State Estimation of the Northern Philippine Sea Circulation
State Estimation of the Northern Philippine Sea Circulation
Abstract:
An ocean state estimate has been developed for the northern Philippine
Sea circulation for April 2010 through April 2011 as part of the North
Pacific Acoustic Laboratory (NPAL) Philippine Sea experiment. These
examples used the MIT general circulation model (MITgcm) and its
adjoint (the Four-Dimensional Variational: 4DVAR method) to match the
model fields to satellite-derived along-track sea surface height
(SSH), gridded sea surface temperature (SST) observations, and Argo
temperature (T) and salinity (S) observations by adjusting model T and
S initial conditions, open boundary conditions, and atmospheric
forcing fields over assimilation windows of one or two months. The
state estimate was evaluated by comparing model hindcasts with the
SSH, SST and Argo observations that were assimilated, since the strong
constraint form is not guaranteed to fit the observations, and
cross-validated with independent T observations spanning from 200 m to
5500 m from a Distributed Vertical Line Array (DVLA) of 149 hydrophones
which included thermistors. The optimized states at the end of the
assimilation periods were also cross-validated by one month forecasts
forced by climatological open boundary conditions, atmospheric
forcing, and run-off fluxes. These forecasts explore the
predictability of the ocean circulation from optimized initial conditions and
cross-validate the state estimate by comparing it to independent
future observations. The model analysis outperformed persistence
(keeping the initial state fixed) and reference (initialized using
assimilated HYCOM/NCODA 1/12 degree global analysis) on average. The
model eddies were used to compute eddy fluxes of heat and salt which
were compared with those from Roemmich and Gilson, 2001.
Sea circulation for April 2010 through April 2011 as part of the North
Pacific Acoustic Laboratory (NPAL) Philippine Sea experiment. These
examples used the MIT general circulation model (MITgcm) and its
adjoint (the Four-Dimensional Variational: 4DVAR method) to match the
model fields to satellite-derived along-track sea surface height
(SSH), gridded sea surface temperature (SST) observations, and Argo
temperature (T) and salinity (S) observations by adjusting model T and
S initial conditions, open boundary conditions, and atmospheric
forcing fields over assimilation windows of one or two months. The
state estimate was evaluated by comparing model hindcasts with the
SSH, SST and Argo observations that were assimilated, since the strong
constraint form is not guaranteed to fit the observations, and
cross-validated with independent T observations spanning from 200 m to
5500 m from a Distributed Vertical Line Array (DVLA) of 149 hydrophones
which included thermistors. The optimized states at the end of the
assimilation periods were also cross-validated by one month forecasts
forced by climatological open boundary conditions, atmospheric
forcing, and run-off fluxes. These forecasts explore the
predictability of the ocean circulation from optimized initial conditions and
cross-validate the state estimate by comparing it to independent
future observations. The model analysis outperformed persistence
(keeping the initial state fixed) and reference (initialized using
assimilated HYCOM/NCODA 1/12 degree global analysis) on average. The
model eddies were used to compute eddy fluxes of heat and salt which
were compared with those from Roemmich and Gilson, 2001.