enhanced arctic precipitation commensurate with what is needed to generate a 400 m thick ice sheet within 500 years. Vardiman (1998) used the same model to test the impact of ocean hot spots over the Mid-Atlantic ridge. This resulted in shifted wind patterns and enhanced downwind precipitation. Gollmer (2013) verified Spelman’s warm sea surface temperature scenario using the Goddard Institute of Space Studies (GISS) ModelE. Unlike CCM1, ModelE has a dynamic ocean coupled to the atmosphere. This allows the ocean temperature to develop over time. With an initial ocean temperature of 30 °C, both at depth and high latitudes, the tropical oceans become unreasonably hot several years into the simulation. Stratospheric aerosols with an optical thickness of 2.0 were added to mitigate heat coming from the warm ocean surface. Even with aerosols, this initial ocean temperature is too high. An additional problem with these simulations is the form and location of precipitation. The elevated temperature of the ocean has a corresponding effect on the atmosphere, resulting in precipitation in the form of rain rather than snow. High precipitation events in the Arctic occur over the ocean, not land. With warm oceans, high pressure systems develop over the cooler land. This results in thermal winds that flow to low pressure systems over the ocean. Doubling the GISS ModelE’s resolution, Gollmer (2018) explored the impact of aerosol distribution on precipitation patterns for the post-flood ice age scenario. Ocean temperatures were initialized at a lower temperature of 24 °C, which is in line with arctic ocean temperatures inferred from marine sediment cores by Sluijs et al. (2006). Although non-uniform aerosol distributions shift the jet stream, and in turn precipitation patterns, it was not enough to provide enhanced precipitation over land at high latitudes. Even with the reduction of initial ocean temperature, during the six-year sensitivity study there was insufficient cooling to allow snowfall or ice accumulation needed for initiating an ice age. Given these results, episodic high precipitation events are seen as a possible source of snowfall needed to initiate an ice age. Vardiman (2003) explored the impact of hypercanes forming over warm oceans using a mesoscale model. These models divide the earth’s surface into cells 30 km on a side or smaller. Operating with cells twenty times larger in length and width, global circulation models fail to capture sub-scale processes, which can generate significant precipitation events. Vardiman and Brewer (2010, 2011, and 2012) used the NCAR Weather Research and Forecasting Model (WRF) to simulate the impact of warm oceans on Yellowstone, the Middle East, and the Eastern United States respectively. Although mesoscale simulations may provide answers to high precipitation events over land, they do not provide a description of what is happening at the global scale. As already seen, warm ocean simulations result in precipitation in the form of rain everywhere except at high altitudes. This necessitates global simulations that extend beyond sensitivity studies. The purpose of this study is to provide a multi-century simulation at the global scale. This provides time for oceans with uniform temperatures to cool and establish circulation patterns due to temperature differences. In turn, the atmosphere cools to the point where snow can accumulate. The next section of this paper describes the climate model used and its boundary and initial conditions. The following section presents the results and commentary of how different oceanic and atmospheric values change over time. This paper concludes with a discussion of relevant observations and proposed future research. II. FOUR CENTURY CLIMATE MODEL RUN Since the 1950s, numerous global circulation models (GCMs) have been developed. A model’s success is gauged by how accurately it predicts long-term temperature, precipitation, and wind patterns for the present climate. In the past seventy years, these models have incorporated sophisticated ocean/atmosphere, land/vegetation, and dynamic sea/land ice interactions. The complexity of these models can no longer be maintained and advanced by a single researcher. Therefore, communities of researchers formed to focus their efforts in developing the best climate models. The Third Coupled Model Intercomparison Project (CMIP3) was released in 2007 in conjunction with the Fourth Assessment Report (AR4) of the Intergovernmental Panel on Climate Change (IPCC). It consisted of twenty-five models from eighteen research groups (PCMDI, 2007). Poorly understood feedbacks involving the carbon cycle and clouds were the focus of CMIP5. By 2013 GCMs were incorporating new parameterizations to provide better results related to these physical processes. CMIP5 involved forty GCMs from twenty research groups (Kamworapan and Surussavadee 2019). This in turn provided projections of climate change used by IPCC AR5. CMIP6 continues to incorporate more accurate representations of biogeochemical processes. To date, there are one hundred models from forty-nine modelling groups contributing to this effort (Hausfather 2019). Results from these models are used in AR6, which is still in development. Recognizing the need to use a vetted GCM, Gollmer (2013) used the Goddard Institute of Space Studies Model E, which was part of AR4. The horizontal resolution of this model was 5° longitude by 4° latitude, with 20 atmospheric layers. It included a coupled-ocean model consisting of the same horizontal resolution and 13 ocean layers. Gollmer (2018) used the AR5 update of the GISS Model E2 (Hansen et al. 1983). This included improvements in calculations of microphysics and other physical effects (Schmidt et al. 2014). At the expense of computational time, the resolution was doubled by representing the atmosphere with cells of 2.5° longitude by 2° latitude and 40 layers. The ocean consisted of 32 layers. Both Gollmer (2013) and (2018) did sensitivity studies, which looked at changes in atmospheric parameters due to changed initial and boundary conditions. These conditions include the following: increased uniform ocean temperature, thick stratospheric aerosols, removal of sea ice, and removal of ice sheets in Greenland and Antarctica. This paper reports on a simulation using the GISS Model E2.1.2, which incorporates improvements linked to CMIP6 (Kelley et al. 2020). With the need to simulate ocean cooling from an initial uniform temperature of 24 °C, a 400-year model run was performed. Given this extended time, it was necessary to drop the resolution back to 5° longitude by 4° latitude with 20 atmospheric layers and 13 ocean layers. Although the resolution was reduced, model improvements added since AR4 were still enabled. In this configuration, one year of model time was simulated in four hours on a 2.5 GHz Intel GOLLMER Rapid ice age 2023 ICC 268
RkJQdWJsaXNoZXIy MTM4ODY=