Modeling of Paleoclimates

PALEOCLIMATE MODELING INCLUDES the development and application of numerical climate models to simulate climates through Earths 4.6 billion-year history. Paleoclimate modeling is a theoretical branch of the natural science fields of paleoclimatology and paleoceanography, which seek to document and understand Earth's climate and ocean history. A fundamental goal of paleoclimate modeling is to determine the causes of past global climate change as a guide to modern and future climate change.

Numerical climate models are mathematical expressions of the theoretical laws and physical processes that govern the climate system, approximated and written in computer code for efficient computation. Climate models can vary tremendously in their complexity, capability, domain, and spatial resolution. Climate models used for paleo-applications range from relatively simple one-dimensional models that are based on the conservation of energy (energy balance models), to models that solve the primitive equations to predict the general circulation of the atmosphere and ocean in three dimensions (general circulation models). The capabilities of these models maybe limited to one component of the Earth system, the atmosphere, for example; or may include interactions between several components, such as the atmosphere, biosphere, ocean, and cryosphere.

The model domain may be global in extent, or limited to a region, such as the western United States, and is discretized into grid cells. The horizontal (usually latitude by longitude) and vertical (usually altitude, depth, or thickness) resolution of the models can vary substantially between models. Regional models of limited domain may have a horizontal grid spacing of approximately 12 mi. (20 km.), while most global models have grid spacing on the order of hundreds of miles (Idiometers). The decision to use a particular model at a specific resolution is based on the scientific problem to be addressed and consideration of the computational cost of running the model.

Paleoclimate models are nearly always initially developed and fine-tuned for the modern climate as a check on the performance of the model. If the model can successfully simulate important aspects of the modern climate, then the model may be modified for paleoclimate applications. It is often assumed that modern climate dynamics and physics are representative of past climates, and thus these aspects do not change for paleoclimate applications. The modifications for paleo-applications are frequently limited to the model's boundary conditions. In mathematics, boundary conditions define the limits of a set of partial differential equations. Similarly, in paleoclimate modeling, the boundary conditions define the limits of the numerical model. Modifications to the boundary conditions are made to represent processes that occur on geological timescales, such as continental drift and seafloor spreading, mountain building and erosion, sea-level change, evolution, and geochemical cycling.

Common boundary conditions required for global paleoclimate modeling include continental positions and topography, ocean bathymetry, vegetation distribution, soil type, solar luminosity, river drainage, atmospheric trace gases (such as carbon dioxide, methane, nitrous oxide, and ozone), and Earth's orbital characteristics (such as obliquity, precession, and eccentricity). The accurate reconstruction of these boundary conditions is a critical aspect, and one [Page 664]of the largest uncertainties, of paleoclimate modeling. For very ancient time periods, geological evidence of past boundary conditions may not exist or may be known only crudely. Numerous paleoclimate-modeling studies have demonstrated that differences within the uncertainty of a boundary condition can have regional or global climatic consequences, as shown by Poulsen, et al.

...