Fitting dynamics to data

Thacker, W. C.; Long, Robert Bryan

doi:10.1029/jc093ic02p01227

Cited by 263 publications

(143 citation statements)

References 35 publications

(23 reference statements)

Supporting

Mentioning

137

Contrasting

Unclassified

Order By: Relevance

“…Such an approach has the advantage that the prognostic components of the ice model, such as thickness and temperature evolution, are accounted for in the model adjoint, thus enabling assimilation of time-dependent data to produce a dynamically consistent state estimate with associated optimized parameters. In the field of ocean modeling, state estimation efforts based on the adjoint method were first introduced by Thacker and Long (1988) and Tziperman and Thacker (1989). Since then, estimation systems, in which the adjoint was derived by means of automated differentiation of full-fledged ocean general circulation models, have provided solutions that are consistent with observational data, suitable for model initializations and in-depth data analysis, as well as a framework for estimating the information content of new observations (Stammer et al, 2002;Wunsch et al, 2009;.…”

Section: Introductionmentioning

confidence: 99%

Parameter and state estimation with a time-dependent adjoint marine ice sheet model

Goldberg

Heimbach

2013

The Cryosphere

View full text Add to dashboard Cite

Abstract.To date, assimilation of observations into largescale ice models has consisted predominantly of timeindependent inversions of surface velocities for basal traction, bed elevation, or ice stiffness, and has relied primarily on analytically derived adjoints of glaciological stress balance models. To overcome limitations of such "snapshot" inversions -i.e., their inability to assimilate time-dependent data for the purpose of constraining transient flow states, or to produce initial states with minimum artificial drift and suitable for time-dependent simulations -we have developed an adjoint of a time-dependent parallel glaciological flow model. The model implements a hybrid shallow shelfshallow ice stress balance, solves the continuity equation for ice thickness evolution, and can represent the floating, fast-sliding, and frozen bed regimes of a marine ice sheet. The adjoint is generated by a combination of analytic methods and the use of algorithmic differentiation (AD) software. Several experiments are carried out with idealized geometries and synthetic observations, including inversion of timedependent surface elevations for past thicknesses, and simultaneous retrieval of basal traction and topography from surface data. Flexible generation of the adjoint for a range of independent uncertain variables is exemplified through sensitivity calculations of grounded ice volume to changes in basal melting of floating and basal sliding of grounded ice. The results are encouraging and suggest the feasibility, using real observations, of improved ice sheet state estimation and comprehensive transient sensitivity assessments.

show abstract

Section: Introductionmentioning

confidence: 99%

Parameter and state estimation with a time-dependent adjoint marine ice sheet model

Goldberg

Heimbach

2013

The Cryosphere

View full text Add to dashboard Cite

show abstract

“…This provides an ideal application for data assimilation in its guise of fitting models to data. 85 Another purpose of data assimilation is that of improving the predictive power of models. This is a familiar application in the context of meteorological forecasting.…”

Section: Assimilation Of Data Into Modelsmentioning

confidence: 99%

Biological Oceanography by Remote Sensing

Srokosz

2000

Encyclopedia of Analytical Chemistry

View full text Add to dashboard Cite

Biological oceanography may be studied from space using sensors on satellites that determine the color of the ocean. The presence of phytoplankton (microscopic algae) in the upper layers of the ocean changes the color of the water as seen from above. In simplified terms, this is due to the selective absorption of blue light by the phytoplankton pigments (primarily chlorophyll) which changes the appearance of the water from blue to green. These changes in color can be observed using a satellite‐borne spectroradiometer that measures the water‐leaving radiance in a number of bands in the visible part of the electromagnetic spectrum. The limitations of the technique are first, that only information on the phytoplankton in the upper layers of the ocean can be obtained (light does not penetrate very far into the ocean). Second, most of the signal measured by the satellite sensor originates in the atmosphere (due to the molecular and aerosol scattering of photons there), so careful correction for atmospheric effects is necessary if good ocean data are to be obtained. Of course, in the presence of clouds the sensor will not “see” the ocean surface at all, and no data will be obtained. Third, only one component of the ocean ecosystem, namely the phytoplankton, can be studied by this means. Despite these limitations, satellite observations of ocean color have given new insights into biological oceanography on a global scale that could not have been obtained by any other means of observation. Observations of ocean color have contributed to a better understanding of the biophysical interactions that determine the phytoplankton productivity, the seasonal and interannual variations of the phytoplankton biomass on global scales, and the role of phytoplankton in the climate system. They have also contributed to the improved modeling of biogeochemical processes in the ocean.

show abstract

“…In general, the required number of iterations is not known a priori, but experience shows that in most problems the number of iterations is roughly proportional to the number of control variables [Thacker and Long, 1988]. So we did some trial runs, with convergence being monitored by following the decrease of cost function from iteration to iteration.…”

Section: Process Of Convergencementioning

confidence: 99%

Inferring coastal upwelling dynamics from a thermistor array

Kelley¹,

Bourque²

1997

J. Geophys. Res.

View full text Add to dashboard Cite

Abstract. Key parameters of coastal upwelling dynamics in the St. Lawrence estuary are inferred from moored thermistor observations, using a data assimilative technique based on a dynamical model that has terms representing upwelling, vertical mixing, and residual heating/advection. The inferred time series of upwelling velocity, w(t), agrees generally with a time series calculated from a more conventional regression calculation based on the upwelling term alone. In each case, w(t) is highly coherent with alongshore wind stress, indicating that coastal upwelling dynamics are predominant in this region. We found data assimilation to be superior to the simpler regression technique in two ways. First, the inferred upwelling time series is more accurate, as gauged by coherence with wind stress. Second, data assimilation allows the additional inference of other dynamical parameters, such as the vertical diffusivity. The thermistor array thus indicates that the dynamics of this region are largely related to Ekman upwelling, with vertical mixing also being important.

show abstract

Fitting dynamics to data

Cited by 263 publications

References 35 publications

Parameter and state estimation with a time-dependent adjoint marine ice sheet model

Parameter and state estimation with a time-dependent adjoint marine ice sheet model

Biological Oceanography by Remote Sensing

Inferring coastal upwelling dynamics from a thermistor array

Contact Info

Product

Resources

About