Research on low-carbon campus based on ecological footprint evaluation and machine learning: A case study in China
Niting Zheng,Sheng Li,Yunpeng Wang,Yuwen Huang,Pietro Bartoccid,Francesco Fantozzid,Junling Huang,Lu Xing,Haiping Yang,Hanping Chen,Qing Yang,Jianlan Li
DOI: https://doi.org/10.1016/j.jclepro.2021.129181
IF: 11.1
2021-01-01
Journal of Cleaner Production
Abstract:Universities, the important locations for scientific research and education, have the responsibility to lead ecological civilization and low carbon transition. Ecological footprint evaluation (EFE) is usually used to measure sustainability of campuses. Although it can provide guidance and reference for overall campus planning, it lacks effective significance for individual behavior, especially when the reduction of carbon emissions is the aim. On the other hand a possible solution can be represented by machine learning. It can identify the key factors that will influence individual's overall carbon emissions caused by students' daily behavior, it can be used to find effective ways to reduce individual carbon emissions. This paper applied EFE and machine learning to comprehensively evaluate campus sustainability and students' carbon emissions. Huazhong University of Science and Technology (HUST), a "University in the Forest", was used as a study case in China. Even if HUST is endowned with a forest coverage of 72%, here we showed that its Ecological Footprint Index was -12.52, indicating strong unsustainability. This is mainly due to the high energy and food consumption, caused by the large population living in the campus and the lacking of energy saving measures. The per capita ecological footprint was relatively high, compared with other universities in the world, which meant more efforts needed to be done on ecological sustainability. Low carbon emission is a key feature for a sustainable campus. Based on the questionnaire survey delivered to 486 students who live in the campus, their daily active data were collected in terms of students' personal clothing, food, housing, consumption and transportation. And their associated carbon emissions were calculated based on emission intensities of Chinese population. Based on 486 detailed datasets, machine learning was then used to identify the key daily behavior to influence students' total carbon emission. Results showed that making behavior changes in air conditioning, food and electric bicycle were the most effective ways to reduce carbon emissions. Finally, while effective suggestions were proposed based on qualitative and quantitative evaluations, it is concluded that it is imperative for universities in China to formulate effective low-carbon policies, to achieve sustainable development and to confront global climate change.