1 |
|
% $Header$ |
2 |
|
% $Name$ |
3 |
\documentclass[12pt]{article} |
\documentclass[12pt]{article} |
4 |
|
|
5 |
\usepackage[]{graphicx} |
\usepackage[]{graphicx} |
52 |
\maketitle |
\maketitle |
53 |
|
|
54 |
\begin{abstract} |
\begin{abstract} |
55 |
Some blabla |
|
56 |
|
As part of ongoing efforts to obtain a best possible synthesis of most |
57 |
|
available, global-scale, ocean and sea ice data, dynamic and thermodynamic |
58 |
|
sea-ice model components have been incorporated in the Massachusetts Institute |
59 |
|
of Technology general circulation model (MITgcm). Sea-ice dynamics use either |
60 |
|
a visco-plastic rheology solved with a line successive relaxation (LSR) |
61 |
|
technique, reformulated on an Arakawa C-grid in order to match the oceanic and |
62 |
|
atmospheric grids of the MITgcm, and modified to permit efficient and accurate |
63 |
|
automatic differentiation of the coupled ocean and sea-ice model |
64 |
|
configurations. |
65 |
|
|
66 |
\end{abstract} |
\end{abstract} |
67 |
|
|
68 |
\section{Introduction} |
\section{Introduction} |
146 |
both thickness $h$ and compactness (concentration) $c$: |
both thickness $h$ and compactness (concentration) $c$: |
147 |
\begin{equation} |
\begin{equation} |
148 |
P_{\max} = P^{*}c\,h\,e^{[C^{*}\cdot(1-c)]}, |
P_{\max} = P^{*}c\,h\,e^{[C^{*}\cdot(1-c)]}, |
149 |
\label{icestrength} |
\label{eq:icestrength} |
150 |
\end{equation} |
\end{equation} |
151 |
with the constants $P^{*}$ and $C^{*}$. The nonlinear bulk and shear |
with the constants $P^{*}$ and $C^{*}$. The nonlinear bulk and shear |
152 |
viscosities $\eta$ and $\zeta$ are functions of ice strain rate |
viscosities $\eta$ and $\zeta$ are functions of ice strain rate |
447 |
which in turn can have a strong effect on solutions in the limit of |
which in turn can have a strong effect on solutions in the limit of |
448 |
nearly rigid regimes (arching and blocking, not shown). |
nearly rigid regimes (arching and blocking, not shown). |
449 |
|
|
450 |
|
\ml{[Say something about performance? This is tricky, as the |
451 |
|
perfomance depends strongly on the configuration. A run with slowly |
452 |
|
changing forcing is favorable for LSR, because then only very few |
453 |
|
iterations are required for convergences while EVP uses its fixed |
454 |
|
number of internal timesteps. If the forcing in changing fast, LSR |
455 |
|
needs far more iterations while EVP still uses the fixed number of |
456 |
|
internal timesteps. I have produces runs where for slow forcing LSR |
457 |
|
is much faster than EVP and for fast forcing, LSR is much slower |
458 |
|
than EVP. EVP is certainly more efficient in terms of vectorization |
459 |
|
and MFLOPS on our SX8, but is that a criterion?]} |
460 |
|
|
461 |
\subsection{C-grid} |
\subsection{C-grid} |
462 |
\begin{itemize} |
\begin{itemize} |
463 |
\item no-slip vs. free-slip for both lsr and evp; |
\item no-slip vs. free-slip for both lsr and evp; |