Local-based mapping of carbon footprint variation in Turkey using artificial neural networks
Küçük Resim Yok
Tarih
2021
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Springer Science and Business Media Deutschland GmbH
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
This study aims to create a classifiable and comparable carbon footprint map of Turkey on the basis of cities with the provincial data inventory and data standard. Therefore, the basic parameters of carbon footprint change were determined by conducting the literature review. The carbon footprint change of each province was modeled using artificial neural networks for 2015 due to data availability. The performance metrics of the model created are (R2 0.97622, MSE 3.19809, RMSE 1.78832). The carbon footprint values of each province were predicted with the lowest error for the years 2014, 2015, 2016, and 2017, for which access data were available with the help of modeling. The spatial distributions of the prediction values were mapped using the geographic information system (GIS). An increase of 7 Mt on average was detected in the carbon footprint change of Turkey in 4 years. Furthermore, it was determined which parameter had a weighted effect in determining the carbon footprint values using the RReliefF algorithm, the feature selection method based on the established model’s prediction values. Accordingly, the parameters affecting the carbon footprint change on Turkey’s model were determined according to their weight as follows (income level 0.057, housing 0.029, the number of motor vehicles 0.028, sectoral industry size 0.025, population density 0.024, energy consumption 0.017, infrastructure 0.005, agricultural area 0.004, and agricultural production value 0.0008), respectively. This study draws a way for local governments to model their climate action plans with a standard data inventory, considering the parameters with weighted effects on carbon footprint change. © 2021, Saudi Society for Geosciences.
Açıklama
Anahtar Kelimeler
Accounting parameters, Artificial neural networks, City carbon map, Climate change, Feature selection, Urban carbon footprint
Kaynak
Arabian Journal of Geosciences
WoS Q Değeri
Scopus Q Değeri
N/A
Cilt
14
Sayı
6
