/[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 1.15 - (hide annotations) (download) (as text)
Tue Feb 26 17:21:48 2008 UTC (17 years, 4 months ago) by mlosch
Branch: MAIN
Changes since 1.14: +48 -32 lines
File MIME type: application/x-tex
fix a few units, spelling etc.
add a paragraph about lateral boundary conditions
add a few comments about advection scheme (and something for JMC to
comment on)

1 mlosch 1.15 % $Header: /u/gcmpack/MITgcm_contrib/articles/ceaice/ceaice.tex,v 1.14 2008/02/26 00:13:20 dimitri Exp $
2 dimitri 1.10 % $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 As part of ongoing efforts to obtain a best possible synthesis of most
56 dimitri 1.12 available, global-scale, ocean and sea ice data, a dynamic and thermodynamic
57     sea-ice model has been coupled to the Massachusetts Institute of Technology
58     general circulation model (MITgcm). Ice mechanics follow a viscous plastic
59     rheology and the ice momentum equations are solved numerically using either
60     line successive relaxation (LSR) or elastic-viscous-plastic (EVP) dynamic
61     models. Ice thermodynamics are represented using either a zero-heat-capacity
62     formulation or a two-layer formulation that conserves enthalpy. The model
63     includes prognostic variables for snow and for sea-ice salinity. The above
64     sea ice model components were borrowed from current-generation climate models
65     but they were reformulated on an Arakawa C-grid in order to match the MITgcm
66     oceanic grid and they were modified in many ways to permit efficient and
67     accurate automatic differentiation. This paper describes the MITgcm sea ice
68     model; it presents example Arctic and Antarctic results from a realistic,
69     eddy-permitting, global ocean and sea-ice configuration; it compares B-grid
70     and C-grid dynamic solvers in a regional Arctic configuration; and it presents
71     example results from coupled ocean and sea-ice adjoint-model integrations.
72 dimitri 1.10
73 dimitri 1.1 \end{abstract}
74    
75     \section{Introduction}
76     \label{sec:intro}
77    
78 dimitri 1.14 The availability of an adjoint model as a powerful research
79     tool complementary to an ocean model was a major design
80     requirement early on in the development of the MIT general
81     circulation model (MITgcm) [Marshall et al. 1997a,
82     Marotzke et al. 1999, Adcroft et al. 2002]. It was recognized
83     that the adjoint permitted very efficient computation
84     of gradients of various scalar-valued model diagnostics,
85     norms or, generally, objective functions with respect
86     to external or independent parameters. Such gradients
87     arise in at least two major contexts. If the objective function
88     is the sum of squared model vs. obervation differences
89     weighted by e.g. the inverse error covariances, the gradient
90     of the objective function can be used to optimize this measure
91     of model vs. data misfit in a least-squares sense. One
92     is then solving a problem of statistical state estimation.
93     If the objective function is a key oceanographic quantity
94     such as meridional heat or volume transport, ocean heat
95     content or mean surface temperature index, the gradient
96     provides a complete set of sensitivities of this quantity
97     with respect to all independent variables simultaneously.
98    
99     References to existing sea-ice adjoint models, explaining that they are either
100     for simplified configurations, for ice-only studies, or for short-duration
101     studies to motivate the present work.
102    
103 dimitri 1.1 Traditionally, probably for historical reasons and the ease of
104     treating the Coriolis term, most standard sea-ice models are
105     discretized on Arakawa-B-grids \citep[e.g.,][]{hibler79, harder99,
106 dimitri 1.14 kreyscher00, zhang98, hunke97}. From the perspective of coupling a
107 dimitri 1.1 sea ice-model to a C-grid ocean model, the exchange of fluxes of heat
108     and fresh-water pose no difficulty for a B-grid sea-ice model
109     \citep[e.g.,][]{timmermann02a}. However, surface stress is defined at
110     velocities points and thus needs to be interpolated between a B-grid
111     sea-ice model and a C-grid ocean model. While the smoothing implicitly
112     associated with this interpolation may mask grid scale noise, it may
113     in two-way coupling lead to a computational mode as will be shown. By
114 dimitri 1.13 choosing a C-grid for the sea-ice model, we circumvent this difficulty
115 dimitri 1.1 altogether and render the stress coupling as consistent as the
116     buoyancy coupling.
117    
118     A further advantage of the C-grid formulation is apparent in narrow
119     straits. In the limit of only one grid cell between coasts there is no
120     flux allowed for a B-grid (with no-slip lateral boundary counditions),
121     whereas the C-grid formulation allows a flux of sea-ice through this
122 dimitri 1.14 passage for all types of lateral boundary conditions. We
123 dimitri 1.1 demonstrate this effect in the Candian archipelago.
124    
125 dimitri 1.14 Talk about problems that make the sea-ice-ocean code very sensitive and
126     changes in the code that reduce these sensitivities.
127    
128     This paper describes the MITgcm sea ice
129     model; it presents example Arctic and Antarctic results from a realistic,
130     eddy-permitting, global ocean and sea-ice configuration; it compares B-grid
131     and C-grid dynamic solvers in a regional Arctic configuration; and it presents
132     example results from coupled ocean and sea-ice adjoint-model integrations.
133    
134 dimitri 1.13 \section{Model}
135     \label{sec:model}
136    
137 dimitri 1.1 \subsection{Dynamics}
138     \label{sec:dynamics}
139    
140 dimitri 1.13 The momentum equation of the sea-ice model is
141 dimitri 1.1 \begin{equation}
142     \label{eq:momseaice}
143     m \frac{D\vek{u}}{Dt} = -mf\vek{k}\times\vek{u} + \vtau_{air} +
144 mlosch 1.15 \vtau_{ocean} - m \nabla{\phi(0)} + \vek{F},
145 dimitri 1.1 \end{equation}
146 dimitri 1.13 where $m=m_{i}+m_{s}$ is the ice and snow mass per unit area;
147     $\vek{u}=u\vek{i}+v\vek{j}$ is the ice velocity vector;
148     $\vek{i}$, $\vek{j}$, and $\vek{k}$ are unit vectors in the $x$, $y$, and $z$
149     directions, respectively;
150     $f$ is the Coriolis parameter;
151     $\vtau_{air}$ and $\vtau_{ocean}$ are the wind-ice and ocean-ice stresses,
152     respectively;
153     $g$ is the gravity accelation;
154     $\nabla\phi(0)$ is the gradient (or tilt) of the sea surface height;
155 mlosch 1.15 $\phi(0) = g\eta + p_{a}/\rho_{0}$ is the sea surface height potential
156     in response to ocean dynamics ($g\eta$) and to atmospheric pressure
157     loading ($p_{a}/\rho_{0}$, where $\rho_{0}$ is a reference density);
158 dimitri 1.13 and $\vek{F}=\nabla\cdot\sigma$ is the divergence of the internal ice stress
159     tensor $\sigma_{ij}$.
160     When using the rescaled vertical coordinate system, z$^\ast$, of
161 mlosch 1.15 \citet{cam08}, $\phi(0)$ also includes a term due to snow and ice
162     loading, $mg/\rho_{0}$.
163 dimitri 1.13 Advection of sea-ice momentum is neglected. The wind and ice-ocean stress
164     terms are given by
165 dimitri 1.1 \begin{align*}
166 dimitri 1.13 \vtau_{air} = & \rho_{air} C_{air} |\vek{U}_{air} -\vek{u}|
167     R_{air} (\vek{U}_{air} -\vek{u}), \\
168     \vtau_{ocean} = & \rho_{ocean}C_{ocean} |\vek{U}_{ocean}-\vek{u}|
169 dimitri 1.1 R_{ocean}(\vek{U}_{ocean}-\vek{u}), \\
170     \end{align*}
171     where $\vek{U}_{air/ocean}$ are the surface winds of the atmosphere
172 dimitri 1.13 and surface currents of the ocean, respectively; $C_{air/ocean}$ are
173     air and ocean drag coefficients; $\rho_{air/ocean}$ are reference
174     densities; and $R_{air/ocean}$ are rotation matrices that act on the
175     wind/current vectors.
176 dimitri 1.1
177 mlosch 1.15 For an isotropic system the stress tensor $\sigma_{ij}$ ($i,j=1,2$) can
178     be related to the ice strain rate and strength by a nonlinear
179     viscous-plastic (VP) constitutive law \citep{hibler79, zhang98}:
180 dimitri 1.1 \begin{equation}
181     \label{eq:vpequation}
182     \sigma_{ij}=2\eta(\dot{\epsilon}_{ij},P)\dot{\epsilon}_{ij}
183     + \left[\zeta(\dot{\epsilon}_{ij},P) -
184     \eta(\dot{\epsilon}_{ij},P)\right]\dot{\epsilon}_{kk}\delta_{ij}
185     - \frac{P}{2}\delta_{ij}.
186     \end{equation}
187     The ice strain rate is given by
188     \begin{equation*}
189     \dot{\epsilon}_{ij} = \frac{1}{2}\left(
190     \frac{\partial{u_{i}}}{\partial{x_{j}}} +
191     \frac{\partial{u_{j}}}{\partial{x_{i}}}\right).
192     \end{equation*}
193 mlosch 1.5 The maximum ice pressure $P_{\max}$, a measure of ice strength, depends on
194     both thickness $h$ and compactness (concentration) $c$:
195 mlosch 1.4 \begin{equation}
196 mlosch 1.5 P_{\max} = P^{*}c\,h\,e^{[C^{*}\cdot(1-c)]},
197 mlosch 1.9 \label{eq:icestrength}
198 mlosch 1.4 \end{equation}
199     with the constants $P^{*}$ and $C^{*}$. The nonlinear bulk and shear
200     viscosities $\eta$ and $\zeta$ are functions of ice strain rate
201     invariants and ice strength such that the principal components of the
202     stress lie on an elliptical yield curve with the ratio of major to
203     minor axis $e$ equal to $2$; they are given by:
204 dimitri 1.1 \begin{align*}
205 mlosch 1.5 \zeta =& \min\left(\frac{P_{\max}}{2\max(\Delta,\Delta_{\min})},
206     \zeta_{\max}\right) \\
207     \eta =& \frac{\zeta}{e^2} \\
208 dimitri 1.1 \intertext{with the abbreviation}
209     \Delta = & \left[
210     \left(\dot{\epsilon}_{11}^2+\dot{\epsilon}_{22}^2\right)
211     (1+e^{-2}) + 4e^{-2}\dot{\epsilon}_{12}^2 +
212     2\dot{\epsilon}_{11}\dot{\epsilon}_{22} (1-e^{-2})
213     \right]^{-\frac{1}{2}}
214     \end{align*}
215 mlosch 1.5 The bulk viscosities are bounded above by imposing both a minimum
216     $\Delta_{\min}=10^{-11}\text{\,s}^{-1}$ (for numerical reasons) and a
217     maximum $\zeta_{\max} = P_{\max}/\Delta^*$, where
218     $\Delta^*=(5\times10^{12}/2\times10^4)\text{\,s}^{-1}$. For stress
219 dimitri 1.13 tensor computation the replacement pressure $P = 2\,\Delta\zeta$
220 mlosch 1.5 \citep{hibler95} is used so that the stress state always lies on the
221     elliptic yield curve by definition.
222    
223 mlosch 1.6 In the so-called truncated ellipse method the shear viscosity $\eta$
224     is capped to suppress any tensile stress \citep{hibler97, geiger98}:
225     \begin{equation}
226     \label{eq:etatem}
227 mlosch 1.15 \eta = \min\left(\frac{\zeta}{e^2},
228 mlosch 1.6 \frac{\frac{P}{2}-\zeta(\dot{\epsilon}_{11}+\dot{\epsilon}_{22})}
229     {\sqrt{(\dot{\epsilon}_{11}+\dot{\epsilon}_{22})^2
230 mlosch 1.15 +4\dot{\epsilon}_{12}^2}}\right).
231 mlosch 1.6 \end{equation}
232    
233 dimitri 1.1 In the current implementation, the VP-model is integrated with the
234     semi-implicit line successive over relaxation (LSOR)-solver of
235     \citet{zhang98}, which allows for long time steps that, in our case,
236 mlosch 1.15 are limited by the explicit treatment of the Coriolis term. The
237 dimitri 1.1 explicit treatment of the Coriolis term does not represent a severe
238     limitation because it restricts the time step to approximately the
239     same length as in the ocean model where the Coriolis term is also
240     treated explicitly.
241    
242     \citet{hunke97}'s introduced an elastic contribution to the strain
243 mlosch 1.15 rate in order to regularize Eq.\refeq{vpequation} in such a way that
244     the resulting elastic-viscous-plastic (EVP) and VP models are
245     identical at steady state,
246 dimitri 1.1 \begin{equation}
247     \label{eq:evpequation}
248     \frac{1}{E}\frac{\partial\sigma_{ij}}{\partial{t}} +
249     \frac{1}{2\eta}\sigma_{ij}
250     + \frac{\eta - \zeta}{4\zeta\eta}\sigma_{kk}\delta_{ij}
251     + \frac{P}{4\zeta}\delta_{ij}
252     = \dot{\epsilon}_{ij}.
253     \end{equation}
254     %In the EVP model, equations for the components of the stress tensor
255     %$\sigma_{ij}$ are solved explicitly. Both model formulations will be
256     %used and compared the present sea-ice model study.
257     The EVP-model uses an explicit time stepping scheme with a short
258     timestep. According to the recommendation of \citet{hunke97}, the
259     EVP-model is stepped forward in time 120 times within the physical
260     ocean model time step (although this parameter is under debate), to
261     allow for elastic waves to disappear. Because the scheme does not
262     require a matrix inversion it is fast in spite of the small timestep
263     \citep{hunke97}. For completeness, we repeat the equations for the
264     components of the stress tensor $\sigma_{1} =
265     \sigma_{11}+\sigma_{22}$, $\sigma_{2}= \sigma_{11}-\sigma_{22}$, and
266     $\sigma_{12}$. Introducing the divergence $D_D =
267     \dot{\epsilon}_{11}+\dot{\epsilon}_{22}$, and the horizontal tension
268     and shearing strain rates, $D_T =
269     \dot{\epsilon}_{11}-\dot{\epsilon}_{22}$ and $D_S =
270 dimitri 1.13 2\dot{\epsilon}_{12}$, respectively, and using the above abbreviations,
271 dimitri 1.1 the equations can be written as:
272     \begin{align}
273     \label{eq:evpstresstensor1}
274     \frac{\partial\sigma_{1}}{\partial{t}} + \frac{\sigma_{1}}{2T} +
275     \frac{P}{2T} &= \frac{P}{2T\Delta} D_D \\
276     \label{eq:evpstresstensor2}
277     \frac{\partial\sigma_{2}}{\partial{t}} + \frac{\sigma_{2} e^{2}}{2T}
278     &= \frac{P}{2T\Delta} D_T \\
279     \label{eq:evpstresstensor12}
280     \frac{\partial\sigma_{12}}{\partial{t}} + \frac{\sigma_{12} e^{2}}{2T}
281     &= \frac{P}{4T\Delta} D_S
282     \end{align}
283     Here, the elastic parameter $E$ is redefined in terms of a damping timescale
284     $T$ for elastic waves \[E=\frac{\zeta}{T}.\]
285     $T=E_{0}\Delta{t}$ with the tunable parameter $E_0<1$ and
286     the external (long) timestep $\Delta{t}$. \citet{hunke97} recommend
287     $E_{0} = \frac{1}{3}$.
288    
289     For details of the spatial discretization, the reader is referred to
290     \citet{zhang98, zhang03}. Our discretization differs only (but
291     importantly) in the underlying grid, namely the Arakawa C-grid, but is
292 mlosch 1.15 otherwise straightforward. The EVP model, in particular, is discretized
293 dimitri 1.1 naturally on the C-grid with $\sigma_{1}$ and $\sigma_{2}$ on the
294     center points and $\sigma_{12}$ on the corner (or vorticity) points of
295     the grid. With this choice all derivatives are discretized as central
296     differences and averaging is only involved in computing $\Delta$ and
297     $P$ at vorticity points.
298    
299     For a general curvilinear grid, one needs in principle to take metric
300 mlosch 1.15 terms into account that arise in the transformation of a curvilinear
301     grid on the sphere. For now, however, we can neglect these metric
302     terms because they are very small on the \ml{[modify following
303     section3:] cubed sphere grids used in this paper; in particular,
304     only near the edges of the cubed sphere grid, we expect them to be
305     non-zero, but these edges are at approximately 35\degS\ or 35\degN\
306     which are never covered by sea-ice in our simulations. Everywhere
307     else the coordinate system is locally nearly cartesian.} However, for
308     last-glacial-maximum or snowball-earth-like simulations the question
309     of metric terms needs to be reconsidered. Either, one includes these
310     terms as in \citet{zhang03}, or one finds a vector-invariant
311     formulation for the sea-ice internal stress term that does not require
312     any metric terms, as it is done in the ocean dynamics of the MITgcm
313     \citep{adcroft04:_cubed_sphere}.
314    
315     Lateral boundary conditions are naturally ``no-slip'' for B-grids, as
316     the tangential velocities points lie on the boundary. For C-grids, the
317     lateral boundary condition for tangential velocities is realized via
318     ``ghost points'', allowing alternatively no-slip or free-slip
319     conditions. In ocean models free-slip boundary conditions in
320     conjunction with piecewise-constant (``castellated'') coastlines have
321     been shown to reduce in effect to no-slip boundary conditions
322     \citep{adcroft98:_slippery_coast}; for sea-ice models the effects of
323     lateral boundary conditions have not yet been studied.
324 dimitri 1.1
325     Moving sea ice exerts a stress on the ocean which is the opposite of
326     the stress $\vtau_{ocean}$ in Eq.\refeq{momseaice}. This stess is
327     applied directly to the surface layer of the ocean model. An
328     alternative ocean stress formulation is given by \citet{hibler87}.
329     Rather than applying $\vtau_{ocean}$ directly, the stress is derived
330     from integrating over the ice thickness to the bottom of the oceanic
331     surface layer. In the resulting equation for the \emph{combined}
332     ocean-ice momentum, the interfacial stress cancels and the total
333     stress appears as the sum of windstress and divergence of internal ice
334     stresses: $\delta(z) (\vtau_{air} + \vek{F})/\rho_0$, \citep[see also
335     Eq.\,2 of][]{hibler87}. The disadvantage of this formulation is that
336     now the velocity in the surface layer of the ocean that is used to
337     advect tracers, is really an average over the ocean surface
338     velocity and the ice velocity leading to an inconsistency as the ice
339     temperature and salinity are different from the oceanic variables.
340    
341     Sea ice distributions are characterized by sharp gradients and edges.
342 mlosch 1.15 For this reason, we employ positive, multidimensional 2nd and 3rd-order
343     advection scheme with flux limiter \citep{roe85, hundsdorfer94} for the
344     thermodynamic variables discussed in the next section.
345 dimitri 1.1
346     \subparagraph{boundary conditions: no-slip, free-slip, half-slip}
347    
348     \begin{itemize}
349     \item transition from existing B-Grid to C-Grid
350     \item boundary conditions: no-slip, free-slip, half-slip
351     \item fancy (multi dimensional) advection schemes
352     \item VP vs.\ EVP \citep{hunke97}
353     \item ocean stress formulation \citep{hibler87}
354     \end{itemize}
355    
356     \subsection{Thermodynamics}
357     \label{sec:thermodynamics}
358    
359     In the original formulation the sea ice model \citep{menemenlis05}
360     uses simple thermodynamics following the appendix of
361     \citet{semtner76}. This formulation does not allow storage of heat
362     (heat capacity of ice is zero, and this type of model is often refered
363     to as a ``zero-layer'' model). Upward heat flux is parameterized
364     assuming a linear temperature profile and together with a constant ice
365     conductivity. It is expressed as $(K/h)(T_{w}-T_{0})$, where $K$ is
366     the ice conductivity, $h$ the ice thickness, and $T_{w}-T_{0}$ the
367     difference between water and ice surface temperatures. The surface
368     heat budget is computed in a similar way to that of
369     \citet{parkinson79} and \citet{manabe79}.
370    
371     There is considerable doubt about the reliability of such a simple
372     thermodynamic model---\citet{semtner84} found significant errors in
373     phase (one month lead) and amplitude ($\approx$50\%\,overestimate) in
374     such models---, so that today many sea ice models employ more complex
375     thermodynamics. A popular thermodynamics model of \citet{winton00} is
376     based on the 3-layer model of \citet{semtner76} and treats brine
377     content by means of enthalphy conservation. This model requires in
378     addition to ice-thickness and compactness (fractional area) additional
379     state variables to be advected by ice velocities, namely enthalphy of
380     the two ice layers and the thickness of the overlying snow layer.
381 mlosch 1.15 \ml{[Jean-Michel, your turf: ]Care must be taken in advecting these
382     quantities in order to ensure conservation of enthalphy. Currently
383     this can only be accomplished with a 2nd-order advection scheme with
384     flux limiter \citep{roe85}.}
385 dimitri 1.1
386 mlosch 1.9
387 dimitri 1.1 \subsection{C-grid}
388     \begin{itemize}
389     \item no-slip vs. free-slip for both lsr and evp;
390     "diagnostics" to look at and use for comparison
391     \begin{itemize}
392     \item ice transport through Fram Strait/Denmark Strait/Davis
393     Strait/Bering strait (these are general)
394     \item ice transport through narrow passages, e.g.\ Nares-Strait
395     \end{itemize}
396     \item compare different advection schemes (if lsr turns out to be more
397     effective, then with lsr otherwise I prefer evp), eg.
398     \begin{itemize}
399     \item default 2nd-order with diff1=0.002
400     \item 1st-order upwind with diff1=0.
401     \item DST3FL (SEAICEadvScheme=33 with diff1=0., should work, works for me)
402     \item 2nd-order wit flux limiter (SEAICEadvScheme=77 with diff1=0.)
403     \end{itemize}
404     That should be enough. Here, total ice mass and location of ice edge
405     is interesting. However, this comparison can be done in an idealized
406     domain, may not require full Arctic Domain?
407     \item
408     Do a little study on the parameters of LSR and EVP
409     \begin{enumerate}
410     \item convergence of LSR, how many iterations do you need to get a
411     true elliptic yield curve
412     \item vary deltaTevp and the relaxation parameter for EVP and see when
413     the EVP solution breaks down (relative to the forcing time scale).
414     For this, it is essential that the evp solver gives use "stripeless"
415     solutions, that is your dtevp = 1sec solutions/or 10sec solutions
416     with SEAICE\_evpDampC = 615.
417     \end{enumerate}
418     \end{itemize}
419    
420     \section{Forward sensitivity experiments}
421     \label{sec:forward}
422    
423     A second series of forward sensitivity experiments have been carried out on an
424     Arctic Ocean domain with open boundaries. Once again the objective is to
425     compare the old B-grid LSR dynamic solver with the new C-grid LSR and EVP
426     solvers. One additional experiment is carried out to illustrate the
427     differences between the two main options for sea ice thermodynamics in the MITgcm.
428    
429     \subsection{Arctic Domain with Open Boundaries}
430     \label{sec:arctic}
431    
432 dimitri 1.12 The Arctic domain of integration is illustrated in Fig.~\ref{fig:arctic1}. It
433 mlosch 1.6 is carved out from, and obtains open boundary conditions from, the
434     global cubed-sphere configuration of the Estimating the Circulation
435     and Climate of the Ocean, Phase II (ECCO2) project
436     \citet{menemenlis05}. The domain size is 420 by 384 grid boxes
437     horizontally with mean horizontal grid spacing of 18 km.
438 dimitri 1.1
439 dimitri 1.12 \begin{figure}
440     %\centerline{{\includegraphics*[width=0.44\linewidth]{\fpath/arctic1.eps}}}
441     \caption{Bathymetry of Arctic Domain.\label{fig:arctic1}}
442     \end{figure}
443    
444 dimitri 1.1 There are 50 vertical levels ranging in thickness from 10 m near the surface
445     to approximately 450 m at a maximum model depth of 6150 m. Bathymetry is from
446     the National Geophysical Data Center (NGDC) 2-minute gridded global relief
447     data (ETOPO2) and the model employs the partial-cell formulation of
448 mlosch 1.6 \citet{adcroft97:_shaved_cells}, which permits accurate representation of the bathymetry. The
449 dimitri 1.1 model is integrated in a volume-conserving configuration using a finite volume
450     discretization with C-grid staggering of the prognostic variables. In the
451 mlosch 1.6 ocean, the non-linear equation of state of \citet{jackett95}. The ocean model is
452 dimitri 1.1 coupled to a sea-ice model described hereinabove.
453    
454 mlosch 1.6 This particular ECCO2 simulation is initialized from rest using the
455     January temperature and salinity distribution from the World Ocean
456     Atlas 2001 (WOA01) [Conkright et al., 2002] and it is integrated for
457     32 years prior to the 1996--2001 period discussed in the study. Surface
458     boundary conditions are from the National Centers for Environmental
459     Prediction and the National Center for Atmospheric Research
460     (NCEP/NCAR) atmospheric reanalysis [Kistler et al., 2001]. Six-hourly
461     surface winds, temperature, humidity, downward short- and long-wave
462     radiations, and precipitation are converted to heat, freshwater, and
463     wind stress fluxes using the \citet{large81, large82} bulk formulae.
464     Shortwave radiation decays exponentially as per Paulson and Simpson
465     [1977]. Additionally the time-mean river run-off from Large and Nurser
466     [2001] is applied and there is a relaxation to the monthly-mean
467     climatological sea surface salinity values from WOA01 with a
468     relaxation time scale of 3 months. Vertical mixing follows
469     \citet{large94} with background vertical diffusivity of
470     $1.5\times10^{-5}\text{\,m$^{2}$\,s$^{-1}$}$ and viscosity of
471     $10^{-3}\text{\,m$^{2}$\,s$^{-1}$}$. A third order, direct-space-time
472     advection scheme with flux limiter is employed \citep{hundsdorfer94}
473     and there is no explicit horizontal diffusivity. Horizontal viscosity
474     follows \citet{lei96} but
475     modified to sense the divergent flow as per Fox-Kemper and Menemenlis
476     [in press]. Shortwave radiation decays exponentially as per Paulson
477     and Simpson [1977]. Additionally, the time-mean runoff of Large and
478     Nurser [2001] is applied near the coastline and, where there is open
479     water, there is a relaxation to monthly-mean WOA01 sea surface
480     salinity with a time constant of 45 days.
481 dimitri 1.1
482     Open water, dry
483     ice, wet ice, dry snow, and wet snow albedo are, respectively, 0.15, 0.85,
484     0.76, 0.94, and 0.8.
485    
486     \begin{itemize}
487     \item Configuration
488     \item OBCS from cube
489     \item forcing
490     \item 1/2 and full resolution
491     \item with a few JFM figs from C-grid LSR no slip
492     ice transport through Canadian Archipelago
493     thickness distribution
494     ice velocity and transport
495     \end{itemize}
496    
497     \begin{itemize}
498     \item Arctic configuration
499     \item ice transport through straits and near boundaries
500     \item focus on narrow straits in the Canadian Archipelago
501     \end{itemize}
502    
503     \begin{itemize}
504     \item B-grid LSR no-slip
505     \item C-grid LSR no-slip
506     \item C-grid LSR slip
507     \item C-grid EVP no-slip
508     \item C-grid EVP slip
509 mlosch 1.6 \item C-grid LSR + TEM (truncated ellipse method, no tensile stress, new flag)
510 dimitri 1.1 \item C-grid LSR no-slip + Winton
511     \item speed-performance-accuracy (small)
512     ice transport through Canadian Archipelago differences
513     thickness distribution differences
514     ice velocity and transport differences
515     \end{itemize}
516    
517     We anticipate small differences between the different models due to:
518     \begin{itemize}
519     \item advection schemes: along the ice-edge and regions with large
520     gradients
521 mlosch 1.6 \item C-grid: less transport through narrow straits for no slip
522     conditons, more for free slip
523 dimitri 1.1 \item VP vs.\ EVP: speed performance, accuracy?
524     \item ocean stress: different water mass properties beneath the ice
525     \end{itemize}
526    
527     \section{Adjoint sensiivities of the MITsim}
528    
529     \subsection{The adjoint of MITsim}
530    
531     The ability to generate tangent linear and adjoint model components
532     of the MITsim has been a main design task.
533     For the ocean the adjoint capability has proven to be an
534     invaluable tool for sensitivity analysis as well as state estimation.
535     In short, the adjoint enables very efficient computation of the gradient
536     of scalar-valued model diagnostics (called cost function or objective function)
537     with respect to many model "variables".
538     These variables can be two- or three-dimensional fields of initial
539     conditions, model parameters such as mixing coefficients, or
540     time-varying surface or lateral (open) boundary conditions.
541     When combined, these variables span a potentially high-dimensional
542     (e.g. O(10$^8$)) so-called control space. Performing parameter perturbations
543     to assess model sensitivities quickly becomes prohibitive at these scales.
544     Alternatively, (time-varying) sensitivities of the objective function
545     to any element of the control space can be computed very efficiently in
546     one single adjoint
547     model integration, provided an efficient adjoint model is available.
548    
549     [REFERENCES]
550    
551    
552     The adjoint operator (ADM) is the transpose of the tangent linear operator (TLM)
553     of the full (in general nonlinear) forward model, i.e. the MITsim.
554     The TLM maps perturbations of elements of the control space
555     (e.g. initial ice thickness distribution)
556     via the model Jacobian
557     to a perturbation in the objective function
558     (e.g. sea-ice export at the end of the integration interval).
559     \textit{Tangent} linearity ensures that the derivatives are evaluated
560     with respect to the underlying model trajectory at each point in time.
561     This is crucial for nonlinear trajectories and the presence of different
562     regimes (e.g. effect of the seaice growth term at or away from the
563     freezing point of the ocean surface).
564     Ensuring tangent linearity can be easily achieved by integrating
565     the full model in sync with the TLM to provide the underlying model state.
566     Ensuring \textit{tangent} adjoints is equally crucial, but much more
567     difficult to achieve because of the reverse nature of the integration:
568     the adjoint accumulates sensitivities backward in time,
569     starting from a unit perturbation of the objective function.
570     The adjoint model requires the model state in reverse order.
571     This presents one of the major complications in deriving an
572     exact, i.e. \textit{tangent} adjoint model.
573    
574     Following closely the development and maintenance of TLM and ADM
575     components of the MITgcm we have relied heavily on the
576     autmomatic differentiation (AD) tool
577     "Transformation of Algorithms in Fortran" (TAF)
578     developed by Fastopt (Giering and Kaminski, 1998)
579     to derive TLM and ADM code of the MITsim.
580     Briefly, the nonlinear parent model is fed to the AD tool which produces
581     derivative code for the specified control space and objective function.
582     Following this approach has (apart from its evident success)
583     several advantages:
584     (1) the adjoint model is the exact adjoint operator of the parent model,
585     (2) the adjoint model can be kept up to date with respect to ongoing
586     development of the parent model, and adjustments to the parent model
587     to extend the automatically generated adjoint are incremental changes
588     only, rather than extensive re-developments,
589     (3) the parallel structure of the parent model is preserved
590     by the adjoint model, ensuring efficient use in high performance
591     computing environments.
592    
593     Some initial code adjustments are required to support dependency analysis
594     of the flow reversal and certain language limitations which may lead
595     to irreducible flow graphs (e.g. GOTO statements).
596     The problem of providing the required model state in reverse order
597     at the time of evaluating nonlinear or conditional
598     derivatives is solved via balancing
599     storing vs. recomputation of the model state in a multi-level
600     checkpointing loop.
601     Again, an initial code adjustment is required to support TAFs
602     checkpointing capability.
603 mlosch 1.6 The code adjustments are sufficiently simple so as not to cause
604 dimitri 1.1 major limitations to the full nonlinear parent model.
605     Once in place, an adjoint model of a new model configuration
606     may be derived in about 10 minutes.
607    
608     [HIGHLIGHT COUPLED NATURE OF THE ADJOINT!]
609    
610     \subsection{Special considerations}
611    
612     * growth term(?)
613    
614     * small active denominators
615    
616     * dynamic solver (implicit function theorem)
617    
618     * approximate adjoints
619    
620    
621     \subsection{An example: sensitivities of sea-ice export through Fram Strait}
622    
623     We demonstrate the power of the adjoint method
624     in the context of investigating sea-ice export sensitivities through Fram Strait
625     (for details of this study see Heimbach et al., 2007).
626 mlosch 1.6 %\citep[for details of this study see][]{heimbach07}. %Heimbach et al., 2007).
627 dimitri 1.1 The domain chosen is a coarsened version of the Arctic face of the
628     high-resolution cubed-sphere configuration of the ECCO2 project
629 mlosch 1.6 \citep[see][]{menemenlis05}. It covers the entire Arctic,
630 dimitri 1.1 extends into the North Pacific such as to cover the entire
631     ice-covered regions, and comprises parts of the North Atlantic
632     down to XXN to enable analysis of remote influences of the
633     North Atlantic current to sea-ice variability and export.
634     The horizontal resolution varies between XX and YY km
635     with 50 unevenly spaced vertical levels.
636     The adjoint models run efficiently on 80 processors
637     (benchmarks have been performed both on an SGI Altix as well as an
638     IBM SP5 at NASA/ARC).
639    
640 mlosch 1.6 Following a 1-year spinup, the model has been integrated for four
641     years between 1992 and 1995. It is forced using realistic 6-hourly
642     NCEP/NCAR atmospheric state variables. Over the open ocean these are
643     converted into air-sea fluxes via the bulk formulae of
644     \citet{large04}. Derivation of air-sea fluxes in the presence of
645     sea-ice is handled by the ice model as described in \refsec{model}.
646 dimitri 1.1 The objective function chosen is sea-ice export through Fram Strait
647 mlosch 1.6 computed for December 1995. The adjoint model computes sensitivities
648     to sea-ice export back in time from 1995 to 1992 along this
649     trajectory. In principle all adjoint model variable (i.e., Lagrange
650     multipliers) of the coupled ocean/sea-ice model are available to
651     analyze the transient sensitivity behaviour of the ocean and sea-ice
652     state. Over the open ocean, the adjoint of the bulk formula scheme
653     computes sensitivities to the time-varying atmospheric state. Over
654     ice-covered parts, the sea-ice adjoint converts surface ocean
655     sensitivities to atmospheric sensitivities.
656    
657     \reffig{4yradjheff}(a--d) depict sensitivities of sea-ice export
658     through Fram Strait in December 1995 to changes in sea-ice thickness
659     12, 24, 36, 48 months back in time. Corresponding sensitivities to
660     ocean surface temperature are depicted in
661     \reffig{4yradjthetalev1}(a--d). The main characteristics is
662     consistency with expected advection of sea-ice over the relevant time
663     scales considered. The general positive pattern means that an
664     increase in sea-ice thickness at location $(x,y)$ and time $t$ will
665     increase sea-ice export through Fram Strait at time $T_e$. Largest
666     distances from Fram Strait indicate fastest sea-ice advection over the
667     time span considered. The ice thickness sensitivities are in close
668     correspondence to ocean surface sentivitites, but of opposite sign.
669     An increase in temperature will incur ice melting, decrease in ice
670     thickness, and therefore decrease in sea-ice export at time $T_e$.
671 dimitri 1.1
672     The picture is fundamentally different and much more complex
673     for sensitivities to ocean temperatures away from the surface.
674 mlosch 1.6 \reffig{4yradjthetalev10??}(a--d) depicts ice export sensitivities to
675 dimitri 1.1 temperatures at roughly 400 m depth.
676     Primary features are the effect of the heat transport of the North
677     Atlantic current which feeds into the West Spitsbergen current,
678     the circulation around Svalbard, and ...
679    
680     \begin{figure}[t!]
681     \centerline{
682     \subfigure[{\footnotesize -12 months}]
683 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim072_cmax2.0E+02.eps}}
684 dimitri 1.1 %\includegraphics*[width=.3\textwidth]{H_c.bin_res_100_lev1.pdf}
685     %
686     \subfigure[{\footnotesize -24 months}]
687 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim145_cmax2.0E+02.eps}}
688 dimitri 1.1 }
689    
690     \centerline{
691     \subfigure[{\footnotesize
692     -36 months}]
693 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim218_cmax2.0E+02.eps}}
694 dimitri 1.1 %
695     \subfigure[{\footnotesize
696     -48 months}]
697 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJheff_arc_lev1_tim292_cmax2.0E+02.eps}}
698 dimitri 1.1 }
699     \caption{Sensitivity of sea-ice export through Fram Strait in December 2005 to
700     sea-ice thickness at various prior times.
701     \label{fig:4yradjheff}}
702     \end{figure}
703    
704    
705     \begin{figure}[t!]
706     \centerline{
707     \subfigure[{\footnotesize -12 months}]
708 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim072_cmax5.0E+01.eps}}
709 dimitri 1.1 %\includegraphics*[width=.3\textwidth]{H_c.bin_res_100_lev1.pdf}
710     %
711     \subfigure[{\footnotesize -24 months}]
712 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim145_cmax5.0E+01.eps}}
713 dimitri 1.1 }
714    
715     \centerline{
716     \subfigure[{\footnotesize
717     -36 months}]
718 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim218_cmax5.0E+01.eps}}
719 dimitri 1.1 %
720     \subfigure[{\footnotesize
721     -48 months}]
722 mlosch 1.4 {\includegraphics*[width=0.44\linewidth]{\fpath/run_4yr_ADJtheta_arc_lev1_tim292_cmax5.0E+01.eps}}
723 dimitri 1.1 }
724 mlosch 1.6 \caption{Same as \reffig{4yradjheff} but for sea surface temperature
725 dimitri 1.1 \label{fig:4yradjthetalev1}}
726     \end{figure}
727    
728    
729    
730     \section{Discussion and conclusion}
731     \label{sec:concl}
732    
733     The story of the paper could be:
734     B-grid ice model + C-grid ocean model does not work properly for us,
735     therefore C-grid ice model with advantages:
736     \begin{enumerate}
737     \item stress coupling simpler (no interpolation required)
738     \item different boundary conditions
739     \item advection schemes carry over trivially from main code
740     \end{enumerate}
741     LSR/EVP solutions are similar with appropriate bcs, evp parameters as
742     a function of forcing time scale (when does VP solution break
743     down). Same for LSR solver, provided that it works (o:
744     Which scheme is more efficient for the resolution/time stepping
745     parameters that we use here. What about other resolutions?
746    
747     \paragraph{Acknowledgements}
748     We thank Jinlun Zhang for providing the original B-grid code and many
749 mlosch 1.6 helpful discussions. ML thanks Elizabeth Hunke for multiple explanations.
750 dimitri 1.1
751 dimitri 1.7 \bibliography{bib/journal_abrvs,bib/seaice,bib/genocean,bib/maths,bib/mitgcmuv,bib/fram}
752 dimitri 1.1
753     \end{document}
754    
755     %%% Local Variables:
756     %%% mode: latex
757     %%% TeX-master: t
758     %%% End:
759 mlosch 1.4
760    
761     A Dynamic-Thermodynamic Sea ice Model for Ocean Climate
762     Estimation on an Arakawa C-Grid
763    
764     Introduction
765    
766     Ice Model:
767     Dynamics formulation.
768     B-C, LSR, EVP, no-slip, slip
769     parallellization
770     Thermodynamics formulation.
771     0-layer Hibler salinity + snow
772     3-layer Winton
773    
774     Idealized tests
775     Funnel Experiments
776     Downstream Island tests
777     B-grid LSR no-slip
778     C-grid LSR no-slip
779     C-grid LSR slip
780     C-grid EVP no-slip
781     C-grid EVP slip
782    
783     Arctic Setup
784     Configuration
785     OBCS from cube
786     forcing
787     1/2 and full resolution
788     with a few JFM figs from C-grid LSR no slip
789     ice transport through Canadian Archipelago
790     thickness distribution
791     ice velocity and transport
792    
793     Arctic forward sensitivity experiments
794     B-grid LSR no-slip
795     C-grid LSR no-slip
796     C-grid LSR slip
797     C-grid EVP no-slip
798     C-grid EVP slip
799     C-grid LSR no-slip + Winton
800     speed-performance-accuracy (small)
801     ice transport through Canadian Archipelago differences
802     thickness distribution differences
803     ice velocity and transport differences
804    
805     Adjoint sensitivity experiment on 1/2-res setup
806     Sensitivity of sea ice volume flow through Fram Strait
807     *** Sensitivity of sea ice volume flow through Canadian Archipelago
808    
809     Summary and conluding remarks

  ViewVC Help
Powered by ViewVC 1.1.22