Optimal selection of predictors for greenhouse gas emissions forecast in Hong Kong

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)

Abstract

This study aims to forecast greenhouse gas (GHG) emissions in Hong Kong with an optimal selection of predictors. A review is made on the time-series of the GHG emissions by different sources in 1992–2021. Twenty potential predictors are examined relating to economics, social developments, energy use, waste production and climatic conditions, based on previous studies on GHG emissions and mitigation. Three principal components produce well-fitting linear-log regression models. The model stability over time is verified by expanding window analysis. The model accuracy is assured by a CV(RMSE) of 6.1508% in the testing sets. Based on hierarchical clustering and stepwise regression, the gross floor area of construction projects is the most significant predictor to forecast GHG emissions, followed by the unemployment rate and the total expense in construction works. High levels of the three predictors facilitate GHG mitigation, but cannot meet the carbon neutral target. Carbon reduction technologies should be strengthened in growing construction projects. Electricity related carbon mitigation should be further enhanced. The novelty of this study is to provide insights to updating and prioritizing decarbonization strategies in the next phase of Hong Kong's Climate Action Plan.

Original languageEnglish
Article number133310
JournalJournal of Cleaner Production
Volume370
DOIs
Publication statusPublished - 10 Oct 2022

Keywords

  • Construction
  • Electricity consumption
  • GHG emissions
  • Linear-log regression
  • Social and economic predictors

Fingerprint

Dive into the research topics of 'Optimal selection of predictors for greenhouse gas emissions forecast in Hong Kong'. Together they form a unique fingerprint.

Cite this