Synthetic generation of monthly sea surface temperatures in "El Niño" regions by means of the Fiering-Svanidze method
Traditional methods, autoregressive moving average models ARMA(p,q), Fiering, Svanidze, among others, have been successfully used in estimating hydrological time series such as flows, inflow volumes to reservoirs, series with a winter component, precipitation; nevertheless, their application to mean...
| Autores: | , , , |
|---|---|
| Formato: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2010 |
| País: | México |
| Recursos: | UNIVERSIDAD NACIONAL AUTÓNOMA DE MÉXICO |
| Repositorio: | Atmósfera |
| Idioma: | inglés |
| OAI Identifier: | oai:ojs.pkp.sfu.ca:article/19486 |
| Acesso em linha: | https://www.revistascca.unam.mx/atm/index.php/atm/article/view/19486 |
| Access Level: | acceso abierto |
| Palavra-chave: | Monthly time series modified Svanidze method ARMA models Fiering method sea surface temperature “El Niño” phenomenon. |
| Resumo: | Traditional methods, autoregressive moving average models ARMA(p,q), Fiering, Svanidze, among others, have been successfully used in estimating hydrological time series such as flows, inflow volumes to reservoirs, series with a winter component, precipitation; nevertheless, their application to mean average sea surface temperatures (SST) in three regions, which jointly explain most of the rainfall and runoff patterns in México (González et al., 2000), did not get to reproduce the high autocorrelations shown by the historical temperature records of each region neither the cross correlations among them. In this paper, a brief description of three traditional methods is presented, followed by their application to the case of study and some comments about the limitations of the obtained results. Finally, a procedure is proposed which allows to take advantage of the virtues of both methods (Fiering and Svanidze), through which it is possible to obtain synthetic records that preserve the statistical characteristics of the historical record, specially the autocorrelations and the cross correlations. |
|---|