The changing surface of the world's roads

The changing surface of the world’s roads Sukanya Randhawa1,*, Guntaj Randhawa2, Clemens Langer1,2, Francis Andorful1,2, Benjamin Herfort1, Daniel Kwakye2, Omer Olchik1, Sven Lautenbach1,2,3, and Alexander Zipf1,2,3 1Heidelberg Institute of Geoinformation Technology (HeiGIT), Berliner Str. 45 (Mathematikon), 69120 Heidelberg, Germany 2GIScience Chair, Institute of Geography, Heidelberg University, Im Neuenheimer Feld 368, 69120 Heidelberg, Germany 3Centre for the Environment, Heidelberg University, Im Neuenheimer Feld 130.1, 69120 Heidelberg, Germany *sukanya.randhawa@heigit.org ABSTRACT Resilient road infrastructure is a cornerstone of the UN Sustainable Development Goals. Yet a primary indicator of network functionality and resilience is critically lacking: a comprehensive global baseline of road surface information. Here, we overcome this gap by applying a deep learning framework to a global mosaic of Planetscope satellite imagery from 2020 and 2024. The result is the first global multi-temporal dataset of road pavedness and width for 9.2 million km of critical arterial roads, achieving 95.5% coverage where nearly half the network was previously unclassified. This dataset reveals a powerful multi-scale geography of human development. At the planetary scale, we show that the rate of change in pavedness is a robust proxy for a country’s development trajectory (correlation with HDI = 0.65). At the national scale, we quantify how unpaved roads constitute a fragile backbone for economic connectivity. We further synthesize our data into a global Humanitarian Passability Matrix with direct implications for humanitarian logistics. At the local scale, case studies demonstrate the framework’s versatility: in Ghana, road quality disparities expose the spatial outcomes of governance; in Pakistan, the data identifies infrastructure vulnerabilities to inform climate resilience planning. Together, this work delivers both a foundational dataset and a multi-scale analytical framework for monitoring global infrastructure, from the dynamics of national development to the realities of local governance, climate adaptation, and equity. Unlike traditional proxies such as nighttime lights, which reflect economic activity, road surface data directly measures the physical infrastructure that underpins prosperity and resilience—at higher spatial resolution. Introduction Road networks are the arteries of the global economy, essential for trade, social integration, and access to fundamental services such as healthcare and education1,2. The quality of this infrastructure, determined largely by its surface type, plays a pivotal role in determining transportation efficiency, disaster resilience, and progress towards the United Nations Sustainable Development Goals3–5. The distinction between paved and unpaved roads shapes not only per capita income levels6 but also a nation’s vulnerability to climate-related disruptions such as floods and extreme weather events7–9. Consequently, accurate, high-resolution and up-to-date global data on road surface conditions is not merely a logistical asset but a fundamental prerequisite for monitoring economic progress, targeting infrastructure investment, and assessing climate vulnerability. While previous research has relied heavily on proxies such as nighttime light intensity to infer economic activity and human development10,11, such radiance-based measures remain limited by their coarse spatial resolution and their indirect relationship to the underlying physical infrastructure that enables growth. In contrast, road surface characteristics offer a direct, high-resolution, and functionally grounded signal of economic investment, accessibility, and resilience—capturing dimensions of human development that nighttime lights cannot resolve. Yet despite this critical need — and substantial progress made by large-scale geospatial AI - a comprehensive global baseline of road surface information remains elusive at the time of writing. Recent years have seen remarkable progress in deriving geospatial datasets of the built environment, driven by large-scale AI initiatives. Recent datasets from organizations like Meta AI12 and Microsoft13 have improved the completeness of global road network geometries, while parallel efforts from Microsoft14 and Google15 have provided comprehensive inventories of building footprints. However, while these datasets have enhanced our understanding of the presence and location of global infrastructure, they do not capture the critical attributes that determine functionality and resilience. This semantic data gap—the absence of physical detail such as surface type—represents a major limitation in current infrastructure intelligence. The current de facto standard, OpenStreetMap (OSM), is a monumental arXiv:2512.04092v1 [physics.soc-ph] 17 Nov 2025 achievement, yet its surface attributes are quite incomplete—covering only 30-40% of the global network—and are often outdated16–18.

This content is AI-processed based on open access ArXiv data.