Nonlinear Processes in Geophysics www.nonlin-processes-geophys.net 1023-5809 1607-7946 14 3 2007 10.5194/npg-14-211-2007 http://www.nonlin-processes-geophys.net/14/211/2007/ http://www.nonlin-processes-geophys.net/14/211/2007/npg-14-211-2007.html http://www.nonlin-processes-geophys.net/14/211/2007/npg-14-211-2007.pdf 211 222 2007-05-25 Prediction of minimum temperatures in an alpine region by linear and non-linear post-processing of meteorological models E. Eccel emanuele.eccel@iasma.it L. Ghielmi P. Granitto R. Barbiero F. Grazzini D. Cesari IASMA Research Centre – Natural Resources Department, Via E. Mach, 1 – 38010 San Michele all'Adige (TN), Italy Instituto de Física Rosario Conicet UNR Bv. 27 de Febrero 210 bis 2000 Rosario, Argentina Autonomous Province of Trento – Meteotrentino, Department of Civil Protection, Via Vannetti, 41 – 38100 Trento, Italy ARPA-SIM Emilia-Romagna, Viale Silvani 6, 40122 Bologna, Italy Model Output Statistics (MOS) refers to a method of post-processing the direct outputs of numerical weather prediction (NWP) models in order to reduce the biases introduced by a coarse horizontal resolution. This technique is especially useful in orographically complex regions, where large differences can be found between the NWP elevation model and the true orography. This study carries out a comparison of linear and non-linear MOS methods, aimed at the prediction of minimum temperatures in a fruit-growing region of the Italian Alps, based on the output of two different NWPs (ECMWF T511–L60 and LAMI-3). Temperature, of course, is a particularly important NWP output; among other roles it drives the local frost forecast, which is of great interest to agriculture. The mechanisms of cold air drainage, a distinctive aspect of mountain environments, are often unsatisfactorily captured by global circulation models. The simplest post-processing technique applied in this work was a correction for the mean bias, assessed at individual model grid points. We also implemented a multivariate linear regression on the output at the grid points surrounding the target area, and two non-linear models based on machine learning techniques: Neural Networks and Random Forest. We compare the performance of all these techniques on four different NWP data sets. Downscaling the temperatures clearly improved the temperature forecasts with respect to the raw NWP output, and also with respect to the basic mean bias correction. Multivariate methods generally yielded better results, but the advantage of using non-linear algorithms was small if not negligible. RF, the best performing method, was implemented on ECMWF prognostic output at 06:00 UTC over the 9 grid points surrounding the target area. Mean absolute errors in the prediction of 2 m temperature at 06:00 UTC were approximately 1.2°C, close to the natural variability inside the area itself. Abdel-Aal, R. E. and Elhadidy, M. A.: A machine-learning approach to modelling and forecasting the minimum temperature at Dharan, Saudi Arabia, Energy, 19(7), 739–749, 1994. Anadreanistakis, M., Lagouvardos, K., Kotroni, V., and Elefteriadis, H.: Correcting temperature and humidity forecasts using Kalman filtering: Potential for agricultural protection in Northern Greece, Atmos. Res., 71(3), 115–125, 2004. André, J. C. and Mahrt, L.: The nocturnal surface inversion and influence of clear-air radiative cooling, J. Atmos. Sci., 39, 864–878, 1982. Arca, B., Benincasa, F., De Vincenzi, M., and Fasano, G.: A neural model to predict the daily minimum of air temperature, 7th International Congress For Computer Technology in Agriculture, Computer technology in agricultural management and risk prevention. Florence, Italy, 15–18 November 1998, Proceedings, 485–493, 1998. Basili, P., Bonafoni, S., and Biondi, R.: Analisi e previsione di temperature minime e di gelate sul bacino del Trasimeno, Rivista Italiana di Agrometeorologia, 2006(1), 46–50, 2006. Bauer, E. and Kohavi, R.: An empirical comparison of voting classification algorithms: Bagging, Boosting, and variants, Machine Learning, 36(1–2), 105–139, 1999. Bishop, C. M.: Neural Networks for Pattern Recognition. Oxford University Press, 475 pp., 1995. Boi, P.: A statistical method for forecasting extreme daily temperature using ECMWF 2-m temperatures and ground station measurements, Meteorol. Appl., 11, 245–251, 2004. Breiman, L.: Bagging predictors, Machine Learning, 26(2), 123–140, 1996. Breiman, L.: Random Forests, Machine Learning, 45(1), 5–32, 2001. Breiman, L., Friedman, J. H., Olshen, R. A., and Stone, C. J.: Classification and regression trees, Belmont, Wadsworth, 368 pp., 1984. Cane, D., Milelli, M., and Gandini, D.: Improvement of the meteorological parameters forecasts for the XX Olympic winter games venue, Geophys. Res. Abstr., 6, 03637, 2004. Carlson M. A. and Stull, R. B.: Subsidence in the nocturnal boundary layer, J. Clim. Appl. Meteorol., 25, 1088–1099, 1986. Casaioli, M., Mantovani, R., Proietti Scorzoni, F., Puca, S., Speranza, A., and Tirozzi, B.: Linear and nonlinear post-processing of numerically forecasted surface temperature, Nonlin. Processes Geophys., 10, 373–383, 2003. Efron, B. and Tibshirani, R. J.: An introduction to the bootstrap, New York: Chapman & Hall, 456 pp., 1983. Freund, Y. and Schapire, R. E.: Experiments with a New Boosting Algorithm, in: Proceedings of the Thirteenth International Conference on Machine Learning, edited by: Kaufmann, M., 148–156, 1996. Galanis, G. and Anadreanistakis, M.: A one-dimensional Kalman filter for the correction of near surface temperature forecasts, Meteorol. Appl., 9(4), 437–441, 2002. Gassmann, M. I. and Mazzeo, N. A.: Nocturnal stable boundary layer height model and ist application, Atmos. Res., 57(4), 247–259, 2001. Geman, S., Bienenstock, E., and Doursat, R.: Neural Networks and the Bias/Variance Dilemma, Neural Computation, 4, 1–58, 1992. Granitto, P., Verdes, P., and Ceccatto, H. A.: Neural Networks Ensembles: Evaluation of Aggregation algorithms, Artificial Intelligence, 163, 139–162, 2005. Ho, T. K.: Multiple Classifier Combination: Lessons and Next Steps, in: Hybrid Methods in Pattern Recognition, edited by: Kandel, A. and Bunke, H., World Scientific, 171–198, 2002. Homleid, M.: Diurnal corrections of short-term surface temperature forecasts using the Kalman filter, Weather and Forecasting, 10(4), 689–707, 1995. Hsieh, W. W. and Tang, B.: Applying Neural Network Models to Prediction and Data Analysis in Meteorology and Oceanography, Bull. Am. Meteorol. Soc., 79(9), 1855–1870, 1998. Huth, R.: Statistical downscaling of daily temperature in Central Europe, J. Climate, 15, 1731–1742, 2002. Huth, R.: Sensitivity of local daily temperature change estimates to the selections of downscaling models and predictors, J. Climate, 17, 640–652, 2004. Liaw, A. and Wiener, M.: Breiman and Cutler's random forests for classification and regression, R-Package "randomForest": http://stat-www.berkeley.edu/users/breiman/RandomForests, 2005. Marzban, C.: Neural networks for postprocessing model output: ARPS, Mon. Wea. Rev., 131, 1103–1111, 2003. Massie, D. R. and Rose, M. A.: Predicting daily maximum temperatures using linear regression and Eta geopotential thickness forecasts, Weather and Forecasting, 12(4), 799–807, 1997. Miksovsky, J. and Raidl, A.: Testing the performance of three nonlinear methods of time series analysis for prediction and downscaling of European daily temperatures, Nonlin. Proceses Geophys., 12, 979–991, 2005. Navone, H. D. and Ceccatto, H. A.: Predicting Indian monsoon rainfall: a neural network approach, Clim. Dyn., 10(6–7), 305–312, 1994. Robinson, C. and Mort, C.: A neural network system for the protection of citrus crop from frost damage, Computers and Electronics in Agriculture, 16, 177–187, 1997. Schizas, C. N., Michaelides, S., Pattichis, C. S., and Livesay, R. R.: Artificial networks in forecasting minimum temperature, Institution of Electrical Engineers, Publ. No. 349, 112–114, 1991. Schoof, J. T. and Pryor, S. C.: Downscaling temperature and precipitation: A comparison of regression-based methods and artificial neural networks, Int. J. Climatol., 21, 773–790, 2001. Schättler U. and Montani, A. (Eds.): Operational Implementations, COSMO Newsletter No.5, Chapter 4. DWD, Offenbach am Main, Germany, available at http://www.cosmo-model.org/, 2005. Sugahara, S.: Uma experi\^encia com modelo estat\'istico (MOS) para a previs\~ao da temperatura m\'inima diária do ar, Brazilian Journal of Geophysics, 18(1), 3–12, 2000. Tang, B., Hsieh, W. W., Monahan, A. H., and Tangang, F.: Skill comparisons between Neural Networks and Canonical Correlation Analysis in predicting the equatorial Pacific sea surface temperature, J. Climate, 13, 287–293, 2000. Tangang, F. T., Tang, B., Monahan, A. H., and Hsieh, W. W.: Forecasting ENSO Events: A Neural Network–Extended EOF Approach, J. Climate, 11(1), 29–41, 1998. Tresp, V.: Committee machines, in: Handbook for Neural Network Signal Processing, edited by: Hu, Y. H. and Hwang, J. N., CRC Press, pp. 5.1–5.14, 2001. Trigo, R. M. and Palutikof, J. P.: Simulation of daily temperatures for climate change scenarios over Portugal: a neural network model approach, Clim. Res., 13, 45–59, 1999. Venables, W. N. and Ripley, B. D.: Modern Applied Statistics with S, Fourth Edition, Springer, New York, 495 pp., 2002. Verdes, P. F., Granitto, P. M., Navone, H. D., and Ceccatto, H. A.: Frost Prediction with Machine Learning Techniques, Proceedings of the VI Argentine Congress on Computer Science, pp. 1423–1433, 2000. Weichert, A. and Bürger, G.: Linear versus nonlinear techniques in downscaling, Clim. Res., 10, 83–93, 1998. Wilks, D. S.: Statistical methods in the atmospheric sciences, Academic Press, 467 pp., 1995. Wilson, L.: Statistical interpretation methods applied to ensemble forecast, Proceedings of the WMO Workshop on Ensemble Prediction, Beijing, 2001. Woodcock, F. and Sputhern, B.: The use of linear regression to improve official temperature forecasts (Adelaide Brisbane Canberra Hobart Melbourne Perth Sydney), Australian Meteorological Magazine, 31(1), 57–62, 1983.