Premium
Modelling and forecasting seasonal precipitation in Florida: A vector time‐domain approach
Author(s) -
Chu PaoShin,
Katz Richard W.,
Ding Ping
Publication year - 1995
Publication title -
international journal of climatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.58
H-Index - 166
eISSN - 1097-0088
pISSN - 0899-8418
DOI - 10.1002/joc.3370150107
Subject(s) - autoregressive model , bivariate analysis , teleconnection , precipitation , climatology , multivariate statistics , environmental science , quantitative precipitation forecast , meteorology , econometrics , autoregressive–moving average model , computer science , statistics , mathematics , geography , geology
Some major stages of vector autoregressive (AR) modelling consisting of specification, order determination, estimation, diagnostic checking, and forecasting are described. To illustrate the utility of multivariate AR models for climate teleconnection research, the seasonal Southern Oscillation Index (SOI) and a precipitation index in Florida are modelled jointly. Two bivariate AR processes, order one and order four, are selected as candidate models to represent the joint series. To make a comparison of model performance, one‐season‐ahead forecasts of Florida precipitation are generated from the bivariate AR models, as well as from a simple regression model, for recent years that are independent from the period on which the model is constructed. Results show that seasonal forecasts from the two‐dimensional, order‐one model are somewhat better than the other two models. One new and important finding of this study is that, in addition to winter precipitation forecast skill, there is also a forecast skill for autumn precipitation. The physical mechanisms by which these teleconnections might arise are briefly discussed. Other applications of vector models are suggested.