Satellite-Based Housing Census: A New Approach to Climate Vulnerability Analysis

The paper presents a method for using satellite data to create a global census of residential buildings, focusing on climate risk assessment. It demonstrates the approach through a case study in Kenya, highlighting how this data can inform climate adaptation and mitigation policies.

Research published by the International Monetary Fund (IMF) in collaboration with DataKind, authored by Andinet Woldemichael and Iyke Maduako proposes a novel method to compile a global census of residential buildings using open-source satellite data. This research, part of the IMF’s ongoing efforts to close data gaps related to climate change, aims to assess the exposure of housing assets to climate hazards, such as floods, and provide a better understanding of how residential buildings contribute to and are affected by climate change. Housing is one of the most significant assets in national economies, both as an investment and a critical part of households’ debt in the form of mortgages. Despite the importance of this sector, many countries, particularly developing nations, lack comprehensive and up-to-date housing data. This gap in information not only affects financial markets but also hampers effective policy formulation in areas such as climate risk management.

Satellite Data: A New Tool for Housing Assessment

The study focuses on utilizing open-source satellite data, specifically Google's Open Buildings dataset, to create a detailed, spatially explicit census of residential buildings. This method offers a cost-effective alternative to traditional housing censuses, which are often expensive, time-consuming, and not regularly updated in many countries. By integrating satellite imagery with socioeconomic data, such as population density and nighttime light intensity, the authors developed a scalable approach to estimate building footprints, classify structures, and assess their exposure to climate risks. The research highlights the use of this data in developing countries like Kenya, where building information is either outdated or nonexistent. The study computes various indicators, including building size, height, and value, alongside measures of climate exposure.

Assessing Flood Risk: A Case Study in Kenya

One of the primary applications of this satellite-based approach is in assessing the vulnerability of residential buildings to climate-related hazards, particularly flooding. The authors present Kenya as a case study, demonstrating how the satellite data can be used to estimate the extent of residential properties exposed to riverine flood risks. By combining building footprint data with flood hazard maps, they calculated the area and monetary value of properties at risk from floods with different return periods, such as 10-year and 20-year floods. The findings show that a significant number of residential properties in Kenya are located in flood-prone areas, with millions of square meters of property exposed to flood risks. These results have important implications for policymakers and planners, who need accurate data to develop effective strategies for climate adaptation and mitigation.

Assigning Property Value Using Downscaling Techniques

In addition to assessing climate risks, the paper introduces a technique for downscaling aggregate capital stock values to the level of individual buildings. This method allows researchers to assign monetary values to residential properties based on their size, location, and other factors, using satellite data combined with socioeconomic indicators. The study uses Kenya as an example, where it estimates the total value of residential buildings and their distribution across different regions. The authors employ a downscaling approach that uses population density and nighttime light data to refine estimates of property values at the pixel level. This innovative approach offers a way to overcome the lack of detailed property data in many countries, particularly in low-income and developing nations.

Challenges in Satellite-Based Housing Censuses

However, the paper also acknowledges several limitations of the proposed method. For instance, the satellite data used for building height estimates often contain missing or inaccurate values, particularly in regions with poor-quality satellite images. In cases where building heights were not available, the authors imputed minimum height values, which may lead to inaccuracies in the estimates. Additionally, the classification of buildings as residential or non-residential is not always reliable, as satellite data sometimes misclassifies large structures like greenhouses or industrial parks. Despite these challenges, the study emphasizes that the satellite-based approach offers a valuable tool for filling critical data gaps in housing and climate risk assessments.

Global Implications for Climate Risk and Sustainable Development

The paper’s findings have broader implications for global efforts to address climate risks and achieve sustainable development goals. The authors argue that detailed, spatially explicit data on residential buildings are essential for quantifying financial exposure to climate hazards and for developing effective policies on climate mitigation and adaptation. The study aligns with initiatives such as the G20 Data Gap Initiative, which has recognized the need for better climate data and has tasked organizations like the IMF with coordinating efforts to close these gaps. The research also highlights the potential for integrating satellite-based building data into macroeconomic models used by institutions like the IMF, which could enhance the ability of policymakers to assess the economic impacts of climate change on housing assets.

The study by Woldemichael and Maduako offers an innovative approach to compiling a global census of residential buildings using satellite data, providing crucial insights into the exposure of housing assets to climate risks. While the method has some limitations, particularly regarding the accuracy of building height and classification data, it represents a significant step forward in addressing the data gaps that hinder effective climate risk assessment and policy formulation. By focusing on a case study in Kenya, the research demonstrates the practical applications of this approach in countries with limited housing data, highlighting its potential to inform climate adaptation and mitigation strategies worldwide.