An artificial neural network technique for downscaling GCM outputs to RCM spatial scale

Abstract. An Artificial Neural Network (ANN) approach is used to downscale ECHAM5 GCM temperature (T) and rainfall (R) fields to RegCM3 regional model scale over Europe. The main inputs to the neural network were the ECHAM5 fields and topography, and RegCM3 topography. An ANN trained for the period 1960–1980 was able to recreate the RegCM3 1981–2000 mean T and R fields with reasonable accuracy. The ANN showed an improvement over a simple lapse-rate correction method for T, although the ANN R field did not capture all the fine-scale detail of the RCM field. An ANN trained over a smaller area of Southern Europe was able to capture this detail with more precision. The ANN was unable to accurately recreate the RCM climate change (CC) signal between 1981–2000 and 2081–2100, and it is suggested that this is because the relationship between the GCM fields, RCM fields and topography is not constant with time and changing climate. An ANN trained with three ten-year "time-slices" was able to better reproduce the RCM CC signal, particularly for the full European domain. This approach shows encouraging results but will need further refinement before becoming a viable supplement to dynamical regional climate modelling of temperature and rainfall.

[1]  Anthony O'Hagan,et al.  Diagnostics for Gaussian Process Emulators , 2009, Technometrics.

[2]  Robert E. Dickinson,et al.  Intercomparison of Bulk Aerodynamic Algorithms for the Computation of Sea Surface Fluxes Using TOGA COARE and TAO Data , 1998 .

[3]  James L. McClelland.  Parallel Distributed Processing , 2005 .

[4]  Richard G. Jones,et al.  Global high resolution versus Limited Area Model climate change projections over Europe: quantifying confidence level from PRUDENCE results , 2005 .

[5]  J. Fritsch,et al.  Numerical Prediction of Convectively Driven Mesoscale Pressure Systems. Part II. Mesoscale Model , 1980 .

[6]  T. Engen-Skaugen,et al.  Refinement of dynamically downscaled precipitation and temperature scenarios , 2007 .

[7]  A. O'Hagan,et al.  Bayesian emulation of complex multi-output and dynamic computer models , 2010 .

[8]  Raquel V. Francisco,et al.  Regional Climate Information—Evaluation and Projections , 2001 .

[9]  Raquel V. Francisco,et al.  Evaluating uncertainties in the prediction of regional climate change , 2000 .

[10]  Erika Coppola,et al.  The Regional Climate Change Hyper-Matrix Framework , 2008 .

[11]  D. Heimann,et al.  Statistical-Dynamical Extrapolation of a Nested Regional Climate Simulation , 2001 .

[12]  Raquel V. Francisco,et al.  Regional Climate Modeling for the Developing World: The ICTP RegCM3 and RegCNET , 2007 .

[13]  Philipp Slusallek,et al.  Introduction to real-time ray tracing , 2005, SIGGRAPH Courses.

[14]  James J. Hack,et al.  Description of the NCAR Community Climate Model (CCM3). Technical note , 1996 .

[15]  William W. Hsieh.  Machine Learning Methods in the Environmental Sciences: Contents , 2009 .

[16]  G. Grell,et al.  A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5) , 1994 .

[17]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[18]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[19]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[20]  Thorsteinn S. Rögnvaldsson,et al.  Using neural networks to identify jets , 1991 .

[21]  F. Giorgi,et al.  Development of a Second-Generation Regional Climate Model (RegCM2). Part II: Convective Processes and Assimilation of Lateral Boundary Conditions , 1993 .

[22]  Sue Ellen Haupt,et al.  Artificial Intelligence Methods in the Environmental Sciences , 2008 .

[23]  Rocco Langone,et al.  Attribution of Precipitation Changes on a Regional Scale by Neural Network Modeling: A Case Study , 2010 .

[24]  T. Wigley,et al.  Statistical downscaling of general circulation model output: A comparison of methods , 1998 .

[25]  D. Maraun,et al.  Precipitation downscaling under climate change: Recent developments to bridge the gap between dynamical models and the end user , 2010 .

[26]  Jeremy S. Pal,et al.  Simulation of regional‐scale water and energy budgets: Representation of subgrid cloud and precipitation processes within RegCM , 2000 .

[27]  P. Linden,et al.  ENSEMBLES: Climate Change and its Impacts - Summary of research and results from the ENSEMBLES project , 2009 .

[28]  R. Dickinson,et al.  Biosphere-Atmosphere Transfer Scheme (BATS) version le as coupled to the NCAR community climate model. Technical note. [NCAR (National Center for Atmospheric Research)] , 1993 .

[29]  A. Pasini,et al.  Neural network modelling for the analysis of forcings/temperatures relationships at different scales in the climate system , 2006 .

[30]  Ann Henderson-Sellers,et al.  Biosphere-atmosphere transfer scheme(BATS) version 1e as coupled to the NCAR community climate model , 1993 .

[31]  F. Giorgi,et al.  Introduction to special section : Regional climate modeling revisited , 1999 .

[32]  Marco Verdecchia,et al.  NN5: A neural network based approach for the downscaling of precipitation fields – Model description and preliminary results , 2009 .

[33]  Robin Chadwick,et al.  An Artificial Neural Network Approach to Multispectral Rainfall Estimation over Africa , 2012 .

[34]  Corinne Le Quéré,et al.  Climate Change 2013: The Physical Science Basis , 2013 .

[35]  Guido Visconti,et al.  Validation of Improved TAMANN Neural Network for Operational Satellite-Derived Rainfall Estimation in Africa , 2006 .