Remote sensing and AI for building climate adaptation applications

Beril Sirmacek,Ricardo Vinuesa

DOI: https://doi.org/10.48550/arXiv.2107.02693

2022-07-12

Abstract:Urban areas are not only one of the biggest contributors to climate change, but also they are one of the most vulnerable areas with high populations who would together experience the negative impacts. In this paper, we address some of the opportunities brought by satellite remote sensing imaging and artificial intelligence (AI) in order to measure climate adaptation of cities automatically. We propose a framework combining AI and simulation which may be useful for extracting indicators from remote-sensing images and may help with predictive estimation of future states of these climate-adaptation-related indicators. When such models become more robust and used in real life applications, they may help decision makers and early responders to choose the best actions to sustain the well-being of society, natural resources and biodiversity. We underline that this is an open field and an on-going area of research for many scientists, therefore we offer an in-depth discussion on the challenges and limitations of data-driven methods and the predictive estimation models in general.

Machine Learning

What problem does this paper attempt to address?

The problems that this paper attempts to solve are: how to use satellite remote - sensing images and artificial intelligence (AI) technology to automatically extract urban climate adaptation indicators and predict the future states of these indicators, thereby helping decision - makers and emergency responders choose the best course of action to maintain social well - being, natural resources and biodiversity.

Specifically, the paper focuses on the following aspects:

1. **The impact of cities on climate change and their vulnerability**:
- Cities are not only one of the main contributors to climate change, but also one of the areas most severely affected by its negative impacts.
- With the acceleration of the urbanization process, it is expected that by 2050, 74% of the European population and 68% of the global population will live in cities. If the current urban climate emergency triggers are not properly identified and changed, a large number of people and the biodiversity inside and outside the cities will face survival pressure.

2. **Smart city data collection and processing**:
- Many smart cities have invested resources in collecting large amounts of data to identify climate stress factors. These data can be analyzed by machine - learning (ML) or artificial intelligence (AI) models to extract indicators related to climate and land use and be used to create prediction models.
- For those cities that are unable to collect big data from distributed sensors and other data - sharing parties, using satellite images to extract indicators can provide a globally applicable method.

3. **Specific applications of using AI and remote - sensing images**:
- Create large - scale rapid maps of land use and environmental resources.
- Predict the changes in indicator future states.
- Simulate hypothetical scenarios for disaster prevention.
- Identify abnormal situations (outlier detection).
- Use explainable artificial intelligence (XAI) methods to explain the impact of indicators.
- Identify the relationships between human activities, climate change and biodiversity changes.
- Make real - time predictions and take actions based on sparse measurement data within the city.

4. **Challenges and limitations in practical applications**:
- Problems such as data collection, feature extraction, model selection, generalization ability, repeatability, maintainability and simulation tools at different scales.
- Ensure the sustainability of the model and its compliance with the United Nations Sustainable Development Goals (SDGs), as well as the ethical and responsible use of AI.

By solving these problems, the paper aims to provide new methods and technical means for urban climate adaptation monitoring, especially in the context of the urgency of dealing with climate change, to ensure that these models can be immediately applied to practical scenarios.