HR: 16:35h
AN: A54A-02    [Abstracts]
TI: Simulation of the Multivariate ENSO index (MEI) as a Markov Chain and estimation of its probability distribution
AU: * Bravo, J L
EM: jlbravo@atmosfera.unam.mx
AF: Centro de Ciencias de la atmósfera, Universidad Nacional Autónoma de México, Cicuito exterior, Ciudad Universitaria, México, DF. 04510, Mexico
AU: Gay, C
EM: cgay@servidor.unam.mx
AF: Centro de Ciencias de la atmósfera, Universidad Nacional Autónoma de México, Cicuito exterior, Ciudad Universitaria, México, DF. 04510, Mexico
AU: Estrada, F
EM: feporrua@atmosfera.unam.mx
AF: Centro de Ciencias de la atmósfera, Universidad Nacional Autónoma de México, Cicuito exterior, Ciudad Universitaria, México, DF. 04510, Mexico
AB: The objectives of this work are the simulation of the behavior of the Multivariate El Niño Southern Oscillation Index (MEI) with the use of a Markov chain with six states. The estimation of the transition matrix was made using observed values obtained from the monthly MEI time series of overlapped bi-monthly means reported by NOAA since 1950; this data are available through the NOAA web page (http:www.cdc.noaa.gov/people/klaus.wolter/MEI/table.html). With the values of the transition matrix it is possible to test the hypothesis that the MEI series can be simulated using a Markov chain calculating the autocorrelation function and the number of transitions among states in simulated ensembles and compared with the corresponding calculated functions obtained using the observed data. The characteristics of the transition matrix indicate that the process is very persistent given the time scale used and because the MEI series is smoothed. The results do not allow to reject the hypothesis that the process could be simulated using a Markov chain. Then we evaluate the probability of the state of the process after a given number of steps (months) to produce a probabilistic forecasting for the MEI state several months in advance. We selected periods with MEI values greater than 1, these represent the realization of ENSO events. With the values reached in these events we estimate the probability for MEI values grater than 1 in steps of 0.25. This is the conditional probability of obtaining grater MEI values once a MEI =1.00 was reached.
UR: http:www.cdc.noaa.gov/people/klaus.wolter/MEI/table.html
DE: 3238 Prediction (3245, 4263)
DE: 3245 Probabilistic forecasting (3238)
SC: Atmospheric Sciences [A]
MN: 2007 Joint Assembly