Author:
Nina Kargapolova
Affiliation:
Laboratory of Stochastic Problems, Institute of Computational Mathematics and Mathematical Geophysics, Pr. Ak. Lavrent’eva 6, Novosibirsk, Russia, Department of Mathematics and Mechanics, Novosibirsk State University, Pirogov St. 2, Novosibirsk and Russia
Keyword(s):
Stochastic Simulation, Non-stationary Random Process, Periodically Correlated Process, Air Temperature, Temperature Extremes, Model Validation.
Related
Ontology
Subjects/Areas/Topics:
Complex Systems Modeling and Simulation
;
Environmental Modeling
;
Formal Methods
;
Mathematical Simulation
;
Simulation and Modeling
;
Stochastic Modeling and Simulation
Abstract:
Two numerical stochastic models of air temperature time-series are considered in this paper. The first model is constructed under the assumption that time-series are nonstationary. In the second model air temperature time-series are considered as a periodically correlated random processes. Data from real observations on weather stations was used for estimation of models’ parameters. On the basis of simulated trajectories, some statistical properties of rare meteorological events, like sharp temperature drops or long-term temperature decreases in summer, are studied.