/[MITgcm]/MITgcm_contrib/articles/ceaice/ceaice.tex
ViewVC logotype

Annotation of /MITgcm_contrib/articles/ceaice/ceaice.tex

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


Revision 1.10 - (hide annotations) (download) (as text)
Mon Feb 25 16:50:56 2008 UTC (17 years, 4 months ago) by dimitri
Branch: MAIN
Changes since 1.9: +13 -3 lines
File MIME type: application/x-tex
addded preliminary abstract

1 dimitri 1.10 % $Header: /u/gcmpack/MITgcm_contrib/articles/ceaice/ceaice.tex,v 1.9 2008/01/21 08:06:00 mlosch Exp $
2     % $Name: $
3 dimitri 1.1 \documentclass[12pt]{article}
4 mlosch 1.4
5 mlosch 1.5 \usepackage[]{graphicx}
6     \usepackage{subfigure}
7 dimitri 1.1
8     \usepackage[round,comma]{natbib}
9 dimitri 1.2 \bibliographystyle{bib/agu04}
10 dimitri 1.1
11     \usepackage{amsmath,amssymb}
12     \newcommand\bmmax{10} \newcommand\hmmax{10}
13     \usepackage{bm}
14    
15     % math abbreviations
16     \newcommand{\vek}[1]{\ensuremath{\mathbf{#1}}}
17     \newcommand{\mat}[1]{\ensuremath{\mathbf{#1}}}
18     \newcommand{\vtau}{\bm{{\tau}}}
19    
20     \newcommand{\degree}{\ensuremath{^\circ}}
21     \newcommand{\degC}{\,\ensuremath{\degree}C}
22     \newcommand{\degE}{\ensuremath{\degree}\,E}
23     \newcommand{\degS}{\ensuremath{\degree}\,S}
24     \newcommand{\degN}{\ensuremath{\degree}\,N}
25     \newcommand{\degW}{\ensuremath{\degree}\,W}
26    
27     % cross reference scheme
28     \newcommand{\reffig}[1]{Figure~\ref{fig:#1}}
29     \newcommand{\reftab}[1]{Table~\ref{tab:#1}}
30     \newcommand{\refapp}[1]{Appendix~\ref{app:#1}}
31     \newcommand{\refsec}[1]{Section~\ref{sec:#1}}
32     \newcommand{\refeq}[1]{\,(\ref{eq:#1})}
33     \newcommand{\refeqs}[2]{\,(\ref{eq:#1})--(\ref{eq:#2})}
34    
35     \newlength{\stdfigwidth}\setlength{\stdfigwidth}{20pc}
36     %\newlength{\stdfigwidth}\setlength{\stdfigwidth}{\columnwidth}
37     \newlength{\mediumfigwidth}\setlength{\mediumfigwidth}{39pc}
38     %\newlength{\widefigwidth}\setlength{\widefigwidth}{39pc}
39     \newlength{\widefigwidth}\setlength{\widefigwidth}{\textwidth}
40 mlosch 1.4 \newcommand{\fpath}{figs}
41    
42     % commenting scheme
43     \newcommand{\ml}[1]{\textsf{\slshape #1}}
44 dimitri 1.1
45     \title{A Dynamic-Thermodynamic Sea ice Model for Ocean Climate
46     Estimation on an Arakawa C-Grid}
47    
48     \author{Martin Losch, Dimitris Menemenlis, Patrick Heimbach, \\
49     Jean-Michel Campin, and Chris Hill}
50     \begin{document}
51    
52     \maketitle
53    
54     \begin{abstract}
55 dimitri 1.10
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 dimitri 1.1 \end{abstract}
67    
68     \section{Introduction}
69     \label{sec:intro}
70    
71     more blabla
72    
73     \section{Model}
74     \label{sec:model}
75    
76     Traditionally, probably for historical reasons and the ease of
77     treating the Coriolis term, most standard sea-ice models are
78     discretized on Arakawa-B-grids \citep[e.g.,][]{hibler79, harder99,
79     kreyscher00, zhang98, hunke97}. From the perspective of coupling a
80     sea ice-model to a C-grid ocean model, the exchange of fluxes of heat
81     and fresh-water pose no difficulty for a B-grid sea-ice model
82     \citep[e.g.,][]{timmermann02a}. However, surface stress is defined at
83     velocities points and thus needs to be interpolated between a B-grid
84     sea-ice model and a C-grid ocean model. While the smoothing implicitly
85     associated with this interpolation may mask grid scale noise, it may
86     in two-way coupling lead to a computational mode as will be shown. By
87     choosing a C-grid for the sea-ice model, we circumvene this difficulty
88     altogether and render the stress coupling as consistent as the
89     buoyancy coupling.
90    
91     A further advantage of the C-grid formulation is apparent in narrow
92     straits. In the limit of only one grid cell between coasts there is no
93     flux allowed for a B-grid (with no-slip lateral boundary counditions),
94     whereas the C-grid formulation allows a flux of sea-ice through this
95     passage for all types of lateral boundary conditions. We (will)
96     demonstrate this effect in the Candian archipelago.
97    
98     \subsection{Dynamics}
99     \label{sec:dynamics}
100    
101     The momentum equations of the sea-ice model are standard with
102     \begin{equation}
103     \label{eq:momseaice}
104     m \frac{D\vek{u}}{Dt} = -mf\vek{k}\times\vek{u} + \vtau_{air} +
105     \vtau_{ocean} - m \nabla{\phi(0)} + \vek{F},
106     \end{equation}
107     where $\vek{u} = u\vek{i}+v\vek{j}$ is the ice velocity vectory, $m$
108     the ice mass per unit area, $f$ the Coriolis parameter, $g$ is the
109     gravity accelation, $\nabla\phi$ is the gradient (tilt) of the sea
110     surface height potential beneath the ice. $\phi$ is the sum of
111     atmpheric pressure $p_{a}$ and loading due to ice and snow
112     $(m_{i}+m_{s})g$. $\vtau_{air}$ and $\vtau_{ocean}$ are the wind and
113     ice-ocean stresses, respectively. $\vek{F}$ is the interaction force
114     and $\vek{i}$, $\vek{j}$, and $\vek{k}$ are the unit vectors in the
115     $x$, $y$, and $z$ directions. Advection of sea-ice momentum is
116     neglected. The wind and ice-ocean stress terms are given by
117     \begin{align*}
118     \vtau_{air} =& \rho_{air} |\vek{U}_{air}|R_{air}(\vek{U}_{air}) \\
119     \vtau_{ocean} =& \rho_{ocean} |\vek{U}_{ocean}-\vek{u}|
120     R_{ocean}(\vek{U}_{ocean}-\vek{u}), \\
121     \end{align*}
122     where $\vek{U}_{air/ocean}$ are the surface winds of the atmosphere
123     and surface currents of the ocean, respectively. $C_{air/ocean}$ are
124     air and ocean drag coefficients, $\rho_{air/ocean}$ reference
125     densities, and $R_{air/ocean}$ rotation matrices that act on the
126     wind/current vectors. $\vek{F} = \nabla\cdot\sigma$ is the divergence
127     of the interal stress tensor $\sigma_{ij}$.
128    
129     For an isotropic system this stress tensor can be related to the ice
130     strain rate and strength by a nonlinear viscous-plastic (VP)
131     constitutive law \citep{hibler79, zhang98}:
132     \begin{equation}
133     \label{eq:vpequation}
134     \sigma_{ij}=2\eta(\dot{\epsilon}_{ij},P)\dot{\epsilon}_{ij}
135     + \left[\zeta(\dot{\epsilon}_{ij},P) -
136     \eta(\dot{\epsilon}_{ij},P)\right]\dot{\epsilon}_{kk}\delta_{ij}
137     - \frac{P}{2}\delta_{ij}.
138     \end{equation}
139     The ice strain rate is given by
140     \begin{equation*}
141     \dot{\epsilon}_{ij} = \frac{1}{2}\left(
142     \frac{\partial{u_{i}}}{\partial{x_{j}}} +
143     \frac{\partial{u_{j}}}{\partial{x_{i}}}\right).
144     \end{equation*}
145 mlosch 1.5 The maximum ice pressure $P_{\max}$, a measure of ice strength, depends on
146     both thickness $h$ and compactness (concentration) $c$:
147 mlosch 1.4 \begin{equation}
148 mlosch 1.5 P_{\max} = P^{*}c\,h\,e^{[C^{*}\cdot(1-c)]},
149 mlosch 1.9 \label{eq:icestrength}
150 mlosch 1.4 \end{equation}
151     with the constants $P^{*}$ and $C^{*}$. The nonlinear bulk and shear
152     viscosities $\eta$ and $\zeta$ are functions of ice strain rate
153     invariants and ice strength such that the principal components of the
154     stress lie on an elliptical yield curve with the ratio of major to
155     minor axis $e$ equal to $2$; they are given by:
156 dimitri 1.1 \begin{align*}
157 mlosch 1.5 \zeta =& \min\left(\frac{P_{\max}}{2\max(\Delta,\Delta_{\min})},
158     \zeta_{\max}\right) \\
159     \eta =& \frac{\zeta}{e^2} \\
160 dimitri 1.1 \intertext{with the abbreviation}
161     \Delta = & \left[
162     \left(\dot{\epsilon}_{11}^2+\dot{\epsilon}_{22}^2\right)
163     (1+e^{-2}) + 4e^{-2}\dot{\epsilon}_{12}^2 +
164     2\dot{\epsilon}_{11}\dot{\epsilon}_{22} (1-e^{-2})
165     \right]^{-\frac{1}{2}}
166     \end{align*}
167 mlosch 1.5 The bulk viscosities are bounded above by imposing both a minimum
168     $\Delta_{\min}=10^{-11}\text{\,s}^{-1}$ (for numerical reasons) and a
169     maximum $\zeta_{\max} = P_{\max}/\Delta^*$, where
170     $\Delta^*=(5\times10^{12}/2\times10^4)\text{\,s}^{-1}$. For stress
171     tensor compuation the replacement pressure $P = 2\,\Delta\zeta$
172     \citep{hibler95} is used so that the stress state always lies on the
173     elliptic yield curve by definition.
174    
175 mlosch 1.6 In the so-called truncated ellipse method the shear viscosity $\eta$
176     is capped to suppress any tensile stress \citep{hibler97, geiger98}:
177     \begin{equation}
178     \label{eq:etatem}
179     \eta = \min(\frac{\zeta}{e^2}
180     \frac{\frac{P}{2}-\zeta(\dot{\epsilon}_{11}+\dot{\epsilon}_{22})}
181     {\sqrt{(\dot{\epsilon}_{11}+\dot{\epsilon}_{22})^2
182     +4\dot{\epsilon}_{12}^2}}
183     \end{equation}
184    
185 dimitri 1.1 In the current implementation, the VP-model is integrated with the
186     semi-implicit line successive over relaxation (LSOR)-solver of
187     \citet{zhang98}, which allows for long time steps that, in our case,
188     is limited by the explicit treatment of the Coriolis term. The
189     explicit treatment of the Coriolis term does not represent a severe
190     limitation because it restricts the time step to approximately the
191     same length as in the ocean model where the Coriolis term is also
192     treated explicitly.
193    
194     \citet{hunke97}'s introduced an elastic contribution to the strain
195     rate elatic-viscous-plastic in order to regularize
196     Eq.\refeq{vpequation} in such a way that the resulting
197     elatic-viscous-plastic (EVP) and VP models are identical at steady
198     state,
199     \begin{equation}
200     \label{eq:evpequation}
201     \frac{1}{E}\frac{\partial\sigma_{ij}}{\partial{t}} +
202     \frac{1}{2\eta}\sigma_{ij}
203     + \frac{\eta - \zeta}{4\zeta\eta}\sigma_{kk}\delta_{ij}
204     + \frac{P}{4\zeta}\delta_{ij}
205     = \dot{\epsilon}_{ij}.
206     \end{equation}
207     %In the EVP model, equations for the components of the stress tensor
208     %$\sigma_{ij}$ are solved explicitly. Both model formulations will be
209     %used and compared the present sea-ice model study.
210     The EVP-model uses an explicit time stepping scheme with a short
211     timestep. According to the recommendation of \citet{hunke97}, the
212     EVP-model is stepped forward in time 120 times within the physical
213     ocean model time step (although this parameter is under debate), to
214     allow for elastic waves to disappear. Because the scheme does not
215     require a matrix inversion it is fast in spite of the small timestep
216     \citep{hunke97}. For completeness, we repeat the equations for the
217     components of the stress tensor $\sigma_{1} =
218     \sigma_{11}+\sigma_{22}$, $\sigma_{2}= \sigma_{11}-\sigma_{22}$, and
219     $\sigma_{12}$. Introducing the divergence $D_D =
220     \dot{\epsilon}_{11}+\dot{\epsilon}_{22}$, and the horizontal tension
221     and shearing strain rates, $D_T =
222     \dot{\epsilon}_{11}-\dot{\epsilon}_{22}$ and $D_S =
223     2\dot{\epsilon}_{12}$, respectively and using the above abbreviations,
224     the equations can be written as:
225     \begin{align}
226     \label{eq:evpstresstensor1}
227     \frac{\partial\sigma_{1}}{\partial{t}} + \frac{\sigma_{1}}{2T} +
228     \frac{P}{2T} &= \frac{P}{2T\Delta} D_D \\
229     \label{eq:evpstresstensor2}
230     \frac{\partial\sigma_{2}}{\partial{t}} + \frac{\sigma_{2} e^{2}}{2T}
231     &= \frac{P}{2T\Delta} D_T \\
232     \label{eq:evpstresstensor12}
233     \frac{\partial\sigma_{12}}{\partial{t}} + \frac{\sigma_{12} e^{2}}{2T}
234     &= \frac{P}{4T\Delta} D_S
235     \end{align}
236     Here, the elastic parameter $E$ is redefined in terms of a damping timescale
237     $T$ for elastic waves \[E=\frac{\zeta}{T}.\]
238     $T=E_{0}\Delta{t}$ with the tunable parameter $E_0<1$ and
239     the external (long) timestep $\Delta{t}$. \citet{hunke97} recommend
240     $E_{0} = \frac{1}{3}$.
241    
242     For details of the spatial discretization, the reader is referred to
243     \citet{zhang98, zhang03}. Our discretization differs only (but
244     importantly) in the underlying grid, namely the Arakawa C-grid, but is
245     otherwise straightforward. The EVP model in particular is discretized
246     naturally on the C-grid with $\sigma_{1}$ and $\sigma_{2}$ on the
247     center points and $\sigma_{12}$ on the corner (or vorticity) points of
248     the grid. With this choice all derivatives are discretized as central
249     differences and averaging is only involved in computing $\Delta$ and
250     $P$ at vorticity points.
251    
252     For a general curvilinear grid, one needs in principle to take metric
253     terms into account that arise in the transformation a curvilinear grid
254     on the sphere. However, for now we can neglect these metric terms
255     because they are very small on the cubed sphere grids used in this
256     paper; in particular, only near the edges of the cubed sphere grid, we
257     expect them to be non-zero, but these edges are at approximately
258     35\degS\ or 35\degN\ which are never covered by sea-ice in our
259     simulations. Everywhere else the coordinate system is locally nearly
260     cartesian. However, for last-glacial-maximum or snowball-earth-like
261     simulations the question of metric terms needs to be reconsidered.
262     Either, one includes these terms as in \citet{zhang03}, or one finds a
263     vector-invariant formulation fo the sea-ice internal stress term that
264     does not require any metric terms, as it is done in the ocean dynamics
265     of the MITgcm \citep{adcroft04:_cubed_sphere}.
266    
267     Moving sea ice exerts a stress on the ocean which is the opposite of
268     the stress $\vtau_{ocean}$ in Eq.\refeq{momseaice}. This stess is
269     applied directly to the surface layer of the ocean model. An
270     alternative ocean stress formulation is given by \citet{hibler87}.
271     Rather than applying $\vtau_{ocean}$ directly, the stress is derived
272     from integrating over the ice thickness to the bottom of the oceanic
273     surface layer. In the resulting equation for the \emph{combined}
274     ocean-ice momentum, the interfacial stress cancels and the total
275     stress appears as the sum of windstress and divergence of internal ice
276     stresses: $\delta(z) (\vtau_{air} + \vek{F})/\rho_0$, \citep[see also
277     Eq.\,2 of][]{hibler87}. The disadvantage of this formulation is that
278     now the velocity in the surface layer of the ocean that is used to
279     advect tracers, is really an average over the ocean surface
280     velocity and the ice velocity leading to an inconsistency as the ice
281     temperature and salinity are different from the oceanic variables.
282    
283     Sea ice distributions are characterized by sharp gradients and edges.
284     For this reason, we employ a positive 3rd-order advection scheme
285     \citep{hundsdorfer94} for the thermodynamic variables discussed in the
286     next section.
287    
288     \subparagraph{boundary conditions: no-slip, free-slip, half-slip}
289    
290     \begin{itemize}
291     \item transition from existing B-Grid to C-Grid
292     \item boundary conditions: no-slip, free-slip, half-slip
293     \item fancy (multi dimensional) advection schemes
294     \item VP vs.\ EVP \citep{hunke97}
295     \item ocean stress formulation \citep{hibler87}
296     \end{itemize}
297    
298     \subsection{Thermodynamics}
299     \label{sec:thermodynamics}
300    
301     In the original formulation the sea ice model \citep{menemenlis05}
302     uses simple thermodynamics following the appendix of
303     \citet{semtner76}. This formulation does not allow storage of heat
304     (heat capacity of ice is zero, and this type of model is often refered
305     to as a ``zero-layer'' model). Upward heat flux is parameterized
306     assuming a linear temperature profile and together with a constant ice
307     conductivity. It is expressed as $(K/h)(T_{w}-T_{0})$, where $K$ is
308     the ice conductivity, $h$ the ice thickness, and $T_{w}-T_{0}$ the
309     difference between water and ice surface temperatures. The surface
310     heat budget is computed in a similar way to that of
311     \citet{parkinson79} and \citet{manabe79}.
312    
313     There is considerable doubt about the reliability of such a simple
314     thermodynamic model---\citet{semtner84} found significant errors in
315     phase (one month lead) and amplitude ($\approx$50\%\,overestimate) in
316     such models---, so that today many sea ice models employ more complex
317     thermodynamics. A popular thermodynamics model of \citet{winton00} is
318     based on the 3-layer model of \citet{semtner76} and treats brine
319     content by means of enthalphy conservation. This model requires in
320     addition to ice-thickness and compactness (fractional area) additional
321     state variables to be advected by ice velocities, namely enthalphy of
322     the two ice layers and the thickness of the overlying snow layer.
323    
324     \section{Funnel Experiments}
325     \label{sec:funnel}
326    
327 mlosch 1.4 For a first/detailed comparison between the different variants of the
328     MIT sea ice model an idealized geometry of a periodic channel,
329     1000\,km long and 500\,m wide on a non-rotating plane, with converging
330     walls forming a symmetric funnel and a narrow strait of 40\,km width
331     is used. The horizontal resolution is 5\,km throughout the domain
332     making the narrow strait 8 grid points wide. The ice model is
333     initialized with a complete ice cover of 50\,cm uniform thickness. The
334     ice model is driven by a constant along channel eastward ocean current
335     of 25\,cm/s that does not see the walls in the domain. All other
336     ice-ocean-atmosphere interactions are turned off, in particular there
337     is no feedback of ice dynamics on the ocean current. All thermodynamic
338     processes are turned off so that ice thickness variations are only
339     caused by convergent or divergent ice flow. Ice volume (effective
340     thickness) and concentration are advected with a third-order scheme
341     with a flux limiter \citep{hundsdorfer94} to avoid undershoots. This
342     scheme is unconditionally stable and does not require additional
343     diffusion. The time step used here is 1\,h.
344    
345     \reffig{funnelf0} compares the dynamic fields ice concentration $c$,
346     effective thickness $h_{eff} = h\cdot{c}$, and velocities $(u,v)$ for
347     five different cases at steady state (after 10\,years of integration):
348     \begin{description}
349     \item[B-LSRns:] LSR solver with no-slip boundary conditions on a B-grid,
350     \item[C-LSRns:] LSR solver with no-slip boundary conditions on a C-grid,
351     \item[C-LSRfs:] LSR solver with free-slip boundary conditions on a C-grid,
352     \item[C-EVPns:] EVP solver with no-slip boundary conditions on a C-grid,
353     \item[C-EVPfs:] EVP solver with free-slip boundary conditions on a C-grid,
354     \end{description}
355     \ml{[We have not implemented the EVP solver on a B-grid.]}
356     \begin{figure*}[htbp]
357 dimitri 1.8 \includegraphics[width=\widefigwidth]{\fpath/all_086280}
358 mlosch 1.4 \caption{Ice concentration, effective thickness [m], and ice
359     velocities [m/s]
360     for 5 different numerical solutions.}
361     \label{fig:funnelf0}
362     \end{figure*}
363     At a first glance, the solutions look similar. This is encouraging as
364     the details of discretization and numerics should not affect the
365     solutions to first order. In all cases the ice-ocean stress pushes the
366     ice cover eastwards, where it converges in the funnel. In the narrow
367     channel the ice moves quickly (nearly free drift) and leaves the
368     channel as narrow band.
369    
370     A close look reveals interesting differences between the B- and C-grid
371     results. The zonal velocity in the narrow channel is nearly the free
372     drift velocity ( = ocean velocity) of 25\,cm/s for the C-grid
373     solutions, regardless of the boundary conditions, while it is just
374     above 20\,cm/s for the B-grid solution. The ice accelerates to
375     25\,cm/s after it exits the channel. Concentrating on the solutions
376     B-LSRns and C-LSRns, the ice volume (effective thickness) along the
377     boundaries in the narrow channel is larger in the B-grid case although
378     the ice concentration is reduces in the C-grid case. The combined
379     effect leads to a larger actual ice thickness at smaller
380     concentrations in the C-grid case. However, since the effective
381     thickness determines the ice strength $P$ in Eq\refeq{icestrength},
382     the ice strength and thus the bulk and shear viscosities are larger in
383     the B-grid case leading to more horizontal friction. This circumstance
384     might explain why the no-slip boundary conditions in the B-grid case
385     appear to be more effective in reducing the flow within the narrow
386     channel, than in the C-grid case. Further, the viscosities are also
387     sensitive to details of the velocity gradients. Via $\Delta$, these
388     gradients enter the viscosities in the denominator so that large
389     gradients tend to reduce the viscosities. This again favors more flow
390     along the boundaries in the C-grid case: larger velocities
391     (\reffig{funnelf0}) on grid points that are closer to the boundary by
392     a factor $\frac{1}{2}$ than in the B-grid case because of the stagger
393     nature of the C-grid lead numerically to larger tangential gradients
394     across the boundary; these in turn make the viscosities smaller for
395     less tangential friction and allow more tangential flow along the
396     boundaries.
397    
398     The above argument can also be invoked to explain the small
399     differences between the free-slip and no-slip solutions on the C-grid.
400     Because of the non-linearities in the ice viscosities, in particular
401 mlosch 1.6 along the boundaries, the no-slip boundary conditions have only a small
402 mlosch 1.4 impact on the solution.
403    
404     The difference between LSR and EVP solutions is largest in the
405 mlosch 1.6 effective thickness and meridional velocity fields. The EVP velocity
406     fields appears to be a little noisy. This noise has been address by
407 mlosch 1.4 \citet{hunke01}. It can be dealt with by reducing EVP's internal time
408 mlosch 1.5 step (increasing the number of iterations along with the computational
409     cost) or by regularizing the bulk and shear viscosities. We revisit
410     the latter option by reproducing some of the results of
411     \citet{hunke01}, namely the experiment described in her section~4, for
412     our C-grid no-slip cases: in a square domain with a few islands the
413     ice model is initialized with constant ice thickness and linearly
414     increasing ice concentration to the east. The model dynamics are
415 mlosch 1.6 forced with a constant anticyclonic ocean gyre and by variable
416     atmospheric wind whose mean direction is diagnonal to the north-east
417 mlosch 1.5 corner of the domain; ice volume and concentration are held constant
418 mlosch 1.6 (no thermodynamics and no advection by ice velocity).
419     \reffig{hunke01} shows the ice velocity field, its divergence, and the
420     bulk viscosity $\zeta$ for the cases C-LRSns and C-EVPns, and for a
421     C-EVPns case, where \citet{hunke01}'s regularization has been
422     implemented; compare to Fig.\,4 in \citet{hunke01}. The regularization
423     contraint limits ice strength and viscosities as a function of damping
424     time scale, resolution and EVP-time step, effectively allowing the
425     elastic waves to damp out more quickly \citep{hunke01}.
426 mlosch 1.4 \begin{figure*}[htbp]
427 dimitri 1.8 \includegraphics[width=\widefigwidth]{\fpath/hun12days}
428 mlosch 1.6 \caption{Ice flow, divergence and bulk viscosities of three
429     experiments with \citet{hunke01}'s test case: C-LSRns (top),
430     C-EVPns (middle), and C-EVPns with damping described in
431     \citet{hunke01} (bottom).}
432 mlosch 1.4 \label{fig:hunke01}
433     \end{figure*}
434    
435 mlosch 1.5 In the far right (``east'') side of the domain the ice concentration
436     is close to one and the ice should be nearly rigid. The applied wind
437     tends to push ice toward the upper right corner. Because the highly
438 mlosch 1.6 compact ice is confined by the boundary, it resists any further
439 mlosch 1.5 compression and exhibits little motion in the rigid region on the
440     right hand side. The C-LSRns solution (top row) allows high
441     viscosities in the rigid region suppressing nearly all flow.
442     \citet{hunke01}'s regularization for the C-EVPns solution (bottom row)
443 mlosch 1.6 clearly suppresses the noise present in $\nabla\cdot\vek{u}$ and
444     $\log_{10}\zeta$ in the
445     unregularized case (middle row), at the cost of reduced viscosities.
446 mlosch 1.5 These reduced viscosities lead to small but finite ice velocities
447     which in turn can have a strong effect on solutions in the limit of
448     nearly rigid regimes (arching and blocking, not shown).
449    
450 mlosch 1.9 \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 dimitri 1.1 \subsection{C-grid}
462     \begin{itemize}
463     \item no-slip vs. free-slip for both lsr and evp;
464     "diagnostics" to look at and use for comparison
465     \begin{itemize}
466     \item ice transport through Fram Strait/Denmark Strait/Davis
467     Strait/Bering strait (these are general)
468     \item ice transport through narrow passages, e.g.\ Nares-Strait
469     \end{itemize}
470     \item compare different advection schemes (if lsr turns out to be more
471     effective, then with lsr otherwise I prefer evp), eg.
472     \begin{itemize}
473     \item default 2nd-order with diff1=0.002
474     \item 1st-order upwind with diff1=0.
475     \item DST3FL (SEAICEadvScheme=33 with diff1=0., should work, works for me)
476     \item 2nd-order wit flux limiter (SEAICEadvScheme=77 with diff1=0.)
477     \end{itemize}
478     That should be enough. Here, total ice mass and location of ice edge
479     is interesting. However, this comparison can be done in an idealized
480     domain, may not require full Arctic Domain?
481     \item
482     Do a little study on the parameters of LSR and EVP
483     \begin{enumerate}
484     \item convergence of LSR, how many iterations do you need to get a
485     true elliptic yield curve
486     \item vary deltaTevp and the relaxation parameter for EVP and see when
487     the EVP solution breaks down (relative to the forcing time scale).
488     For this, it is essential that the evp solver gives use "stripeless"
489     solutions, that is your dtevp = 1sec solutions/or 10sec solutions
490     with SEAICE\_evpDampC = 615.
491     \end{enumerate}
492     \end{itemize}
493    
494     \section{Forward sensitivity experiments}
495     \label{sec:forward}
496    
497     A second series of forward sensitivity experiments have been carried out on an
498     Arctic Ocean domain with open boundaries. Once again the objective is to
499     compare the old B-grid LSR dynamic solver with the new C-grid LSR and EVP
500     solvers. One additional experiment is carried out to illustrate the
501     differences between the two main options for sea ice thermodynamics in the MITgcm.
502    
503     \subsection{Arctic Domain with Open Boundaries}
504     \label{sec:arctic}
505    
506 mlosch 1.6 The Arctic domain of integration is illustrated in Fig.~\ref{???}. It
507     is carved out from, and obtains open boundary conditions from, the
508     global cubed-sphere configuration of the Estimating the Circulation
509     and Climate of the Ocean, Phase II (ECCO2) project
510     \citet{menemenlis05}. The domain size is 420 by 384 grid boxes
511     horizontally with mean horizontal grid spacing of 18 km.
512 dimitri 1.1
513     There are 50 vertical levels ranging in thickness from 10 m near the surface
514     to approximately 450 m at a maximum model depth of 6150 m. Bathymetry is from
515     the National Geophysical Data Center (NGDC) 2-minute gridded global relief
516     data (ETOPO2) and the model employs the partial-cell formulation of
517 mlosch 1.6 \citet{adcroft97:_shaved_cells}, which permits accurate representation of the bathymetry. The
518 dimitri 1.1 model is integrated in a volume-conserving configuration using a finite volume
519     discretization with C-grid staggering of the prognostic variables. In the
520 mlosch 1.6 ocean, the non-linear equation of state of \citet{jackett95}. The ocean model is
521 dimitri 1.1 coupled to a sea-ice model described hereinabove.
522    
523 mlosch 1.6 This particular ECCO2 simulation is initialized from rest using the
524     January temperature and salinity distribution from the World Ocean
525     Atlas 2001 (WOA01) [Conkright et al., 2002] and it is integrated for
526     32 years prior to the 1996--2001 period discussed in the study. Surface
527     boundary conditions are from the National Centers for Environmental
528     Prediction and the National Center for Atmospheric Research
529     (NCEP/NCAR) atmospheric reanalysis [Kistler et al., 2001]. Six-hourly
530     surface winds, temperature, humidity, downward short- and long-wave
531     radiations, and precipitation are converted to heat, freshwater, and
532     wind stress fluxes using the \citet{large81, large82} bulk formulae.
533     Shortwave radiation decays exponentially as per Paulson and Simpson
534     [1977]. Additionally the time-mean river run-off from Large and Nurser
535     [2001] is applied and there is a relaxation to the monthly-mean
536     climatological sea surface salinity values from WOA01 with a
537     relaxation time scale of 3 months. Vertical mixing follows
538     \citet{large94} with background vertical diffusivity of
539     $1.5\times10^{-5}\text{\,m$^{2}$\,s$^{-1}$}$ and viscosity of
540     $10^{-3}\text{\,m$^{2}$\,s$^{-1}$}$. A third order, direct-space-time
541     advection scheme with flux limiter is employed \citep{hundsdorfer94}
542     and there is no explicit horizontal diffusivity. Horizontal viscosity
543     follows \citet{lei96} but
544     modified to sense the divergent flow as per Fox-Kemper and Menemenlis
545     [in press]. Shortwave radiation decays exponentially as per Paulson
546     and Simpson [1977]. Additionally, the time-mean runoff of Large and
547     Nurser [2001] is applied near the coastline and, where there is open
548     water, there is a relaxation to monthly-mean WOA01 sea surface
549     salinity with a time constant of 45 days.
550 dimitri 1.1
551     Open water, dry
552     ice, wet ice, dry snow, and wet snow albedo are, respectively, 0.15, 0.85,
553     0.76, 0.94, and 0.8.
554    
555     \begin{itemize}
556     \item Configuration
557     \item OBCS from cube
558     \item forcing
559     \item 1/2 and full resolution
560     \item with a few JFM figs from C-grid LSR no slip
561     ice transport through Canadian Archipelago
562     thickness distribution
563     ice velocity and transport
564     \end{itemize}
565    
566     \begin{itemize}
567     \item Arctic configuration
568     \item ice transport through straits and near boundaries
569     \item focus on narrow straits in the Canadian Archipelago
570     \end{itemize}
571    
572     \begin{itemize}
573     \item B-grid LSR no-slip
574     \item C-grid LSR no-slip
575     \item C-grid LSR slip
576     \item C-grid EVP no-slip
577     \item C-grid EVP slip
578 mlosch 1.6 \item C-grid LSR + TEM (truncated ellipse method, no tensile stress, new flag)
579 dimitri 1.1 \item C-grid LSR no-slip + Winton
580     \item speed-performance-accuracy (small)
581     ice transport through Canadian Archipelago differences
582     thickness distribution differences
583     ice velocity and transport differences
584     \end{itemize}
585    
586     We anticipate small differences between the different models due to:
587     \begin{itemize}
588     \item advection schemes: along the ice-edge and regions with large
589     gradients
590 mlosch 1.6 \item C-grid: less transport through narrow straits for no slip
591     conditons, more for free slip
592 dimitri 1.1 \item VP vs.\ EVP: speed performance, accuracy?
593     \item ocean stress: different water mass properties beneath the ice
594     \end{itemize}
595    
596     \section{Adjoint sensiivities of the MITsim}
597    
598     \subsection{The adjoint of MITsim}
599    
600     The ability to generate tangent linear and adjoint model components
601     of the MITsim has been a main design task.
602     For the ocean the adjoint capability has proven to be an
603     invaluable tool for sensitivity analysis as well as state estimation.
604     In short, the adjoint enables very efficient computation of the gradient
605     of scalar-valued model diagnostics (called cost function or objective function)
606     with respect to many model "variables".
607     These variables can be two- or three-dimensional fields of initial
608     conditions, model parameters such as mixing coefficients, or
609     time-varying surface or lateral (open) boundary conditions.
610     When combined, these variables span a potentially high-dimensional
611     (e.g. O(10$^8$)) so-called control space. Performing parameter perturbations
612     to assess model sensitivities quickly becomes prohibitive at these scales.
613     Alternatively, (time-varying) sensitivities of the objective function
614     to any element of the control space can be computed very efficiently in
615     one single adjoint
616     model integration, provided an efficient adjoint model is available.
617    
618     [REFERENCES]
619    
620    
621     The adjoint operator (ADM) is the transpose of the tangent linear operator (TLM)
622     of the full (in general nonlinear) forward model, i.e. the MITsim.
623     The TLM maps perturbations of elements of the control space
624     (e.g. initial ice thickness distribution)
625     via the model Jacobian
626     to a perturbation in the objective function
627     (e.g. sea-ice export at the end of the integration interval).
628     \textit{Tangent} linearity ensures that the derivatives are evaluated
629     with respect to the underlying model trajectory at each point in time.
630     This is crucial for nonlinear trajectories and the presence of different
631     regimes (e.g. effect of the seaice growth term at or away from the
632     freezing point of the ocean surface).
633     Ensuring tangent linearity can be easily achieved by integrating
634     the full model in sync with the TLM to provide the underlying model state.
635     Ensuring \textit{tangent} adjoints is equally crucial, but much more
636     difficult to achieve because of the reverse nature of the integration:
637     the adjoint accumulates sensitivities backward in time,
638     starting from a unit perturbation of the objective function.
639     The adjoint model requires the model state in reverse order.
640     This presents one of the major complications in deriving an
641     exact, i.e. \textit{tangent} adjoint model.
642    
643     Following closely the development and maintenance of TLM and ADM
644     components of the MITgcm we have relied heavily on the
645     autmomatic differentiation (AD) tool
646     "Transformation of Algorithms in Fortran" (TAF)
647     developed by Fastopt (Giering and Kaminski, 1998)
648     to derive TLM and ADM code of the MITsim.
649     Briefly, the nonlinear parent model is fed to the AD tool which produces
650     derivative code for the specified control space and objective function.
651     Following this approach has (apart from its evident success)
652     several advantages:
653     (1) the adjoint model is the exact adjoint operator of the parent model,
654     (2) the adjoint model can be kept up to date with respect to ongoing
655     development of the parent model, and adjustments to the parent model
656     to extend the automatically generated adjoint are incremental changes
657     only, rather than extensive re-developments,
658     (3) the parallel structure of the parent model is preserved
659     by the adjoint model, ensuring efficient use in high performance
660     computing environments.
661    
662     Some initial code adjustments are required to support dependency analysis
663     of the flow reversal and certain language limitations which may lead
664     to irreducible flow graphs (e.g. GOTO statements).
665     The problem of providing the required model state in reverse order
666     at the time of evaluating nonlinear or conditional
667     derivatives is solved via balancing
668     storing vs. recomputation of the model state in a multi-level
669     checkpointing loop.
670     Again, an initial code adjustment is required to support TAFs
671     checkpointing capability.
672 mlosch 1.6 The code adjustments are sufficiently simple so as not to cause
673 dimitri 1.1 major limitations to the full nonlinear parent model.
674     Once in place, an adjoint model of a new model configuration
675     may be derived in about 10 minutes.
676    
677     [HIGHLIGHT COUPLED NATURE OF THE ADJOINT!]
678    
679     \subsection{Special considerations}
680    
681     * growth term(?)
682    
683     * small active denominators
684    
685     * dynamic solver (implicit function theorem)
686    
687     * approximate adjoints
688    
689    
690     \subsection{An example: sensitivities of sea-ice export through Fram Strait}
691    
692     We demonstrate the power of the adjoint method
693     in the context of investigating sea-ice export sensitivities through Fram Strait
694     (for details of this study see Heimbach et al., 2007).
695 mlosch 1.6 %\citep[for details of this study see][]{heimbach07}. %Heimbach et al., 2007).
696 dimitri 1.1 The domain chosen is a coarsened version of the Arctic face of the
697     high-resolution cubed-sphere configuration of the ECCO2 project
698 mlosch 1.6 \citep[see][]{menemenlis05}. It covers the entire Arctic,
699 dimitri 1.1 extends into the North Pacific such as to cover the entire
700     ice-covered regions, and comprises parts of the North Atlantic
701     down to XXN to enable analysis of remote influences of the
702     North Atlantic current to sea-ice variability and export.
703     The horizontal resolution varies between XX and YY km
704     with 50 unevenly spaced vertical levels.
705     The adjoint models run efficiently on 80 processors
706     (benchmarks have been performed both on an SGI Altix as well as an
707     IBM SP5 at NASA/ARC).
708    
709 mlosch 1.6 Following a 1-year spinup, the model has been integrated for four
710     years between 1992 and 1995. It is forced using realistic 6-hourly
711     NCEP/NCAR atmospheric state variables. Over the open ocean these are
712     converted into air-sea fluxes via the bulk formulae of
713     \citet{large04}. Derivation of air-sea fluxes in the presence of
714     sea-ice is handled by the ice model as described in \refsec{model}.
715 dimitri 1.1 The objective function chosen is sea-ice export through Fram Strait
716 mlosch 1.6 computed for December 1995. The adjoint model computes sensitivities
717     to sea-ice export back in time from 1995 to 1992 along this
718     trajectory. In principle all adjoint model variable (i.e., Lagrange
719     multipliers) of the coupled ocean/sea-ice model are available to
720     analyze the transient sensitivity behaviour of the ocean and sea-ice
721     state. Over the open ocean, the adjoint of the bulk formula scheme
722     computes sensitivities to the time-varying atmospheric state. Over
723     ice-covered parts, the sea-ice adjoint converts surface ocean
724     sensitivities to atmospheric sensitivities.
725    
726     \reffig{4yradjheff}(a--d) depict sensitivities of sea-ice export
727     through Fram Strait in December 1995 to changes in sea-ice thickness
728     12, 24, 36, 48 months back in time. Corresponding sensitivities to
729     ocean surface temperature are depicted in
730     \reffig{4yradjthetalev1}(a--d). The main characteristics is
731     consistency with expected advection of sea-ice over the relevant time
732     scales considered. The general positive pattern means that an
733     increase in sea-ice thickness at location $(x,y)$ and time $t$ will
734     increase sea-ice export through Fram Strait at time $T_e$. Largest
735     distances from Fram Strait indicate fastest sea-ice advection over the
736     time span considered. The ice thickness sensitivities are in close
737     correspondence to ocean surface sentivitites, but of opposite sign.
738     An increase in temperature will incur ice melting, decrease in ice
739     thickness, and therefore decrease in sea-ice export at time $T_e$.
740 dimitri 1.1
741     The picture is fundamentally different and much more complex
742     for sensitivities to ocean temperatures away from the surface.
743 mlosch 1.6 \reffig{4yradjthetalev10??}(a--d) depicts ice export sensitivities to
744 dimitri 1.1 temperatures at roughly 400 m depth.
745     Primary features are the effect of the heat transport of the North
746     Atlantic current which feeds into the West Spitsbergen current,
747     the circulation around Svalbard, and ...
748    
749     \begin{figure}[t!]
750     \centerline{
751     \subfigure[{\footnotesize -12 months}]
752 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim072_cmax2.0E+02.eps}}
753 dimitri 1.1 %\includegraphics*[width=.3\textwidth]{H_c.bin_res_100_lev1.pdf}
754     %
755     \subfigure[{\footnotesize -24 months}]
756 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim145_cmax2.0E+02.eps}}
757 dimitri 1.1 }
758    
759     \centerline{
760     \subfigure[{\footnotesize
761     -36 months}]
762 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim218_cmax2.0E+02.eps}}
763 dimitri 1.1 %
764     \subfigure[{\footnotesize
765     -48 months}]
766 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim292_cmax2.0E+02.eps}}
767 dimitri 1.1 }
768     \caption{Sensitivity of sea-ice export through Fram Strait in December 2005 to
769     sea-ice thickness at various prior times.
770     \label{fig:4yradjheff}}
771     \end{figure}
772    
773    
774     \begin{figure}[t!]
775     \centerline{
776     \subfigure[{\footnotesize -12 months}]
777 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim072_cmax5.0E+01.eps}}
778 dimitri 1.1 %\includegraphics*[width=.3\textwidth]{H_c.bin_res_100_lev1.pdf}
779     %
780     \subfigure[{\footnotesize -24 months}]
781 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim145_cmax5.0E+01.eps}}
782 dimitri 1.1 }
783    
784     \centerline{
785     \subfigure[{\footnotesize
786     -36 months}]
787 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim218_cmax5.0E+01.eps}}
788 dimitri 1.1 %
789     \subfigure[{\footnotesize
790     -48 months}]
791 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim292_cmax5.0E+01.eps}}
792 dimitri 1.1 }
793 mlosch 1.6 \caption{Same as \reffig{4yradjheff} but for sea surface temperature
794 dimitri 1.1 \label{fig:4yradjthetalev1}}
795     \end{figure}
796    
797    
798    
799     \section{Discussion and conclusion}
800     \label{sec:concl}
801    
802     The story of the paper could be:
803     B-grid ice model + C-grid ocean model does not work properly for us,
804     therefore C-grid ice model with advantages:
805     \begin{enumerate}
806     \item stress coupling simpler (no interpolation required)
807     \item different boundary conditions
808     \item advection schemes carry over trivially from main code
809     \end{enumerate}
810     LSR/EVP solutions are similar with appropriate bcs, evp parameters as
811     a function of forcing time scale (when does VP solution break
812     down). Same for LSR solver, provided that it works (o:
813     Which scheme is more efficient for the resolution/time stepping
814     parameters that we use here. What about other resolutions?
815    
816     \paragraph{Acknowledgements}
817     We thank Jinlun Zhang for providing the original B-grid code and many
818 mlosch 1.6 helpful discussions. ML thanks Elizabeth Hunke for multiple explanations.
819 dimitri 1.1
820 dimitri 1.7 \bibliography{bib/journal_abrvs,bib/seaice,bib/genocean,bib/maths,bib/mitgcmuv,bib/fram}
821 dimitri 1.1
822     \end{document}
823    
824     %%% Local Variables:
825     %%% mode: latex
826     %%% TeX-master: t
827     %%% End:
828 mlosch 1.4
829    
830     A Dynamic-Thermodynamic Sea ice Model for Ocean Climate
831     Estimation on an Arakawa C-Grid
832    
833     Introduction
834    
835     Ice Model:
836     Dynamics formulation.
837     B-C, LSR, EVP, no-slip, slip
838     parallellization
839     Thermodynamics formulation.
840     0-layer Hibler salinity + snow
841     3-layer Winton
842    
843     Idealized tests
844     Funnel Experiments
845     Downstream Island tests
846     B-grid LSR no-slip
847     C-grid LSR no-slip
848     C-grid LSR slip
849     C-grid EVP no-slip
850     C-grid EVP slip
851    
852     Arctic Setup
853     Configuration
854     OBCS from cube
855     forcing
856     1/2 and full resolution
857     with a few JFM figs from C-grid LSR no slip
858     ice transport through Canadian Archipelago
859     thickness distribution
860     ice velocity and transport
861    
862     Arctic forward sensitivity experiments
863     B-grid LSR no-slip
864     C-grid LSR no-slip
865     C-grid LSR slip
866     C-grid EVP no-slip
867     C-grid EVP slip
868     C-grid LSR no-slip + Winton
869     speed-performance-accuracy (small)
870     ice transport through Canadian Archipelago differences
871     thickness distribution differences
872     ice velocity and transport differences
873    
874     Adjoint sensitivity experiment on 1/2-res setup
875     Sensitivity of sea ice volume flow through Fram Strait
876     *** Sensitivity of sea ice volume flow through Canadian Archipelago
877    
878     Summary and conluding remarks

  ViewVC Help
Powered by ViewVC 1.1.22