Interpolating monthly precipitation by self-organizing map (SOM) and multilayer perceptron (MLP)

Forskningsoutput: TidskriftsbidragArtikel i vetenskaplig tidskrift

Bibtex

@article{48578446196a4ae78f565abd496721a9,
title = "Interpolating monthly precipitation by self-organizing map (SOM) and multilayer perceptron (MLP)",
abstract = "There are needs to find better and more efficient methods to interpolate precipitation data in space and time. Interpolation of precipitation is explored using a self-organizing map (SOM) in a region with large complexity of precipitation mechanisms (northern Iran). The technique is used both for regionalization and for interpolating monthly precipitation for stations with missing data for 1-, 2-, 5- and 10-year periods using a jack-knife procedure to obtain objective results. The SOM is able both to find regions with similar precipitation mechanisms and to interpolate with accuracy. The results show that precipitation interpolation can be improved considerably by taking into account the regionalization properties in the SOM modelling. The SOM results are compared with those from a well-defined multilayer perceptron (MLP). The findings suggest that, without regionalization, MLP modelling is generally better than SOM. However, when regionalization is included, SOM performs better than MLP. Il est n{\'e}cessaire de trouver des m{\'e}thodes meilleures et plus efficaces pour interpoler des donn{\'e}es de pr{\'e}cipitation dans l'espace et le temps. L'utilisation d'une carte auto-organis{\'e}e (SOM) pour l'interpolation des pr{\'e}cipitations est explor{\'e}e dans une r{\'e}gion aux m{\'e}canismes de pr{\'e}cipitation tr{\`e}s complexes (nord de l'Iran). La technique est utilis{\'e}e pour la r{\'e}gionalisation et pour l'interpolation des pr{\'e}cipitations mensuelles de stations qui pr{\'e}sentent des lacunes pour des p{\'e}riodes de 1, 2, 5 et 10 ans, {\`a} l'aide d'une proc{\'e}dure jack-knife pour obtenir des r{\'e}sultats objectifs. La SOM est capable d'identifier les r{\'e}gions dont les m{\'e}canismes de pr{\'e}cipitation sont similaires et d'interpoler avec pr{\'e}cision. Les r{\'e}sultats montrent que l'interpolation des pr{\'e}cipitations peut {\^e}tre consid{\'e}rablement am{\'e}lior{\'e}e en tenant compte des propri{\'e}t{\'e}s de r{\'e}gionalisation dans la mod{\'e}lisation SOM. Les r{\'e}sultats de SOM sont compar{\'e}s avec ceux d'un perceptron multi-couches (PMC) bien d{\'e}fini. Les r{\'e}sultats sugg{\`e}rent que, sans r{\'e}gionalisation, la mod{\'e}lisation par PMC est g{\'e}n{\'e}ralement meilleure que par SOM. Cependant, lorsque la r{\'e}gionalisation est introduite, la SOM donne de meilleurs r{\'e}sultats que le PMC.",
keywords = "missing data, interpolation, multilayer perceptron (MLP), northern Iran, self-organizing map (SOM), precipitation",
author = "Kalteh, {Aman Mohammad} and Ronny Berndtsson",
year = "2007",
doi = "10.1623/hysj.52.2.305",
language = "English",
volume = "52",
pages = "305--317",
journal = "Hydrological Sciences Journal",
issn = "0262-6667",
publisher = "Taylor & Francis",
number = "2",

}