Modelo de “ensemble” para incrementar la robusted de la predicción de avenidas utilizando redes neuronales artificiales: aplicación a la cuenca Gardon (sureste de Francia)

Autores/as

  • T. Darras LGEI, Ecole des Mines d’Alès, 6 Avenue de Clavières, Univ Montpellier - Hydrosciences, Univ Montpellier
  • A. Johannet LGEI, Ecole des Mines d’Alès, 6 Avenue de Clavières, Univ Montpellier
  • B. Vayssade LGEI, Ecole des Mines d’Alès, 6 Avenue de Clavières, Univ Montpellier
  • L. Kong-A-Siou ENSEGID
  • S. Pistre Hydrosciences, Univ Montpellier

DOI:

https://doi.org/10.21701/bolgeomin.129.3.007

Palabras clave:

riada de ciclo rápido, previsión, redes neuronales artificiales, selección de modelos, validación cruzada

Resumen


En las últimas décadas el uso de redes neuronales en la modelización hidrológica ha aumentado, debido a su propiedad fundamental como aproximador universal y parsimónico de funciones no lineales. En el campo de la previsión de inundaciones, los perceptrones de alimentación directa (feed-forward) y de tipo multicapa recurrentes (recurrent multilayer) han confirmado su eficiencia. Sin embargo, la capacidad predictiva depende de los parámetros de inicialización del sistema neuronal. La eficacia del método de validación cruzada para seleccionar las condiciones óptimas de inicialización que conducen a las mejores predicciones ha sido analizada. La dependencia de la inicialización en los modelos de retroalimentación y de multicapa recurrente ha sido comparada para las predicciones con antelación de una hora. Nuestro trabajo demuestra que la validación cruzada no permite la selección de la mejor inicialización. Un modelo más robusto ha sido desarrollado gracias al uso de la mediana de los resultados de varios modelos; en ese contexto, este trabajo analiza la estructura de los meta modelos tanto para los sistemas basados en redes retroalimentadas como para aquellos basados en redes multicapa recurrentes.

Descargas

Los datos de descargas todavía no están disponibles.

Citas

Abrahart, R. J. and See, L. M. 1997. Neural network model-ling of non-linear hydrological relationships,Hydrological Earth System Sciences, 11 (5), 1563-1579. https://doi.org/10.5194/hess-11-1563-2007

Anctil, F., Lauzon, N. and Filion, M. 2008. Added gains of soilmoisture content observations for streamflow predic-tions using neural networks, Journal of Hydrology,359(3-4), 225-234. https://doi.org/10.1016/j.jhydrol.2008.07.003

Artigue, G., Johannet, A., Borrell, V. and Pistre, S. 2012.Flash flood forecasting in poorly gauged basins usingneural networks: case study of the Gardon de Mialetbasin (southern France), Natural Hazards Earth SystemSciences, 12 (11), 3307-3324. https://doi.org/10.5194/nhess-12-3307-2012

Ayral, P. A. 2005. Contribution à la spatialisation du modèleopérationnel de prévision des crues éclair ALHTAÏR,Université de Provence Aix-Marseille.

Barron, A. R. 1993. Universal approximation bounds forsuperpositions of a sigmoidal function, IEEETransactions on Information Theory, 39 (3), 930-945. https://doi.org/10.1109/18.256500

Borga, M., Anagnostou, E. N., Blöschl, G. and Creutin, J. D.2011. Flash flood forecasting, warning and risk manage-ment: the HYDRATE project, Environmental SciencePolicy, 14 (7), 834-844. https://doi.org/10.1016/j.envsci.2011.05.017

Bornancin-Plantier, A., Johannet, A., Borrell Estupina, V.,Roussel-Ragot, P. and Dreyfus, G. 2011. Conception demodèles de prévision des crues éclair par apprentissageartificiel, in EGU2011-1794, 2011, vol. 13.C

osandey, C. and Robinson, M.: Hydrologie continentale, A.Colin., 2000.Dawson, C. W. and Wilby, R. L. 2001. Hydrological modellingusing artificial neural networks, Progress in PhysicalGeography, 25 (1), 80-108. https://doi.org/10.1191/030913301674775671

Diettrich, T. G. 2015. Ensemble Methods in MachineLearning, in Lecture Notes in Computer Science, p. 115,Springer Verlag, New-York. [online] Available from:http://www.eecs.wsu.edu/~holder/courses/CptS570/fall07/papers/Dietterich00.pdf (Accessed 16 June 2015).

Dreyfus, G. 2005. Neural Networks: Methodology andApplications, Softcover reprint of hardcover 1st ed. 2005edition., Springer, Berlin; New York.

Garambois, P. A., Larnier, K., Roux, H., Labat, D. and Dartus,D. 2014. Analysis of flash flood-triggering rainfall for aprocess-oriented hydrological model, AtmosphericResearch, 137, 14-24. https://doi.org/10.1016/j.atmosres.2013.09.016

Gaume, E., Bain, V., Bernardara, P., Newinger, O., Barbuc,M., Bateman, A., Blaškoviová, L., Blöschl, G., Borga, M.,Dumitrescu, A., Daliakopoulos, I., Garcia, J., Irimescu, A.,Kohnova, S., Koutroulis, A., Marchi, L., Matreata, S.,Medina, V., Preciso, E., Sempere-Torres, D., Stancalie, G.,Szolgay, J., Tsanis, I., Velasco, D. and Viglione, A. 2009. Acompilation of data on European flash floods, Journal ofHydrology, 367 (1-2), 70-78. https://doi.org/10.1016/j.jhydrol.2008.12.028

Geman, S., Bienenstock, E. and Doursat, R. 1992. NeuralNetworks and the Bias/Variance Dilemma, NeuralComputing, 4 (1), 1-58. https://doi.org/10.1162/neco.1992.4.1.1

Hornik, K., Stinchcombe, M. and White, H. 1989. Multilayerfeedforward networks are universal approximators,Neural Networks, 2 (5), 359-366. https://doi.org/10.1016/0893-6080(89)90020-8

Huet, P., Martin, X., Prime, J.-L., Foin, P., Laurain, C. andCannard, P. 2003. Retour d'expériences des crues deseptembre 2002 dans les départements du Gard, del'Hérault, du Vaucluse, des Bouches-du-Rhône, del'Ardèche et de la Drôme., Inspection générale del'Environnement, Paris, France. [online] Available from:http://cgedd.documentation.developpement-durable.gouv.fr/document.xsp?id=Cgpc-OUV00000419(Accessed 24 March 2015).

Kitanidis, P. K. and Bras, R. L. 1980. Real-time forecastingwith a conceptual hydrologic model: 2. Applications andresults, Water Resources Research, 16(6), 1034-1044. https://doi.org/10.1029/WR016i006p01034

Kong-A-Siou, L., Johannet, A., Borrell, V. and Pistre, S. 2011.Complexity selection of a neural network model forkarst flood forecasting: The case of the Lez Basin (south-ern France). Journal of Hydrology. 403 (3-4), 367-380. https://doi.org/10.1016/j.jhydrol.2011.04.015

Kong-A-Siou, L., Johannet, A., Valérie, B. E. and Pistre, S.2012. Optimization of the generalization capability forrainfall-runoff modeling by neural networks: the case ofthe Lez aquifer (southern France), Environmental EarthSciences, 65 (8), 2365-2375. https://doi.org/10.1007/s12665-011-1450-9

Kong-A-Siou, L., Fleury, P., Johannet, A., Borrell Estupina, V.,Pistre, S. and Dörfliger N. , 2014. Performance and com-plementarity of two systemic models (reservoir andneural networks) used to simulate spring discharge andpiezometry for a karst aquifer. Journal of Hydrology, 519(D), 3178-3192. https://doi.org/10.1016/j.jhydrol.2014.10.041

Le Lay, M. and Saulnier, G. M. 2007. Exploring the signatureof climate and landscape spatial variabilities in flashflood events: Case of the 8-9 September 2002 Cévennes-Vivarais catastrophic event, Geophysics Res.Letters, 34(13) [online] https://doi.org/10.1029/2007GL029746

Llasat, M. C., Llasat-Botija, M., Prat, M. A., Porcú, F., Price, C.,Mugnai, A., Lagouvardos, K., Kotroni, V., Katsanos, D.,Michaelides, S. and others 2010. High-impact floods andflash floods in Mediterranean countries: the FLASH pre-liminary database, Advances in Geosciences, 23 (23),47-55. https://doi.org/10.5194/adgeo-23-47-2010

Llasat, M. C., Marcos, R., Llasat-Botija, M., Gilabert, J.,Turco, M. and Quintana-Seguí, P. 2014. Flash flood evo-lution in North-Western Mediterranean, AtmosphericResearch, 149, 230-243. https://doi.org/10.1016/j.atmosres.2014.05.024

Marchandise, A. 2007. Modélisation hydrologique distribuéesur le Gardon d'Anduze; étude comparative de différentsmodèles pluie-débit, extrapolation de la normale à l'ex-trême et tests d'hypothèses sur les processushydrologiques, Université Montpellier II-Sciences etTechniques du Languedoc. [online] Available from:http://www.ohmcv.fr/Documents/theses/these_marchan-dise-old.pdf (Accessed 8 December 2014).

Montz, B. E. and Gruntfest, E. 2002. Flash flood mitigation:recommendations for research and applications, GlobalEnvironmental Change Part B Environmental Hazards, 4(1), 15-22. https://doi.org/10.1016/S1464-2867(02)00011-6

Moussa, R. 2010. When monstrosity can be beautiful whilenormality can be ugly: assessing the performance ofevent-based flood models, Hydrology Sciences Journal,55, 1074-1084. https://doi.org/10.1080/02626667.2010.505893

Moussa, R., Chahinian, N. and Bocquillon, C. 2007.Distributed hydrological modelling of a Mediterraneanmountainous catchment - Model construction andmulti-site validation, Journal of Hydrology, 337 (1-2),35-51 https://doi.org/10.1016/j.jhydrol.2007.01.028

Nash, Je. and Sutcliffe, J. V., 1970. River flow forecastingthrough conceptual models part I-A discussion of prin-ciples, Journal of Hydrology, 10 (3), 282-290. https://doi.org/10.1016/0022-1694(70)90255-6

Nikolopoulos, E. I., Anagnostou, E. N., Borga, M., Vivoni, E.R. and Papadopoulos, A. 2011. Sensitivity of a mountainbasin flash flood to initial wetness condition and rainfallvariability, Journal of Hydrology, 402 (3-4), 165-178 https://doi.org/10.1016/j.jhydrol.2010.12.020

Price, C., Yair, Y., Mugnai, A., Lagouvardos, K., Llasat, M. C.,Michaelides, S., Dayan, U., Dietrich, S., Galanti, E.,Garrote, L., Harats, N., Katsanos, D., Kohn, M., Kotroni,V., Llasat-Botija, M., Lynn, B., Mediero, L., Morin, E.,Nicolaides, K., Rozalis, S., Savvidou, K. and Ziv, B. 2011.The FLASH Project: using lightning data to better under-stand and predict flash floods, Environmental SciencePolicy, 14 (7), 898-911. https://doi.org/10.1016/j.envsci.2011.03.004

Schmidhuber, J., 2015. Deep Learning in Neural Networks:An Overview. Neural Networks,61, 85-117. https://doi.org/10.1016/j.neunet.2014.09.003 PMid:25462637

SIEE, 2004. Validation des relevés hydrométriques del'événement des 8 & 9 septembre 2002, DirectionDépartementale de l'Equipement du Gard.

Sjöberg, J., Zhang, Q., Ljung, L., Benveniste, A., Delyon, B.,Glorennec, P.-Y., Hjalmarsson, H. and Juditsky, A. 1995.Nonlinear black-box modeling in system identification: aunified overview, Automatica, 31 (12), 1691-1724. https://doi.org/10.1016/0005-1098(95)00120-8

Stone, M. 1974. Cross-Validatory Choice and Assessment ofStatistical Predictions, Journal of the Royal StatisticalSociety Series B Methodology, 36 (2), 111-147. https://doi.org/10.1111/j.2517-6161.1974.tb00994.x

Toukourou, M., Johannet, A., Dreyfus, G. and Ayral, P.-A.2011. Rainfall-runoff modeling of flash floods in theabsence of rainfall forecasts: the case of "Cévenol flashfloods," Applied Intelligence, 35 (2), 178-189. https://doi.org/10.1007/s10489-010-0210-y

Tramblay, Y., Bouvier, C., Martin, C., Didon-Lescot, J.-F.,Todorovik, D. and Domergue, J.-M. 2010. Assessment ofinitial soil moisture conditions for event-based rain-fall-runoff modelling, Journal of Hydrology, 387 (3-4),176-187 https://doi.org/10.1016/j.jhydrol.2010.04.006

Descargas

Publicado

2024-05-14

Cómo citar

Darras, T., Johannet, A., Vayssade, B. ., Kong-A-Siou, L. ., & Pistre, S. (2024). Modelo de “ensemble” para incrementar la robusted de la predicción de avenidas utilizando redes neuronales artificiales: aplicación a la cuenca Gardon (sureste de Francia). Boletín Geológico Y Minero, 129(3), 565–578. https://doi.org/10.21701/bolgeomin.129.3.007

Número

Sección

Artículos