Relating precipitation to climatic factors at Central Queensland, Australia : a data engineering approach

This paper contributes a data engineering framework to relate precipitation at Central Queensland in Australia to other climatic factors and ENSO. Advanced data engineering concepts including computational intelligence techniques are used to model precipitation characteristics for areas within the region. A seasonal stratification process based on standardized precipitation index, predictor selection based on mutual information, a multiple imputation technique and a computational intelligence based approach to examine the influence of ENSO have been demonstrated. An ensemble based regression approach has also been highlighted to characterize the relation between predictors and precipitation. Results indicate that a data engineering framework is effective in unraveling the inter-relationships between different factors and precipitation, and characteristics of the relation vary spatially. The outcomes are expected to aid design of regional forecasting model and relevant statistical downscaling.