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

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

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


Revision 1.3 - (hide annotations) (download) (as text)
Tue Sep 25 20:13:42 2001 UTC (23 years, 9 months ago) by adcroft
Branch: MAIN
Changes since 1.2: +374 -29 lines
File MIME type: application/x-tex
Doc. about tracer advection schemes.

1 adcroft 1.3 % $Header: /u/gcmpack/mitgcmdoc/part2/tracer.tex,v 1.2 2001/08/09 20:45:27 adcroft Exp $
2 adcroft 1.2 % $Name: $
3 adcroft 1.1
4     \section{Tracer equations}
5    
6 adcroft 1.2 The basic discretization used for the tracer equations is the second
7     order piece-wise constant finite volume form of the forced
8     advection-diussion equations. There are many alternatives to second
9     order method for advection and alternative parameterizations for the
10     sub-grid scale processes. The Gent-McWilliams eddy parameterization,
11     KPP mixing scheme and PV flux parameterization are all dealt with in
12     separate sections. The basic discretization of the advection-diffusion
13     part of the tracer equations and the various advection schemes will be
14     described here.
15    
16 adcroft 1.3 \subsection{Time-stepping of tracers: ABII}
17    
18     The default advection scheme is the centered second order method which
19     requires a second order or quasi-second order time-stepping scheme to
20     be stable. Historically this has been the quasi-second order
21     Adams-Bashforth method (ABII) and applied to all terms. For an
22     arbitrary tracer, $\tau$, the forced advection-diffusion equation
23     reads:
24     \begin{equation}
25     \partial_t \tau + G_{adv}^\tau = G_{diff}^\tau + G_{forc}^\tau
26     \end{equation}
27     where $G_{adv}^\tau$, $G_{diff}^\tau$ and $G_{forc}^\tau$ are the
28     tendencies due to advection, diffusion and forcing, respectively,
29     namely:
30     \begin{eqnarray}
31     G_{adv}^\tau & = & \partial_x u \tau + \partial_y v \tau + \partial_r w \tau
32     - \tau \nabla \cdot {\bf v} \\
33     G_{diff}^\tau & = & \nabla \cdot {\bf K} \nabla \tau
34     \end{eqnarray}
35     and the forcing can be some arbitrary function of state, time and
36     space.
37    
38     The term, $\tau \nabla \cdot {\bf v}$, is required to retain local
39     conservation in conjunction with the linear implicit free-surface. It
40     only affects the surface layer since the flow is non-divergent
41     everywhere else. This term is therefore referred to as the surface
42     correction term. Global conservation is not possible using the
43     flux-form (as here) and a linearized free-surface
44     (\cite{Griffies00,Campin02}).
45    
46     The continuity equation can be recovered by setting
47     $G_{diff}=G_{forc}=0$ and $\tau=1$.
48    
49     The driver routine that calls the routines to calculate tendancies are
50     {\em S/R CALC\_GT} and {\em S/R CALC\_GS} for temperature and salt
51     (moisture), respectively. These in turn call a generic advection
52     diffusion routine {\em S/R GAD\_CALC\_RHS} that is called with the
53     flow field and relevent tracer as arguments and returns the collective
54     tendancy due to advection and diffusion. Forcing is add subsequently
55     in {\em S/R CALC\_GT} or {\em S/R CALC\_GS} to the same tendancy
56     array.
57    
58     \fbox{ \begin{minipage}{4.75in}
59     {\em S/R GAD\_CALC\_RHS} ({\em pkg/generic\_advdiff/gad\_calc\_rhs.F})
60    
61     $\tau$: {\bf tracer} (argument)
62    
63     $G^{(n)}$: {\bf gTracer} (argument)
64    
65     $F_r$: {\bf fVerT} (argument)
66    
67     \end{minipage} }
68    
69    
70     The space and time discretizations are treated seperately (method of
71     lines). The Adams-Bashforth time discretization reads:
72     \marginpar{$\epsilon$: {\bf AB\_eps}}
73     \marginpar{$\Delta t$: {\bf deltaTtracer}}
74     \begin{equation}
75     \tau^{(n+1)} = \tau^{(n)} + \Delta t \left(
76     (\frac{3}{2} + \epsilon) G^{(n)} - (\frac{1}{2} + \epsilon) G^{(n-1)}
77     \right)
78     \end{equation}
79     where $G^{(n)} = G_{adv}^\tau + G_{diff}^\tau + G_{src}^\tau$ at time
80     step $n$.
81    
82     Strictly speaking the ABII scheme should be applied only to the
83     advection terms. However, this scheme is only used in conjuction with
84     the standard second, third and fourth order advection
85     schemes. Selection of any other advection scheme disables
86     Adams-Bashforth for tracers so that explicit diffusion and forcing use
87     the forward method.
88    
89     \fbox{ \begin{minipage}{4.75in}
90     {\em S/R TIMESTEP\_TRACER} ({\em model/src/timestep\_tracer.F})
91    
92     $\tau$: {\bf tracer} (argument)
93    
94     $G^{(n)}$: {\bf gTracer} (argument)
95    
96     $G^{(n-1)}$: {\bf gTrNm1} (argument)
97    
98     $\Delta t$: {\bf deltaTtracer} (PARAMS.h)
99    
100     \end{minipage} }
101    
102     \begin{figure}
103     \resizebox{5.5in}{!}{\includegraphics{part2/advect-1d-lo.eps}}
104     \caption{
105     Comparison of 1-D advection schemes. Courant number is 0.05 with 60
106     points and solutions are shown for T=1 (one complete period).
107     a) Shows the upwind biased schemes; first order upwind, DST3,
108     third order upwind and second order upwind.
109     b) Shows the centered schemes; Lax-Wendroff, DST4, centered second order,
110     centered fourth order and finite volume fourth order.
111     c) Shows the second order flux limiters: minmod, Superbee,
112     MC limiter and the van Leer limiter.
113     d) Shows the DST3 method with flux limiters due to Sweby with
114     $\mu=1$, $\mu=c/(1-c)$ and a fourth order DST method with Sweby limiter,
115     $\mu=c/(1-c)$.
116     \label{fig:advect-1d-lo}
117     }
118     \end{figure}
119    
120     \begin{figure}
121     \resizebox{5.5in}{!}{\includegraphics{part2/advect-1d-hi.eps}}
122     \caption{
123     Comparison of 1-D advection schemes. Courant number is 0.89 with 60
124     points and solutions are shown for T=1 (one complete period).
125     a) Shows the upwind biased schemes; first order upwind and DST3.
126     Third order upwind and second order upwind are unstable at this Courant number.
127     b) Shows the centered schemes; Lax-Wendroff, DST4. Centered second order,
128     centered fourth order and finite volume fourth order and unstable at this
129     Courant number.
130     c) Shows the second order flux limiters: minmod, Superbee,
131     MC limiter and the van Leer limiter.
132     d) Shows the DST3 method with flux limiters due to Sweby with
133     $\mu=1$, $\mu=c/(1-c)$ and a fourth order DST method with Sweby limiter,
134     $\mu=c/(1-c)$.
135     \label{fig:advect-1d-hi}
136     }
137     \end{figure}
138    
139     \section{Linear advection schemes}
140    
141     The advection schemes known as centered second order, centered fourth
142     order, first order upwind and upwind biased third order are known as
143     linear advection schemes because the coefficient for interpolation of
144     the advected tracer are linear and a function only of the flow, not
145     the tracer field it self. We discuss these first since they are most
146     commonly used in the field and most familiar.
147    
148 adcroft 1.2 \subsection{Centered second order advection-diffusion}
149    
150     The basic discretization, centered second order, is the default. It is
151     designed to be consistant with the continuity equation to facilitate
152 adcroft 1.3 conservation properties analogous to the continuum. However, centered
153     second order advection is notoriously noisey and must be used in
154     conjuction with some finite amount of diffusion to produce a sensible
155     solution.
156    
157     The advection operator is discretized:
158 adcroft 1.1 \begin{equation}
159 adcroft 1.3 {\cal A}_c \Delta r_f h_c G_{adv}^\tau =
160     \delta_i F_x + \delta_j F_y + \delta_k F_r
161 adcroft 1.1 \end{equation}
162 adcroft 1.2 where the area integrated fluxes are given by:
163     \begin{eqnarray}
164 adcroft 1.3 F_x & = & U \overline{ \tau }^i \\
165     F_y & = & V \overline{ \tau }^j \\
166     F_r & = & W \overline{ \tau }^k
167 adcroft 1.2 \end{eqnarray}
168 adcroft 1.1 The quantities $U$, $V$ and $W$ are volume fluxes defined:
169     \marginpar{$U$: {\bf uTrans} }
170     \marginpar{$V$: {\bf vTrans} }
171     \marginpar{$W$: {\bf rTrans} }
172     \begin{eqnarray}
173     U & = & \Delta y_g \Delta r_f h_w u \\
174     V & = & \Delta x_g \Delta r_f h_s v \\
175     W & = & {\cal A}_c w
176     \end{eqnarray}
177    
178 adcroft 1.3 For non-divergent flow, this discretization can be shown to conserve
179     the tracer both locally and globally and to globally conserve tracer
180     variance, $\tau^2$. The proof is given in \cite{Adcroft95,Adcroft97}.
181    
182     \fbox{ \begin{minipage}{4.75in}
183     {\em S/R GAD\_C2\_ADV\_X} ({\em gad\_c2\_adv\_x.F})
184    
185     $F_x$: {\bf uT} (argument)
186    
187     $U$: {\bf uTrans} (argument)
188    
189     $\tau$: {\bf tracer} (argument)
190    
191     {\em S/R GAD\_C2\_ADV\_Y} ({\em gad\_c2\_adv\_y.F})
192    
193     $F_y$: {\bf vT} (argument)
194    
195     $V$: {\bf vTrans} (argument)
196    
197     $\tau$: {\bf tracer} (argument)
198    
199     {\em S/R GAD\_C2\_ADV\_R} ({\em gad\_c2\_adv\_r.F})
200    
201     $F_r$: {\bf wT} (argument)
202    
203     $W$: {\bf rTrans} (argument)
204    
205     $\tau$: {\bf tracer} (argument)
206    
207     \end{minipage} }
208    
209    
210     \subsection{Third order upwind bias advection}
211    
212     Upwind biased third order advection offers a relatively good
213     compromise between accuracy and smoothness. It is not a ``positive''
214     scheme meaning false extrema are permitted but the amplitude of such
215     are significantly reduced over the centered second order method.
216    
217     The third order upwind fluxes are discretized:
218     \begin{eqnarray}
219     F_x & = & U \overline{\tau - \frac{1}{6} \delta_{ii} \tau}^i
220     + \frac{1}{2} |U| \delta_i \frac{1}{6} \delta_{ii} \tau \\
221     F_y & = & V \overline{\tau - \frac{1}{6} \delta_{ii} \tau}^j
222     + \frac{1}{2} |V| \delta_j \frac{1}{6} \delta_{jj} \tau \\
223     F_r & = & W \overline{\tau - \frac{1}{6} \delta_{ii} \tau}^k
224     + \frac{1}{2} |W| \delta_k \frac{1}{6} \delta_{kk} \tau
225     \end{eqnarray}
226    
227     At boundaries, $\delta_{\hat{n}} \tau$ is set to zero allowing
228     $\delta_{nn}$ to be evaluated. We are currently examing the accuracy
229     of this boundary condition and the effect on the solution.
230    
231    
232     \subsection{Centered fourth order advection}
233    
234     Centered fourth order advection is formally the most accurate scheme
235     we have implemented and can be used to great effect in high resolution
236     simultation where dynamical scales are well resolved. However, the
237     scheme is noisey like the centered second order method and so must be
238     used with some finite amount of diffusion. Bi-harmonic is recommended
239     since it is more scale selective and less likely to diffuse away the
240     well resolved gradient the fourth order scheme worked so hard to
241     create.
242    
243     The centered fourth order fluxes are discretized:
244     \begin{eqnarray}
245     F_x & = & U \overline{\tau - \frac{1}{6} \delta_{ii} \tau}^i \\
246     F_y & = & V \overline{\tau - \frac{1}{6} \delta_{ii} \tau}^j \\
247     F_r & = & W \overline{\tau - \frac{1}{6} \delta_{ii} \tau}^k
248     \end{eqnarray}
249    
250     As for the third order scheme, the best discretization near boundaries
251     is under investigation but currenlty $\delta_i \tau=0$ on a boundary.
252    
253     \subsection{First order upwind advection}
254    
255     Although the upwind scheme is the underlying scheme for the robust or
256     non-linear methods given later, we haven't actually supplied this
257     method for general use. It would be very diffusive and it is unlikely
258     that it could ever produce more useful results than the positive
259     higher order schemes.
260    
261     Upwind bias is introduced into many schemes using the {\em abs}
262     function and is allows the first order upwind flux to be written:
263     \begin{eqnarray}
264     F_x & = & U \overline{ \tau }^i - \frac{1}{2} |U| \delta_i \tau \\
265     F_y & = & V \overline{ \tau }^j - \frac{1}{2} |V| \delta_j \tau \\
266     F_r & = & W \overline{ \tau }^k - \frac{1}{2} |W| \delta_k \tau
267     \end{eqnarray}
268    
269     If for some reason, the above method is required, then the second
270     order flux limiter scheme described later reduces to the above scheme
271     if the limiter is set to zero.
272    
273    
274     \section{Non-linear advection schemes}
275    
276     Non-linear advection schemes invoke non-linear interpolation and are
277     widely used in computational fluid dynamics (non-linear does not refer
278     to the non-linearity of the advection operator). The flux limited
279     advection schemes belong to the class of finite volume methods which
280     neatly ties into the spatial discretization of the model.
281    
282     When employing the flux limited schemes, first order upwind or
283     direct-space-time method the time-stepping is switched to forward in
284     time.
285    
286     \subsection{Second order flux limiters}
287    
288     The second order flux limiter method can be cast in several ways but
289     is generally expressed in terms of other flux approximations. For
290     example, in terms of a first order upwind flux and second order
291     Lax-Wendroff flux, the limited flux is given as:
292     \begin{equation}
293     F = F_1 + \psi(r) F_{LW}
294     \end{equation}
295     where $\psi(r)$ is the limiter function,
296     \begin{equation}
297     F_1 = u \overline{\tau}^i - \frac{1}{2} |u| \delta_i \tau
298     \end{equation}
299     is the upwind flux,
300     \begin{equation}
301     F_{LW} = F_1 + \frac{|u|}{2} (1-c) \delta_i \tau
302     \end{equation}
303     is the Lax-Wendroff flux and $c = \frac{u \Delta t}{\Delta x}$ is the
304     Courant (CFL) number.
305    
306     The limiter function, $\psi(r)$, takes the slope ratio
307     \begin{eqnarray}
308     r = \frac{ \tau_{i-1} - \tau_{i-2} }{ \tau_{i} - \tau_{i-1} } & \forall & u > 0
309     \\
310     r = \frac{ \tau_{i+1} - \tau_{i} }{ \tau_{i} - \tau_{i-1} } & \forall & u < 0
311     \end{eqnarray}
312     as it's argument. There are many choices of limiter function but we
313     only provide the Superbee limiter \cite{Roe85}:
314     \begin{equation}
315     \psi(r) = \max[0,\min[1,2r],\min[2,r]]
316     \end{equation}
317    
318    
319     \subsection{Third order direct space time}
320    
321     The direct-space-time method deals with space and time discretization
322     together (other methods that treat space and time seperately are known
323     collectively as the ``Method of Lines''). The Lax-Wendroff scheme
324     falls into this category; it adds sufficient diffusion to a second
325     order flux that the forward-in-time method is stable. The upwind
326     biased third order DST scheme is:
327     \begin{eqnarray}
328     F = u \left( \tau_{i-1}
329     + d_0 (\tau_{i}-\tau_{i-1}) + d_1 (\tau_{i-1}-\tau_{i-2}) \right)
330     & \forall & u > 0 \\
331     F = u \left( \tau_{i}
332     - d_0 (\tau_{i}-\tau_{i-1}) - d_1 (\tau_{i+1}-\tau_{i}) \right)
333     & \forall & u < 0
334     \end{eqnarray}
335     where
336     \begin{eqnarray}
337     d_1 & = & \frac{1}{6} ( 2 - |c| ) ( 1 - |c| ) \\
338     d_2 & = & \frac{1}{6} ( 1 - |c| ) ( 1 + |c| )
339     \end{eqnarray}
340     The coefficients $d_0$ and $d_1$ approach $1/3$ and $1/6$ repectively
341     as the Courant number, $c$, vanishes. In this limit, the conventional
342     third order upwind method is recovered. For finite Courant number, the
343     deviations from the linear method are analogous to the diffusion added
344     to centered second order advection in the Lax-Wendroff scheme.
345    
346     The DST3 method described above must be used in a forward-in-time
347     manner and is stable for $0 \le |c| \le 1$. Although the scheme
348     appears to be forward-in-time, it is in fact second order in time and
349     the accuracy increases with the Courant number! For low Courant
350     number, DST3 produces very similar results (indistinguishable in
351     Fig.~\ref{fig:advect-1d-lo}) to the linear third order method but for
352     large Courant number, where the linear upwind third order method is
353     unstable, the scheme is extremely accurate
354     (Fig.~\ref{fig:advect-1d-hi}) with only minor overshoots.
355    
356     \subsection{Third order direct space time with flux limiting}
357    
358     The overshoots in the DST3 method can be controlled with a flux limiter.
359     The limited flux is written:
360     \begin{equation}
361     F =
362     \frac{1}{2}(u+|u|)\left( \tau_{i-1} + \psi(r^+)(\tau_{i} - \tau_{i-1} )\right)
363     +
364     \frac{1}{2}(u-|u|)\left( \tau_{i-1} + \psi(r^-)(\tau_{i} - \tau_{i-1} )\right)
365     \end{equation}
366     where
367     \begin{eqnarray}
368     r^+ & = & \frac{\tau_{i-1} - \tau_{i-2}}{\tau_{i} - \tau_{i-1}} \\
369     r^- & = & \frac{\tau_{i+1} - \tau_{i}}{\tau_{i} - \tau_{i-1}}
370     \end{eqnarray}
371     and the limiter is the Sweby limiter:
372     \begin{equation}
373     \psi(r) = \max[0, \min[\min(1,d_0+d_1r],\frac{1-c}{c}r ]]
374     \end{equation}
375 adcroft 1.1
376 adcroft 1.3 \subsection{Multi-dimensional advection}
377 adcroft 1.1
378 adcroft 1.3 In many of the aforementioned advection schemes the behaviour in
379     multiple dimensions is not necessarily as good as the one dimensional
380     behaviour. For instance, a shape preserving monotonic scheme in one
381     dimension can have severe shape distortion in two dimensions if the
382     two components of horizontal fluxes are treated independently. There
383     is a large body of literature on the subject dealing with this problem
384     and among the fixes are operator and flux splitting methods, corner
385     flux methods and more. We have adopted a variant on the standard
386     splitting methods that allows the flux calculations to be implemented
387     as if in one dimension:
388     \begin{eqnarray}
389     \tau^{n+1/3} & = & \tau^{n}
390     - \Delta t \left( \frac{1}{\Delta x} \delta_i F^x(\tau^{n})
391     + \tau^{n} \frac{1}{\Delta x} \delta_i u \right) \\
392     \tau^{n+2/3} & = & \tau^{n}
393     - \Delta t \left( \frac{1}{\Delta y} \delta_j F^y(\tau^{n+1/3})
394     + \tau^{n} \frac{1}{\Delta y} \delta_i v \right) \\
395     \tau^{n+3/3} & = & \tau^{n}
396     - \Delta t \left( \frac{1}{\Delta r} \delta_k F^x(\tau^{n+2/3})
397     + \tau^{n} \frac{1}{\Delta r} \delta_i w \right)
398     \end{eqnarray}
399 adcroft 1.1
400 adcroft 1.3 In order to incorporate this method into the general model algorithm,
401     we compute the effective tendancy rather than update the tracer so
402     that other terms such as diffusion are using the $n$ time-level and
403     not the updated $n+3/3$ quantities:
404     \begin{equation}
405     G^{n+1/2}_{adv} = \frac{1}{\Delta t} ( \tau^{n+3/3} - \tau^{n} )
406     \end{equation}
407     So that the over all time-stepping looks likes:
408     \begin{equation}
409     \tau^{n+1} = \tau^{n} + \Delta t \left( G^{n+1/2}_{adv} + G_{diff}(\tau^{n}) + G^{n}_{forcing} \right)
410     \end{equation}

  ViewVC Help
Powered by ViewVC 1.1.22