<abstract xmlns="http://www.w3.org/1999/xhtml">Agricultural greenhouse gases emissions are mainly produced in direct emissions from plant and animal production as well as those associated with land use changes. Agriculture is a major source of atmospheric nitrous oxide (N2O). N2O emissions from agricultural production has the source primarily in soil fertilized by mineral and organic fertilizers. In Poland, agricultural soils are responsible for 77.1% of emissions. Emissions associated with the animal manner farming amount 22.8%. Studies attempt to modeling and predicting of N2O emissions from Direct Soil Emissions in relation to the use of crops and livestock population. In the analysis an artificial neural networks were used. The best values showing the quality of neural regression model were obtained by multilayer perceptrons MLP. Based on the sensitivity analysis, attempts were taken to determine the extent of the contribution of each selected variables on the estimate of the direct emissions of N2O from agricultural soils. The sensitivity analysis of designed network on the structure MLP 9-4-1 shows that the amount of nitrogen fertilizer consumption has the biggest share in the shaping of N2O emissions from Direct Soil Emissions. The sensitivity analysis of network on the structure MLP 16-5-1 pointed to participate cattle and pigs as the most important in the formation of N2O emissions from Direct Soil Emissions. Among the crops in Poland, which may affect the release of N2O stands out rapeseed and rye. The study was conducted using the statistical package Statistica v. 10.0.</abstract>

Agricultural greenhouse gases emissions are mainly produced in direct emissions from plant and animal production as well as those associated with land use changes. Agriculture is a major source of atmospheric nitrous oxide (N2O). N2O emissions from agricultural production has the source primarily in soil fertilized by mineral and organic fertilizers. In Poland, agricultural soils are responsible for 77.1% of emissions. Emissions associated with the animal manner farming amount 22.8%. Studies attempt to modeling and predicting of N2O emissions from Direct Soil Emissions in relation to the use of crops and livestock population. In the analysis an artificial neural networks were used. The best values showing the quality of neural regression model were obtained by multilayer perceptrons MLP. Based on the sensitivity analysis, attempts were taken to determine the extent of the contribution of each selected variables on the estimate of the direct emissions of N2O from agricultural soils. The sensitivity analysis of designed network on the structure MLP 9-4-1 shows that the amount of nitrogen fertilizer consumption has the biggest share in the shaping of N2O emissions from Direct Soil Emissions. The sensitivity analysis of network on the structure MLP 16-5-1 pointed to participate cattle and pigs as the most important in the formation of N2O emissions from Direct Soil Emissions. Among the crops in Poland, which may affect the release of N2O stands out rapeseed and rye. The study was conducted using the statistical package Statistica v. 10.0.

Use of Artificial Neural Networks in Predicting Direct Nitrous Oxide Emissions from Agricultural Soils

Department of Economics and Regional Research, Faculty of Economy and Management, Opole University of Technology, ul. L. Waryńskiego 4, 45-047 Opole, Poland, phone +48 77 449 87 91, fax +48 77 449 88 16

Ecological Chemistry and Engineering S

Agricultural greenhouse gases emissions are mainly produced in direct emissions from plant and animal production as well as those associated with land use...

Use of Artificial Neural Networks in Predicting Direct Nitrous Oxide Emissions from Agricultural Soils

Publicado en línea: 29 may 2013

Páginas: 419 - 428

DOI: https://doi.org/10.2478/eces-2013-0030

This content is open access.