/[MITgcm]/manual/s_algorithm/text/tracer.tex
ViewVC logotype

Diff of /manual/s_algorithm/text/tracer.tex

Parent Directory Parent Directory | Revision Log Revision Log | View Revision Graph Revision Graph | View Patch Patch

revision 1.3 by adcroft, Tue Sep 25 20:13:42 2001 UTC revision 1.13 by adcroft, Mon May 6 19:19:31 2002 UTC
# Line 2  Line 2 
2  % $Name$  % $Name$
3    
4  \section{Tracer equations}  \section{Tracer equations}
5    \label{sect:tracer_equations}
6    
7  The basic discretization used for the tracer equations is the second  The basic discretization used for the tracer equations is the second
8  order piece-wise constant finite volume form of the forced  order piece-wise constant finite volume form of the forced
9  advection-diussion equations. There are many alternatives to second  advection-diffusion equations. There are many alternatives to second
10  order method for advection and alternative parameterizations for the  order method for advection and alternative parameterizations for the
11  sub-grid scale processes. The Gent-McWilliams eddy parameterization,  sub-grid scale processes. The Gent-McWilliams eddy parameterization,
12  KPP mixing scheme and PV flux parameterization are all dealt with in  KPP mixing scheme and PV flux parameterization are all dealt with in
# Line 14  part of the tracer equations and the var Line 15  part of the tracer equations and the var
15  described here.  described here.
16    
17  \subsection{Time-stepping of tracers: ABII}  \subsection{Time-stepping of tracers: ABII}
18    \label{sect:tracer_equations_abII}
19    
20  The default advection scheme is the centered second order method which  The default advection scheme is the centered second order method which
21  requires a second order or quasi-second order time-stepping scheme to  requires a second order or quasi-second order time-stepping scheme to
# Line 41  only affects the surface layer since the Line 43  only affects the surface layer since the
43  everywhere else. This term is therefore referred to as the surface  everywhere else. This term is therefore referred to as the surface
44  correction term. Global conservation is not possible using the  correction term. Global conservation is not possible using the
45  flux-form (as here) and a linearized free-surface  flux-form (as here) and a linearized free-surface
46  (\cite{Griffies00,Campin02}).  (\cite{griffies:00,campin:02}).
47    
48  The continuity equation can be recovered by setting  The continuity equation can be recovered by setting
49  $G_{diff}=G_{forc}=0$ and $\tau=1$.  $G_{diff}=G_{forc}=0$ and $\tau=1$.
50    
51  The driver routine that calls the routines to calculate tendancies are  The driver routine that calls the routines to calculate tendencies are
52  {\em S/R CALC\_GT} and {\em S/R CALC\_GS} for temperature and salt  {\em S/R CALC\_GT} and {\em S/R CALC\_GS} for temperature and salt
53  (moisture), respectively. These in turn call a generic advection  (moisture), respectively. These in turn call a generic advection
54  diffusion routine {\em S/R GAD\_CALC\_RHS} that is called with the  diffusion routine {\em S/R GAD\_CALC\_RHS} that is called with the
55  flow field and relevent tracer as arguments and returns the collective  flow field and relevant tracer as arguments and returns the collective
56  tendancy due to advection and diffusion. Forcing is add subsequently  tendency due to advection and diffusion. Forcing is add subsequently
57  in {\em S/R CALC\_GT} or {\em S/R CALC\_GS} to the same tendancy  in {\em S/R CALC\_GT} or {\em S/R CALC\_GS} to the same tendency
58  array.  array.
59    
60  \fbox{ \begin{minipage}{4.75in}  \fbox{ \begin{minipage}{4.75in}
# Line 66  $F_r$: {\bf fVerT} (argument) Line 68  $F_r$: {\bf fVerT} (argument)
68    
69  \end{minipage} }  \end{minipage} }
70    
71    The space and time discretization are treated separately (method of
72  The space and time discretizations are treated seperately (method of  lines). Tendencies are calculated at time levels $n$ and $n-1$ and
73  lines). The Adams-Bashforth time discretization reads:  extrapolated to $n+1/2$ using the Adams-Bashforth method:
74  \marginpar{$\epsilon$: {\bf AB\_eps}}  \marginpar{$\epsilon$: {\bf AB\_eps}}
 \marginpar{$\Delta t$: {\bf deltaTtracer}}  
75  \begin{equation}  \begin{equation}
76  \tau^{(n+1)} = \tau^{(n)} + \Delta t \left(  G^{(n+1/2)} =
77  (\frac{3}{2} + \epsilon) G^{(n)} - (\frac{1}{2} + \epsilon) G^{(n-1)}  (\frac{3}{2} + \epsilon) G^{(n)} - (\frac{1}{2} + \epsilon) G^{(n-1)}
 \right)  
78  \end{equation}  \end{equation}
79  where $G^{(n)} = G_{adv}^\tau + G_{diff}^\tau + G_{src}^\tau$ at time  where $G^{(n)} = G_{adv}^\tau + G_{diff}^\tau + G_{src}^\tau$ at time
80  step $n$.  step $n$. The tendency at $n-1$ is not re-calculated but rather the
81    tendency at $n$ is stored in a global array for later re-use.
82    
83  Strictly speaking the ABII scheme should be applied only to the  \fbox{ \begin{minipage}{4.75in}
84  advection terms. However, this scheme is only used in conjuction with  {\em S/R ADAMS\_BASHFORTH2} ({\em model/src/adams\_bashforth2.F})
85  the standard second, third and fourth order advection  
86  schemes. Selection of any other advection scheme disables  $G^{(n+1/2)}$: {\bf gTracer} (argument on exit)
87  Adams-Bashforth for tracers so that explicit diffusion and forcing use  
88  the forward method.  $G^{(n)}$: {\bf gTracer} (argument on entry)
89    
90    $G^{(n-1)}$: {\bf gTrNm1} (argument)
91    
92    $\epsilon$: {\bf ABeps} (PARAMS.h)
93    
94    \end{minipage} }
95    
96    The tracers are stepped forward in time using the extrapolated tendency:
97    \begin{equation}
98    \tau^{(n+1)} = \tau^{(n)} + \Delta t G^{(n+1/2)}
99    \end{equation}
100    \marginpar{$\Delta t$: {\bf deltaTtracer}}
101    
102  \fbox{ \begin{minipage}{4.75in}  \fbox{ \begin{minipage}{4.75in}
103  {\em S/R TIMESTEP\_TRACER} ({\em model/src/timestep\_tracer.F})  {\em S/R TIMESTEP\_TRACER} ({\em model/src/timestep\_tracer.F})
104    
105  $\tau$: {\bf tracer} (argument)  $\tau^{(n+1)}$: {\bf gTracer} (argument on exit)
106    
107  $G^{(n)}$: {\bf gTracer} (argument)  $\tau^{(n)}$: {\bf tracer} (argument on entry)
108    
109  $G^{(n-1)}$: {\bf gTrNm1} (argument)  $G^{(n+1/2)}$: {\bf gTracer} (argument)
110    
111  $\Delta t$: {\bf deltaTtracer} (PARAMS.h)  $\Delta t$: {\bf deltaTtracer} (PARAMS.h)
112    
113  \end{minipage} }  \end{minipage} }
114    
115    Strictly speaking the ABII scheme should be applied only to the
116    advection terms. However, this scheme is only used in conjunction with
117    the standard second, third and fourth order advection
118    schemes. Selection of any other advection scheme disables
119    Adams-Bashforth for tracers so that explicit diffusion and forcing use
120    the forward method.
121    
122    
123    
124    
125    \section{Linear advection schemes}
126    \label{sect:tracer-advection}
127    
128  \begin{figure}  \begin{figure}
129  \resizebox{5.5in}{!}{\includegraphics{part2/advect-1d-lo.eps}}  \resizebox{5.5in}{!}{\includegraphics{part2/advect-1d-lo.eps}}
130  \caption{  \caption{
# Line 136  $\mu=c/(1-c)$. Line 162  $\mu=c/(1-c)$.
162  }  }
163  \end{figure}  \end{figure}
164    
 \section{Linear advection schemes}  
   
165  The advection schemes known as centered second order, centered fourth  The advection schemes known as centered second order, centered fourth
166  order, first order upwind and upwind biased third order are known as  order, first order upwind and upwind biased third order are known as
167  linear advection schemes because the coefficient for interpolation of  linear advection schemes because the coefficient for interpolation of
# Line 148  commonly used in the field and most fami Line 172  commonly used in the field and most fami
172  \subsection{Centered second order advection-diffusion}  \subsection{Centered second order advection-diffusion}
173    
174  The basic discretization, centered second order, is the default. It is  The basic discretization, centered second order, is the default. It is
175  designed to be consistant with the continuity equation to facilitate  designed to be consistent with the continuity equation to facilitate
176  conservation properties analogous to the continuum. However, centered  conservation properties analogous to the continuum. However, centered
177  second order advection is notoriously noisey and must be used in  second order advection is notoriously noisy and must be used in
178  conjuction with some finite amount of diffusion to produce a sensible  conjunction with some finite amount of diffusion to produce a sensible
179  solution.  solution.
180    
181  The advection operator is discretized:  The advection operator is discretized:
# Line 177  W & = & {\cal A}_c w Line 201  W & = & {\cal A}_c w
201    
202  For non-divergent flow, this discretization can be shown to conserve  For non-divergent flow, this discretization can be shown to conserve
203  the tracer both locally and globally and to globally conserve tracer  the tracer both locally and globally and to globally conserve tracer
204  variance, $\tau^2$. The proof is given in \cite{Adcroft95,Adcroft97}.  variance, $\tau^2$. The proof is given in \cite{adcroft:95,adcroft:97}.
205    
206  \fbox{ \begin{minipage}{4.75in}  \fbox{ \begin{minipage}{4.75in}
207  {\em S/R GAD\_C2\_ADV\_X} ({\em gad\_c2\_adv\_x.F})  {\em S/R GAD\_C2\_ADV\_X} ({\em gad\_c2\_adv\_x.F})
# Line 225  F_r & = & W \overline{\tau - \frac{1}{6} Line 249  F_r & = & W \overline{\tau - \frac{1}{6}
249  \end{eqnarray}  \end{eqnarray}
250    
251  At boundaries, $\delta_{\hat{n}} \tau$ is set to zero allowing  At boundaries, $\delta_{\hat{n}} \tau$ is set to zero allowing
252  $\delta_{nn}$ to be evaluated. We are currently examing the accuracy  $\delta_{nn}$ to be evaluated. We are currently examine the accuracy
253  of this boundary condition and the effect on the solution.  of this boundary condition and the effect on the solution.
254    
255    \fbox{ \begin{minipage}{4.75in}
256    {\em S/R GAD\_U3\_ADV\_X} ({\em gad\_u3\_adv\_x.F})
257    
258    $F_x$: {\bf uT} (argument)
259    
260    $U$: {\bf uTrans} (argument)
261    
262    $\tau$: {\bf tracer} (argument)
263    
264    {\em S/R GAD\_U3\_ADV\_Y} ({\em gad\_u3\_adv\_y.F})
265    
266    $F_y$: {\bf vT} (argument)
267    
268    $V$: {\bf vTrans} (argument)
269    
270    $\tau$: {\bf tracer} (argument)
271    
272    {\em S/R GAD\_U3\_ADV\_R} ({\em gad\_u3\_adv\_r.F})
273    
274    $F_r$: {\bf wT} (argument)
275    
276    $W$: {\bf rTrans} (argument)
277    
278    $\tau$: {\bf tracer} (argument)
279    
280    \end{minipage} }
281    
282  \subsection{Centered fourth order advection}  \subsection{Centered fourth order advection}
283    
284  Centered fourth order advection is formally the most accurate scheme  Centered fourth order advection is formally the most accurate scheme
285  we have implemented and can be used to great effect in high resolution  we have implemented and can be used to great effect in high resolution
286  simultation where dynamical scales are well resolved. However, the  simulation where dynamical scales are well resolved. However, the
287  scheme is noisey like the centered second order method and so must be  scheme is noisy like the centered second order method and so must be
288  used with some finite amount of diffusion. Bi-harmonic is recommended  used with some finite amount of diffusion. Bi-harmonic is recommended
289  since it is more scale selective and less likely to diffuse away the  since it is more scale selective and less likely to diffuse away the
290  well resolved gradient the fourth order scheme worked so hard to  well resolved gradient the fourth order scheme worked so hard to
# Line 248  F_r & = & W \overline{\tau - \frac{1}{6} Line 298  F_r & = & W \overline{\tau - \frac{1}{6}
298  \end{eqnarray}  \end{eqnarray}
299    
300  As for the third order scheme, the best discretization near boundaries  As for the third order scheme, the best discretization near boundaries
301  is under investigation but currenlty $\delta_i \tau=0$ on a boundary.  is under investigation but currently $\delta_i \tau=0$ on a boundary.
302    
303    \fbox{ \begin{minipage}{4.75in}
304    {\em S/R GAD\_C4\_ADV\_X} ({\em gad\_c4\_adv\_x.F})
305    
306    $F_x$: {\bf uT} (argument)
307    
308    $U$: {\bf uTrans} (argument)
309    
310    $\tau$: {\bf tracer} (argument)
311    
312    {\em S/R GAD\_C4\_ADV\_Y} ({\em gad\_c4\_adv\_y.F})
313    
314    $F_y$: {\bf vT} (argument)
315    
316    $V$: {\bf vTrans} (argument)
317    
318    $\tau$: {\bf tracer} (argument)
319    
320    {\em S/R GAD\_C4\_ADV\_R} ({\em gad\_c4\_adv\_r.F})
321    
322    $F_r$: {\bf wT} (argument)
323    
324    $W$: {\bf rTrans} (argument)
325    
326    $\tau$: {\bf tracer} (argument)
327    
328    \end{minipage} }
329    
330    
331  \subsection{First order upwind advection}  \subsection{First order upwind advection}
332    
# Line 310  r = \frac{ \tau_{i-1} - \tau_{i-2} }{ \t Line 388  r = \frac{ \tau_{i-1} - \tau_{i-2} }{ \t
388  r = \frac{ \tau_{i+1} - \tau_{i} }{ \tau_{i} - \tau_{i-1} } & \forall & u < 0  r = \frac{ \tau_{i+1} - \tau_{i} }{ \tau_{i} - \tau_{i-1} } & \forall & u < 0
389  \end{eqnarray}  \end{eqnarray}
390  as it's argument. There are many choices of limiter function but we  as it's argument. There are many choices of limiter function but we
391  only provide the Superbee limiter \cite{Roe85}:  only provide the Superbee limiter \cite{roe:85}:
392  \begin{equation}  \begin{equation}
393  \psi(r) = \max[0,\min[1,2r],\min[2,r]]  \psi(r) = \max[0,\min[1,2r],\min[2,r]]
394  \end{equation}  \end{equation}
395    
396    \fbox{ \begin{minipage}{4.75in}
397    {\em S/R GAD\_FLUXLIMIT\_ADV\_X} ({\em gad\_fluxlimit\_adv\_x.F})
398    
399    $F_x$: {\bf uT} (argument)
400    
401    $U$: {\bf uTrans} (argument)
402    
403    $\tau$: {\bf tracer} (argument)
404    
405    {\em S/R GAD\_FLUXLIMIT\_ADV\_Y} ({\em gad\_fluxlimit\_adv\_y.F})
406    
407    $F_y$: {\bf vT} (argument)
408    
409    $V$: {\bf vTrans} (argument)
410    
411    $\tau$: {\bf tracer} (argument)
412    
413    {\em S/R GAD\_FLUXLIMIT\_ADV\_R} ({\em gad\_fluxlimit\_adv\_r.F})
414    
415    $F_r$: {\bf wT} (argument)
416    
417    $W$: {\bf rTrans} (argument)
418    
419    $\tau$: {\bf tracer} (argument)
420    
421    \end{minipage} }
422    
423    
424  \subsection{Third order direct space time}  \subsection{Third order direct space time}
425    
426  The direct-space-time method deals with space and time discretization  The direct-space-time method deals with space and time discretization
427  together (other methods that treat space and time seperately are known  together (other methods that treat space and time separately are known
428  collectively as the ``Method of Lines''). The Lax-Wendroff scheme  collectively as the ``Method of Lines''). The Lax-Wendroff scheme
429  falls into this category; it adds sufficient diffusion to a second  falls into this category; it adds sufficient diffusion to a second
430  order flux that the forward-in-time method is stable. The upwind  order flux that the forward-in-time method is stable. The upwind
# Line 337  where Line 442  where
442  d_1 & = & \frac{1}{6} ( 2 - |c| ) ( 1 - |c| ) \\  d_1 & = & \frac{1}{6} ( 2 - |c| ) ( 1 - |c| ) \\
443  d_2 & = & \frac{1}{6} ( 1 - |c| ) ( 1 + |c| )  d_2 & = & \frac{1}{6} ( 1 - |c| ) ( 1 + |c| )
444  \end{eqnarray}  \end{eqnarray}
445  The coefficients $d_0$ and $d_1$ approach $1/3$ and $1/6$ repectively  The coefficients $d_0$ and $d_1$ approach $1/3$ and $1/6$ respectively
446  as the Courant number, $c$, vanishes. In this limit, the conventional  as the Courant number, $c$, vanishes. In this limit, the conventional
447  third order upwind method is recovered. For finite Courant number, the  third order upwind method is recovered. For finite Courant number, the
448  deviations from the linear method are analogous to the diffusion added  deviations from the linear method are analogous to the diffusion added
# Line 345  to centered second order advection in th Line 450  to centered second order advection in th
450    
451  The DST3 method described above must be used in a forward-in-time  The DST3 method described above must be used in a forward-in-time
452  manner and is stable for $0 \le |c| \le 1$. Although the scheme  manner and is stable for $0 \le |c| \le 1$. Although the scheme
453  appears to be forward-in-time, it is in fact second order in time and  appears to be forward-in-time, it is in fact third order in time and
454  the accuracy increases with the Courant number! For low Courant  the accuracy increases with the Courant number! For low Courant
455  number, DST3 produces very similar results (indistinguishable in  number, DST3 produces very similar results (indistinguishable in
456  Fig.~\ref{fig:advect-1d-lo}) to the linear third order method but for  Fig.~\ref{fig:advect-1d-lo}) to the linear third order method but for
# Line 353  large Courant number, where the linear u Line 458  large Courant number, where the linear u
458  unstable, the scheme is extremely accurate  unstable, the scheme is extremely accurate
459  (Fig.~\ref{fig:advect-1d-hi}) with only minor overshoots.  (Fig.~\ref{fig:advect-1d-hi}) with only minor overshoots.
460    
461    \fbox{ \begin{minipage}{4.75in}
462    {\em S/R GAD\_DST3\_ADV\_X} ({\em gad\_dst3\_adv\_x.F})
463    
464    $F_x$: {\bf uT} (argument)
465    
466    $U$: {\bf uTrans} (argument)
467    
468    $\tau$: {\bf tracer} (argument)
469    
470    {\em S/R GAD\_DST3\_ADV\_Y} ({\em gad\_dst3\_adv\_y.F})
471    
472    $F_y$: {\bf vT} (argument)
473    
474    $V$: {\bf vTrans} (argument)
475    
476    $\tau$: {\bf tracer} (argument)
477    
478    {\em S/R GAD\_DST3\_ADV\_R} ({\em gad\_dst3\_adv\_r.F})
479    
480    $F_r$: {\bf wT} (argument)
481    
482    $W$: {\bf rTrans} (argument)
483    
484    $\tau$: {\bf tracer} (argument)
485    
486    \end{minipage} }
487    
488    
489  \subsection{Third order direct space time with flux limiting}  \subsection{Third order direct space time with flux limiting}
490    
491  The overshoots in the DST3 method can be controlled with a flux limiter.  The overshoots in the DST3 method can be controlled with a flux limiter.
# Line 373  and the limiter is the Sweby limiter: Line 506  and the limiter is the Sweby limiter:
506  \psi(r) = \max[0, \min[\min(1,d_0+d_1r],\frac{1-c}{c}r ]]  \psi(r) = \max[0, \min[\min(1,d_0+d_1r],\frac{1-c}{c}r ]]
507  \end{equation}  \end{equation}
508    
509    \fbox{ \begin{minipage}{4.75in}
510    {\em S/R GAD\_DST3FL\_ADV\_X} ({\em gad\_dst3\_adv\_x.F})
511    
512    $F_x$: {\bf uT} (argument)
513    
514    $U$: {\bf uTrans} (argument)
515    
516    $\tau$: {\bf tracer} (argument)
517    
518    {\em S/R GAD\_DST3FL\_ADV\_Y} ({\em gad\_dst3\_adv\_y.F})
519    
520    $F_y$: {\bf vT} (argument)
521    
522    $V$: {\bf vTrans} (argument)
523    
524    $\tau$: {\bf tracer} (argument)
525    
526    {\em S/R GAD\_DST3FL\_ADV\_R} ({\em gad\_dst3\_adv\_r.F})
527    
528    $F_r$: {\bf wT} (argument)
529    
530    $W$: {\bf rTrans} (argument)
531    
532    $\tau$: {\bf tracer} (argument)
533    
534    \end{minipage} }
535    
536    
537  \subsection{Multi-dimensional advection}  \subsection{Multi-dimensional advection}
538    
539  In many of the aforementioned advection schemes the behaviour in  \begin{figure}
540    \resizebox{5.5in}{!}{\includegraphics{part2/advect-2d-lo-diag.eps}}
541    \caption{
542    Comparison of advection schemes in two dimensions; diagonal advection
543    of a resolved Gaussian feature. Courant number is 0.01 with
544    30$\times$30 points and solutions are shown for T=1/2. White lines
545    indicate zero crossing (ie. the presence of false minima).  The left
546    column shows the second order schemes; top) centered second order with
547    Adams-Bashforth, middle) Lax-Wendroff and bottom) Superbee flux
548    limited. The middle column shows the third order schemes; top) upwind
549    biased third order with Adams-Bashforth, middle) third order direct
550    space-time method and bottom) the same with flux limiting. The top
551    right panel shows the centered fourth order scheme with
552    Adams-Bashforth and right middle panel shows a fourth order variant on
553    the DST method. Bottom right panel shows the Superbee flux limiter
554    (second order) applied independently in each direction (method of
555    lines).
556    \label{fig:advect-2d-lo-diag}
557    }
558    \end{figure}
559    
560    \begin{figure}
561    \resizebox{5.5in}{!}{\includegraphics{part2/advect-2d-mid-diag.eps}}
562    \caption{
563    Comparison of advection schemes in two dimensions; diagonal advection
564    of a resolved Gaussian feature. Courant number is 0.27 with
565    30$\times$30 points and solutions are shown for T=1/2. White lines
566    indicate zero crossing (ie. the presence of false minima).  The left
567    column shows the second order schemes; top) centered second order with
568    Adams-Bashforth, middle) Lax-Wendroff and bottom) Superbee flux
569    limited. The middle column shows the third order schemes; top) upwind
570    biased third order with Adams-Bashforth, middle) third order direct
571    space-time method and bottom) the same with flux limiting. The top
572    right panel shows the centered fourth order scheme with
573    Adams-Bashforth and right middle panel shows a fourth order variant on
574    the DST method. Bottom right panel shows the Superbee flux limiter
575    (second order) applied independently in each direction (method of
576    lines).
577    \label{fig:advect-2d-mid-diag}
578    }
579    \end{figure}
580    
581    \begin{figure}
582    \resizebox{5.5in}{!}{\includegraphics{part2/advect-2d-hi-diag.eps}}
583    \caption{
584    Comparison of advection schemes in two dimensions; diagonal advection
585    of a resolved Gaussian feature. Courant number is 0.47 with
586    30$\times$30 points and solutions are shown for T=1/2. White lines
587    indicate zero crossings and initial maximum values (ie. the presence
588    of false extrema).  The left column shows the second order schemes;
589    top) centered second order with Adams-Bashforth, middle) Lax-Wendroff
590    and bottom) Superbee flux limited. The middle column shows the third
591    order schemes; top) upwind biased third order with Adams-Bashforth,
592    middle) third order direct space-time method and bottom) the same with
593    flux limiting. The top right panel shows the centered fourth order
594    scheme with Adams-Bashforth and right middle panel shows a fourth
595    order variant on the DST method. Bottom right panel shows the Superbee
596    flux limiter (second order) applied independently in each direction
597    (method of lines).
598    \label{fig:advect-2d-hi-diag}
599    }
600    \end{figure}
601    
602    
603    
604    In many of the aforementioned advection schemes the behavior in
605  multiple dimensions is not necessarily as good as the one dimensional  multiple dimensions is not necessarily as good as the one dimensional
606  behaviour. For instance, a shape preserving monotonic scheme in one  behavior. For instance, a shape preserving monotonic scheme in one
607  dimension can have severe shape distortion in two dimensions if the  dimension can have severe shape distortion in two dimensions if the
608  two components of horizontal fluxes are treated independently. There  two components of horizontal fluxes are treated independently. There
609  is a large body of literature on the subject dealing with this problem  is a large body of literature on the subject dealing with this problem
# Line 398  as if in one dimension: Line 624  as if in one dimension:
624  \end{eqnarray}  \end{eqnarray}
625    
626  In order to incorporate this method into the general model algorithm,  In order to incorporate this method into the general model algorithm,
627  we compute the effective tendancy rather than update the tracer so  we compute the effective tendency rather than update the tracer so
628  that other terms such as diffusion are using the $n$ time-level and  that other terms such as diffusion are using the $n$ time-level and
629  not the updated $n+3/3$ quantities:  not the updated $n+3/3$ quantities:
630  \begin{equation}  \begin{equation}
# Line 408  So that the over all time-stepping looks Line 634  So that the over all time-stepping looks
634  \begin{equation}  \begin{equation}
635  \tau^{n+1} = \tau^{n} + \Delta t \left( G^{n+1/2}_{adv} + G_{diff}(\tau^{n}) + G^{n}_{forcing} \right)  \tau^{n+1} = \tau^{n} + \Delta t \left( G^{n+1/2}_{adv} + G_{diff}(\tau^{n}) + G^{n}_{forcing} \right)
636  \end{equation}  \end{equation}
637    
638    \fbox{ \begin{minipage}{4.75in}
639    {\em S/R GAD\_ADVECTION} ({\em gad\_advection.F})
640    
641    $\tau$: {\bf Tracer} (argument)
642    
643    $G^{n+1/2}_{adv}$: {\bf Gtracer} (argument)
644    
645    $F_x, F_y, F_r$: {\bf af} (local)
646    
647    $U$: {\bf uTrans} (local)
648    
649    $V$: {\bf vTrans} (local)
650    
651    $W$: {\bf rTrans} (local)
652    
653    \end{minipage} }
654    
655    
656    \section{Comparison of advection schemes}
657    
658    Figs.~\ref{fig:advect-2d-lo-diag}, \ref{fig:advect-2d-mid-diag} and
659    \ref{fig:advect-2d-hi-diag} show solutions to a simple diagonal
660    advection problem using a selection of schemes for low, moderate and
661    high Courant numbers, respectively. The top row shows the linear
662    schemes, integrated with the Adams-Bashforth method. Theses schemes
663    are clearly unstable for the high Courant number and weakly unstable
664    for the moderate Courant number. The presence of false extrema is very
665    apparent for all Courant numbers. The middle row shows solutions
666    obtained with the unlimited but multi-dimensional schemes. These
667    solutions also exhibit false extrema though the pattern now shows
668    symmetry due to the multi-dimensional scheme. Also, the schemes are
669    stable at high Courant number where the linear schemes weren't. The
670    bottom row (left and middle) shows the limited schemes and most
671    obvious is the absence of false extrema. The accuracy and stability of
672    the unlimited non-linear schemes is retained at high Courant number
673    but at low Courant number the tendency is to loose amplitude in sharp
674    peaks due to diffusion. The one dimensional tests shown in
675    Figs.~\ref{fig:advect-1d-lo} and \ref{fig:advect-1d-hi} showed this
676    phenomenon.
677    
678    Finally, the bottom left and right panels use the same advection
679    scheme but the right does not use the multi-dimensional method. At low
680    Courant number this appears to not matter but for moderate Courant
681    number severe distortion of the feature is apparent. Moreover, the
682    stability of the multi-dimensional scheme is determined by the maximum
683    Courant number applied of each dimension while the stability of the
684    method of lines is determined by the sum. Hence, in the high Courant
685    number plot, the scheme is unstable.
686    
687    With many advection schemes implemented in the code two questions
688    arise: ``Which scheme is best?'' and ``Why don't you just offer the
689    best advection scheme?''. Unfortunately, no one advection scheme is
690    ``the best'' for all particular applications and for new applications
691    it is often a matter of trial to determine which is most
692    suitable. Here are some guidelines but these are not the rule;
693    \begin{itemize}
694    \item If you have a coarsely resolved model, using a
695    positive or upwind biased scheme will introduce significant diffusion
696    to the solution and using a centered higher order scheme will
697    introduce more noise. In this case, simplest may be best.
698    \item If you have a high resolution model, using a higher order
699    scheme will give a more accurate solution but scale-selective
700    diffusion might need to be employed. The flux limited methods
701    offer similar accuracy in this regime.
702    \item If your solution has shocks or propagating fronts then a
703    flux limited scheme is almost essential.
704    \item If your time-step is limited by advection, the multi-dimensional
705    non-linear schemes have the most stability (up to Courant number 1).
706    \item If you need to know how much diffusion/dissipation has occurred you
707    will have a lot of trouble figuring it out with a non-linear method.
708    \item The presence of false extrema is non-physical and this alone is the
709    strongest argument for using a positive scheme.
710    \end{itemize}

Legend:
Removed from v.1.3  
changed lines
  Added in v.1.13

  ViewVC Help
Powered by ViewVC 1.1.22