Heinrich Events

Session 4: Ice-sheets and Sea-level

Session Description

This session aims to investigate whether there was a significant rise in eustatic sea-level during Heinrich events and whether we see the impact of this on marginal basins, e.g., do we see a change in watermass communication. We will also assess whether the associated sea-level rise/freshwater fluxes had an impact on AMOC stability, as indicated by results from ocean-climate models.

Session Talks

Heinrich Events: a selection of impacts on subtropical marginal seas

Eelco Rohling

Eelco J Rohling

National Oceanography Centre, Southampton

The Mediterranean Sea and Red Sea are both highly evaporative basins, with very small connecting straits to the open ocean. The Mediterranean connection with the North Atlantic through the Strait of Gibraltar is 284 m deep, and the Red Sea's connection to the Indian Ocean at Bab-el-Mandab is only 137 m deep. Water-mass exchange through both straits is predominantly controlled by changes in the buoyancy forcing (freshwater and heat budget) and in the strait dimensions (sea level). I will discuss some of the most important influences on records of change through times of Heinrich Events in both basins, to highlight similarities and differences. I will focus especially on changes in the Mediterranean Outflow plume, which affect its settling and hydrographic importance within the interior of the North Atlantic.

Constraining the size of Heinrich Events using an Iceberg/Sediment model and a 3D ice sheet model

William Roberts

William H.G. Roberts1*, P.J.Valdes1, A.J.Payne2

1BRIDGE, School of Geographical Sciences, University of Bristol, U.K.

2Bristol Glaciology Centre, School of Geographical Sciences, University of Bristol, U.K.

Heinrich Layers, anomalously thick layers of ice borne sediment in the north Atlantic ocean, and the events that caused them have long been associated with abrupt climate changes during glacial times. However, there is still no consensus about either how much ice is needed to transport this sediment or how such a large volume of ice could be produced. Estimates for this ice volume do exist and may be broadly separated into two categories: estimates derived from ice sheet models, and estimates derived from isotope records. There is a wide discrepancy between these two sets of estimates with the isotope derived estimates being at least one, and sometimes two, orders of magnitude larger than those from ice sheet models.

We shall describe here two different methods to further constrain these events. First, we use an iceberg model that includes sediment to simulate the delivery of sediment to the north Atlantic during a Heinrich Event. Second we use a three dimensional ice sheet model (Glimmer) with realistic topography to determine the volume of ice that leaves Hudson Strait during the thermo-mechanical surging events that the model simulates.

We show that the iceberg model can simulate the pattern of ice raft debris from a Heinrich Event and that we can simulate the sediment layer thickness that would result from the volume of ice released by the different estimates. We show that to best fit the observed Heinrich layer sediment thickness, 60x104 km3 of ice needs to be released during the event. This matches the icesheet derived estimates better than the isotope derived estimates and suggests that Heinrich Events released relatively small volumes of ice. The surges from the 3-D ice sheet produce a larger volume of ice for each Heinrich Event than the iceberg model suggest is needed to form the Heinrich Layers, but the volume is consistent with other ice sheet models and significantly smaller than the volume that the isotopes suggest.

Coupled ice sheet - climate modelling of ice sheet collapses

Florian Ziemen

Florian Ziemen, Christian Rodehacke, Uwe Mikolajewicz

Max Planck Institute for Meteorology, Hamburg

One major challenge in predicting future climate change is the validation of the numerical models used for the predictions. By studying past periods of rapid climate change, we can improve our understanding of the climate system, and test our numerical models by comparing the model output with proxy data. A particular good example of these ice sheet - climate interactions are Heinrich events. They caused large climate changes that are well represented in the proxies. We use a coarse resolution complex climate model coupled with an ice sheet model to study the ice sheet collapses and their consequences in the climate system.

Our model comprises of the atmosphere-ocean-vegetation general circulation model ECHAM5/MPIOM/LPJ interactively coupled with the ice sheet model mPISM. mPISM is a modified version of the Parallel Ice Sheet Model from the University of Alaska, Fairbanks. We run ECHAM5 in T31 resolution (3.75°), and mPISM on a 20 km grid covering most of the northern hemisphere. We couple the models directly without any flux correction or anomaly maps. For the surface mass balance, we use a positive degree day scheme with lapse rate correction and height desertification effect.

We force our model with constant last glacial maximum orbital conditions and obtain recurring ice sheet collapses. We show results from asynchronously (1:10) coupled experiments. In these experiments, the collapses cause freshwater pulses reaching peak values close to 0.1 Sv. The typical duration of a collapse is about 2000 years. The total amount of ice released is about 2 Mio km3 (5m of sea level equivalent). In the ocean, a salinity anomaly develops and is propagated southward by the gyre systems. The North Atlantic deep water cell weakens by 2-3 Sv and the Antarctic bottom water cell strengthens. The surface air temperature is locally reduced by up to 3 K and a there is a slight cooling over large parts of the northern hemisphere.

Dr. Jennifer D. Stanford, Geography & Environment, University of Southampton