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

Contents of /manual/s_algorithm/text/nonlin_visc.tex

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


Revision 1.6 - (show annotations) (download) (as text)
Mon Jun 27 02:10:49 2011 UTC (14 years ago) by dfer
Branch: MAIN
CVS Tags: checkpoint01, HEAD
Changes since 1.5: +6 -6 lines
File MIME type: application/x-tex
Error occurred while calculating annotation data.
By product of the GM/Redi fix.

1 % $Header: /u/gcmpack/manual/s_algorithm/text/nonlin_visc.tex,v 1.5 2010/08/30 23:09:18 jmc Exp $
2 % $Name: $
3
4 \section{Nonlinear Viscosities for Large Eddy Simulation}
5 \label{sec:nonlin-visc}
6
7 In Large Eddy Simulations (LES), a turbulent closure needs to be
8 provided that accounts for the effects of subgridscale motions on the
9 large scale. With sufficiently powerful computers, we could resolve
10 the entire flow down to the molecular viscosity scales
11 {($L_{\nu}\approx 1 \rm cm$)}. Current computation allows perhaps
12 four decades to be resolved, so the largest problem computationally
13 feasible would be about 10m. Most oceanographic problems are much
14 larger in scale, so some form of LES is required, where only the
15 largest scales of motion are resolved, and the subgridscale's effects
16 on the large-scale are parameterized.
17
18 To formalize this process, we can introduce a filter over the
19 subgridscale L: $u_\alpha\rightarrow \BFKav u_\alpha$ and $L:
20 b\rightarrow \BFKav b$. This filter has some intrinsic length and time
21 scales, and we assume that the flow at that scale can be characterized
22 with a single velocity scale ($V$) and vertical buoyancy gradient
23 ($N^2$). The filtered equations of motion in a local Mercator
24 projection about the gridpoint in question (see Appendix for notation
25 and details of approximation) are:
26 \begin{eqnarray}
27 \BFKaDt \BFKatu- \frac{\BFKatv
28 \sin\theta}{\BFKRo\sin\theta_0}+\frac{\BFKMr}{\BFKRo}\BFKpd{x}{\BFKav\pi}
29 & = & -\left({\BFKav{\BFKDt \BFKtu}}-{\BFKaDt \BFKatu}\right)
30 +\frac{\nabla^2{\BFKatu}}{\BFKRe}\label{eq:mercat}\\
31 \BFKaDt\BFKatv+\frac{\BFKatu\sin\theta}{\BFKRo\sin\theta_0}
32 +\frac{\BFKMr}{\BFKRo}\BFKpd{y}{\BFKav\pi}
33 & = & -\left({\BFKav{\BFKDt \BFKtv}}-{\BFKaDt \BFKatv}\right)
34 +\frac{\nabla^2{\BFKatv}}{\BFKRe}\nonumber\\
35 \BFKaDt {\BFKav w} +\frac{\BFKpd{z}{\BFKav\pi}-\BFKav b}{\BFKFr^2\lambda^2}
36 & = & -\left(\BFKav{\BFKDt w}-\BFKaDt {\BFKav{w}}\right)
37 +\frac{\nabla^2\BFKav w}{\BFKRe}\nonumber\\
38 \BFKaDt{\ \BFKav b}+\BFKav w & = &
39 -\left(\BFKav{\BFKDt{b}}-\BFKaDt{\ \BFKav b} \right)
40 +\frac{\nabla^2 \BFKav b}{\Pr\BFKRe}\nonumber \\
41 \mu^2\left(\BFKpd x\BFKatu + \BFKpd y\BFKatv \right)+\BFKpd z {\BFKav w}
42 & = & 0\label{eq:cont}
43 \end{eqnarray}
44 Tildes denote multiplication by $\cos\theta/\cos\theta_0$ to account
45 for converging meridians.
46
47 The ocean is usually turbulent, and an operational definition of
48 turbulence is that the terms in parentheses (the 'eddy' terms) on the
49 right of (\ref{eq:mercat}) are of comparable magnitude to the terms on
50 the left-hand side. The terms proportional to the inverse of \BFKRe,
51 instead, are many orders of magnitude smaller than all of the other
52 terms in virtually every oceanic application.
53
54 \subsection{Eddy Viscosity}
55 A turbulent closure provides an approximation to the 'eddy' terms on
56 the right of the preceding equations. The simplest form of LES is
57 just to increase the viscosity and diffusivity until the viscous and
58 diffusive scales are resolved. That is, we approximate:
59 \begin{eqnarray}
60 \left({\BFKav{\BFKDt \BFKtu}}-{\BFKaDt \BFKatu}\right)
61 \approx\frac{\nabla^2_h{\BFKatu}}{\BFKRe_h}
62 +\frac{\BFKpds{z}{\BFKatu}}{\BFKRe_v}\label{eq:eddyvisc}, & &
63 \left({\BFKav{\BFKDt \BFKtv}}-{\BFKaDt \BFKatv}\right)
64 \approx\frac{\nabla^2_h{\BFKatv}}{\BFKRe_h}
65 +\frac{\BFKpds{z}{\BFKatv}}{\BFKRe_v}\nonumber\\
66 \left(\BFKav{\BFKDt w}-\BFKaDt {\BFKav{w}}\right)
67 \approx\frac{\nabla^2_h\BFKav w}{\BFKRe_h}
68 +\frac{\BFKpds{z}{\BFKav w}}{\BFKRe_v}\nonumber, & &
69 \left(\BFKav{\BFKDt{b}}-\BFKaDt{\ \BFKav b} \right)
70 \approx\frac{\nabla^2_h \BFKav b}{\Pr\BFKRe_h}
71 +\frac{\BFKpds{z} {\BFKav b}}{\Pr\BFKRe_v}\nonumber
72 \end{eqnarray}
73
74 \subsubsection{Reynolds-Number Limited Eddy Viscosity}
75 One way of ensuring that the gridscale is sufficiently viscous
76 (\textit{ie.} resolved) is to choose the eddy viscosity $A_h$ so that
77 the gridscale horizontal Reynolds number based on this eddy viscosity,
78 $\BFKRe_h$, to is O(1). That is, if the gridscale is to be viscous,
79 then the viscosity should be chosen to make the viscous terms as large
80 as the advective ones. Bryan \textit{et al} \cite{Bryanetal75} notes
81 that a computational mode is squelched by using $\BFKRe_h<$2.
82
83 MITgcm users can select horizontal eddy viscosities based on
84 $\BFKRe_h$ using two methods. 1) The user may estimate the velocity
85 scale expected from the calculation and grid spacing and set the {\sf
86 viscAh} to satisfy $\BFKRe_h<2$. 2) The user may use {\sf
87 viscAhReMax}, which ensures that the viscosity is always chosen so
88 that $\BFKRe_h<{\sf viscAhReMax}$. This last option should be used
89 with caution, however, since it effectively implies that viscous terms
90 are fixed in magnitude relative to advective terms. While it may be a
91 useful method for specifying a minimum viscosity with little effort,
92 tests \cite{Bryanetal75} have shown that setting {\sf viscAhReMax}=2
93 often tends to increase the viscosity substantially over other more
94 'physical' parameterizations below, especially in regions where
95 gradients of velocity are small (and thus turbulence may be weak), so
96 perhaps a more liberal value should be used, \textit{eg.} {\sf
97 viscAhReMax}=10.
98
99 While it is certainly necessary that viscosity be active at the
100 gridscale, the wavelength where dissipation of energy or enstrophy
101 occurs is not necessarily $L=A_h/U$. In fact, it is by ensuring that
102 the either the dissipation of energy in a 3-d turbulent cascade
103 (Smagorinsky) or dissipation of enstrophy in a 2-d turbulent cascade
104 (Leith) is resolved that these parameterizations derive their physical
105 meaning.
106
107 \subsubsection{Vertical Eddy Viscosities}
108 Vertical eddy viscosities are often chosen in a more subjective way,
109 as model stability is not usually as sensitive to vertical viscosity.
110 Usually the 'observed' value from finescale measurements, etc., is
111 used (\textit{eg.} {\sf viscAr}$\approx1\times10^{-4} m^2/s$). However,
112 Smagorinsky \cite{Smagorinsky93} notes that the Smagorinsky
113 parameterization of isotropic turbulence implies a value of the
114 vertical viscosity as well as the horizontal viscosity (see below).
115
116 \subsubsection{Smagorinsky Viscosity}
117 Some \cite{sm63,Smagorinsky93} suggest choosing a viscosity
118 that \emph{depends on the resolved motions}. Thus, the overall
119 viscous operator has a nonlinear dependence on velocity. Smagorinsky
120 chose his form of viscosity by considering Kolmogorov's ideas about
121 the energy spectrum of 3-d isotropic turbulence.
122
123 Kolmogorov suppposed that is that energy is injected into the flow at
124 large scales (small $k$) and is 'cascaded' or transferred
125 conservatively by nonlinear processes to smaller and smaller scales
126 until it is dissipated near the viscous scale. By setting the energy
127 flux through a particular wavenumber $k$, $\epsilon$, to be a constant
128 in $k$, there is only one combination of viscosity and energy flux
129 that has the units of length, the Kolmogorov wavelength. It is
130 $L_\epsilon(\nu)\propto\pi\epsilon^{-1/4}\nu^{3/4}$ (the $\pi$ stems
131 from conversion from wavenumber to wavelength). To ensure that this
132 viscous scale is resolved in a numerical model, the gridscale should
133 be decreased until $L_\epsilon(\nu)>L$ (so-called Direct Numerical
134 Simulation, or DNS). Alternatively, an eddy viscosity can be used and
135 the corresponding Kolmogorov length can be made larger than the
136 gridscale, $L_\epsilon(A_h)\propto\pi\epsilon^{-1/4}A_h^{3/4}$ (for
137 Large Eddy Simulation or LES).
138
139 There are two methods of ensuring that the Kolmogorov length is
140 resolved in MITgcm. 1) The user can estimate the flux of energy
141 through spectral space for a given simulation and adjust grid spacing
142 or {\sf viscAh} to ensure that $L_\epsilon(A_h)>L$. 2) The user may
143 use the approach of Smagorinsky with {\sf viscC2Smag}, which estimates
144 the energy flux at every grid point, and adjusts the viscosity
145 accordingly.
146
147 Smagorinsky formed the energy equation from the momentum equations by
148 dotting them with velocity. There are some complications when using
149 the hydrostatic approximation as described by Smagorinsky
150 \cite{Smagorinsky93}. The positive definite energy dissipation by
151 horizontal viscosity in a hydrostatic flow is $\nu D^2$, where D is
152 the deformation rate at the viscous scale. According to Kolmogorov's
153 theory, this should be a good approximation to the energy flux at any
154 wavenumber $\epsilon\approx\nu D^2$. Kolmogorov and Smagorinsky noted
155 that using an eddy viscosity that exceeds the molecular value $\nu$
156 should ensure that the energy flux through viscous scale set by the
157 eddy viscosity is the same as it would have been had we resolved all
158 the way to the true viscous scale. That is, $\epsilon\approx
159 A_{hSmag} \BFKav D^2$. If we use this approximation to estimate the
160 Kolmogorov viscous length, then
161 \begin{equation}
162 L_\epsilon(A_{hSmag})\propto\pi\epsilon^{-1/4}A_{hSmag}^{3/4}\approx\pi(A_{hSmag}
163 \BFKav D^2)^{-1/4}A_{hSmag}^{3/4} = \pi A_{hSmag}^{1/2}\BFKav D^{-1/2}
164 \end{equation}
165 To make $L_\epsilon(A_{hSmag})$ scale with the gridscale, then
166 \begin{equation}
167 A_{hSmag} = \left(\frac{{\sf viscC2Smag}}{\pi}\right)^2L^2|\BFKav D|
168 \end{equation}
169 Where the deformation rate appropriate for hydrostatic flows with
170 shallow-water scaling is
171 \begin{equation}
172 |\BFKav D|=\sqrt{\left(\BFKpd{x}{\BFKav \BFKtu}-\BFKpd{y}{\BFKav \BFKtv}\right)^2
173 +\left(\BFKpd{y}{\BFKav \BFKtu}+\BFKpd{x}{\BFKav \BFKtv}\right)^2}
174 \end{equation}
175 The coefficient {\sf viscC2Smag} is what an MITgcm user sets, and it
176 replaces the proportionality in the Kolmogorov length with an
177 equality. Others \cite{griffies:00} suggest values of {\sf viscC2Smag}
178 from 2.2 to 4 for oceanic problems. Smagorinsky \cite{Smagorinsky93}
179 shows that values from 0.2 to 0.9 have been used in atmospheric
180 modeling.
181
182 Smagorinsky \cite{Smagorinsky93} shows that a corresponding vertical
183 viscosity should be used:
184 \begin{equation}
185 A_{vSmag}=\left(\frac{{\sf viscC2Smag}}{\pi}\right)^2H^2
186 \sqrt{\left(\BFKpd{z}{\BFKav \BFKtu}\right)^2
187 +\left(\BFKpd{z}{\BFKav \BFKtv}\right)^2}
188 \end{equation}
189 This vertical viscosity is currently not implemented in MITgcm
190 (although it may be soon).
191
192 \subsubsection{Leith Viscosity}
193 Leith \cite{Leith68,Leith96} notes that 2-d turbulence is quite
194 different from 3-d. In two-dimensional turbulence, energy cascades to
195 larger scales, so there is no concern about resolving the scales of
196 energy dissipation. Instead, another quantity, enstrophy, (which is
197 the vertical component of vorticity squared) is conserved in 2-d
198 turbulence, and it cascades to smaller scales where it is dissipated.
199
200 Following a similar argument to that above about energy flux, the
201 enstrophy flux is estimated to be equal to the positive-definite
202 gridscale dissipation rate of enstrophy $\eta\approx A_{hLeith}
203 |\nabla\BFKav \omega_3|^2$. By dimensional analysis, the
204 enstrophy-dissipation scale is $L_\eta(A_{hLeith})\propto\pi
205 A_{hLeith}^{1/2}\eta^{-1/6}$. Thus, the Leith-estimated length scale
206 of enstrophy-dissipation and the resulting eddy viscosity are
207 \begin{eqnarray}
208 L_\eta(A_{hLeith})\propto\pi A_{hLeith}^{1/2}\eta^{-1/6}
209 & = & \pi A_{hLeith}^{1/3}|\nabla \BFKav \omega_3|^{-1/3} \\
210 A_{hLeith} & = &
211 \left(\frac{{\sf viscC2Leith}}{\pi}\right)^3L^3|\nabla \BFKav\omega_3| \\
212 |\nabla\omega_3| & \equiv &
213 \sqrt{\left[\BFKpd{x}{\ }
214 \left(\BFKpd{x}{\BFKav \BFKtv}-\BFKpd{y}{\BFKav
215 \BFKtu}\right)\right]^2
216 +\left[\BFKpd{y}{\ }\left(\BFKpd{x}{\BFKav \BFKtv}
217 -\BFKpd{y}{\BFKav \BFKtu}\right)\right]^2}
218 \end{eqnarray}
219
220 \subsubsection{Modified Leith Viscosity}
221 The argument above for the Leith viscosity parameterization uses
222 concepts from purely 2-dimensional turbulence, where the horizontal
223 flow field is assumed to be divergenceless. However, oceanic flows
224 are only quasi-two dimensional. While the barotropic flow, or the
225 flow within isopycnal layers may behave nearly as two-dimensional
226 turbulence, there is a possibility that these flows will be divergent.
227 In a high-resolution numerical model, these flows may be substantially
228 divergent near the grid scale, and in fact, numerical instabilities
229 exist which are only horizontally divergent and have little vertical
230 vorticity. This causes a difficulty with the Leith viscosity, which
231 can only responds to buildup of vorticity at the grid scale.
232
233 MITgcm offers two options for dealing with this problem. 1) The
234 Smagorinsky viscosity can be used instead of Leith, or in conjunction
235 with Leith--a purely divergent flow does cause an increase in
236 Smagorinsky viscosity. 2) The {\sf viscC2LeithD} parameter can be
237 set. This is a damping specifically targeting purely divergent
238 instabilities near the gridscale. The combined viscosity has the
239 form:
240 \begin{eqnarray}
241 A_{hLeith} & = &
242 L^3\sqrt{\left(\frac{{\sf viscC2Leith}}{\pi}\right)^6
243 |\nabla \BFKav \omega_3|^2
244 +\left(\frac{{\sf viscC2LeithD}}{\pi}\right)^6
245 |\nabla \nabla\cdot \BFKav {\tilde u}_h|^2} \\
246 |\nabla \nabla\cdot \BFKav {\tilde u}_h| & \equiv &
247 \sqrt{\left[\BFKpd{x}{\ }\left(\BFKpd{x}{\BFKav \BFKtu}
248 +\BFKpd{y}{\BFKav \BFKtv}\right)\right]^2
249 +\left[\BFKpd{y}{\ }\left(\BFKpd{x}{\BFKav \BFKtu}
250 +\BFKpd{y}{\BFKav \BFKtv}\right)\right]^2}
251 \end{eqnarray}
252 Whether there is any physical rationale for this correction is unclear
253 at the moment, but the numerical consequences are good. The
254 divergence in flows with the grid scale larger or comparable to the
255 Rossby radius is typically much smaller than the vorticity, so this
256 adjustment only rarely adjusts the viscosity if ${\sf
257 viscC2LeithD}={\sf viscC2Leith}$. However, the rare regions where
258 this viscosity acts are often the locations for the largest vales of
259 vertical velocity in the domain. Since the CFL condition on vertical
260 velocity is often what sets the maximum timestep, this viscosity may
261 substantially increase the allowable timestep without severely
262 compromising the verity of the simulation. Tests have shown that in
263 some calculations, a timestep three times larger was allowed when
264 ${\sf viscC2LeithD}={\sf viscC2Leith}$.
265
266 \subsubsection{Courant--Freidrichs--Lewy Constraint on Viscosity}
267 Whatever viscosities are used in the model, the choice is constrained
268 by gridscale and timestep by the Courant--Freidrichs--Lewy (CFL)
269 constraint on stability:
270 \begin{eqnarray}
271 A_h & < & \frac{L^2}{4\Delta t} \\
272 A_4 & \le & \frac{L^4}{32\Delta t}
273 \end{eqnarray}
274 The viscosities may be automatically limited to be no greater than
275 these values in MITgcm by specifying {\sf viscAhGridMax}$<1$ and
276 {\sf viscA4GridMax}$<1$. Similarly-scaled minimum values of
277 viscosities are provided by {\sf viscAhGridMin} and {\sf
278 viscA4GridMin}, which if used, should be set to values $\ll 1$. $L$
279 is roughly the gridscale (see below).
280
281 Following \cite{griffies:00}, we note that there is a factor of $\Delta
282 x^2/8$ difference between the harmonic and biharmonic viscosities.
283 Thus, whenever a non-dimensional harmonic coefficient is used in the
284 MITgcm (\textit{eg.} {\sf viscAhGridMax}$<1$), the biharmonic equivalent is
285 scaled so that the same non-dimensional value can be used (\textit{eg.} {\sf
286 viscA4GridMax}$<1$).
287
288 \subsubsection{Biharmonic Viscosity}
289 \cite{ho78} suggested that eddy viscosities ought to be focuses on
290 the dynamics at the grid scale, as larger motions would be 'resolved'.
291 To enhance the scale selectivity of the viscous operator, he suggested
292 a biharmonic eddy viscosity instead of a harmonic (or Laplacian)
293 viscosity:
294 \begin{eqnarray}
295 \left({\BFKav{\BFKDt \BFKtu}}-{\BFKaDt \BFKatu}\right)\approx
296 \frac{-\nabla^4_h{\BFKatu}}{\BFKRe_4}
297 +\frac{\BFKpds{z}{\BFKatu}}{\BFKRe_v}\label{eq:bieddyvisc}, & &
298 \left({\BFKav{\BFKDt \BFKtv}}-{\BFKaDt \BFKatv}\right)\approx
299 \frac{-\nabla^4_h{\BFKatv}}{\BFKRe_4}
300 +\frac{\BFKpds{z}{\BFKatv}}{\BFKRe_v}\nonumber\\
301 \left(\BFKav{\BFKDt w}-\BFKaDt
302 {\BFKav{w}}\right)\approx\frac{-\nabla^4_h\BFKav
303 w}{\BFKRe_4}+\frac{\BFKpds{z}{\BFKav w}}{\BFKRe_v}\nonumber, & &
304 \left(\BFKav{\BFKDt{b}}-\BFKaDt{\ \BFKav b} \right)\approx
305 \frac{-\nabla^4_h \BFKav b}{\Pr\BFKRe_4}
306 +\frac{\BFKpds{z} {\BFKav b}}{\Pr\BFKRe_v}\nonumber
307 \end{eqnarray}
308 \cite{griffies:00} propose that if one scales the biharmonic viscosity by
309 stability considerations, then the biharmonic viscous terms will be
310 similarly active to harmonic viscous terms at the gridscale of the
311 model, but much less active on larger scale motions. Similarly, a
312 biharmonic diffusivity can be used for less diffusive flows.
313
314 In practice, biharmonic viscosity and diffusivity allow a less
315 viscous, yet numerically stable, simulation than harmonic viscosity
316 and diffusivity. However, there is no physical rationale for such
317 operators being of leading order, and more boundary conditions must be
318 specified than for the harmonic operators. If one considers the
319 approximations of \ref{eq:eddyvisc} and \ref{eq:bieddyvisc} to be
320 terms in the Taylor series expansions of the eddy terms as functions
321 of the large-scale gradient, then one can argue that both harmonic and
322 biharmonic terms would occur in the series, and the only question is
323 the choice of coefficients. Using biharmonic viscosity alone implies
324 that one zeros the first non-vanishing term in the Taylor series,
325 which is unsupported by any fluid theory or observation.
326
327 Nonetheless, MITgcm supports a plethora of biharmonic viscosities
328 and diffusivities, which are controlled with parameters named
329 similarly to the harmonic viscosities and diffusivities with the
330 substitution $h\rightarrow 4$. MITgcm also supports a biharmonic
331 Leith and Smagorinsky viscosities:
332 \begin{eqnarray}
333 A_{4Smag} & = &
334 \left(\frac{{\sf viscC4Smag}}{\pi}\right)^2\frac{L^4}{8}|D| \\
335 A_{4Leith} & = &
336 \frac{L^5}{8}\sqrt{\left(\frac{{\sf viscC4Leith}}{\pi}\right)^6
337 |\nabla \BFKav \omega_3|^2
338 +\left(\frac{{\sf viscC4LeithD}}{\pi}\right)^6
339 |\nabla \nabla\cdot \BFKav {\bf \BFKtu}_h|^2}
340 \end{eqnarray}
341 However, it should be noted that unlike the harmonic forms, the
342 biharmonic scaling does not easily relate to whether
343 energy-dissipation or enstrophy-dissipation scales are resolved. If
344 similar arguments are used to estimate these scales and scale them to
345 the gridscale, the resulting biharmonic viscosities should be:
346 \begin{eqnarray}
347 A_{4Smag} & = &
348 \left(\frac{{\sf viscC4Smag}}{\pi}\right)^5L^5
349 |\nabla^2\BFKav {\bf \BFKtu}_h| \\
350 A_{4Leith} & = &
351 L^6\sqrt{\left(\frac{{\sf viscC4Leith}}{\pi}\right)^{12}
352 |\nabla^2 \BFKav \omega_3|^2
353 +\left(\frac{{\sf viscC4LeithD}}{\pi}\right)^{12}
354 |\nabla^2 \nabla\cdot \BFKav {\bf \BFKtu}_h|^2}
355 \end{eqnarray}
356 Thus, the biharmonic scaling suggested by \cite{griffies:00} implies:
357 \begin{eqnarray}
358 |D| & \propto & L|\nabla^2\BFKav {\bf \BFKtu}_h|\\
359 |\nabla \BFKav \omega_3| & \propto & L|\nabla^2 \BFKav \omega_3|
360 \end{eqnarray}
361 It is not at all clear that these assumptions ought to hold. Only the
362 \cite{griffies:00} forms are currently implemented in MITgcm.
363
364 \subsubsection{Selection of Length Scale}
365 Above, the length scale of the grid has been denoted $L$. However, in
366 strongly anisotropic grids, $L_x$ and $L_y$ will be quite different in
367 some locations. In that case, the CFL condition suggests that the
368 minimum of $L_x$ and $L_y$ be used. On the other hand, other
369 viscosities which involve whether a particular wavelength is
370 'resolved' might be better suited to use the maximum of $L_x$ and
371 $L_y$. Currently, MITgcm uses {\sf useAreaViscLength} to select
372 between two options. If false, the geometric mean of $L^2_x$ and
373 $L^2_y$ is used for all viscosities, which is closer to the minimum
374 and occurs naturally in the CFL constraint. If {\sf
375 useAreaViscLength} is true, then the square root of the area of the
376 grid cell is used.
377
378 \subsection{Mercator, Nondimensional Equations}
379 The rotating, incompressible, Boussinesq equations of motion
380 \cite{Gill1982} on a sphere can be written in Mercator projection
381 about a latitude $\theta_0$ and geopotential height $z=r-r_0$. The
382 nondimensional form of these equations is:
383 \begin{equation}
384 \BFKRo\BFKDt\BFKtu- \frac{\BFKtv
385 \sin\theta}{\sin\theta_0}+\BFKMr\BFKpd{x}{\pi}
386 +\frac{\lambda\BFKFr^2\BFKMr\cos \theta}{\mu\sin\theta_0} w
387 = -\frac{\BFKFr^2\BFKMr \BFKtu w}{r/H}
388 +\frac{\BFKRo{\bf \hat x}\cdot\nabla^2{\bf u}}{\BFKRe}
389 \end{equation}
390 \begin{equation}
391 \BFKRo\BFKDt\BFKtv+
392 \frac{\BFKtu\sin\theta}{\sin\theta_0}+\BFKMr\BFKpd{y}{\pi}
393 = -\frac{\mu\BFKRo\tan\theta(\BFKtu^2+\BFKtv^2)}{r/L}
394 -\frac{\BFKFr^2\BFKMr \BFKtv w}{r/H}
395 +\frac{\BFKRo{\bf \hat y}\cdot\nabla^2{\bf u}}{\BFKRe}
396 \end{equation}
397 \begin{eqnarray}
398 \BFKFr^2\lambda^2\BFKDt w -b+\BFKpd{z}{\pi}
399 -\frac{\lambda\cot \theta_0 \BFKtu}{\BFKMr}
400 & = & \frac{\lambda\mu^2(\BFKtu^2+\BFKtv^2)}{\BFKMr(r/L)}
401 +\frac{\BFKFr^2\lambda^2{\bf \hat z}\cdot\nabla^2{\bf u}}{\BFKRe} \\
402 \BFKDt b+w & = & \frac{\nabla^2 b}{\Pr\BFKRe}\nonumber \\
403 \mu^2\left(\BFKpd x\BFKtu + \BFKpd y\BFKtv \right)+\BFKpd z w
404 & = & 0
405 \end{eqnarray}
406 Where
407 \begin{equation}
408 \mu\equiv\frac{\cos\theta_0}{\cos\theta},\ \ \
409 \BFKtu=\frac{u^*}{V\mu},\ \ \ \BFKtv=\frac{v^*}{V\mu}
410 \end{equation}
411 %% EH3 :: This is the key bit thats messed up in the next equation
412 %% \Dt\ \equiv \mu^2\left(\tu\pd x\ +\tv \pd y\ \right)+\frac{\Fr^2\Mr}{\Ro} w\pd z
413 \begin{equation}
414 f_0\equiv2\Omega\sin\theta_0,\ \ \
415 %,\ \ \ \BFKDt\ \equiv \mu^2\left(\BFKtu\BFKpd x\
416 %+\BFKtv \BFKpd y\ \right)+\frac{\BFKFr^2\BFKMr}{\BFKRo} w\BFKpd z\
417 \frac{D}{Dt} \equiv \mu^2\left(\BFKtu\frac{\partial}{\partial x}
418 +\BFKtv \frac{\partial}{\partial y} \right)
419 +\frac{\BFKFr^2\BFKMr}{\BFKRo} w\frac{\partial}{\partial z}
420 \end{equation}
421 \begin{equation}
422 x\equiv \frac{r}{L} \phi \cos \theta_0, \ \ \
423 y\equiv \frac{r}{L} \int_{\theta_0}^\theta
424 \frac{\cos \theta_0 \BFKd \theta'}{\cos\theta'}, \ \ \
425 z\equiv \lambda\frac{r-r_0}{L}
426 \end{equation}
427 \begin{equation}
428 t^*=t \frac{L}{V},\ \ \ b^*= b\frac{V f_0\BFKMr}{\lambda}
429 \end{equation}
430 \begin{equation}
431 \pi^*=\pi V f_0 L\BFKMr,\ \ \
432 w^*=w V \frac{\BFKFr^2\lambda\BFKMr}{\BFKRo}
433 \end{equation}
434 \begin{equation}
435 \BFKRo\equiv\frac{V}{f_0 L},\ \ \ \BFKMr\equiv \max[1,\BFKRo]
436 \end{equation}
437 \begin{equation}
438 \BFKFr\equiv\frac{V}{N \lambda L}, \ \ \
439 \BFKRe\equiv\frac{VL}{\nu}, \ \ \
440 \BFKPr\equiv\frac{\nu}{\kappa}
441 \end{equation}
442 Dimensional variables are denoted by an asterisk where necessary. If
443 we filter over a grid scale typical for ocean models ($1m<L<100km$,
444 $0.0001<\lambda<1$, $0.001m/s <V<1 m/s$, $f_0<0.0001 s^{-1}$, $0.01
445 s^{-1}<N<0.0001 s^{-1}$), these equations are very well approximated
446 by
447 \begin{eqnarray}
448 \BFKRo{\BFKDt\BFKtu}- \frac{\BFKtv
449 \sin\theta}{\sin\theta_0}+\BFKMr\BFKpd{x}{\pi}
450 & =& -\frac{\lambda\BFKFr^2\BFKMr\cos \theta}{\mu\sin\theta_0} w
451 +\frac{\BFKRo\nabla^2{\BFKtu}}{\BFKRe} \\
452 \BFKRo\BFKDt\BFKtv+
453 \frac{\BFKtu\sin\theta}{\sin\theta_0}+\BFKMr\BFKpd{y}{\pi}
454 & = & \frac{\BFKRo\nabla^2{\BFKtv}}{\BFKRe} \\
455 \BFKFr^2\lambda^2\BFKDt w -b+\BFKpd{z}{\pi}
456 & = & \frac{\lambda\cot \theta_0 \BFKtu}{\BFKMr}
457 +\frac{\BFKFr^2\lambda^2\nabla^2w}{\BFKRe} \\
458 \BFKDt b+w & = & \frac{\nabla^2 b}{\Pr\BFKRe} \\
459 \mu^2\left(\BFKpd x\BFKtu + \BFKpd y\BFKtv \right)+\BFKpd z w
460 & = & 0 \\
461 \nabla^2 & \approx & \left(\frac{\partial^2}{\partial x^2}
462 +\frac{\partial^2}{\partial y^2}
463 +\frac{\partial^2}{\lambda^2\partial z^2}\right)
464 \end{eqnarray}
465 Neglecting the non-frictional terms on the right-hand side is usually
466 called the 'traditional' approximation. It is appropriate, with
467 either large aspect ratio or far from the tropics. This approximation
468 is used here, as it does not affect the form of the eddy stresses
469 which is the main topic. The frictional terms are preserved in this
470 approximate form for later comparison with eddy stresses.

  ViewVC Help
Powered by ViewVC 1.1.22