RESEARCH PAPER PRESENTATION

Global stocktake of 53,000 urban climate studies

AI maps 53,295 urban climate studies, exposing major evidence gaps in small, fast-growing cities, especially in Africa and Asia.

Research Paper
MitigationAdaptationBehavior and capacityRegulation & governance

AI-enabled evidence mapping can systematically reveal the scale, biases and policy relevance of global urban climate research.

The authors investigate whether supervised and unsupervised machine learning can identify and classify the full body of urban climate research more comprehensively than manual approaches. They assess how this evidence is distributed across topics, places and city types, and compare it with what the IPCC has actually used. The paper also examines whether current assessment processes amplify existing research biases or help balance them.

Paper hypothesis
1

The problem

Urban climate evidence is expanding too quickly and across too many disciplines for conventional reviews to track effectively. As a result, policymakers and assessment bodies such as the IPCC risk relying on a fragmented and biased knowledge base, especially for cities in the Global South. This matters because many of the most climate-vulnerable and fastest-growing urban areas have the least research coverage, limiting evidence-based planning where it is most needed.

2

Key findings

  • 53,295 urban climate studies were identified, including 19,733 city-specific case studies, showing a much larger evidence base than previously estimated.
  • 84% of the literature was published from 2012–2022, and city case study literature is now 239 times larger than in 1990.
  • Research is heavily skewed toward larger, wealthier, coastal and higher-emitting cities, while small and fast-growing cities in Africa and Asia remain sparsely studied.
  • The IPCC AR6 directly covered only 4.6% of the full urban climate literature, and only 6.8% of available reviews were used.

53,295

URBAN CLIMATE STUDIES IDENTIFIED GLOBALLY

84%

OF STUDIES PUBLISHED BETWEEN 2012 AND 2022

239x

GROWTH IN CITY CASE STUDY LITERATURE SINCE 1990

4.6%

OF TOTAL URBAN CLIMATE LITERATURE COVERED BY IPCC AR6

3

What cities should do

Overall assessment: This is most useful for fast-growing, under-studied cities and for cities seeking to adapt lessons from better-studied peers rather than waiting for bespoke local evidence. Cities and supporting institutions should use evidence more strategically, while also helping close major knowledge gaps in underserved regions.

  • Use city typologies to transfer lessons from similar cities when local studies are scarce.
  • Prioritize planning support for small and fast-growing cities in Africa, Asia and inland regions.
  • Pair adaptation and mitigation planning with better evidence synthesis, not just new case studies.
  • Expand attention to solution-oriented topics such as planning outcomes and implementation, especially in smaller cities.
  • Draw on interdisciplinary evidence, but also scan technical sectors like cooling, transport and smart energy.
  • Work with researchers to co-produce locally relevant evidence where major gaps persist.
4

Implementation Approach

The paper recommends building a more coordinated urban climate evidence system that combines continuous AI-based evidence mapping, targeted review efforts, and cross-city learning. Implementation should focus on making underused evidence accessible to practitioners and assessment bodies, while directing new research toward the most underserved city types and regions.

  1. Years 1-2

    Adopt searchable evidence platforms and identify major local evidence gaps by city type, topic and region.

  2. Years 1-3

    Develop city typologies to match under-studied cities with better-studied peers for transfer learning.

  3. Years 2-4

    Commission systematic reviews in weakly synthesized areas such as transport, energy, negative emissions and adaptation.

  4. Years 3-5

    Co-produce new case studies in small, fast-growing Global South cities where transferability is limited.

  5. Ongoing

    Continuously update evidence databases and feed synthesized findings into IPCC and city decision processes.