5:15 PM - 6:45 PM
[MIS10-P13] Understanding the Southern Ocean through model-data synthesis
Keywords:Antarctica, Southern Ocean, Model-data synthesis
How fast Antarctica will lose ice and contribute to sea-level rise has become one of today’s most urgent scientific questions. Antarctic ice-sheet mass loss is primarily triggered by the ocean, but relatively little is understood about the drivers of ocean warming where it matters – at the ice sheet margin – nor about the impact of ice-shelf melting on the ocean. We (a) conduct oceanographic observations, (b) improve ocean simulations achieving high model-data agreement, and (c) develop ice-ocean coupled simulations to better understand the past, present, and future state of the Southern Ocean and identify the causes of ocean warming and Antarctic ice loss on various time scales. We present the recent progress of our two approaches to understanding such ocean processes.
First, to study circum-Antarctic-scale ocean circulation, we evaluate existing ocean reanalysis based on the Massachusetts Institute of Technology general circulation model (MITgcm). Specifically, we investigate large-scale ocean processes modulating cross-shelf exchange, and thus possibly impacting ice shelf melting, in ECCOv4r5, ECCO LLC270, SOSE, and GECCO3. The MITgcm-based ocean reanalyses show Antarctic Circumpolar Currents (ACC) measuring approximately 149±11 Sv. The simulated 2ºC isotherms, which are located in positions similar to the ACC, roughly represent the southern extent of the current. Simulated Weddell and Ross Gyre strengths are 51±11 Sv and 25±8 Sv, respectively, consistent with observation-based estimates. However, all of these large-scale ocean state estimates fail to reproduce water mass properties and variability over the continental shelves around the Antarctic continent.
Second, we develop regional ocean simulations by downscaling these ocean analyses. We further apply optimization techniques (Green’s functions and the adjoint method) to improve the model-data agreement. Our recent work, for example, employs the adjoint-model estimation method, for the first time with explicit representation of sub-ice shelf cavities, to develop an ocean state estimate for the Amundsen and Bellingshausen Seas. The finer grid spacing of the regional simulations and the joint estimation of ocean circulation and ice shelf melt allow us to achieve much closer agreement with observations on the Antarctic continental shelf than is possible in the large-scale state estimates. In this presentation, we will summarize our ongoing development of downscaled, Antarctic regional simulations (e.g., Amundsen, Bellingshausen, Weddell, and Ross Seas; and along the East Antarctic coast).
First, to study circum-Antarctic-scale ocean circulation, we evaluate existing ocean reanalysis based on the Massachusetts Institute of Technology general circulation model (MITgcm). Specifically, we investigate large-scale ocean processes modulating cross-shelf exchange, and thus possibly impacting ice shelf melting, in ECCOv4r5, ECCO LLC270, SOSE, and GECCO3. The MITgcm-based ocean reanalyses show Antarctic Circumpolar Currents (ACC) measuring approximately 149±11 Sv. The simulated 2ºC isotherms, which are located in positions similar to the ACC, roughly represent the southern extent of the current. Simulated Weddell and Ross Gyre strengths are 51±11 Sv and 25±8 Sv, respectively, consistent with observation-based estimates. However, all of these large-scale ocean state estimates fail to reproduce water mass properties and variability over the continental shelves around the Antarctic continent.
Second, we develop regional ocean simulations by downscaling these ocean analyses. We further apply optimization techniques (Green’s functions and the adjoint method) to improve the model-data agreement. Our recent work, for example, employs the adjoint-model estimation method, for the first time with explicit representation of sub-ice shelf cavities, to develop an ocean state estimate for the Amundsen and Bellingshausen Seas. The finer grid spacing of the regional simulations and the joint estimation of ocean circulation and ice shelf melt allow us to achieve much closer agreement with observations on the Antarctic continental shelf than is possible in the large-scale state estimates. In this presentation, we will summarize our ongoing development of downscaled, Antarctic regional simulations (e.g., Amundsen, Bellingshausen, Weddell, and Ross Seas; and along the East Antarctic coast).