Search

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. • A review of main initiatives of sustainable campus construction worldwide. • An evaluation combining campus ecological footprint and individual carbon footprint. • The ecological footprint of "China Forest University" reflects unsustainable. • Machine learning identifies internal driving force of campus carbon emissions. • Targeted guidelines on campus low-carbon construction are provided. [ABSTRACT FROM AUTHOR]

In order to mitigate global climate change and air pollution, the Chinese government has assigned high priority to expanding low-carbon power generation in China. Recent studies have shown that wind power is one of the most promising renewable energy option in China. Although many studies have estimated the generation potential of onshore wind power, their results vary widely from 1783 TWh to 39,000 TWh. Therefore, we examine the different assumptions in these papers and identify three main factors influencing the results. The three influencing factors are: weather data set, land utilisation factor, and wind turbine configuration. For our model-based analysis, we define a reference scenario which is used to compare the results. Our analysis shows using a different weather data set increases the generation potential to roughly 35,000 TWh. This is 54% higher than the generation potential of the reference scenario. The land utilisation factor also has a large influence, ranging between −10% and −51%. The studies' assumptions and data should be subjected to careful scrutiny, as the calculated wind power potentials are widely used to develop decarbonisation strategies for the energy system. • Study results on wind power potentials in China vary from 1783 TWh to 39,000 TWh • Influencing factors: weather data set, land utilisation, and wind turbine configuration. • Model-based sensitivity analysis with nine scenario settings. • Chosen weather data set has the highest impact on the calculated results with a variation of 54%. [ABSTRACT FROM AUTHOR]