On the Variability and Trend of Weddell Sea Polynyas in High-Resolution CESM-iHESP simulations

Since the beginning of satellite passive-microwave remote sensing of sea-ice concentration, Weddell Sea Polynyas (WSPs) occurred only in years 1974-1976. From 1977 to 2020, such large polynyas have not been observed; the fresher and more stratified Weddell Sea may be a reason for suppressing the return of WSPs. More recently, the largest and most prolonged Maud Rise Polynyas (MRPs) recurred in the austral winter of 2016-2017. Distinct and multi-year polynyas may thus reemerge in nature with a periodicity about 40-50 years. While most Earth System Models (ESMs) produce quasi permanent WSPs, the Community ESM (CESM) produces WSPs intermittently only when running at high-resolution (atmosphere and land 0.25°, ocean and sea ice 0.1°). The international Laboratory for High-resolution Earth System Prediction (iHESP) partnership recently produced an unprecedented set of high-resolution CESM simulations. Here we analyze the occurrence of MRPs and WSPs in a 500-year preindustrial control (PI-CTRL) simulation and in a 250-year historical and future transient (HF-TNST) simulation from 1850 to 2100.

The PI-CTRL simulation produces open ocean polynyas in the Weddell Sea intermittently with a period of about 40-50 years; a multidecadal periodic mixed-layer salinity increase upstream of the polynyas appears to be the main precursor for their periodic emergence in this simulation; ultimately, it is the seasonal brine rejection effect superimposed on the multidecadal increase that raises the mixed-layer salinity beyond the tipping point for triggering deep convection and a polynya event. The initiation of polynya events is thus controlled by surface properties while the location is determined by bathymetric features. The persistent Taylor cap effect, which is well represented by the high-resolution topography of the Maud Rise seamount and the Astrid Ridge, preconditions this region by weakening the stratification. Polynyas thus occur predominantly first at these two locations, while WSPs grow out of MRPs that are associated with strong convection and thus high surface salinity anomalies. The wind-stress curl, the SAM index, and the strength of the Weddell Gyre are all correlated with the polynya events. The simulation suffers from a climate drift of Warm Deep Water which, however, does not disrupt the intermittent occurrence of polynyas.

The 40-50 years periodicity of WSP occurrence is also prominent in the HF-TNST simulation, while MRPs occur more frequently than observed (i.e., since year 1973). Furthermore, while the mixed-layer depth co-varies with the occurrence of polynyas in the Weddell Sea throughout the 500 years of the PI-CTRL simulation, it does so in the HF-TNST simulation only up to about year 2040. After that point in time, WSPs develop mostly in open embayments (not all surrounded by sea ice), and are not associated with deep mixed layers anymore, suggesting that warming from above will reduce the winter ice cover rather than warming from below due to deep convection. We will elaborate on the mechanism of this new mode of operation.