CR 46:67-76 (2011)  -  DOI: https://doi.org/10.3354/cr00956

ABSTRACT: Satellite imagery of weather was a major breakthrough in the technology of synoptic weather forecasts. Historically, the Earth’s climate has experienced dramatic changes, and researchers have found that long-term weather patterns do exist. Such long-term changes in weather patterns have very small variations in total atmospheric variables and cannot be observed directly. Complicated space-time detection of low energy signals is required and can be done only by computationally processing space-time data. In this paper, new statistical methods for processing such data are discussed and applied so as to create images of long-term changes in climate over space. Using the historical Central England temperature (CET) time series, we clearly identified temporal scales of 2 to 5 yr and longer than 13 yr. Using monthly global temperature records obtained from the National Climatic Data Center’s Global Historical Climatology Network, a long-term average temperature profile along latitude was identified. Global maps of trends of deviations from the average temperature profile display slowly increasing temperatures over a major part of the world. Maps of 2 to 5 yr scales display deviations similar to those observed during an El Niño event and provide the opportunity for explanation and prediction of weather anomalies in different regions of the world. In this paper, we utilize recent achievements in the technology of processing 3-dimensional data, i.e. in the separation of scales greater than 1 yr in monthly global temperature records. The Kolmogorov-Zurbenko spline filter allows for direct selection of scales in time and space to obtain a smooth outcome without the application of any models.

KEY WORDS: Separation of scales · Kolmogorov-Zurbenko filtration in time and space · Kolmogorov-Zurbenko spline · KZS · El Niño