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

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

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


Revision 1.2 - (hide annotations) (download) (as text)
Sat Mar 1 01:01:52 2008 UTC (17 years, 4 months ago) by dimitri
Branch: MAIN
Changes since 1.1: +15 -0 lines
File MIME type: application/x-tex
added some words to introduction

1 dimitri 1.1 \section{Introduction}
2     \label{sec:intro}
3    
4 dimitri 1.2 In the past five years, oceanographic state estimation has matured to the
5     extent that estimates of the evolving circulation of the ocean constrained by
6     in-situ and remotely sensed global observations are now routinely available
7     and being applied to myriad scientific problems \citep{wun07}. Ocean state
8     estimation is the process of fitting an ocean general circulation model (GCM)
9     to a multitude of observations. As formulated by the consortium Estimating
10     the Circulation and Climate of the Ocean (ECCO), an automatic differentiation
11     tool is used to calculate the so-called adjoint code of a GCM. The method of
12     Lagrange multipliers is then used to render the problem one of unconstrained
13     least-squares minimization. Although much has been achieved, the existing
14     ECCO estimates lack intercative sea ice. This limits the ability of ECCO to
15     utilize satellite data constraints over sea-ice covered regions. This also
16     limits the usefulness of the ECCO ocean state estimates for describing and
17     studying polar-subpolar interactions.
18    
19 dimitri 1.1 The availability of an adjoint model as a powerful research tool
20     complementary to an ocean model was a major design requirement early
21     on in the development of the MIT general circulation model (MITgcm)
22     [Marshall et al. 1997a, Marotzke et al. 1999, Adcroft et al. 2002]. It
23     was recognized that the adjoint model permitted computing the
24     gradients of various scalar-valued model diagnostics, norms or,
25     generally, objective functions with respect to external or independent
26     parameters very efficiently. The information associtated with these
27     gradients is useful in at least two major contexts. First, for state
28     estimation problems, the objective function is the sum of squared
29     differences between observations and model results weighted by the
30     inverse error covariances. The gradient of such an objective function
31     can be used to reduce this measure of model-data misfit to find the
32     optimal model solution in a least-squares sense. Second, the
33     objective function can be a key oceanographic quantity such as
34     meridional heat or volume transport, ocean heat content or mean
35     surface temperature index. In this case the gradient provides a
36     complete set of sensitivities of this quantity to all independent
37     variables simultaneously. These sensitivities can be used to address
38     the cause of, say, changing net transports accurately.
39    
40     References to existing sea-ice adjoint models, explaining that they are either
41     for simplified configurations, for ice-only studies, or for short-duration
42     studies to motivate the present work.
43    
44     Traditionally, probably for historical reasons and the ease of
45     treating the Coriolis term, most standard sea-ice models are
46     discretized on Arakawa-B-grids \citep[e.g.,][]{hibler79, harder99,
47     kreyscher00, zhang98, hunke97}. From the perspective of coupling a
48     sea ice-model to a C-grid ocean model, the exchange of fluxes of heat
49     and fresh-water pose no difficulty for a B-grid sea-ice model
50     \citep[e.g.,][]{timmermann02a}. However, surface stress is defined at
51     velocities points and thus needs to be interpolated between a B-grid
52     sea-ice model and a C-grid ocean model. Smoothing implicitly
53     associated with this interpolation may mask grid scale noise and may
54     contribute to stabilizing the solution. On the other hand, by
55     smoothing the stress signals are damped which could lead to reduced
56     variability of the system. By choosing a C-grid for the sea-ice model,
57     we circumvent this difficulty altogether and render the stress
58     coupling as consistent as the buoyancy coupling.
59    
60     A further advantage of the C-grid formulation is apparent in narrow
61     straits. In the limit of only one grid cell between coasts there is no
62     flux allowed for a B-grid (with no-slip lateral boundary counditions),
63     and models have used topographies artificially widened straits to
64     avoid this problem \citep{holloway07}. The C-grid formulation on the
65     other hand allows a flux of sea-ice through narrow passages if
66     free-slip along the boundaries is allowed. We demonstrate this effect
67     in the Candian archipelago.
68    
69     Talk about problems that make the sea-ice-ocean code very sensitive and
70     changes in the code that reduce these sensitivities.
71    
72     This paper describes the MITgcm sea ice
73     model; it presents example Arctic and Antarctic results from a realistic,
74     eddy-permitting, global ocean and sea-ice configuration; it compares B-grid
75     and C-grid dynamic solvers in a regional Arctic configuration; and it presents
76     example results from coupled ocean and sea-ice adjoint-model integrations.

  ViewVC Help
Powered by ViewVC 1.1.22