GuNm1} ({\em DYNVARS.h}) |
248 |
|
|
249 |
|
$G_v^n$: {\bf Gv} ({\em DYNVARS.h}) |
250 |
|
|
251 |
|
$G_v^{n-1}, v^*$: {\bf GvNm1} ({\em DYNVARS.h}) |
252 |
|
|
253 |
|
$G_u^{(n+1/2)}$: {\bf GuTmp} (local) |
254 |
|
|
255 |
|
$G_v^{(n+1/2)}$: {\bf GvTmp} (local) |
256 |
|
|
257 |
|
\end{minipage} } |
258 |
|
|
259 |
|
|
260 |
|
|
261 |
|
|
262 |
|
|
263 |
\subsection{Stagger baroclinic time stepping} |
\subsection{Stagger baroclinic time stepping} |
320 |
\label{eq-solve2D_rhs} |
\label{eq-solve2D_rhs} |
321 |
\end{eqnarray} |
\end{eqnarray} |
322 |
|
|
323 |
|
\fbox{ \begin{minipage}{4.75in} |
324 |
|
{\em S/R SOLVE\_FOR\_PRESSURE} ({\em solve\_for\_pressure.F}) |
325 |
|
|
326 |
|
$u^*$: {\bf GuNm1} ({\em DYNVARS.h}) |
327 |
|
|
328 |
|
$v^*$: {\bf GvNm1} ({\em DYNVARS.h}) |
329 |
|
|
330 |
|
$\eta^*$: {\bf cg2d\_b} (\em SOLVE\_FOR\_PRESSURE.h) |
331 |
|
|
332 |
|
$\eta^{n+1}$: {\bf etaN} (\em DYNVARS.h) |
333 |
|
|
334 |
|
\end{minipage} } |
335 |
|
|
336 |
|
|
337 |
Once ${\eta}^{n+1}$ has been found, substituting into |
Once ${\eta}^{n+1}$ has been found, substituting into |
338 |
\ref{eq-tDsC-Hmom} yields $\vec{\bf v}^{n+1}$ if the model is |
\ref{eq-tDsC-Hmom} yields $\vec{\bf v}^{n+1}$ if the model is |
339 |
hydrostatic ($\epsilon_{nh}=0$): |
hydrostatic ($\epsilon_{nh}=0$): |
372 |
been moved to the beginning of the time step (instead of at the end), |
been moved to the beginning of the time step (instead of at the end), |
373 |
without any consequence on the solution. |
without any consequence on the solution. |
374 |
|
|
375 |
|
\fbox{ \begin{minipage}{4.75in} |
376 |
|
{\em S/R CORRECTION\_STEP} ({\em correction\_step.F}) |
377 |
|
|
378 |
|
$\eta^{n+1}$: {\bf etaN} (\em DYNVARS.h) |
379 |
|
|
380 |
|
$\phi_{nh}^{n+1}$: {\bf phi\_nh} (\em DYNVARS.h) |
381 |
|
|
382 |
|
$u^*$: {\bf GuNm1} ({\em DYNVARS.h}) |
383 |
|
|
384 |
|
$v^*$: {\bf GvNm1} ({\em DYNVARS.h}) |
385 |
|
|
386 |
|
$u^{n+1}$: {\bf uVel} ({\em DYNVARS.h}) |
387 |
|
|
388 |
|
$v^{n+1}$: {\bf vVel} ({\em DYNVARS.h}) |
389 |
|
|
390 |
|
\end{minipage} } |
391 |
|
|
392 |
|
|
393 |
|
|
394 |
Regarding the implementation of the surface pressure solver, all |
Regarding the implementation of the surface pressure solver, all |
395 |
computation are done within the routine {\it SOLVE\_FOR\_PRESSURE} and |
computation are done within the routine {\it SOLVE\_FOR\_PRESSURE} and |
396 |
its dependent calls. The standard method to solve the 2D elliptic |
its dependent calls. The standard method to solve the 2D elliptic |
407 |
at the same point in the code. |
at the same point in the code. |
408 |
|
|
409 |
|
|
410 |
|
|
411 |
\subsection{Crank-Nickelson barotropic time stepping} |
\subsection{Crank-Nickelson barotropic time stepping} |
412 |
|
|
413 |
The full implicit time stepping described previously is |
The full implicit time stepping described previously is |