The impact of human behaviors and the built environment on energy use and heat in cities

Abstract

Globally, urbanization has intensified climate change, increasing the urgency of monitoring energy consumption and environmental impacts driven by human activities, and of formulating evidence-based environmental policies, particularly at a fine-grained level in areas where data are scarce or nonexistent. This is critical because the emissions, primarily from energy use, are a major driver of climate change (Intergovernmental Panel on Climate Change (IPCC), 2023). The heating and cooling system is the largest source of energy use in urban infrastructure (IEA, 2022a), prompting policy efforts to decrease energy demand driven by human activities (IEA, 2022b). To support effective policymaking, it is essential to understand the impacts of human behavior and the built environment on energy use and heat in cities. ☐ Urban metabolism has been studied as a framework to identify internal flows between energy, anthropogenic activities, and the physical and social environments in cities (Wolman, 1965). Approaching the topic from a scientific observational perspective, the urban metabolism framework is viewed as a composite of three main pillars: urban energy use, environmental impacts, and human factors (Dobler, Bianco, et al., 2021). To reduce energy demand by policymaking, evidence such as consistent, accurate, and timely energy data and statistics is a key element in the long-term planning for investment in the energy sector for both federal and state governments. ☐ Evaluating residential energy policies often lacks flexibility and effectiveness due to limited data availability and uncertainty at fine spatial scales. For instance, data on HVAC operations at the unit level and cooling infrastructure at the census block level are often insufficient to support evidence-based energy and environmental policymaking. ☐ Observational methods using imagery, such as satellite images and proximal remote sensing, can enhance evidence-based policymaking, particularly in the fields of energy and environmental policy. Rooftop observatory cameras have provided critical data. For instance, Qamar et al. (2022) demonstrated how ground-based hyperspectral imaging can capture temporal changes in vegetation health alongside air quality measurements. With advancements in artificial intelligence (AI) techniques, these observational methods are increasingly used to generate fine-grained evidence. This enables the development of policies informed by ground-truth data on human behavior, which are often inaccessible through conventional approaches such as surveys and interviews. ☐ This study explored a novel and non-intrusive method of using observation from images, such as infrared camera images and satellite images, to provide information that can be used for evidence-based or data-driven policymaking in the energy and environmental policy field. This dissertation provides three aspects of using proximal remote sensing and machine learning techniques, focusing on the urban metabolism framework: urban energy use, environmental impacts, and human behavior. ☐ First, how can proximal remote thermal imaging be used to quantify energy end-use behaviors in a way that could be used to assess the efficacy of heating and cooling policies designed to reduce the frequency of air conditioner use? Second, can we develop an AI model that can be applied to satellite imagery of a small city, using Wilmington, Delaware, as a case study, to detect heat-mitigating white roofs, with the long-term goal of using these measurements to inform future mitigation strategies? Lastly, what are the benefits and precautions associated with observational methods, particularly in determining whether AI-based algorithms can identify and de-identify behavioral information within aggregate data to evaluate individual privacy protection? ☐ Combining these methods and frameworks, this study demonstrates the feasibility of using remote sensing for energy and environmental policymaking. For instance, the first study contributes by providing methods to monitor heating and cooling usage at the unit level, supporting evidence for New York City's Local Law 97, which requires building owners to self-report energy usage, and the Cooling Guideline, which discourages continuous air conditioning use. The second study supports the U.S. Environmental Protection Agency (EPA)’s Cool Roof Program. The final study contributes by reviewing existing laws and policies, evaluating the feasibility of data aggregation to anonymize behavioral energy use, and assessing whether AI can de-anonymize such data using remote sensing as a proxy for energy consumption. ☐ The results of this study not only provide evidence for practical policy solutions but also reveal scientific phenomena about urban energy metabolism through observational methods. This study aims to provide policymaking based on scientific phenomena. This research contributes to the study on the impact of human behaviors and the built environment on energy use and heat in cities.