Modelagem de séries temporais em dados de temperatura do ar por semivariografia univariada : interpolação spline e validação cruzada

Abstract

The model implies a process of transcribing and create a series of funcionamento organizations or simplifications of real systems. The use of mathematical modeling of a variable microclimate, with the aim of describing the behavior of the series using geostatistics, which is characterized by the analysis of the spatial / temporal distribution is determined through variogram, the distance or the time dependence of period them. This feature has great application, especially to perform interpolation, characterization, estimates or simulations. The survey was conducted in experimental site of the lattice towers of the Great Experiment Scale Biosphere-Atmosphere Experiment in Amazonia (LBA), the Biological Reserve of Jaru (REBIO Jaru), JiParaná. The temperature data were collected in 2010, and developed routines in Matlab 2012b software. The geostatistical analysis was used to define the model of the time series. The time dependence was analyzed by adjusting the classical semivariogram.