2001: Ocean-Atmosphere Models: Connecting Regional and Global Climate

The ocean is a vast reservoir of heat and moisture and has a profound effect upon the behavior of the overlying atmosphere. Thus, the climate simulated by an atmospheric general circulation model (AGCM) depends upon its representation of the ocean. In the simplest case, temperature at the ocean surface is prescribed, either using the observed values at a particular time or the climatological average; variations in ocean temperature in response to atmospheric changes are prohibited. However, many interesting climate phenomena, such as ENSO, or longer-term climate variability, arise from interactions between the atmosphere and ocean. Therefore, an ocean model allowing adjustment of the ocean to the atmosphere is needed in order to model these events. "Mixed-layer" or "q-flux" ocean models prescribe the accumulation of heat by currents into the upper ocean, but allow its temperature to vary through exchange of heat with the atmosphere. Finally, the atmospheric model can be coupled to a full "dynamic" ocean model, where everything --- including temperature and currents --- is allowed to vary in response to the atmospheric forcing. Each configuration has advantages and disadvantages that depend on the science questions being asked.

In this project, we will contrast the climate simulated by the NASA GISS AGCM coupled to these three different representations of the ocean: prescribed climatological ocean temperature, a q-flux model, and a full ocean GCM (OGCM). The students will contribute to research at GISS by evaluating the quality of each model simulation through comparison to observations, while highlighting the effect of ocean interactions upon the atmosphere by contrasting the response to the different ocean treatments. Initially the team will focus on evaluating the models' treatment of two variables: radiation and precipitation.

The particular questions that will guide our research include:

• How do the model simulations compare with observations? What ocean model results in the most "realistic" simulation of the climate? (How should "realistic" be defined?) Does the answer vary according to the variable examined?
• Are there consistent relationships between certain variables independent of the ocean treatment? (e.g. is rainfall always associated with low sea level pressure?)
• How large is the year-to-year variability in each model? How does that compare to the differences between models?

By focusing on these specific questions we hope to address a number of basic concepts in climate research:

• Understanding how the oceans influence the climate.
• Understanding the complexity of climate and the necessity of taking a holistic view (i.e. to what extent are fields consistent with, or dependent on, other fields)
• Understanding the role of "noise" or variability in climate, and how this can be distinguished from variations in behavior due to model differences.
• Judging which model or tool is appropriate for a particular scientific question. How can a hierarchy of models be used together to infer conclusions about the climate?