Using Big Data and Data Analytics to intelligently monitor and model meteorological systems

Abstract

The main purpose of this work is a systematic analysis of the modern state of monitoring (goals, capabilities, technologies) of meteorological data, in particular, big data, analytics (mining) and operational-analytical data storefronts, their information sections (profiles), ontological representation of intellectual support for meteorological decision making. The tasks of modeling using time series in conditions of their noisiness (only "white," Gaussian noise is considered), data consolidation and implementation of data management models and meteorological monitoring process are also considered. The study uses methods of analysis, decomposition, situational modeling, parametric identification, mathematical statistics, classification and risk management. The main results of the work: 1) monitoring analytics as a system infrastructure of the "ecosystem," taking into account the demand for information objects (three-level for simplicity); 2) a model of calculations that ensures the storage, analysis and adaptability of data according to subject needs, "on the fly"; 3) a structural ontological diagram of the presentation of decision making based on monitoring, as well as a formal mathematical model corresponding to it. The obtained results in practical monitoring will allow you to solve both complex tasks (based on Big Data), and set (evaluate) the risks allowed in monitoring, build identification algorithms and programs (plans) for safe, controlled monitoring in conditions of poorly defined, fuzzy situations.