Multifactor non-linear regression model with feedbacks – regional climate simulation

Abstract

The article analyzes the improved multifactor non-linear regression model with feedbacks and overdetermined expressions for coefficients of system of equations. It studies probable situations occurring when sampling of influencing factors takes place after selection process is complete. It also shows change of type of generating functions and expressions for determining the elements of matrixes and vectors of free terms required in order to get coefficients of feedbacks. The presented system of equations of non-linear regression model takes into account the influence of feedbacks between responses as independent term of equations. General significance of influencing factors and their powers is defined using Fisher’s test. To determine connection between zones of intensive interaction of hydrometeorological characteristics of the North Atlantic and uniform regions of Eastern Europe the model was tested as a simulation one. To reveal zones of active interaction between the atmosphere and the ocean in the North Atlantic region and to determine the regions causing a significant impact on formation of peculiarities of climatic regimes in the regions of response, methods of cluster and component analysis were applied to influencing factors serving as characteristics of heat and moisture exchange in the near-surface layer and also characteristics of heat, moisture content and circulation properties of air at the 850 hPa and 700 hPa levels before including the latter in the model. Main components of interactions of hydrometeorological characteristics were included into initial sets of samples of the model’s influencing factors and responses. Significant multiple correlation coefficients which characterize the degree of adequacy of the model prove the possibility of its practical use when solving similar problems.

Authors and Affiliations

E. N. Serga

Keywords

Related Articles

Entropy analysis for local structure of granular matter

We apply phenomenological approach based on statistical mechanical concept in the form of Kirkwood-Buff arguments to describe the structural parameters of 2D binary granular mixture. By use of the scale-invariant model f...

Current water balance of Khadzhibeyskyi liman

The article evaluates inputs of various natural and anthropogenic factors to the formation of water balance of Khadzhibeyskyi Liman under the present-day conditions. We discovered that in the latest decades, due to clima...

Spatiotemporal distribution and meteorological characteriatics of hot dry winds in Ukraine at the end of XX ‒ beginning of XXI century

The hot dry winds in Ukraine are dangerous phenomenon, which can significantly impact to crop yields. This winds damages of crops, which active development occurs in the spring and summer time. Their impacts during the s...

Non-equillibrium green’s functions method in matrix represenatation. 4. Quantum interference and dephasing

The following topics are discussed in the frame of the «bottom – up» approach of modern nanoelectronics: a super-brief but hopefully self-containing introduction to the Hamiltonian matrix whose eigenvalues tell us the al...

Thermoelectric phenomena and devices in the generalized model of electron transport

In the tutorial review article intended for researchers, university lecturers and students the thermoelectric Seebeck and Peltier phenomena are considered in the framework of a generalized transport model due to R. Landa...

Download PDF file
  • EP ID EP328267
  • DOI -
  • Views 84
  • Downloads 0

How To Cite

E. N. Serga (2017). Multifactor non-linear regression model with feedbacks – regional climate simulation. Вісник Одеського державного екологічного університету, 21(21), -. https://europub.co.uk./articles/-A-328267