HR: 0800h
AN: C11B-0436    [Abstracts]
TI: Empirical Sea Ice Thickness Estimation in the Arctic Ocean
AU: * Platonov, N G
EM: belchans@eimb.ru
AF: Russian Academy of Sciences, Institute of Ecology and Evolution Leninsky Prospect 33, Moscow, 119071, Russian Federation
AU: Douglas, D C
EM: ddouglas@usgs.gov
AF: USGS Alaska Science Center, 3100 National Park Rd, Juneau, AK 99801, United States
AU: Eremeev, V A
EM: belchans@eimb.ru
AF: Russian Academy of Sciences, Institute of Ecology and Evolution Leninsky Prospect 33, Moscow, 119071, Russian Federation
AU: Mordvintsev, I N
EM: belchans@eimb.ru
AF: Russian Academy of Sciences, Institute of Ecology and Evolution Leninsky Prospect 33, Moscow, 119071, Russian Federation
AB: This study evaluates methods to improve a recently developed neural network (NN) algorithm that estimates sea ice thickness with spatial resolution nearby 100 km at monthly intervals during 1982 - 2003 (Belchansky et al., accepted, J. Climate). For any grid cell, at each position along its drift trajectory, sea ice thickness changes are controlled by geophysical inputs that include dynamic and thermodynamic forcing parameters such as short- and long-wave radiation, cumulative freeze-degree days, ice drift velocity, and an ice-drift derived divergence/convergence index. Improvements to the original method included: 1) expanding the learning data with updated submarine draft data from NSIDC; 2) partitioning all learning data into non-overlapping categories of ice thickness; 3) learning the NN independently for each ice thickness category, and then combining fractions of ice categories to derive a sea ice thickness distribution for each grid cell; 4) replacing and expanding the original NCEP-NCAR Reanalysis radiation inputs with their analogs from the NCEP-DOE Reanalysis-2 data sets; 5) reconstructing the ice divergence-convergence index; and 6) separating the learning data into level ice and ridged ice categories. The contributions of dynamical and thermodynamical components to sea ice volume change in the central Arctic were examined. The influence of ice thickness to the sea ice volume balance is predominant for high latitudes, while for low latitudes, ice volume is related to ice extent.
DE: 0758 Remote sensing
DE: 0798 Modeling
DE: 4540 Ice mechanics and air/sea/ice exchange processes (0700, 0750, 0752, 0754)
SC: Cryosphere [C]
MN: 2007 Fall Meeting