| 1 |
% $Header$ |
% $Header$ |
| 2 |
% $Name$ |
% $Name$ |
| 3 |
|
|
| 4 |
The convention used in this section is as follows: |
This chapter lays out the numerical schemes that are |
| 5 |
Time is "discretize" using a time step $\Delta t$ |
employed in the core MITgcm algorithm. Whenever possible |
| 6 |
and $\Phi^n$ refers to the variable $\Phi$ |
links are made to actual program code in the MITgcm implementation. |
| 7 |
at time $t = n \Delta t$ . We used the notation $\Phi^{(n)}$ |
The chapter begins with a discussion of the temporal discretization |
| 8 |
when time interpolation is required to estimate the value of $\phi$ |
used in MITgcm. This discussion is followed by sections that |
| 9 |
at the time $n \Delta t$. |
describe the spatial discretization. The schemes employed for momentum |
| 10 |
|
terms are described first, afterwards the schemes that apply to |
| 11 |
\section{Time integration} |
passive and dynamically active tracers are described. |
| 12 |
|
|
| 13 |
The discretization in time of the model equations (cf section I ) |
\input{part2/notation} |
| 14 |
does not depend of the discretization in space of each |
|
| 15 |
term, so that this section can be read independently. |
\section{Time-stepping} |
| 16 |
Also for this purpose, we will refers to the continuous |
\begin{rawhtml} |
| 17 |
space-derivative form of model equations, and focus on |
<!-- CMIREDIR:time-stepping: --> |
| 18 |
the time discretization. |
\end{rawhtml} |
| 19 |
|
|
| 20 |
The continuous form of the model equations is: |
The equations of motion integrated by the model involve four |
| 21 |
|
prognostic equations for flow, $u$ and $v$, temperature, $\theta$, and |
| 22 |
\begin{eqnarray} |
salt/moisture, $S$, and three diagnostic equations for vertical flow, |
| 23 |
\partial_t \theta & = & G_\theta |
$w$, density/buoyancy, $\rho$/$b$, and pressure/geo-potential, |
| 24 |
\\ |
$\phi_{hyd}$. In addition, the surface pressure or height may by |
| 25 |
\partial_t S & = & G_s |
described by either a prognostic or diagnostic equation and if |
| 26 |
\\ |
non-hydrostatics terms are included then a diagnostic equation for |
| 27 |
b' & = & b'(\theta,S,r) |
non-hydrostatic pressure is also solved. The combination of prognostic |
| 28 |
\\ |
and diagnostic equations requires a model algorithm that can march |
| 29 |
\partial_r \phi'_{hyd} & = & -b' |
forward prognostic variables while satisfying constraints imposed by |
| 30 |
\label{eq-r-split-hyd} |
diagnostic equations. |
| 31 |
\\ |
|
| 32 |
\partial_t \vec{\bf v} |
Since the model comes in several flavors and formulation, it would be |
| 33 |
+ {\bf \nabla}_h b_s \eta |
confusing to present the model algorithm exactly as written into code |
| 34 |
+ \epsilon_{nh} {\bf \nabla}_h \phi'_{nh} |
along with all the switches and optional terms. Instead, we present |
| 35 |
& = & \vec{\bf G}_{\vec{\bf v}} |
the algorithm for each of the basic formulations which are: |
| 36 |
- {\bf \nabla}_h \phi'_{hyd} |
\begin{enumerate} |
| 37 |
\label{eq-r-split-hmom} |
\item the semi-implicit pressure method for hydrostatic equations |
| 38 |
\\ |
with a rigid-lid, variables co-located in time and with |
| 39 |
\epsilon_{nh} \frac {\partial{\dot{r}}}{\partial{t}} |
Adams-Bashforth time-stepping, \label{it:a} |
| 40 |
+ \epsilon_{nh} \partial_r \phi'_{nh} |
\item as \ref{it:a}. but with an implicit linear free-surface, \label{it:b} |
| 41 |
& = & \epsilon_{nh} G_{\dot{r}} |
\item as \ref{it:a}. or \ref{it:b}. but with variables staggered in time, |
| 42 |
\label{eq-r-split-rdot} |
\label{it:c} |
| 43 |
\\ |
\item as \ref{it:a}. or \ref{it:b}. but with non-hydrostatic terms included, |
| 44 |
{\bf \nabla}_h \cdot \vec{\bf v} + \partial_r \dot{r} |
\item as \ref{it:b}. or \ref{it:c}. but with non-linear free-surface. |
| 45 |
& = & 0 |
\end{enumerate} |
| 46 |
\label{eq-r-cont} |
|
| 47 |
|
In all the above configurations it is also possible to substitute the |
| 48 |
|
Adams-Bashforth with an alternative time-stepping scheme for terms |
| 49 |
|
evaluated explicitly in time. Since the over-arching algorithm is |
| 50 |
|
independent of the particular time-stepping scheme chosen we will |
| 51 |
|
describe first the over-arching algorithm, known as the pressure |
| 52 |
|
method, with a rigid-lid model in section |
| 53 |
|
\ref{sect:pressure-method-rigid-lid}. This algorithm is essentially |
| 54 |
|
unchanged, apart for some coefficients, when the rigid lid assumption |
| 55 |
|
is replaced with a linearized implicit free-surface, described in |
| 56 |
|
section \ref{sect:pressure-method-linear-backward}. These two flavors |
| 57 |
|
of the pressure-method encompass all formulations of the model as it |
| 58 |
|
exists today. The integration of explicit in time terms is out-lined |
| 59 |
|
in section \ref{sect:adams-bashforth} and put into the context of the |
| 60 |
|
overall algorithm in sections \ref{sect:adams-bashforth-sync} and |
| 61 |
|
\ref{sect:adams-bashforth-staggered}. Inclusion of non-hydrostatic |
| 62 |
|
terms requires applying the pressure method in three dimensions |
| 63 |
|
instead of two and this algorithm modification is described in section |
| 64 |
|
\ref{sect:non-hydrostatic}. Finally, the free-surface equation may be |
| 65 |
|
treated more exactly, including non-linear terms, and this is |
| 66 |
|
described in section \ref{sect:nonlinear-freesurface}. |
| 67 |
|
|
| 68 |
|
|
| 69 |
|
\section{Pressure method with rigid-lid} |
| 70 |
|
\label{sect:pressure-method-rigid-lid} |
| 71 |
|
\begin{rawhtml} |
| 72 |
|
<!-- CMIREDIR:pressure_method_rigid_lid: --> |
| 73 |
|
\end{rawhtml} |
| 74 |
|
|
| 75 |
|
\begin{figure} |
| 76 |
|
\begin{center} |
| 77 |
|
\resizebox{4.0in}{!}{\includegraphics{part2/pressure-method-rigid-lid.eps}} |
| 78 |
|
\end{center} |
| 79 |
|
\caption{ |
| 80 |
|
A schematic of the evolution in time of the pressure method |
| 81 |
|
algorithm. A prediction for the flow variables at time level $n+1$ is |
| 82 |
|
made based only on the explicit terms, $G^{(n+^1/_2)}$, and denoted |
| 83 |
|
$u^*$, $v^*$. Next, a pressure field is found such that $u^{n+1}$, |
| 84 |
|
$v^{n+1}$ will be non-divergent. Conceptually, the $*$ quantities |
| 85 |
|
exist at time level $n+1$ but they are intermediate and only |
| 86 |
|
temporary.} |
| 87 |
|
\label{fig:pressure-method-rigid-lid} |
| 88 |
|
\end{figure} |
| 89 |
|
|
| 90 |
|
\begin{figure} |
| 91 |
|
\begin{center} \fbox{ \begin{minipage}{4.5in} \begin{tabbing} |
| 92 |
|
aaa \= aaa \= aaa \= aaa \= aaa \= aaa \kill |
| 93 |
|
\filelink{FORWARD\_STEP}{model-src-forward_step.F} \\ |
| 94 |
|
\> DYNAMICS \\ |
| 95 |
|
\>\> TIMESTEP \` $u^*$,$v^*$ (\ref{eq:ustar-rigid-lid},\ref{eq:vstar-rigid-lid}) \\ |
| 96 |
|
\> SOLVE\_FOR\_PRESSURE \\ |
| 97 |
|
\>\> CALC\_DIV\_GHAT \` $H\widehat{u^*}$,$H\widehat{v^*}$ (\ref{eq:elliptic}) \\ |
| 98 |
|
\>\> CG2D \` $\eta^{n+1}$ (\ref{eq:elliptic}) \\ |
| 99 |
|
\> MOMENTUM\_CORRECTION\_STEP \\ |
| 100 |
|
\>\> CALC\_GRAD\_PHI\_SURF \` $\nabla \eta^{n+1}$ \\ |
| 101 |
|
\>\> CORRECTION\_STEP \` $u^{n+1}$,$v^{n+1}$ (\ref{eq:un+1-rigid-lid},\ref{eq:vn+1-rigid-lid}) |
| 102 |
|
\end{tabbing} \end{minipage} } \end{center} |
| 103 |
|
\caption{Calling tree for the pressure method algorithm |
| 104 |
|
(\filelink{FORWARD\_STEP}{model-src-forward_step.F})} |
| 105 |
|
\label{fig:call-tree-pressure-method} |
| 106 |
|
\end{figure} |
| 107 |
|
|
| 108 |
|
The horizontal momentum and continuity equations for the ocean |
| 109 |
|
(\ref{eq:ocean-mom} and \ref{eq:ocean-cont}), or for the atmosphere |
| 110 |
|
(\ref{eq:atmos-mom} and \ref{eq:atmos-cont}), can be summarized by: |
| 111 |
|
\begin{eqnarray} |
| 112 |
|
\partial_t u + g \partial_x \eta & = & G_u \\ |
| 113 |
|
\partial_t v + g \partial_y \eta & = & G_v \\ |
| 114 |
|
\partial_x u + \partial_y v + \partial_z w & = & 0 |
| 115 |
\end{eqnarray} |
\end{eqnarray} |
| 116 |
where |
where we are adopting the oceanic notation for brevity. All terms in |
| 117 |
\begin{eqnarray*} |
the momentum equations, except for surface pressure gradient, are |
| 118 |
G_\theta & = & |
encapsulated in the $G$ vector. The continuity equation, when |
| 119 |
- \vec{\bf v} \cdot {\bf \nabla} \theta + {\cal Q}_\theta |
integrated over the fluid depth, $H$, and with the rigid-lid/no normal |
| 120 |
\\ |
flow boundary conditions applied, becomes: |
| 121 |
G_S & = & |
\begin{equation} |
| 122 |
- \vec{\bf v} \cdot {\bf \nabla} S + {\cal Q}_S |
\partial_x H \widehat{u} + \partial_y H \widehat{v} = 0 |
| 123 |
\\ |
\label{eq:rigid-lid-continuity} |
| 124 |
\vec{\bf G}_{\vec{\bf v}} |
\end{equation} |
| 125 |
& = & |
Here, $H\widehat{u} = \int_H u dz$ is the depth integral of $u$, |
| 126 |
- \vec{\bf v} \cdot {\bf \nabla} \vec{\bf v} |
similarly for $H\widehat{v}$. The rigid-lid approximation sets $w=0$ |
| 127 |
- f \hat{\bf k} \wedge \vec{\bf v} |
at the lid so that it does not move but allows a pressure to be |
| 128 |
+ \vec{\cal F}_{\vec{\bf v}} |
exerted on the fluid by the lid. The horizontal momentum equations and |
| 129 |
\\ |
vertically integrated continuity equation are be discretized in time |
| 130 |
G_{\dot{r}} |
and space as follows: |
| 131 |
& = & |
\begin{eqnarray} |
| 132 |
- \vec{\bf v} \cdot {\bf \nabla} \dot{r} |
u^{n+1} + \Delta t g \partial_x \eta^{n+1} |
| 133 |
+ {\cal F}_{\dot{r}} |
& = & u^{n} + \Delta t G_u^{(n+1/2)} |
| 134 |
\end{eqnarray*} |
\label{eq:discrete-time-u} |
| 135 |
The exact form of all the "{\it G}"s terms is described in the next |
\\ |
| 136 |
section (). Here its sufficient to mention that they contains |
v^{n+1} + \Delta t g \partial_y \eta^{n+1} |
| 137 |
all the RHS terms except the pressure / geo- potential terms. |
& = & v^{n} + \Delta t G_v^{(n+1/2)} |
| 138 |
|
\label{eq:discrete-time-v} |
| 139 |
The switch $\epsilon_{nh}$ allows to activate the non hydrostatic |
\\ |
| 140 |
mode ($\epsilon_{nh}=1$) for the ocean model. Otherwise, |
\partial_x H \widehat{u^{n+1}} |
| 141 |
in the hydrostatic limit $\epsilon_{nh} = 0$ |
+ \partial_y H \widehat{v^{n+1}} & = & 0 |
| 142 |
and equation \ref{eq-r-split-rdot} vanishes. |
\label{eq:discrete-time-cont-rigid-lid} |
| 143 |
|
\end{eqnarray} |
| 144 |
The equation for $\eta$ is obtained by integrating the |
As written here, terms on the LHS all involve time level $n+1$ and are |
| 145 |
continuity equation over the entire depth of the fluid, |
referred to as implicit; the implicit backward time stepping scheme is |
| 146 |
from $R_{min}(x,y)$ up to $R_o(x,y)$ |
being used. All other terms in the RHS are explicit in time. The |
| 147 |
(Linear free surface): |
thermodynamic quantities are integrated forward in time in parallel |
| 148 |
\begin{eqnarray} |
with the flow and will be discussed later. For the purposes of |
| 149 |
\epsilon_{fs} \partial_t \eta = |
describing the pressure method it suffices to say that the hydrostatic |
| 150 |
\left. \dot{r} \right|_{r=r_{surf}} + \epsilon_{fw} (P-E) = |
pressure gradient is explicit and so can be included in the vector |
| 151 |
- {\bf \nabla} \cdot \int_{R_{min}}^{R_o} \vec{\bf v} dr |
$G$. |
| 152 |
+ \epsilon_{fw} (P-E) |
|
| 153 |
\label{eq-cont-2D} |
Substituting the two momentum equations into the depth integrated |
| 154 |
\end{eqnarray} |
continuity equation eliminates $u^{n+1}$ and $v^{n+1}$ yielding an |
| 155 |
|
elliptic equation for $\eta^{n+1}$. Equations |
| 156 |
Where $\epsilon_{fs}$,$\epsilon_{fw}$ are two flags to |
\ref{eq:discrete-time-u}, \ref{eq:discrete-time-v} and |
| 157 |
distinguish between a free-surface equation ($\epsilon_{fs}=1$) |
\ref{eq:discrete-time-cont-rigid-lid} can then be re-arranged as follows: |
| 158 |
or the rigid-lid approximation ($\epsilon_{fs}=0$); |
\begin{eqnarray} |
| 159 |
and to distinguish when exchange of Fresh-Water is included |
u^{*} & = & u^{n} + \Delta t G_u^{(n+1/2)} \label{eq:ustar-rigid-lid} \\ |
| 160 |
at the ocean surface (natural BC) ($\epsilon_{fw} = 1$) |
v^{*} & = & v^{n} + \Delta t G_v^{(n+1/2)} \label{eq:vstar-rigid-lid} \\ |
| 161 |
or not ($\epsilon_{fw} = 0$). |
\partial_x \Delta t g H \partial_x \eta^{n+1} |
| 162 |
|
+ \partial_y \Delta t g H \partial_y \eta^{n+1} |
|
The hydrostatic potential is found by |
|
|
integrating \ref{eq-r-split-hyd} with the boundary condition that |
|
|
$\phi'_{hyd}(r=R_o) = 0$: |
|
|
\begin{eqnarray*} |
|
|
& & |
|
|
\int_{r'}^{R_o} \partial_r \phi'_{hyd} dr = |
|
|
\left[ \phi'_{hyd} \right]_{r'}^{R_o} = |
|
|
\int_{r'}^{R_o} - b' dr |
|
|
\\ |
|
|
\Rightarrow & & |
|
|
\phi'_{hyd}(x,y,r') = \int_{r'}^{R_o} b' dr |
|
|
\end{eqnarray*} |
|
|
|
|
|
\subsection{General method} |
|
|
|
|
|
The general algorithm consist in a "predictor step" that computes |
|
|
the forward tendencies ("G" terms") and all |
|
|
the "first guess" values star notation): |
|
|
$\vec{\bf v}^*, \theta^*, S^*$ (and $\dot{r}^*$ |
|
|
in non-hydrostatic mode). This is done in the routine {\it DYNAMICS}. |
|
|
|
|
|
Then the implicit terms that appear here on the left hand side (LHS), |
|
|
are solved as follows: |
|
|
Since implicit vertical diffusion and viscosity terms |
|
|
are independent from the barotropic flow adjustment, |
|
|
they are computed first, solving a 3 diagonal Nr x Nr linear system, |
|
|
and incorporated at the end of the {\it DYNAMICS} routines. |
|
|
Then the surface pressure and non hydrostatic pressure |
|
|
are evaluated ({\it SOLVE\_FOR\_PRESSURE}); |
|
|
|
|
|
Finally, within a "corrector step', |
|
|
(routine {\it THE\_CORRECTION\_STEP}) |
|
|
the new values of $u,v,w,\theta,S$ |
|
|
are derived according to the above equations |
|
|
(see details in II.1.3). |
|
|
|
|
|
At this point, the regular time step is over, but |
|
|
the correction step contains also other optional |
|
|
adjustments such as convective adjustment algorithm, or filters |
|
|
(zonal FFT filter, shapiro filter) |
|
|
that applied on both momentum and tracer fields. |
|
|
just before setting the $n+1$ new time step value. |
|
|
|
|
|
Since the pressure solver precision is of the order of |
|
|
the "target residual" that could be lower than the |
|
|
the computer truncation error, and also because some filters |
|
|
might alter the divergence part of the flow field, |
|
|
a final evaluation of the surface r anomaly $\eta^{n+1}$ |
|
|
is performed, according to \ref{eq-rtd-eta} ({\it CALC\_EXACT\_ETA}). |
|
|
This ensures a perfect volume conservation. |
|
|
Note that there is no need for an equivalent Non-hydrostatic |
|
|
"exact conservation" step, since W is already computed after |
|
|
the filters applied. |
|
|
|
|
|
optionnal forcing terms (package):\\ |
|
|
Regarding optional forcing terms (usually part of a "package"), |
|
|
that a account for a specific source or sink term (e.g.: condensation |
|
|
as a sink of water vapor Q), they are generally incorporated |
|
|
in the main algorithm as follows; |
|
|
At the the beginning of the time step, |
|
|
the additionnal tendencies are computed |
|
|
as function of the present state (time step $n$) and external forcing ; |
|
|
Then within the main part of model, |
|
|
only those new tendencies are added to the model variables. |
|
|
|
|
|
[more details needed]\\ |
|
|
The atmospheric physics follows this general scheme. |
|
|
|
|
|
\subsection{Standard synchronous time stepping} |
|
|
|
|
|
In the standard formulation, the surface pressure is |
|
|
evaluated at time step n+1 (implicit method). |
|
|
The above set of equations is then discretized in time |
|
|
as follows: |
|
|
\begin{eqnarray} |
|
|
\left[ 1 - \partial_r \kappa_v^\theta \partial_r \right] |
|
|
\theta^{n+1} & = & \theta^* |
|
|
\\ |
|
|
\left[ 1 - \partial_r \kappa_v^S \partial_r \right] |
|
|
S^{n+1} & = & S^* |
|
|
\\ |
|
|
%{b'}^{n} & = & b'(\theta^{n},S^{n},r) |
|
|
%\partial_r {\phi'_{hyd}}^{n} & = & {-b'}^{n} |
|
|
%\\ |
|
|
{\phi'_{hyd}}^{n} & = & \int_{r'}^{R_o} b'(\theta^{n},S^{n},r) dr |
|
|
\label{eq-rtd-hyd} |
|
|
\\ |
|
|
\vec{\bf v} ^{n+1} |
|
|
+ \Delta t {\bf \nabla}_h b_s {\eta}^{n+1} |
|
|
+ \epsilon_{nh} \Delta t {\bf \nabla} {\phi'_{nh}}^{n+1} |
|
|
- \partial_r A_v \partial_r \vec{\bf v}^{n+1} |
|
| 163 |
& = & |
& = & |
| 164 |
\vec{\bf v}^* |
\partial_x H \widehat{u^{*}} |
| 165 |
\label{eq-rtd-hmom} |
+ \partial_y H \widehat{v^{*}} \label{eq:elliptic} |
| 166 |
\\ |
\\ |
| 167 |
\epsilon_{fs} {\eta}^{n+1} + \Delta t |
u^{n+1} & = & u^{*} - \Delta t g \partial_x \eta^{n+1} \label{eq:un+1-rigid-lid}\\ |
| 168 |
{\bf \nabla}_h \cdot \int_{R_{min}}^{R_o} \vec{\bf v}^{n+1} dr |
v^{n+1} & = & v^{*} - \Delta t g \partial_y \eta^{n+1} \label{eq:vn+1-rigid-lid} |
|
& = & |
|
|
\epsilon_{fs} {\eta}^{n} + \epsilon_{fw} \Delta_t (P-E)^{n} |
|
|
\nonumber |
|
|
\\ |
|
|
% = \epsilon_{fs} {\eta}^{n} & + & \epsilon_{fw} \Delta_t (P-E)^{n} |
|
|
\label{eq-rtd-eta} |
|
|
\\ |
|
|
\epsilon_{nh} \left( \dot{r} ^{n+1} |
|
|
+ \Delta t \partial_r {\phi'_{nh}} ^{n+1} |
|
|
\right) |
|
|
& = & \epsilon_{nh} \dot{r}^* |
|
|
\label{eq-rtd-rdot} |
|
|
\\ |
|
|
{\bf \nabla}_h \cdot \vec{\bf v}^{n+1} + \partial_r \dot{r}^{n+1} |
|
|
& = & 0 |
|
|
\label{eq-rtd-cont} |
|
| 169 |
\end{eqnarray} |
\end{eqnarray} |
| 170 |
where |
Equations \ref{eq:ustar-rigid-lid} to \ref{eq:vn+1-rigid-lid}, solved |
| 171 |
|
sequentially, represent the pressure method algorithm used in the |
| 172 |
|
model. The essence of the pressure method lies in the fact that any |
| 173 |
|
explicit prediction for the flow would lead to a divergence flow field |
| 174 |
|
so a pressure field must be found that keeps the flow non-divergent |
| 175 |
|
over each step of the integration. The particular location in time of |
| 176 |
|
the pressure field is somewhat ambiguous; in |
| 177 |
|
Fig.~\ref{fig:pressure-method-rigid-lid} we depicted as co-located |
| 178 |
|
with the future flow field (time level $n+1$) but it could equally |
| 179 |
|
have been drawn as staggered in time with the flow. |
| 180 |
|
|
| 181 |
|
The correspondence to the code is as follows: |
| 182 |
|
\begin{itemize} |
| 183 |
|
\item |
| 184 |
|
the prognostic phase, equations \ref{eq:ustar-rigid-lid} and \ref{eq:vstar-rigid-lid}, |
| 185 |
|
stepping forward $u^n$ and $v^n$ to $u^{*}$ and $v^{*}$ is coded in |
| 186 |
|
\filelink{TIMESTEP()}{model-src-timestep.F} |
| 187 |
|
\item |
| 188 |
|
the vertical integration, $H \widehat{u^*}$ and $H |
| 189 |
|
\widehat{v^*}$, divergence and inversion of the elliptic operator in |
| 190 |
|
equation \ref{eq:elliptic} is coded in |
| 191 |
|
\filelink{SOLVE\_FOR\_PRESSURE()}{model-src-solve_for_pressure.F} |
| 192 |
|
\item |
| 193 |
|
finally, the new flow field at time level $n+1$ given by equations |
| 194 |
|
\ref{eq:un+1-rigid-lid} and \ref{eq:vn+1-rigid-lid} is calculated in |
| 195 |
|
\filelink{CORRECTION\_STEP()}{model-src-correction_step.F}. |
| 196 |
|
\end{itemize} |
| 197 |
|
The calling tree for these routines is given in |
| 198 |
|
Fig.~\ref{fig:call-tree-pressure-method}. |
| 199 |
|
|
| 200 |
|
|
| 201 |
|
%\paragraph{Need to discuss implicit viscosity somewhere:} |
| 202 |
|
In general, the horizontal momentum time-stepping can contain some terms |
| 203 |
|
that are treated implicitly in time, |
| 204 |
|
such as the vertical viscosity when using the backward time-stepping scheme |
| 205 |
|
(\varlink{implicitViscosity}{implicitViscosity} {\it =.TRUE.}). |
| 206 |
|
The method used to solve those implicit terms is provided in |
| 207 |
|
section \ref{sect:implicit-backward-stepping}, and modifies |
| 208 |
|
equations \ref{eq:discrete-time-u} and \ref{eq:discrete-time-v} to |
| 209 |
|
give: |
| 210 |
\begin{eqnarray} |
\begin{eqnarray} |
| 211 |
\theta^* & = & |
u^{n+1} - \Delta t \partial_z A_v \partial_z u^{n+1} |
| 212 |
\theta ^{n} + \Delta t G_{\theta} ^{(n+1/2)} |
+ \Delta t g \partial_x \eta^{n+1} & = & u^{n} + \Delta t G_u^{(n+1/2)} |
|
\\ |
|
|
S^* & = & |
|
|
S ^{n} + \Delta t G_{S} ^{(n+1/2)} |
|
|
\\ |
|
|
\vec{\bf v}^* & = & |
|
|
\vec{\bf v}^{n} + \Delta t \vec{\bf G}_{\vec{\bf v}} ^{(n+1/2)} |
|
|
+ \Delta t {\bf \nabla}_h {\phi'_{hyd}}^{(n+1/2)} |
|
| 213 |
\\ |
\\ |
| 214 |
\dot{r}^* & = & |
v^{n+1} - \Delta t \partial_z A_v \partial_z v^{n+1} |
| 215 |
\dot{r} ^{n} + \Delta t G_{\dot{r}} ^{(n+1/2)} |
+ \Delta t g \partial_y \eta^{n+1} & = & v^{n} + \Delta t G_v^{(n+1/2)} |
| 216 |
\end{eqnarray} |
\end{eqnarray} |
| 217 |
|
|
|
Note that implicit vertical terms (viscosity and diffusivity) are |
|
|
not considered as part of the "{\it G}" terms, but are |
|
|
written separately here. |
|
|
|
|
|
To ensure a second order time discretization for both |
|
|
momentum and tracer, |
|
|
The "G" terms are "extrapolated" forward in time |
|
|
(Adams Bashforth time stepping) |
|
|
from the values computed at time step $n$ and $n-1$ |
|
|
to the time $n+1/2$: |
|
|
$$G^{(n+1/2)} = G^n + (1/2+\epsilon_{AB}) (G^n - G^{n-1})$$ |
|
|
A small number for the parameter $\epsilon_{AB}$ is generally used |
|
|
to stabilize this time stepping. |
|
|
|
|
|
In the standard non-stagger formulation, |
|
|
the Adams-Bashforth time stepping is also applied to the |
|
|
hydrostatic (pressure / geo-) potential term $\nabla_h \Phi'_{hyd}$. |
|
|
Note that presently, this term is in fact incorporated to the |
|
|
$\vec{\bf G}_{\vec{\bf v}}$ arrays ({\bf gU,gV}). |
|
|
|
|
|
\subsection{Stagger baroclinic time stepping} |
|
|
|
|
|
An alternative is to evaluate $\phi'_{hyd}$ with the |
|
|
new tracer fields, and step forward the momentum after. |
|
|
This option is known as stagger baroclinic time stepping, |
|
|
since tracer and momentum are step forward in time one after the other. |
|
|
It can be activated turning on a running flag parameter |
|
|
{\bf staggerTimeStep} in file "{\it data}"). |
|
|
|
|
|
The main advantage of this time stepping compared to a synchronous one, |
|
|
is a better stability, specially regarding internal gravity waves, |
|
|
and a very natural implementation of a 2nd order in time |
|
|
hydrostatic pressure / geo- potential term. |
|
|
In the other hand, a synchronous time step might be better |
|
|
for convection problems; Its also make simpler time dependent forcing |
|
|
and diagnostic implementation ; and allows a more efficient threading. |
|
|
|
|
|
Although the stagger time step does not affect deeply the |
|
|
structure of the code --- a switch allows to evaluate the |
|
|
hydrostatic pressure / geo- potential from new $\theta,S$ |
|
|
instead of the Adams-Bashforth estimation --- |
|
|
this affect the way the time discretization is presented : |
|
| 218 |
|
|
| 219 |
\begin{eqnarray*} |
\section{Pressure method with implicit linear free-surface} |
| 220 |
\left[ 1 - \partial_r \kappa_v^\theta \partial_r \right] |
\label{sect:pressure-method-linear-backward} |
| 221 |
\theta^{n+1/2} & = & \theta^* |
\begin{rawhtml} |
| 222 |
\\ |
<!-- CMIREDIR:pressure_method_linear_backward: --> |
| 223 |
\left[ 1 - \partial_r \kappa_v^S \partial_r \right] |
\end{rawhtml} |
| 224 |
S^{n+1/2} & = & S^* |
|
| 225 |
\end{eqnarray*} |
The rigid-lid approximation filters out external gravity waves |
| 226 |
with |
subsequently modifying the dispersion relation of barotropic Rossby |
| 227 |
\begin{eqnarray*} |
waves. The discrete form of the elliptic equation has some zero |
| 228 |
\theta^* & = & |
eigen-values which makes it a potentially tricky or inefficient |
| 229 |
\theta ^{(n-1/2)} + \Delta t G_{\theta} ^{(n)} |
problem to solve. |
| 230 |
\\ |
|
| 231 |
S^* & = & |
The rigid-lid approximation can be easily replaced by a linearization |
| 232 |
S ^{(n-1/2)} + \Delta t G_{S} ^{(n)} |
of the free-surface equation which can be written: |
| 233 |
\end{eqnarray*} |
\begin{equation} |
| 234 |
And |
\partial_t \eta + \partial_x H \widehat{u} + \partial_y H \widehat{v} = P-E+R |
| 235 |
\begin{eqnarray*} |
\label{eq:linear-free-surface=P-E} |
| 236 |
%{b'}^{n+1/2} & = & b'(\theta^{n+1/2},S^{n+1/2},r) |
\end{equation} |
| 237 |
%\\ |
which differs from the depth integrated continuity equation with |
| 238 |
%\partial_r {\phi'_{hyd}}^{n+1/2} & = & {-b'}^{n+1/2} |
rigid-lid (\ref{eq:rigid-lid-continuity}) by the time-dependent term |
| 239 |
{\phi'_{hyd}}^{n+1/2} & = & \int_{r'}^{R_o} b'(\theta^{n+1/2},S^{n+1/2},r) dr |
and fresh-water source term. |
| 240 |
%\label{eq-rtd-hyd} |
|
| 241 |
\\ |
Equation \ref{eq:discrete-time-cont-rigid-lid} in the rigid-lid |
| 242 |
\vec{\bf v} ^{n+1} |
pressure method is then replaced by the time discretization of |
| 243 |
+ \Delta t {\bf \nabla}_h b_s {\eta}^{n+1} |
\ref{eq:linear-free-surface=P-E} which is: |
| 244 |
+ \epsilon_{nh} \Delta t {\bf \nabla}_h {\phi'_{nh}}^{n+1} |
\begin{equation} |
| 245 |
- \partial_r A_v \partial_r \vec{\bf v}^{n+1} |
\eta^{n+1} |
| 246 |
& = & |
+ \Delta t \partial_x H \widehat{u^{n+1}} |
| 247 |
\vec{\bf v}^* |
+ \Delta t \partial_y H \widehat{v^{n+1}} |
| 248 |
%\label{eq-rtd-hmom} |
= |
| 249 |
|
\eta^{n} |
| 250 |
|
+ \Delta t ( P - E ) |
| 251 |
|
\label{eq:discrete-time-backward-free-surface} |
| 252 |
|
\end{equation} |
| 253 |
|
where the use of flow at time level $n+1$ makes the method implicit |
| 254 |
|
and backward in time. This is the preferred scheme since it still |
| 255 |
|
filters the fast unresolved wave motions by damping them. A centered |
| 256 |
|
scheme, such as Crank-Nicholson (see section \ref{sect:freesurf-CrankNick}), |
| 257 |
|
would alias the energy of the fast modes onto slower modes of motion. |
| 258 |
|
|
| 259 |
|
As for the rigid-lid pressure method, equations |
| 260 |
|
\ref{eq:discrete-time-u}, \ref{eq:discrete-time-v} and |
| 261 |
|
\ref{eq:discrete-time-backward-free-surface} can be re-arranged as follows: |
| 262 |
|
\begin{eqnarray} |
| 263 |
|
u^{*} & = & u^{n} + \Delta t G_u^{(n+1/2)} \label{eq:ustar-backward-free-surface} \\ |
| 264 |
|
v^{*} & = & v^{n} + \Delta t G_v^{(n+1/2)} \label{eq:vstar-backward-free-surface} \\ |
| 265 |
|
\eta^* & = & \epsilon_{fs} \left( \eta^{n} + \Delta t (P-E) \right)- \Delta t |
| 266 |
|
\partial_x H \widehat{u^{*}} |
| 267 |
|
+ \partial_y H \widehat{v^{*}} |
| 268 |
|
\\ |
| 269 |
|
\partial_x g H \partial_x \eta^{n+1} |
| 270 |
|
& + & \partial_y g H \partial_y \eta^{n+1} |
| 271 |
|
- \frac{\epsilon_{fs} \eta^{n+1}}{\Delta t^2} |
| 272 |
|
= |
| 273 |
|
- \frac{\eta^*}{\Delta t^2} |
| 274 |
|
\label{eq:elliptic-backward-free-surface} |
| 275 |
\\ |
\\ |
| 276 |
\epsilon_{fs} {\eta}^{n+1} + \Delta t |
u^{n+1} & = & u^{*} - \Delta t g \partial_x \eta^{n+1} \label{eq:un+1-backward-free-surface}\\ |
| 277 |
{\bf \nabla}_h \cdot \int_{R_{min}}^{R_o} \vec{\bf v}^{n+1} dr |
v^{n+1} & = & v^{*} - \Delta t g \partial_y \eta^{n+1} \label{eq:vn+1-backward-free-surface} |
| 278 |
& = & |
\end{eqnarray} |
| 279 |
\epsilon_{fs} {\eta}^{n} + \epsilon_{fw} \Delta_t (P-E)^{n} |
Equations~\ref{eq:ustar-backward-free-surface} |
| 280 |
\\ |
to~\ref{eq:vn+1-backward-free-surface}, solved sequentially, represent |
| 281 |
\epsilon_{nh} \left( \dot{r} ^{n+1} |
the pressure method algorithm with a backward implicit, linearized |
| 282 |
+ \Delta t \partial_r {\phi'_{nh}} ^{n+1} |
free surface. The method is still formerly a pressure method because |
| 283 |
\right) |
in the limit of large $\Delta t$ the rigid-lid method is |
| 284 |
& = & \epsilon_{nh} \dot{r}^* |
recovered. However, the implicit treatment of the free-surface allows |
| 285 |
%\label{eq-rtd-rdot} |
the flow to be divergent and for the surface pressure/elevation to |
| 286 |
\\ |
respond on a finite time-scale (as opposed to instantly). To recover |
| 287 |
{\bf \nabla}_h \cdot \vec{\bf v}^{n+1} + \partial_r \dot{r}^{n+1} |
the rigid-lid formulation, we introduced a switch-like parameter, |
| 288 |
& = & 0 |
$\epsilon_{fs}$ (\varlink{freesurfFac}{freesurfFac}), |
| 289 |
%\label{eq-rtd-cont} |
which selects between the free-surface and rigid-lid; |
| 290 |
\end{eqnarray*} |
$\epsilon_{fs}=1$ allows the free-surface to evolve; $\epsilon_{fs}=0$ |
| 291 |
with |
imposes the rigid-lid. The evolution in time and location of variables |
| 292 |
|
is exactly as it was for the rigid-lid model so that |
| 293 |
|
Fig.~\ref{fig:pressure-method-rigid-lid} is still |
| 294 |
|
applicable. Similarly, the calling sequence, given in |
| 295 |
|
Fig.~\ref{fig:call-tree-pressure-method}, is as for the |
| 296 |
|
pressure-method. |
| 297 |
|
|
| 298 |
|
|
| 299 |
|
\section{Explicit time-stepping: Adams-Bashforth} |
| 300 |
|
\label{sect:adams-bashforth} |
| 301 |
|
\begin{rawhtml} |
| 302 |
|
<!-- CMIREDIR:adams_bashforth: --> |
| 303 |
|
\end{rawhtml} |
| 304 |
|
|
| 305 |
|
In describing the the pressure method above we deferred describing the |
| 306 |
|
time discretization of the explicit terms. We have historically used |
| 307 |
|
the quasi-second order Adams-Bashforth method for all explicit terms |
| 308 |
|
in both the momentum and tracer equations. This is still the default |
| 309 |
|
mode of operation but it is now possible to use alternate schemes for |
| 310 |
|
tracers (see section \ref{sect:tracer-advection}). |
| 311 |
|
|
| 312 |
|
\begin{figure} |
| 313 |
|
\begin{center} \fbox{ \begin{minipage}{4.5in} \begin{tabbing} |
| 314 |
|
aaa \= aaa \= aaa \= aaa \= aaa \= aaa \kill |
| 315 |
|
FORWARD\_STEP \\ |
| 316 |
|
\> THERMODYNAMICS \\ |
| 317 |
|
\>\> CALC\_GT \\ |
| 318 |
|
\>\>\> GAD\_CALC\_RHS \` $G_\theta^n = G_\theta( u, \theta^n )$ \\ |
| 319 |
|
\>either\>\> EXTERNAL\_FORCING \` $G_\theta^n = G_\theta^n + {\cal Q}$ \\ |
| 320 |
|
\>\>\> ADAMS\_BASHFORTH2 \` $G_\theta^{(n+1/2)}$ (\ref{eq:adams-bashforth2}) \\ |
| 321 |
|
\>or\>\> EXTERNAL\_FORCING \` $G_\theta^{(n+1/2)} = G_\theta^{(n+1/2)} + {\cal Q}$ \\ |
| 322 |
|
\>\> TIMESTEP\_TRACER \` $\tau^*$ (\ref{eq:taustar}) \\ |
| 323 |
|
\>\> IMPLDIFF \` $\tau^{(n+1)}$ (\ref{eq:tau-n+1-implicit}) |
| 324 |
|
\end{tabbing} \end{minipage} } \end{center} |
| 325 |
|
\caption{ |
| 326 |
|
Calling tree for the Adams-Bashforth time-stepping of temperature with |
| 327 |
|
implicit diffusion. |
| 328 |
|
(\filelink{THERMODYNAMICS}{model-src-thermodynamics.F}, |
| 329 |
|
\filelink{ADAMS\_BASHFORTH2}{model-src-adams_bashforth2.F})} |
| 330 |
|
\label{fig:call-tree-adams-bashforth} |
| 331 |
|
\end{figure} |
| 332 |
|
|
| 333 |
|
In the previous sections, we summarized an explicit scheme as: |
| 334 |
|
\begin{equation} |
| 335 |
|
\tau^{*} = \tau^{n} + \Delta t G_\tau^{(n+1/2)} |
| 336 |
|
\label{eq:taustar} |
| 337 |
|
\end{equation} |
| 338 |
|
where $\tau$ could be any prognostic variable ($u$, $v$, $\theta$ or |
| 339 |
|
$S$) and $\tau^*$ is an explicit estimate of $\tau^{n+1}$ and would be |
| 340 |
|
exact if not for implicit-in-time terms. The parenthesis about $n+1/2$ |
| 341 |
|
indicates that the term is explicit and extrapolated forward in time |
| 342 |
|
and for this we use the quasi-second order Adams-Bashforth method: |
| 343 |
|
\begin{equation} |
| 344 |
|
G_\tau^{(n+1/2)} = ( 3/2 + \epsilon_{AB}) G_\tau^n |
| 345 |
|
- ( 1/2 + \epsilon_{AB}) G_\tau^{n-1} |
| 346 |
|
\label{eq:adams-bashforth2} |
| 347 |
|
\end{equation} |
| 348 |
|
This is a linear extrapolation, forward in time, to |
| 349 |
|
$t=(n+1/2+{\epsilon_{AB}})\Delta t$. An extrapolation to the mid-point |
| 350 |
|
in time, $t=(n+1/2)\Delta t$, corresponding to $\epsilon_{AB}=0$, |
| 351 |
|
would be second order accurate but is weakly unstable for oscillatory |
| 352 |
|
terms. A small but finite value for $\epsilon_{AB}$ stabilizes the |
| 353 |
|
method. Strictly speaking, damping terms such as diffusion and |
| 354 |
|
dissipation, and fixed terms (forcing), do not need to be inside the |
| 355 |
|
Adams-Bashforth extrapolation. However, in the current code, it is |
| 356 |
|
simpler to include these terms and this can be justified if the flow |
| 357 |
|
and forcing evolves smoothly. Problems can, and do, arise when forcing |
| 358 |
|
or motions are high frequency and this corresponds to a reduced |
| 359 |
|
stability compared to a simple forward time-stepping of such terms. |
| 360 |
|
The model offers the possibility to leave the tracer and momentum |
| 361 |
|
forcing terms and the dissipation terms outside the |
| 362 |
|
Adams-Bashforth extrapolation, by turning off the logical flags |
| 363 |
|
\varlink{forcing\_In\_AB}{forcing_In_AB} |
| 364 |
|
(parameter file {\em data}, namelist {\em PARM01}, default value = True). |
| 365 |
|
(\varlink{tracForcingOutAB}{tracForcingOutAB}, default=0, |
| 366 |
|
\varlink{momForcingOutAB}{momForcingOutAB}, default=0) |
| 367 |
|
and \varlink{momDissip\_In\_AB}{momDissip_In_AB} |
| 368 |
|
(parameter file {\em data}, namelist {\em PARM01}, default value = True). |
| 369 |
|
respectively. |
| 370 |
|
|
| 371 |
|
A stability analysis for an oscillation equation should be given at this point. |
| 372 |
|
\marginpar{AJA needs to find his notes on this...} |
| 373 |
|
|
| 374 |
|
A stability analysis for a relaxation equation should be given at this point. |
| 375 |
|
\marginpar{...and for this too.} |
| 376 |
|
|
| 377 |
|
\begin{figure} |
| 378 |
|
\begin{center} |
| 379 |
|
\resizebox{5.5in}{!}{\includegraphics{part2/oscil+damp_AB2.eps}} |
| 380 |
|
\end{center} |
| 381 |
|
\caption{ |
| 382 |
|
Oscillatory and damping response of |
| 383 |
|
quasi-second order Adams-Bashforth scheme for different values |
| 384 |
|
of the $\epsilon_{AB}$ parameter (0., 0.1, 0.25, from top to bottom) |
| 385 |
|
The analytical solution (in black), the physical mode (in blue) |
| 386 |
|
and the numerical mode (in red) are represented with a CFL |
| 387 |
|
step of 0.1. |
| 388 |
|
The left column represents the oscillatory response |
| 389 |
|
on the complex plane for CFL ranging from 0.1 up to 0.9. |
| 390 |
|
The right column represents the damping response amplitude |
| 391 |
|
(y-axis) function of the CFL (x-axis). |
| 392 |
|
} |
| 393 |
|
\label{fig:adams-bashforth-respons} |
| 394 |
|
\end{figure} |
| 395 |
|
|
| 396 |
|
|
| 397 |
|
|
| 398 |
|
\section{Implicit time-stepping: backward method} |
| 399 |
|
\label{sect:implicit-backward-stepping} |
| 400 |
|
\begin{rawhtml} |
| 401 |
|
<!-- CMIREDIR:implicit_time-stepping_backward: --> |
| 402 |
|
\end{rawhtml} |
| 403 |
|
|
| 404 |
|
Vertical diffusion and viscosity can be treated implicitly in time |
| 405 |
|
using the backward method which is an intrinsic scheme. |
| 406 |
|
Recently, the option to treat the vertical advection |
| 407 |
|
implicitly has been added, but not yet tested; therefore, |
| 408 |
|
the description hereafter is limited to diffusion and viscosity. |
| 409 |
|
For tracers, |
| 410 |
|
the time discretized equation is: |
| 411 |
|
\begin{equation} |
| 412 |
|
\tau^{n+1} - \Delta t \partial_r \kappa_v \partial_r \tau^{n+1} = |
| 413 |
|
\tau^{n} + \Delta t G_\tau^{(n+1/2)} |
| 414 |
|
\label{eq:implicit-diffusion} |
| 415 |
|
\end{equation} |
| 416 |
|
where $G_\tau^{(n+1/2)}$ is the remaining explicit terms extrapolated |
| 417 |
|
using the Adams-Bashforth method as described above. Equation |
| 418 |
|
\ref{eq:implicit-diffusion} can be split split into: |
| 419 |
|
\begin{eqnarray} |
| 420 |
|
\tau^* & = & \tau^{n} + \Delta t G_\tau^{(n+1/2)} |
| 421 |
|
\label{eq:taustar-implicit} \\ |
| 422 |
|
\tau^{n+1} & = & {\cal L}_\tau^{-1} ( \tau^* ) |
| 423 |
|
\label{eq:tau-n+1-implicit} |
| 424 |
|
\end{eqnarray} |
| 425 |
|
where ${\cal L}_\tau^{-1}$ is the inverse of the operator |
| 426 |
|
\begin{equation} |
| 427 |
|
{\cal L} = \left[ 1 + \Delta t \partial_r \kappa_v \partial_r \right] |
| 428 |
|
\end{equation} |
| 429 |
|
Equation \ref{eq:taustar-implicit} looks exactly as \ref{eq:taustar} |
| 430 |
|
while \ref{eq:tau-n+1-implicit} involves an operator or matrix |
| 431 |
|
inversion. By re-arranging \ref{eq:implicit-diffusion} in this way we |
| 432 |
|
have cast the method as an explicit prediction step and an implicit |
| 433 |
|
step allowing the latter to be inserted into the over all algorithm |
| 434 |
|
with minimal interference. |
| 435 |
|
|
| 436 |
|
Fig.~\ref{fig:call-tree-adams-bashforth} shows the calling sequence for |
| 437 |
|
stepping forward a tracer variable such as temperature. |
| 438 |
|
|
| 439 |
|
In order to fit within the pressure method, the implicit viscosity |
| 440 |
|
must not alter the barotropic flow. In other words, it can only |
| 441 |
|
redistribute momentum in the vertical. The upshot of this is that |
| 442 |
|
although vertical viscosity may be backward implicit and |
| 443 |
|
unconditionally stable, no-slip boundary conditions may not be made |
| 444 |
|
implicit and are thus cast as a an explicit drag term. |
| 445 |
|
|
| 446 |
|
\section{Synchronous time-stepping: variables co-located in time} |
| 447 |
|
\label{sect:adams-bashforth-sync} |
| 448 |
|
\begin{rawhtml} |
| 449 |
|
<!-- CMIREDIR:adams_bashforth_sync: --> |
| 450 |
|
\end{rawhtml} |
| 451 |
|
|
| 452 |
|
\begin{figure} |
| 453 |
|
\begin{center} |
| 454 |
|
\resizebox{5.0in}{!}{\includegraphics{part2/adams-bashforth-sync.eps}} |
| 455 |
|
\end{center} |
| 456 |
|
\caption{ |
| 457 |
|
A schematic of the explicit Adams-Bashforth and implicit time-stepping |
| 458 |
|
phases of the algorithm. All prognostic variables are co-located in |
| 459 |
|
time. Explicit tendencies are evaluated at time level $n$ as a |
| 460 |
|
function of the state at that time level (dotted arrow). The explicit |
| 461 |
|
tendency from the previous time level, $n-1$, is used to extrapolate |
| 462 |
|
tendencies to $n+1/2$ (dashed arrow). This extrapolated tendency |
| 463 |
|
allows variables to be stably integrated forward-in-time to render an |
| 464 |
|
estimate ($*$-variables) at the $n+1$ time level (solid |
| 465 |
|
arc-arrow). The operator ${\cal L}$ formed from implicit-in-time terms |
| 466 |
|
is solved to yield the state variables at time level $n+1$. } |
| 467 |
|
\label{fig:adams-bashforth-sync} |
| 468 |
|
\end{figure} |
| 469 |
|
|
| 470 |
|
\begin{figure} |
| 471 |
|
\begin{center} \fbox{ \begin{minipage}{4.7in} \begin{tabbing} |
| 472 |
|
aaa \= aaa \= aaa \= aaa \= aaa \= aaa \kill |
| 473 |
|
FORWARD\_STEP \\ |
| 474 |
|
\>\> EXTERNAL\_FIELDS\_LOAD\\ |
| 475 |
|
\>\> DO\_ATMOSPHERIC\_PHYS \\ |
| 476 |
|
\>\> DO\_OCEANIC\_PHYS \\ |
| 477 |
|
\> THERMODYNAMICS \\ |
| 478 |
|
\>\> CALC\_GT \\ |
| 479 |
|
\>\>\> GAD\_CALC\_RHS \` $G_\theta^n = G_\theta( u, \theta^n )$ (\ref{eq:Gt-n-sync})\\ |
| 480 |
|
\>\>\> EXTERNAL\_FORCING \` $G_\theta^n = G_\theta^n + {\cal Q}$ \\ |
| 481 |
|
\>\>\> ADAMS\_BASHFORTH2 \` $G_\theta^{(n+1/2)}$ (\ref{eq:Gt-n+5-sync}) \\ |
| 482 |
|
\>\> TIMESTEP\_TRACER \` $\theta^*$ (\ref{eq:tstar-sync}) \\ |
| 483 |
|
\>\> IMPLDIFF \` $\theta^{(n+1)}$ (\ref{eq:t-n+1-sync}) \\ |
| 484 |
|
\> DYNAMICS \\ |
| 485 |
|
\>\> CALC\_PHI\_HYD \` $\phi_{hyd}^n$ (\ref{eq:phi-hyd-sync}) \\ |
| 486 |
|
\>\> MOM\_FLUXFORM or MOM\_VECINV \` $G_{\vec{\bf v}}^n$ (\ref{eq:Gv-n-sync})\\ |
| 487 |
|
\>\> TIMESTEP \` $\vec{\bf v}^*$ (\ref{eq:Gv-n+5-sync}, \ref{eq:vstar-sync}) \\ |
| 488 |
|
\>\> IMPLDIFF \` $\vec{\bf v}^{**}$ (\ref{eq:vstarstar-sync}) \\ |
| 489 |
|
\> UPDATE\_R\_STAR or UPDATE\_SURF\_DR \` (NonLin-FS only)\\ |
| 490 |
|
\> SOLVE\_FOR\_PRESSURE \\ |
| 491 |
|
\>\> CALC\_DIV\_GHAT \` $\eta^*$ (\ref{eq:nstar-sync}) \\ |
| 492 |
|
\>\> CG2D \` $\eta^{n+1}$ (\ref{eq:elliptic-sync}) \\ |
| 493 |
|
\> MOMENTUM\_CORRECTION\_STEP \\ |
| 494 |
|
\>\> CALC\_GRAD\_PHI\_SURF \` $\nabla \eta^{n+1}$ \\ |
| 495 |
|
\>\> CORRECTION\_STEP \` $u^{n+1}$,$v^{n+1}$ (\ref{eq:v-n+1-sync})\\ |
| 496 |
|
\> TRACERS\_CORRECTION\_STEP \\ |
| 497 |
|
\>\> CYCLE\_TRACER \` $\theta^{n+1}$ \\ |
| 498 |
|
\>\> FILTER \` Shapiro Filter, Zonal Filter (FFT) \\ |
| 499 |
|
\>\> CONVECTIVE\_ADJUSTMENT \` \\ |
| 500 |
|
\end{tabbing} \end{minipage} } \end{center} |
| 501 |
|
\caption{ |
| 502 |
|
Calling tree for the overall synchronous algorithm using |
| 503 |
|
Adams-Bashforth time-stepping. |
| 504 |
|
The place where the model geometry |
| 505 |
|
({\bf hFac} factors) is updated is added here but is only relevant |
| 506 |
|
for the non-linear free-surface algorithm. |
| 507 |
|
For completeness, the external forcing, |
| 508 |
|
ocean and atmospheric physics have been added, although they are mainly |
| 509 |
|
optional} |
| 510 |
|
\label{fig:call-tree-adams-bashforth-sync} |
| 511 |
|
\end{figure} |
| 512 |
|
|
| 513 |
|
The Adams-Bashforth extrapolation of explicit tendencies fits neatly |
| 514 |
|
into the pressure method algorithm when all state variables are |
| 515 |
|
co-located in time. Fig.~\ref{fig:adams-bashforth-sync} illustrates |
| 516 |
|
the location of variables in time and the evolution of the algorithm |
| 517 |
|
with time. The algorithm can be represented by the sequential solution |
| 518 |
|
of the follow equations: |
| 519 |
|
\begin{eqnarray} |
| 520 |
|
G_{\theta,S}^{n} & = & G_{\theta,S} ( u^n, \theta^n, S^n ) |
| 521 |
|
\label{eq:Gt-n-sync} \\ |
| 522 |
|
G_{\theta,S}^{(n+1/2)} & = & (3/2+\epsilon_{AB}) G_{\theta,S}^{n}-(1/2+\epsilon_{AB}) G_{\theta,S}^{n-1} |
| 523 |
|
\label{eq:Gt-n+5-sync} \\ |
| 524 |
|
(\theta^*,S^*) & = & (\theta^{n},S^{n}) + \Delta t G_{\theta,S}^{(n+1/2)} |
| 525 |
|
\label{eq:tstar-sync} \\ |
| 526 |
|
(\theta^{n+1},S^{n+1}) & = & {\cal L}^{-1}_{\theta,S} (\theta^*,S^*) |
| 527 |
|
\label{eq:t-n+1-sync} \\ |
| 528 |
|
\phi^n_{hyd} & = & \int b(\theta^n,S^n) dr |
| 529 |
|
\label{eq:phi-hyd-sync} \\ |
| 530 |
|
\vec{\bf G}_{\vec{\bf v}}^{n} & = & \vec{\bf G}_{\vec{\bf v}} ( \vec{\bf v}^n, \phi^n_{hyd} ) |
| 531 |
|
\label{eq:Gv-n-sync} \\ |
| 532 |
|
\vec{\bf G}_{\vec{\bf v}}^{(n+1/2)} & = & (3/2 + \epsilon_{AB} ) \vec{\bf G}_{\vec{\bf v}}^{n} - (1/2 + \epsilon_{AB} ) \vec{\bf G}_{\vec{\bf v}}^{n-1} |
| 533 |
|
\label{eq:Gv-n+5-sync} \\ |
| 534 |
|
\vec{\bf v}^{*} & = & \vec{\bf v}^{n} + \Delta t \vec{\bf G}_{\vec{\bf v}}^{(n+1/2)} |
| 535 |
|
\label{eq:vstar-sync} \\ |
| 536 |
|
\vec{\bf v}^{**} & = & {\cal L}_{\vec{\bf v}}^{-1} ( \vec{\bf v}^* ) |
| 537 |
|
\label{eq:vstarstar-sync} \\ |
| 538 |
|
\eta^* & = & \epsilon_{fs} \left( \eta^{n} + \Delta t (P-E) \right)- \Delta t |
| 539 |
|
\nabla \cdot H \widehat{ \vec{\bf v}^{**} } |
| 540 |
|
\label{eq:nstar-sync} \\ |
| 541 |
|
\nabla \cdot g H \nabla \eta^{n+1} & - & \frac{\epsilon_{fs} \eta^{n+1}}{\Delta t^2} |
| 542 |
|
~ = ~ - \frac{\eta^*}{\Delta t^2} |
| 543 |
|
\label{eq:elliptic-sync} \\ |
| 544 |
|
\vec{\bf v}^{n+1} & = & \vec{\bf v}^{*} - \Delta t g \nabla \eta^{n+1} |
| 545 |
|
\label{eq:v-n+1-sync} |
| 546 |
|
\end{eqnarray} |
| 547 |
|
Fig.~\ref{fig:adams-bashforth-sync} illustrates the location of |
| 548 |
|
variables in time and evolution of the algorithm with time. The |
| 549 |
|
Adams-Bashforth extrapolation of the tracer tendencies is illustrated |
| 550 |
|
by the dashed arrow, the prediction at $n+1$ is indicated by the |
| 551 |
|
solid arc. Inversion of the implicit terms, ${\cal |
| 552 |
|
L}^{-1}_{\theta,S}$, then yields the new tracer fields at $n+1$. All |
| 553 |
|
these operations are carried out in subroutine {\em THERMODYNAMICS} an |
| 554 |
|
subsidiaries, which correspond to equations \ref{eq:Gt-n-sync} to |
| 555 |
|
\ref{eq:t-n+1-sync}. |
| 556 |
|
Similarly illustrated is the Adams-Bashforth extrapolation of |
| 557 |
|
accelerations, stepping forward and solving of implicit viscosity and |
| 558 |
|
surface pressure gradient terms, corresponding to equations |
| 559 |
|
\ref{eq:Gv-n-sync} to \ref{eq:v-n+1-sync}. |
| 560 |
|
These operations are carried out in subroutines {\em DYNAMCIS}, {\em |
| 561 |
|
SOLVE\_FOR\_PRESSURE} and {\em MOMENTUM\_CORRECTION\_STEP}. This, then, |
| 562 |
|
represents an entire algorithm for stepping forward the model one |
| 563 |
|
time-step. The corresponding calling tree is given in |
| 564 |
|
\ref{fig:call-tree-adams-bashforth-sync}. |
| 565 |
|
|
| 566 |
|
\section{Staggered baroclinic time-stepping} |
| 567 |
|
\label{sect:adams-bashforth-staggered} |
| 568 |
|
\begin{rawhtml} |
| 569 |
|
<!-- CMIREDIR:adams_bashforth_staggered: --> |
| 570 |
|
\end{rawhtml} |
| 571 |
|
|
| 572 |
|
\begin{figure} |
| 573 |
|
\begin{center} |
| 574 |
|
\resizebox{5.5in}{!}{\includegraphics{part2/adams-bashforth-staggered.eps}} |
| 575 |
|
\end{center} |
| 576 |
|
\caption{ |
| 577 |
|
A schematic of the explicit Adams-Bashforth and implicit time-stepping |
| 578 |
|
phases of the algorithm but with staggering in time of thermodynamic |
| 579 |
|
variables with the flow. |
| 580 |
|
Explicit momentum tendencies are evaluated at time level $n-1/2$ as a |
| 581 |
|
function of the flow field at that time level $n-1/2$. |
| 582 |
|
The explicit tendency from the previous time level, $n-3/2$, is used to |
| 583 |
|
extrapolate tendencies to $n$ (dashed arrow). |
| 584 |
|
The hydrostatic pressure/geo-potential $\phi_{hyd}$ is evaluated directly |
| 585 |
|
at time level $n$ (vertical arrows) and used with the extrapolated tendencies |
| 586 |
|
to step forward the flow variables from $n-1/2$ to $n+1/2$ (solid arc-arrow). |
| 587 |
|
The implicit-in-time operator ${\cal L}_{\bf u,v}$ (vertical arrows) is |
| 588 |
|
then applied to the previous estimation of the the flow field ($*$-variables) |
| 589 |
|
and yields to the two velocity components $u,v$ at time level $n+1/2$. |
| 590 |
|
These are then used to calculate the advection term (dashed arc-arrow) |
| 591 |
|
of the thermo-dynamics tendencies at time step $n$. |
| 592 |
|
The extrapolated thermodynamics tendency, from time level $n-1$ and $n$ |
| 593 |
|
to $n+1/2$, allows thermodynamic variables to be stably integrated |
| 594 |
|
forward-in-time (solid arc-arrow) up to time level $n+1$. |
| 595 |
|
} |
| 596 |
|
\label{fig:adams-bashforth-staggered} |
| 597 |
|
\end{figure} |
| 598 |
|
|
| 599 |
|
For well stratified problems, internal gravity waves may be the |
| 600 |
|
limiting process for determining a stable time-step. In the |
| 601 |
|
circumstance, it is more efficient to stagger in time the |
| 602 |
|
thermodynamic variables with the flow |
| 603 |
|
variables. Fig.~\ref{fig:adams-bashforth-staggered} illustrates the |
| 604 |
|
staggering and algorithm. The key difference between this and |
| 605 |
|
Fig.~\ref{fig:adams-bashforth-sync} is that the thermodynamic variables |
| 606 |
|
are solved after the dynamics, using the recently updated flow field. |
| 607 |
|
This essentially allows the gravity wave terms to leap-frog in |
| 608 |
|
time giving second order accuracy and more stability. |
| 609 |
|
|
| 610 |
|
The essential change in the staggered algorithm is that the |
| 611 |
|
thermodynamics solver is delayed from half a time step, |
| 612 |
|
allowing the use of the most recent velocities to compute |
| 613 |
|
the advection terms. Once the thermodynamics fields are |
| 614 |
|
updated, the hydrostatic pressure is computed |
| 615 |
|
to step forwrad the dynamics. |
| 616 |
|
Note that the pressure gradient must also be taken out of the |
| 617 |
|
Adams-Bashforth extrapolation. Also, retaining the integer time-levels, |
| 618 |
|
$n$ and $n+1$, does not give a user the sense of where variables are |
| 619 |
|
located in time. Instead, we re-write the entire algorithm, |
| 620 |
|
\ref{eq:Gt-n-sync} to \ref{eq:v-n+1-sync}, annotating the |
| 621 |
|
position in time of variables appropriately: |
| 622 |
|
\begin{eqnarray} |
| 623 |
|
\phi^{n}_{hyd} & = & \int b(\theta^{n},S^{n}) dr |
| 624 |
|
\label{eq:phi-hyd-staggered} \\ |
| 625 |
|
\vec{\bf G}_{\vec{\bf v}}^{n-1/2} & = & \vec{\bf G}_{\vec{\bf v}} ( \vec{\bf v}^{n-1/2} ) |
| 626 |
|
\label{eq:Gv-n-staggered} \\ |
| 627 |
|
\vec{\bf G}_{\vec{\bf v}}^{(n)} & = & (3/2 + \epsilon_{AB} ) \vec{\bf G}_{\vec{\bf v}}^{n-1/2} - (1/2 + \epsilon_{AB} ) \vec{\bf G}_{\vec{\bf v}}^{n-3/2} |
| 628 |
|
\label{eq:Gv-n+5-staggered} \\ |
| 629 |
|
\vec{\bf v}^{*} & = & \vec{\bf v}^{n-1/2} + \Delta t \left( \vec{\bf G}_{\vec{\bf v}}^{(n)} - \nabla \phi_{hyd}^{n} \right) |
| 630 |
|
\label{eq:vstar-staggered} \\ |
| 631 |
|
\vec{\bf v}^{**} & = & {\cal L}_{\vec{\bf v}}^{-1} ( \vec{\bf v}^* ) |
| 632 |
|
\label{eq:vstarstar-staggered} \\ |
| 633 |
|
\eta^* & = & \epsilon_{fs} \left( \eta^{n-1/2} + \Delta t (P-E)^n \right)- \Delta t |
| 634 |
|
\nabla \cdot H \widehat{ \vec{\bf v}^{**} } |
| 635 |
|
\label{eq:nstar-staggered} \\ |
| 636 |
|
\nabla \cdot g H \nabla \eta^{n+1/2} & - & \frac{\epsilon_{fs} \eta^{n+1/2}}{\Delta t^2} |
| 637 |
|
~ = ~ - \frac{\eta^*}{\Delta t^2} |
| 638 |
|
\label{eq:elliptic-staggered} \\ |
| 639 |
|
\vec{\bf v}^{n+1/2} & = & \vec{\bf v}^{*} - \Delta t g \nabla \eta^{n+1/2} |
| 640 |
|
\label{eq:v-n+1-staggered} \\ |
| 641 |
|
G_{\theta,S}^{n} & = & G_{\theta,S} ( u^{n+1/2}, \theta^{n}, S^{n} ) |
| 642 |
|
\label{eq:Gt-n-staggered} \\ |
| 643 |
|
G_{\theta,S}^{(n+1/2)} & = & (3/2+\epsilon_{AB}) G_{\theta,S}^{n}-(1/2+\epsilon_{AB}) G_{\theta,S}^{n-1} |
| 644 |
|
\label{eq:Gt-n+5-staggered} \\ |
| 645 |
|
(\theta^*,S^*) & = & (\theta^{n},S^{n}) + \Delta t G_{\theta,S}^{(n+1/2)} |
| 646 |
|
\label{eq:tstar-staggered} \\ |
| 647 |
|
(\theta^{n+1},S^{n+1}) & = & {\cal L}^{-1}_{\theta,S} (\theta^*,S^*) |
| 648 |
|
\label{eq:t-n+1-staggered} |
| 649 |
|
\end{eqnarray} |
| 650 |
|
The corresponding calling tree is given in |
| 651 |
|
\ref{fig:call-tree-adams-bashforth-staggered}. |
| 652 |
|
The staggered algorithm is activated with the run-time flag |
| 653 |
|
{\bf staggerTimeStep}{\em=.TRUE.} in parameter file {\em data}, |
| 654 |
|
namelist {\em PARM01}. |
| 655 |
|
|
| 656 |
|
\begin{figure} |
| 657 |
|
\begin{center} \fbox{ \begin{minipage}{4.7in} \begin{tabbing} |
| 658 |
|
aaa \= aaa \= aaa \= aaa \= aaa \= aaa \kill |
| 659 |
|
FORWARD\_STEP \\ |
| 660 |
|
\>\> EXTERNAL\_FIELDS\_LOAD\\ |
| 661 |
|
\>\> DO\_ATMOSPHERIC\_PHYS \\ |
| 662 |
|
\>\> DO\_OCEANIC\_PHYS \\ |
| 663 |
|
\> DYNAMICS \\ |
| 664 |
|
\>\> CALC\_PHI\_HYD \` $\phi_{hyd}^n$ (\ref{eq:phi-hyd-staggered}) \\ |
| 665 |
|
\>\> MOM\_FLUXFORM or MOM\_VECINV \` $G_{\vec{\bf v}}^{n-1/2}$ |
| 666 |
|
(\ref{eq:Gv-n-staggered})\\ |
| 667 |
|
\>\> TIMESTEP \` $\vec{\bf v}^*$ (\ref{eq:Gv-n+5-staggered}, |
| 668 |
|
\ref{eq:vstar-staggered}) \\ |
| 669 |
|
\>\> IMPLDIFF \` $\vec{\bf v}^{**}$ (\ref{eq:vstarstar-staggered}) \\ |
| 670 |
|
\> UPDATE\_R\_STAR or UPDATE\_SURF\_DR \` (NonLin-FS only)\\ |
| 671 |
|
\> SOLVE\_FOR\_PRESSURE \\ |
| 672 |
|
\>\> CALC\_DIV\_GHAT \` $\eta^*$ (\ref{eq:nstar-staggered}) \\ |
| 673 |
|
\>\> CG2D \` $\eta^{n+1/2}$ (\ref{eq:elliptic-staggered}) \\ |
| 674 |
|
\> MOMENTUM\_CORRECTION\_STEP \\ |
| 675 |
|
\>\> CALC\_GRAD\_PHI\_SURF \` $\nabla \eta^{n+1/2}$ \\ |
| 676 |
|
\>\> CORRECTION\_STEP \` $u^{n+1/2}$,$v^{n+1/2}$ (\ref{eq:v-n+1-staggered})\\ |
| 677 |
|
\> THERMODYNAMICS \\ |
| 678 |
|
\>\> CALC\_GT \\ |
| 679 |
|
\>\>\> GAD\_CALC\_RHS \` $G_\theta^n = G_\theta( u, \theta^n )$ |
| 680 |
|
(\ref{eq:Gt-n-staggered})\\ |
| 681 |
|
\>\>\> EXTERNAL\_FORCING \` $G_\theta^n = G_\theta^n + {\cal Q}$ \\ |
| 682 |
|
\>\>\> ADAMS\_BASHFORTH2 \` $G_\theta^{(n+1/2)}$ (\ref{eq:Gt-n+5-staggered}) \\ |
| 683 |
|
\>\> TIMESTEP\_TRACER \` $\theta^*$ (\ref{eq:tstar-staggered}) \\ |
| 684 |
|
\>\> IMPLDIFF \` $\theta^{(n+1)}$ (\ref{eq:t-n+1-staggered}) \\ |
| 685 |
|
\> TRACERS\_CORRECTION\_STEP \\ |
| 686 |
|
\>\> CYCLE\_TRACER \` $\theta^{n+1}$ \\ |
| 687 |
|
\>\> FILTER \` Shapiro Filter, Zonal Filter (FFT) \\ |
| 688 |
|
\>\> CONVECTIVE\_ADJUSTMENT \` \\ |
| 689 |
|
\end{tabbing} \end{minipage} } \end{center} |
| 690 |
|
\caption{ |
| 691 |
|
Calling tree for the overall staggered algorithm using |
| 692 |
|
Adams-Bashforth time-stepping. |
| 693 |
|
The place where the model geometry |
| 694 |
|
({\bf hFac} factors) is updated is added here but is only relevant |
| 695 |
|
for the non-linear free-surface algorithm. |
| 696 |
|
} |
| 697 |
|
\label{fig:call-tree-adams-bashforth-staggered} |
| 698 |
|
\end{figure} |
| 699 |
|
|
| 700 |
|
The only difficulty with this approach is apparent in equation |
| 701 |
|
\ref{eq:Gt-n-staggered} and illustrated by the dotted arrow |
| 702 |
|
connecting $u,v^{n+1/2}$ with $G_\theta^{n}$. The flow used to advect |
| 703 |
|
tracers around is not naturally located in time. This could be avoided |
| 704 |
|
by applying the Adams-Bashforth extrapolation to the tracer field |
| 705 |
|
itself and advecting that around but this approach is not yet |
| 706 |
|
available. We're not aware of any detrimental effect of this |
| 707 |
|
feature. The difficulty lies mainly in interpretation of what |
| 708 |
|
time-level variables and terms correspond to. |
| 709 |
|
|
| 710 |
|
|
| 711 |
|
\section{Non-hydrostatic formulation} |
| 712 |
|
\label{sect:non-hydrostatic} |
| 713 |
|
\begin{rawhtml} |
| 714 |
|
<!-- CMIREDIR:non-hydrostatic_formulation: --> |
| 715 |
|
\end{rawhtml} |
| 716 |
|
|
| 717 |
|
The non-hydrostatic formulation re-introduces the full vertical |
| 718 |
|
momentum equation and requires the solution of a 3-D elliptic |
| 719 |
|
equations for non-hydrostatic pressure perturbation. We still |
| 720 |
|
intergrate vertically for the hydrostatic pressure and solve a 2-D |
| 721 |
|
elliptic equation for the surface pressure/elevation for this reduces |
| 722 |
|
the amount of work needed to solve for the non-hydrostatic pressure. |
| 723 |
|
|
| 724 |
|
The momentum equations are discretized in time as follows: |
| 725 |
|
\begin{eqnarray} |
| 726 |
|
\frac{1}{\Delta t} u^{n+1} + g \partial_x \eta^{n+1} + \partial_x \phi_{nh}^{n+1} |
| 727 |
|
& = & \frac{1}{\Delta t} u^{n} + G_u^{(n+1/2)} \label{eq:discrete-time-u-nh} \\ |
| 728 |
|
\frac{1}{\Delta t} v^{n+1} + g \partial_y \eta^{n+1} + \partial_y \phi_{nh}^{n+1} |
| 729 |
|
& = & \frac{1}{\Delta t} v^{n} + G_v^{(n+1/2)} \label{eq:discrete-time-v-nh} \\ |
| 730 |
|
\frac{1}{\Delta t} w^{n+1} + \partial_r \phi_{nh}^{n+1} |
| 731 |
|
& = & \frac{1}{\Delta t} w^{n} + G_w^{(n+1/2)} \label{eq:discrete-time-w-nh} |
| 732 |
|
\end{eqnarray} |
| 733 |
|
which must satisfy the discrete-in-time depth integrated continuity, |
| 734 |
|
equation~\ref{eq:discrete-time-backward-free-surface} and the local continuity equation |
| 735 |
|
\begin{equation} |
| 736 |
|
\partial_x u^{n+1} + \partial_y v^{n+1} + \partial_r w^{n+1} = 0 |
| 737 |
|
\label{eq:non-divergence-nh} |
| 738 |
|
\end{equation} |
| 739 |
|
As before, the explicit predictions for momentum are consolidated as: |
| 740 |
\begin{eqnarray*} |
\begin{eqnarray*} |
| 741 |
\vec{\bf v}^* & = & |
u^* & = & u^n + \Delta t G_u^{(n+1/2)} \\ |
| 742 |
\vec{\bf v} ^{n} + \Delta t \vec{\bf G}_{\vec{\bf v}} ^{(n+1/2)} |
v^* & = & v^n + \Delta t G_v^{(n+1/2)} \\ |
| 743 |
+ \Delta t {\bf \nabla}_h {\phi'_{hyd}}^{n+1/2} |
w^* & = & w^n + \Delta t G_w^{(n+1/2)} |
|
\\ |
|
|
\dot{r}^* & = & |
|
|
\dot{r} ^{n} + \Delta t G_{\dot{r}} ^{(n+1/2)} |
|
| 744 |
\end{eqnarray*} |
\end{eqnarray*} |
| 745 |
|
but this time we introduce an intermediate step by splitting the |
| 746 |
|
tendancy of the flow as follows: |
| 747 |
|
\begin{eqnarray} |
| 748 |
|
u^{n+1} = u^{**} - \Delta t \partial_x \phi_{nh}^{n+1} |
| 749 |
|
& & |
| 750 |
|
u^{**} = u^{*} - \Delta t g \partial_x \eta^{n+1} \\ |
| 751 |
|
v^{n+1} = v^{**} - \Delta t \partial_y \phi_{nh}^{n+1} |
| 752 |
|
& & |
| 753 |
|
v^{**} = v^{*} - \Delta t g \partial_y \eta^{n+1} |
| 754 |
|
\end{eqnarray} |
| 755 |
|
Substituting into the depth integrated continuity |
| 756 |
|
(equation~\ref{eq:discrete-time-backward-free-surface}) gives |
| 757 |
|
\begin{equation} |
| 758 |
|
\partial_x H \partial_x \left( g \eta^{n+1} + \widehat{\phi}_{nh}^{n+1} \right) |
| 759 |
|
+ |
| 760 |
|
\partial_y H \partial_y \left( g \eta^{n+1} + \widehat{\phi}_{nh}^{n+1} \right) |
| 761 |
|
- \frac{\epsilon_{fs}\eta^*}{\Delta t^2} |
| 762 |
|
= - \frac{\eta^*}{\Delta t^2} |
| 763 |
|
\end{equation} |
| 764 |
|
which is approximated by equation |
| 765 |
|
\ref{eq:elliptic-backward-free-surface} on the basis that i) |
| 766 |
|
$\phi_{nh}^{n+1}$ is not yet known and ii) $\nabla \widehat{\phi}_{nh} |
| 767 |
|
<< g \nabla \eta$. If \ref{eq:elliptic-backward-free-surface} is |
| 768 |
|
solved accurately then the implication is that $\widehat{\phi}_{nh} |
| 769 |
|
\approx 0$ so that thet non-hydrostatic pressure field does not drive |
| 770 |
|
barotropic motion. |
| 771 |
|
|
| 772 |
|
The flow must satisfy non-divergence |
| 773 |
|
(equation~\ref{eq:non-divergence-nh}) locally, as well as depth |
| 774 |
|
integrated, and this constraint is used to form a 3-D elliptic |
| 775 |
|
equations for $\phi_{nh}^{n+1}$: |
| 776 |
|
\begin{equation} |
| 777 |
|
\partial_{xx} \phi_{nh}^{n+1} + \partial_{yy} \phi_{nh}^{n+1} + |
| 778 |
|
\partial_{rr} \phi_{nh}^{n+1} = |
| 779 |
|
\partial_x u^{**} + \partial_y v^{**} + \partial_r w^{*} |
| 780 |
|
\end{equation} |
| 781 |
|
|
| 782 |
%--------------------------------------------------------------------- |
The entire algorithm can be summarized as the sequential solution of |
| 783 |
|
the following equations: |
| 784 |
|
\begin{eqnarray} |
| 785 |
|
u^{*} & = & u^{n} + \Delta t G_u^{(n+1/2)} \label{eq:ustar-nh} \\ |
| 786 |
|
v^{*} & = & v^{n} + \Delta t G_v^{(n+1/2)} \label{eq:vstar-nh} \\ |
| 787 |
|
w^{*} & = & w^{n} + \Delta t G_w^{(n+1/2)} \label{eq:wstar-nh} \\ |
| 788 |
|
\eta^* ~ = ~ \epsilon_{fs} \left( \eta^{n} + \Delta t (P-E) \right) |
| 789 |
|
& - & \Delta t |
| 790 |
|
\partial_x H \widehat{u^{*}} |
| 791 |
|
+ \partial_y H \widehat{v^{*}} |
| 792 |
|
\\ |
| 793 |
|
\partial_x g H \partial_x \eta^{n+1} |
| 794 |
|
+ \partial_y g H \partial_y \eta^{n+1} |
| 795 |
|
& - & \frac{\epsilon_{fs} \eta^{n+1}}{\Delta t^2} |
| 796 |
|
~ = ~ |
| 797 |
|
- \frac{\eta^*}{\Delta t^2} |
| 798 |
|
\label{eq:elliptic-nh} |
| 799 |
|
\\ |
| 800 |
|
u^{**} & = & u^{*} - \Delta t g \partial_x \eta^{n+1} \label{eq:unx-nh}\\ |
| 801 |
|
v^{**} & = & v^{*} - \Delta t g \partial_y \eta^{n+1} \label{eq:vnx-nh}\\ |
| 802 |
|
\partial_{xx} \phi_{nh}^{n+1} + \partial_{yy} \phi_{nh}^{n+1} + |
| 803 |
|
\partial_{rr} \phi_{nh}^{n+1} & = & |
| 804 |
|
\partial_x u^{**} + \partial_y v^{**} + \partial_r w^{*} \\ |
| 805 |
|
u^{n+1} & = & u^{**} - \Delta t \partial_x \phi_{nh}^{n+1} \label{eq:un+1-nh}\\ |
| 806 |
|
v^{n+1} & = & v^{**} - \Delta t \partial_y \phi_{nh}^{n+1} \label{eq:vn+1-nh}\\ |
| 807 |
|
\partial_r w^{n+1} & = & - \partial_x u^{n+1} - \partial_y v^{n+1} |
| 808 |
|
\end{eqnarray} |
| 809 |
|
where the last equation is solved by vertically integrating for |
| 810 |
|
$w^{n+1}$. |
| 811 |
|
|
|
\subsection{Surface pressure} |
|
| 812 |
|
|
| 813 |
Substituting \ref{eq-rtd-hmom} into \ref{eq-rtd-cont}, assuming |
|
| 814 |
$\epsilon_{nh} = 0$ yields a Helmholtz equation for ${\eta}^{n+1}$: |
\section{Variants on the Free Surface} |
| 815 |
|
\label{sect:free-surface} |
| 816 |
|
|
| 817 |
|
We now describe the various formulations of the free-surface that |
| 818 |
|
include non-linear forms, implicit in time using Crank-Nicholson, |
| 819 |
|
explicit and [one day] split-explicit. First, we'll reiterate the |
| 820 |
|
underlying algorithm but this time using the notation consistent with |
| 821 |
|
the more general vertical coordinate $r$. The elliptic equation for |
| 822 |
|
free-surface coordinate (units of $r$), corresponding to |
| 823 |
|
\ref{eq:discrete-time-backward-free-surface}, and |
| 824 |
|
assuming no non-hydrostatic effects ($\epsilon_{nh} = 0$) is: |
| 825 |
\begin{eqnarray} |
\begin{eqnarray} |
| 826 |
\epsilon_{fs} {\eta}^{n+1} - |
\epsilon_{fs} {\eta}^{n+1} - |
| 827 |
{\bf \nabla}_h \cdot \Delta t^2 (R_o-R_{min}) |
{\bf \nabla}_h \cdot \Delta t^2 (R_o-R_{fixed}) {\bf \nabla}_h b_s |
| 828 |
{\bf \nabla}_h b_s {\eta}^{n+1} |
{\eta}^{n+1} = {\eta}^* |
| 829 |
= {\eta}^* |
\label{eq-solve2D} |
|
\label{solve_2d} |
|
| 830 |
\end{eqnarray} |
\end{eqnarray} |
| 831 |
where |
where |
| 832 |
\begin{eqnarray} |
\begin{eqnarray} |
| 833 |
{\eta}^* = \epsilon_{fs} \: {\eta}^{n} - |
{\eta}^* = \epsilon_{fs} \: {\eta}^{n} - |
| 834 |
\Delta t {\bf \nabla}_h \cdot \int_{R_{min}}^{R_o} \vec{\bf v}^* dr |
\Delta t {\bf \nabla}_h \cdot \int_{R_{fixed}}^{R_o} \vec{\bf v}^* dr |
| 835 |
\: + \: \epsilon_{fw} \Delta_t (P-E)^{n} |
\: + \: \epsilon_{fw} \Delta t (P-E)^{n} |
| 836 |
\label{solve_2d_rhs} |
\label{eq-solve2D_rhs} |
| 837 |
\end{eqnarray} |
\end{eqnarray} |
| 838 |
|
|
| 839 |
Once ${\eta}^{n+1}$ has been found substituting into \ref{eq-rtd-hmom} |
\fbox{ \begin{minipage}{4.75in} |
| 840 |
would yield $\vec{\bf v}^{n+1}$ if the model is hydrostatic |
{\em S/R SOLVE\_FOR\_PRESSURE} ({\em solve\_for\_pressure.F}) |
| 841 |
($\epsilon_{nh}=0$): |
|
| 842 |
|
$u^*$: {\bf gU} ({\em DYNVARS.h}) |
| 843 |
|
|
| 844 |
|
$v^*$: {\bf gV} ({\em DYNVARS.h}) |
| 845 |
|
|
| 846 |
|
$\eta^*$: {\bf cg2d\_b} (\em SOLVE\_FOR\_PRESSURE.h) |
| 847 |
|
|
| 848 |
|
$\eta^{n+1}$: {\bf etaN} (\em DYNVARS.h) |
| 849 |
|
|
| 850 |
|
\end{minipage} } |
| 851 |
|
|
| 852 |
|
|
| 853 |
|
Once ${\eta}^{n+1}$ has been found, substituting into |
| 854 |
|
\ref{eq:discrete-time-u}, \ref{eq:discrete-time-v} yields $\vec{\bf v}^{n+1}$ |
| 855 |
|
if the model is hydrostatic ($\epsilon_{nh}=0$): |
| 856 |
$$ |
$$ |
| 857 |
\vec{\bf v}^{n+1} = \vec{\bf v}^{*} |
\vec{\bf v}^{n+1} = \vec{\bf v}^{*} |
| 858 |
- \Delta t {\bf \nabla}_h b_s {\eta}^{n+1} |
- \Delta t {\bf \nabla}_h b_s {\eta}^{n+1} |
| 860 |
|
|
| 861 |
This is known as the correction step. However, when the model is |
This is known as the correction step. However, when the model is |
| 862 |
non-hydrostatic ($\epsilon_{nh}=1$) we need an additional step and an |
non-hydrostatic ($\epsilon_{nh}=1$) we need an additional step and an |
| 863 |
additional equation for $\phi'_{nh}$. This is obtained by |
additional equation for $\phi'_{nh}$. This is obtained by substituting |
| 864 |
substituting \ref{eq-rtd-hmom} and \ref{eq-rtd-rdot} into |
\ref{eq:discrete-time-u-nh}, \ref{eq:discrete-time-v-nh} and \ref{eq:discrete-time-w-nh} |
| 865 |
\ref{eq-rtd-cont}: |
into continuity: |
| 866 |
\begin{equation} |
\begin{equation} |
| 867 |
\left[ {\bf \nabla}_h^2 + \partial_{rr} \right] {\phi'_{nh}}^{n+1} |
\left[ {\bf \nabla}_h^2 + \partial_{rr} \right] {\phi'_{nh}}^{n+1} |
| 868 |
= \frac{1}{\Delta t} \left( |
= \frac{1}{\Delta t} \left( |
| 883 |
- \epsilon_{nh} \Delta t {\bf \nabla}_h {\phi'_{nh}}^{n+1} |
- \epsilon_{nh} \Delta t {\bf \nabla}_h {\phi'_{nh}}^{n+1} |
| 884 |
\end{equation} |
\end{equation} |
| 885 |
and the vertical velocity is found by integrating the continuity |
and the vertical velocity is found by integrating the continuity |
| 886 |
equation vertically. |
equation vertically. Note that, for the convenience of the restart |
| 887 |
Note that, for convenience regarding the restart procedure, |
procedure, the vertical integration of the continuity equation has |
| 888 |
the integration of the continuity equation has been |
been moved to the beginning of the time step (instead of at the end), |
|
moved at the beginning of the time step (instead of at the end), |
|
| 889 |
without any consequence on the solution. |
without any consequence on the solution. |
| 890 |
|
|
| 891 |
Regarding the implementation, all those computation are done |
\fbox{ \begin{minipage}{4.75in} |
| 892 |
within the routine {\it SOLVE\_FOR\_PRESSURE} and its dependent |
{\em S/R CORRECTION\_STEP} ({\em correction\_step.F}) |
| 893 |
{\it CALL}s. |
|
| 894 |
The standard method to solve the 2D elliptic problem (\ref{solve_2d}) |
$\eta^{n+1}$: {\bf etaN} (\em DYNVARS.h) |
| 895 |
uses the conjugate gradient method (routine {\it CG2D}); The |
|
| 896 |
the solver matrix and conjugate gradient operator are only function |
$\phi_{nh}^{n+1}$: {\bf phi\_nh} (\em NH\_VARS.h) |
| 897 |
of the discretized domain and are therefore evaluated separately, |
|
| 898 |
before the time iteration loop, within {\it INI\_CG2D}. |
$u^*$: {\bf gU} ({\em DYNVARS.h}) |
| 899 |
The computation of the RHS $\eta^*$ is partly |
|
| 900 |
done in {\it CALC\_DIV\_GHAT} and in {\it SOLVE\_FOR\_PRESSURE}. |
$v^*$: {\bf gV} ({\em DYNVARS.h}) |
| 901 |
|
|
| 902 |
The same method is applied for the non hydrostatic part, using |
$u^{n+1}$: {\bf uVel} ({\em DYNVARS.h}) |
| 903 |
a conjugate gradient 3D solver ({\it CG3D}) that is initialized |
|
| 904 |
in {\it INI\_CG3D}. The RHS terms of 2D and 3D problems |
$v^{n+1}$: {\bf vVel} ({\em DYNVARS.h}) |
| 905 |
are computed together, within the same part of the code. |
|
| 906 |
|
\end{minipage} } |
| 907 |
|
|
| 908 |
|
|
| 909 |
|
|
| 910 |
|
Regarding the implementation of the surface pressure solver, all |
| 911 |
|
computation are done within the routine {\it SOLVE\_FOR\_PRESSURE} and |
| 912 |
|
its dependent calls. The standard method to solve the 2D elliptic |
| 913 |
|
problem (\ref{eq-solve2D}) uses the conjugate gradient method (routine |
| 914 |
|
{\it CG2D}); the solver matrix and conjugate gradient operator are |
| 915 |
|
only function of the discretized domain and are therefore evaluated |
| 916 |
|
separately, before the time iteration loop, within {\it INI\_CG2D}. |
| 917 |
|
The computation of the RHS $\eta^*$ is partly done in {\it |
| 918 |
|
CALC\_DIV\_GHAT} and in {\it SOLVE\_FOR\_PRESSURE}. |
| 919 |
|
|
| 920 |
|
The same method is applied for the non hydrostatic part, using a |
| 921 |
|
conjugate gradient 3D solver ({\it CG3D}) that is initialized in {\it |
| 922 |
|
INI\_CG3D}. The RHS terms of 2D and 3D problems are computed together |
| 923 |
|
at the same point in the code. |
| 924 |
|
|
| 925 |
|
|
| 926 |
|
|
|
\newpage |
|
|
%----------------------------------------------------------------------------------- |
|
| 927 |
\subsection{Crank-Nickelson barotropic time stepping} |
\subsection{Crank-Nickelson barotropic time stepping} |
| 928 |
|
\label{sect:freesurf-CrankNick} |
| 929 |
|
|
| 930 |
The full implicit time stepping described previously is unconditionally stable |
The full implicit time stepping described previously is |
| 931 |
but damps the fast gravity waves, resulting in a loss of |
unconditionally stable but damps the fast gravity waves, resulting in |
| 932 |
gravity potential energy. |
a loss of potential energy. The modification presented now allows one |
| 933 |
The modification presented hereafter allows to combine an implicit part |
to combine an implicit part ($\beta,\gamma$) and an explicit part |
| 934 |
($\beta,\gamma$) and an explicit part ($1-\beta,1-\gamma$) for the surface |
($1-\beta,1-\gamma$) for the surface pressure gradient ($\beta$) and |
| 935 |
pressure gradient ($\beta$) and for the barotropic flow divergence ($\gamma$). |
for the barotropic flow divergence ($\gamma$). |
| 936 |
\\ |
\\ |
| 937 |
For instance, $\beta=\gamma=1$ is the previous fully implicit scheme; |
For instance, $\beta=\gamma=1$ is the previous fully implicit scheme; |
| 938 |
$\beta=\gamma=1/2$ is the non damping (energy conserving), unconditionally |
$\beta=\gamma=1/2$ is the non damping (energy conserving), unconditionally |
| 940 |
corresponds to the forward - backward scheme that conserves energy but is |
corresponds to the forward - backward scheme that conserves energy but is |
| 941 |
only stable for small time steps.\\ |
only stable for small time steps.\\ |
| 942 |
In the code, $\beta,\gamma$ are defined as parameters, respectively |
In the code, $\beta,\gamma$ are defined as parameters, respectively |
| 943 |
{\it implicSurfPress}, {\it implicDiv2DFlow}. They are read from |
{\bf implicSurfPress}, {\bf implicDiv2DFlow}. They are read from |
| 944 |
the main data file "{\it data}" and are set by default to 1,1. |
the main parameter file "{\em data}" and are set by default to 1,1. |
| 945 |
|
|
| 946 |
Equations \ref{eq-rtd-hmom} and \ref{eq-rtd-eta} are modified as follows: |
Equations \ref{eq:ustar-backward-free-surface} -- |
| 947 |
$$ |
\ref{eq:vn+1-backward-free-surface} are modified as follows: |
| 948 |
|
\begin{eqnarray*} |
| 949 |
\frac{ \vec{\bf v}^{n+1} }{ \Delta t } |
\frac{ \vec{\bf v}^{n+1} }{ \Delta t } |
| 950 |
+ {\bf \nabla}_h b_s [ \beta {\eta}^{n+1} + (1-\beta) {\eta}^{n} ] |
+ {\bf \nabla}_h b_s [ \beta {\eta}^{n+1} + (1-\beta) {\eta}^{n} ] |
| 951 |
+ \epsilon_{nh} {\bf \nabla}_h {\phi'_{nh}}^{n+1} |
+ \epsilon_{nh} {\bf \nabla}_h {\phi'_{nh}}^{n+1} |
| 952 |
= \frac{ \vec{\bf v}^* }{ \Delta t } |
= \frac{ \vec{\bf v}^* }{ \Delta t } |
| 953 |
$$ |
\end{eqnarray*} |
| 954 |
$$ |
\begin{eqnarray} |
| 955 |
\epsilon_{fs} \frac{ {\eta}^{n+1} - {\eta}^{n} }{ \Delta t} |
\epsilon_{fs} \frac{ {\eta}^{n+1} - {\eta}^{n} }{ \Delta t} |
| 956 |
+ {\bf \nabla}_h \cdot \int_{R_{min}}^{R_o} |
+ {\bf \nabla}_h \cdot \int_{R_{fixed}}^{R_o} |
| 957 |
[ \gamma \vec{\bf v}^{n+1} + (1-\gamma) \vec{\bf v}^{n}] dr |
[ \gamma \vec{\bf v}^{n+1} + (1-\gamma) \vec{\bf v}^{n}] dr |
| 958 |
= \epsilon_{fw} (P-E) |
= \epsilon_{fw} (P-E) |
| 959 |
$$ |
\label{eq:eta-n+1-CrankNick} |
| 960 |
|
\end{eqnarray} |
| 961 |
where: |
where: |
| 962 |
\begin{eqnarray*} |
\begin{eqnarray*} |
| 963 |
\vec{\bf v}^* & = & |
\vec{\bf v}^* & = & |
| 967 |
\\ |
\\ |
| 968 |
{\eta}^* & = & |
{\eta}^* & = & |
| 969 |
\epsilon_{fs} {\eta}^{n} + \epsilon_{fw} \Delta t (P-E) |
\epsilon_{fs} {\eta}^{n} + \epsilon_{fw} \Delta t (P-E) |
| 970 |
- \Delta t {\bf \nabla}_h \cdot \int_{R_{min}}^{R_o} |
- \Delta t {\bf \nabla}_h \cdot \int_{R_{fixed}}^{R_o} |
| 971 |
[ \gamma \vec{\bf v}^* + (1-\gamma) \vec{\bf v}^{n}] dr |
[ \gamma \vec{\bf v}^* + (1-\gamma) \vec{\bf v}^{n}] dr |
| 972 |
\end{eqnarray*} |
\end{eqnarray*} |
| 973 |
\\ |
\\ |
| 974 |
In the hydrostatic case ($\epsilon_{nh}=0$), |
In the hydrostatic case ($\epsilon_{nh}=0$), allowing us to find |
| 975 |
this allow to find ${\eta}^{n+1}$, according to: |
${\eta}^{n+1}$, thus: |
| 976 |
$$ |
$$ |
| 977 |
\epsilon_{fs} {\eta}^{n+1} - |
\epsilon_{fs} {\eta}^{n+1} - |
| 978 |
{\bf \nabla}_h \cdot \beta\gamma \Delta t^2 b_s (R_o - R_{min}) |
{\bf \nabla}_h \cdot \beta\gamma \Delta t^2 b_s (R_o - R_{fixed}) |
| 979 |
{\bf \nabla}_h {\eta}^{n+1} |
{\bf \nabla}_h {\eta}^{n+1} |
| 980 |
= {\eta}^* |
= {\eta}^* |
| 981 |
$$ |
$$ |
| 982 |
and then to compute (correction step): |
and then to compute ({\em CORRECTION\_STEP}): |
| 983 |
$$ |
$$ |
| 984 |
\vec{\bf v}^{n+1} = \vec{\bf v}^{*} |
\vec{\bf v}^{n+1} = \vec{\bf v}^{*} |
| 985 |
- \beta \Delta t {\bf \nabla}_h b_s {\eta}^{n+1} |
- \beta \Delta t {\bf \nabla}_h b_s {\eta}^{n+1} |
| 986 |
$$ |
$$ |
| 987 |
|
|
| 988 |
The non-hydrostatic part is solved as described previously. |
%The non-hydrostatic part is solved as described previously. |
| 989 |
\\ \\ |
|
| 990 |
N.B: |
\noindent |
| 991 |
\\ |
Notes: |
| 992 |
a) The non-hydrostatic part of the code has not yet been |
\begin{enumerate} |
| 993 |
updated, %since it falls out of the purpose of this test, |
\item The RHS term of equation \ref{eq:eta-n+1-CrankNick} |
| 994 |
so that this option cannot be used with $(\beta,\gamma) \neq (1,1)$. |
corresponds the contribution of fresh water flux (P-E) |
| 995 |
\\ |
to the free-surface variations ($\epsilon_{fw}=1$, |
| 996 |
b) To remind, the stability criteria with the Crank-Nickelson time stepping |
{\bf useRealFreshWater}{\em=TRUE} in parameter file {\em data}). |
| 997 |
for the pure linear gravity wave problem in cartesian coordinate is: |
In order to remain consistent with the tracer equation, specially in |
| 998 |
\\ |
the non-linear free-surface formulation, this term is also |
| 999 |
$\star$~ $\beta + \gamma < 1$ : unstable |
affected by the Crank-Nickelson time stepping. The RHS reads: |
| 1000 |
\\ |
$\epsilon_{fw} ( \gamma (P-E)^{n+1/2} + (1-\gamma) (P-E)^{n-1/2} )$ |
| 1001 |
$\star$~ $\beta \geq 1/2$ and $ \gamma \geq 1/2$ : stable |
\item The non-hydrostatic part of the code has not yet been |
| 1002 |
\\ |
updated, and therefore cannot be used with $(\beta,\gamma) \neq (1,1)$. |
| 1003 |
$\star$~ $\beta + \gamma \geq 1$ : stable if |
\item The stability criteria with Crank-Nickelson time stepping |
| 1004 |
%, for all wave length $(k\Delta x,l\Delta y)$ |
for the pure linear gravity wave problem in cartesian coordinates is: |
| 1005 |
|
\begin{itemize} |
| 1006 |
|
\item $\beta + \gamma < 1$ : unstable |
| 1007 |
|
\item $\beta \geq 1/2$ and $ \gamma \geq 1/2$ : stable |
| 1008 |
|
\item $\beta + \gamma \geq 1$ : stable if |
| 1009 |
$$ |
$$ |
| 1010 |
c_{max}^2 (\beta - 1/2)(\gamma - 1/2) + 1 \geq 0 |
c_{max}^2 (\beta - 1/2)(\gamma - 1/2) + 1 \geq 0 |
| 1011 |
$$ |
$$ |
| 1020 |
c_{max} = 2 \Delta t \: \sqrt{g H} \: |
c_{max} = 2 \Delta t \: \sqrt{g H} \: |
| 1021 |
\sqrt{ \frac{1}{\Delta x^2} + \frac{1}{\Delta y^2} } |
\sqrt{ \frac{1}{\Delta x^2} + \frac{1}{\Delta y^2} } |
| 1022 |
$$ |
$$ |
| 1023 |
|
\end{itemize} |
| 1024 |
|
\end{enumerate} |
| 1025 |
|
|
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch