HR: 1340h
AN: A23B-0804    [Abstracts]
TI: Optimizing climate model parameter choice using Linear Inverse Modeling (LIM)
AU: Farrell, B F
AF: Harvard University, Division of Engineering and Applied Sciences, 12 Oxford Street, Cambridge, MA 02138
AU: * Dykema, J A
EM: dykema@fas.harvard.edu
AF: Harvard University, Division of Engineering and Applied Sciences, 12 Oxford Street, Cambridge, MA 02138
AU: Anderson, J G
AF: Harvard University, Division of Engineering and Applied Sciences, 12 Oxford Street, Cambridge, MA 02138
AB: Currently, a range of parameterizations exists for physical processes in the climate system, such as evaporation, convection, and cloud formation. These parameterizations are often based on physical principles, but the range of temperature projections among model simulations of future climate, as illustrated in the most recent IPCC report (Climate Change 2001), suggests much room for improvement. Within these parameterizations are coefficients that control the net effects of small-scale processes. Examples include coefficients of drag, precipitation efficiencies and rates of evaporation. These coefficients may be optimized by bringing the statistical and dynamical properties of climate model output into agreement with the statistical and dynamical properties of climate observations. For this purpose the observations must be carefully chosen to possess high information content, coverage, homogeneity, and reproducibility. We demonstrate the application of this technique by optimizing the parameters of a simple model based on observations. The relevant dynamics are extracted directly from space observations by Linear Inverse Modeling (LIM). LIM is a method for extracting the intrinsic linear dynamics that govern the climatology of a complex system directly from observations of the system. Here we apply LIM to observations of spectrally resolved infrared radiances. Spectrally resolved infrared radiances measured at the top of the atmosphere contain information on temperature, water vapor, clouds, and other absorbers, including aerosols and trace gases, at all altitudes from the surface through the stratosphere, covering of the entire earth. Infrared radiances, as they would be observed at the top of the atmosphere, may be calculated from a climate model via a radiative transfer code. The model parameter values are optimized by bringing the dynamical properties of infrared radiances calculated from model output into agreement with the dynamical properties of infrared radiance observations determined by LIM.
DE: 1694 Instruments and techniques
DE: 1610 Atmosphere (0315, 0325)
DE: 1620 Climate dynamics (3309)
DE: 1640 Remote sensing
SC: Atmospheric Sciences [A]
MN: 2004 AGU Fall Meeting