A new NN approach has been introduced and preliminary investigated by the authors for calculating physical processes in atmospheric and ocean models. It is based on application of advanced NN techniques to emulation of model physics for NCAR CAM-2 (Krasnopolsky et al. 2003, 2004a-d; Fox-Rabinovitz et al. 2004), and for ocean models (Krasnopolsky et al. 2002). This approach is a novel synergetic combination of NN techniques and deterministic modeling. The approach uses NNs as a statistical technique for accurate and fast emulation or approximation of model physics components. NN emulations have been developed for existing parameterizations of atmospheric and oceanic physics that has already been very carefully tested and validated by its developers off-line and then on-line through experimentation with the entire model. Due to the capability of NNs to provide an unprecedented accuracy for approximation of complex multidimensional systems like model physics, our NN emulations of model physics parameterizations are practically identical to original physical parameterizations. In the context of our approach, the accuracy and speed-up of NN emulations is always measured against the original parameterization. In this paper the approach is applied to the radiation block of the NCAR CAM. The performance and the accuracy of the NN emulations are estimated and discussed. For example, the application of this approach allowed us to accelerate the calculation of the LW radiation parameterization by about 50-80 times (or takes only 1% - 2% of the original parameterization computation time) without compromising the accuracy and integrity of the original long wave radiation parameterization. Deferent statistical measures and characteristics are used for the estimation of the performance and accuracy. An additional important evaluation of the developed NN emulations was also performed in NCAR CAM where parallel runs were performed with the original radiation parameterizations and with their NN emulations. The impact of using NN emulation on climate simulation has been assessed by a comparison of some basic climate characteristics of parallel NCAR CAM simulations, calculated with the original LW radiation parameterization and its NN emulations. The differences between the simulations with the original LW radiation parameterization and its NN emulations appear to be very small for simulated fields, well within observational error.

REFERENCES Fox-Rabinovitz, M. S., V.M. Krasnopolsky, D.V. Chalikov, 2004: Decadal climate simulations using NN emulations for long wave radiation parameterization, to be submitted Krasnopolsky V., D.V. Chalikov, and H.L. Tolman (2002) A neural network technique to improve computational efficiency of numerical oceanic models, Ocean Modelling, v. 4, 363-383 Krasnopolsky V.M. and F. Chevallier (2003), Some neural network applications in environmental sciences. Part II: Advancing the computational efficiency of environmental numerical models. Neural Networks, 16, 335-348. Krasnopolsky V.M., M.S. Fox-Rabinovitz, and D.V. Chalikov, (2004a): Using neural networks for fast and accurate approximation of the long wave radiation parameterization in the NCAR community atmospheric model: evaluation of computational performance and accuracy of approximation, Proc., 15th Symposium on Global Change and Climate Variations, 84th AMS Annual Meeting, Seattle, Washington, CD-ROM, P1.20 Krasnopolsky V.M., M.S. Fox-Rabinovitz, and D.V. Chalikov, (2004b): Fast and Accurate Approximation of the Long Wave Radiation Parameterization in a GCM Using Neural Networks: Evaluation of Computational Performance and Accuracy of Approximation in the NCAR CAM-2, Proc. CIMSA 2004 – IEEE International Conference on Computational Intelligence for Measurement Systems and Applications, Boston, MD, USA, 14-16 July 2004, CD-ROM, pp. 57-62 Krasnopolsky V.M., M.S. Fox-Rabinovitz, and D.V. Chalikov, (2004c): New Approach to Calculation of Atmospheric Model Physics: Accurate and Fast Neural Network Emulation of Long Wave Radiation in a Climate Model, submitted Krasnopolsky V.M. and M.S. Fox-Rabinovitz, (2004d): A New Synergetic Paradigm in Environmental Numerical Modeling: Hybrid Environmental Numerical Models Consisting of Deterministic and Machine Learning Components, submitted.