ArticlePublished: 13 October 2025Yang Chen ORCID: orcid.org/0009-0005-7600-88471,2,Dayang Wang ORCID: orcid.org/0009-0006-8499-27811,2,Luchen Zhang ORCID: orcid.org/0000-0001-6738-47991,2,Chongxiao Wang3,Cheng Sun ORCID: orcid.org/0000-0003-1365-27801,2 &…Qi Dong ORCID: orcid.org/0000-0001-6587-80741,2 Nature Climate Change (2025)Cite this articleSubjectsClimate changeDecision makingInterdisciplinary studiesPoliticsAbstractClimate change poses a rapidly growing threat to World Heritage sites, making effective adaptation strategies essential. However, the extent to which climate adaptation strategies have been integrated into heritage conservation frameworks remains underexplored. Applying text mining and large language models to analyse 535 World Heritage sites, we assess climate awareness (vulnerability, adaptation and resilience) and local actions (policy, process, planning and management) across 1,868 World Heritage documents. We observe spatial differences in climate awareness, influenced by regional contexts and document characteristics. Climate awareness does not align neatly with national political or economic features. Local heritage management and planning actions are negatively associated with vulnerability awareness, while policies show positive associations with overall climate awareness. Our findings demonstrate that large language models are effective tools for text classification in the interdisciplinary field of heritage and climate and offer practical perspectives into the gap between awareness and action in heritage conservation.This is a preview of subscription content, access via your institutionAccess optionsAccess Nature and 54 other Nature Portfolio journalsGet Nature+, our best-value online-access subscription27,99 € / 30 dayscancel any timeLearn moreSubscribe to this journalReceive 12 print issues and online access269,00 € per yearonly 22,42 € per issueLearn moreBuy this articlePurchase on SpringerLinkInstant access to full article PDFBuy nowPrices may be subject to local taxes which are calculated during checkoutFig. 1: Climate awareness of WHS.Fig. 2: Spatial differences in climate awareness.Data availabilityAll the data used in this research are available via figshare at https://doi.org/10.6084/m9.figshare.28823297 (ref. 59). LCT and ABE can be found on the website of UNESCO WHC (https://whc.unesco.org/) as of December 2024.Code availabilityThe code for data collection, processing and analysis is available via figshare at https://doi.org/10.6084/m9.figshare.28823297 (ref. 59).ReferencesSustainable Development Goal 11.4 (UNESCO, 2023); https://uis.unesco.org/en/topic/sustainable-development-goal-11-4Operational Guidelines for the Implementation of the World Heritage Convention (UNESCO WHC, 1977); https://whc.unesco.org/archive/opguide77a.pdfJokilehto, J. The World Heritage List. What is OUV? Defining the Outstanding Universal Value of Cultural World Heritage Properties (Hendrik Bäßler Verlag, 2008).Court, S. et al. Guidance and Toolkit for Impact Assessments in a World Heritage Context (UNESCO, 2022).Chen, G. et al. Natural world heritage sites are at risk from climate change globally. Commun. Earth Environ. 5, 760 (2024).Article Google Scholar Reimann, L., Vafeidis, A. T., Brown, S., Hinkel, J. & Tol, R. S. J. Mediterranean UNESCO World Heritage at risk from coastal flooding and erosion due to sea-level rise. Nat. Commun. 9, 4161 (2018).Article Google Scholar Perry, J. & Gordon, I. J. Adaptive heritage: is this creative thinking or abandoning our values?. Climate 9, 128 (2021).Article Google Scholar Lin, B. B. et al. Holistic climate change adaptation for World Heritage. Nat. Sustain. 6, 1157–1165 (2023).Article Google Scholar Chen, Y., Zhang, L. & Dong, Q. Using natural language processing to evaluate local conservation text: a study of 624 documents from 303 sites of the World Heritage Cities Programme. J. Cult. Herit. 70, 259–270 (2024).Article Google Scholar Lamb, W. F., Creutzig, F., Callaghan, M. W. & Minx, J. C. Learning about urban climate solutions from case studies. Nat. Clim. Change 9, 279–287 (2019).Article Google Scholar Callaghan, M. et al. Machine-learning-based evidence and attribution mapping of 100,000 climate impact studies. Nat. Clim. Change 11, 966–972 (2021).Article Google Scholar Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. BERT: pre-training of deep bidirectional transformers for language understanding. In Proc. 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (eds Burstein, J. et al.) 4171–4186 (Association for Computational Linguistics, 2019).Fu, X., Wang, R. & Li, C. Can ChatGPT evaluate plans?. J. Am. Plan. Assoc. 1, 12 (2023).Google Scholar Fu, X. Natural language rocessing in urban planning: a research agenda. J. Plan. Lit. 39, 395–407 (2024).Nicola, J. ‘In awe’: scientists impressed by latest ChatGPT model o1. Nature 634, 275–276 (2024).IPCC. Climate Change 2022: Impacts, Adaptation, and Vulnerability (eds Pörtner, H.-O. et al.) (Cambridge Univ. Press, 2022).Climate Change and World Heritage (UNESCO WHC, 2023); https://whc.unesco.org/en/climatechange/Format for the Nomination of Properties for Inscription on the World Heritage List (UNESCO WHC, 2005); https://whc.unesco.org/document/191594State of Conservation Information System (UNESCO WHC, accessed 20 February 2025); https://whc.unesco.org/en/soc/Periodic Reporting (UNESCO WHC, accessed 20 February 2025); https://whc.unesco.org/en/periodicreporting/Climate Change Adaptation for Natural World Heritage Sites—A Practical Guide (UNESCO WHC, 2014); https://whc.unesco.org/document/129276The Future of our Pasts: Engaging Cultural Heritage in Climate Action (ICOMOS, 2019); https://civvih.icomos.org/wp-content/uploads/Future-of-Our-Pasts-Report-min.pdfWang, X. M., Li, H. M., Wang, Y. F. & Zhao, X. F. Assessing climate risk related to precipitation on cultural heritage at the provincial level in China. Sci. Total Environ. 835, 155489 (2022).Article CAS Google Scholar Banda, A. M., Banda, K., Sakala, E., Chomba, M. & Nyambe, I. A. Assessment of land use change in the wetland of Barotse Floodplain, Zambezi River Sub-Basin, Zambia. Nat. Hazards 115, 1193–1211 (2023).Article Google Scholar Vandemeulebroucke, I., Vanderschelden, B., Janssens, K., Caluwaerts, S. & Van Den Bossche, N. The impact of climate change on degradation in historical building envelopes: progress in research using hygrothermal models. J. Cult. Herit. 70, 345–363 (2024).Article Google Scholar Wartman, M. et al. Assessing blue carbon in mangrove ecosystems of Seychelles. J. Environ. Manag. 374, 123967 (2025).Article CAS Google Scholar Lelièvre, Y. et al. Taxonomic and functional diversity of subtidal benthic communities associated with hard substrates at Crozet archipelago (sub-Antarctic, Southern Ocean). Front. Mar. Sci. 10, 1291038 (2023).Liang, Q. F., Zhai, J. & Li, C. From separation to incorporation: development of a unifying framework that integrated bird habitats with public recreation spaces within the wetland park system. J. Clean. Prod. 430, 139647 (2023).Khalaf, R. W. Integrity: enabling a future-oriented approach to cultural heritage. Hist. Environ. Policy 13, 5–27 (2022).Article Google Scholar Richards, J., Brimblecombe, P. & Engelstaedter, S. Modelling temperature–precipitation pressures on African timber heritage. Int. J. Climatol. 43, 7447–7462 (2023).Article Google Scholar Boellinghaus, T., Wentland, E., Jassmann, R., Keller, R. & Wolfrum, A. Conservation of the shaft #1 headgear at the Tsumeb Mine, Namibia: corrosion protection. Stud. Conserv. 68, 228–242 (2023).Article Google Scholar Collareta, A. et al. Reviewing the palaeontological and palaeoenvironmental heritage of the Monti Pisani Massif (Italy): a compelling history of animals, plants and climates through three geological eras. Geosciences 13, 332 (2023).Cerrillo-Cuenca, E. et al. Emergent heritage: the digital conservation of archaeological sites in reservoirs and the case of the Dolmen de Guadalperal (Spain). Herit. Sci. 9, 114 (2021).Manunza, A. et al. Genetic diversity and historical demography of underutilised goat breeds in North-Western Europe. Sci. Rep. 13, 20728 (2023).Chen, Y., Dong, Q., Zhang, L., Shao, L. & Tang, Y. Using natural language processing to identify and evaluate heritage value for buffer zone delineation in Hengdaohezi Town. npj Herit. Sci. 13, 262 (2025).Article Google Scholar Chen, H., Zheng, Y., Xing, J. & Zhou, C. Urban planning priorities analysis based on urban heat risk supply–demand mismatch: a case study of Hangzhou, China. Sustain. Cities Soc. 130, 106502 (2025).Article Google Scholar Zhang, L., Tang, Y. & Chen, Y. Integrating historic preservation into a wider framework of urban development through comprehensive planning: an evaluation of local comprehensive plans in U.S. state capitals. Herit. Sci. 10, 192 (2022).Article Google Scholar Operational Guidelines for the Implementation of the World Heritage Convention (UNESCO WHC, 2005); https://whc.unesco.org/archive/opguide05-en.pdfFeng, C. et al. Sustainably developing global blue carbon for climate change mitigation and economic benefits through international cooperation. Nat. Commun. 14, 6144 (2023).Article CAS Google Scholar Johnathan, G., Esther, S. & Jonas, M. Author correction: National models of climate governance among major emitters. Nat. Clim. Change 13, 748 (2023).Callaghan, M. et al. Machine learning map of climate policy literature reveals disparities between scientific attention, policy density, and emissions. npj Clim. Action 4, 7 (2025).Article Google Scholar Kim, J. S. & Kim, S. K. Ageing population and green space dynamics for climate change adaptation in Southeast Asia. Nat. Clim. Change 14, 490–495 (2024).Article Google Scholar Chen, Z., Guo, W., Dai, L.-R., Ling, Z.-H. & Du, J. Neural text clustering with document-level attention based on dynamic soft labels. Proc. Interspeech 2019, 4225–4229 (2019).Fornaciari, T., Uma, A., Poesio, M. & Hovy, D. Hard and soft evaluation of NLP models with BOOtSTrap SAmpling—BooStSa. In Proc. 60th Annual Meeting of the Association for Computational Linguistics: System Demonstrations (eds Basile, V. et al.) 127–134 (Association for Computational Linguistics, 2022).Vaismoradi, M., Turunen, H. & Bondas, T. Content analysis and thematic analysis: implications for conducting a qualitative descriptive study. Nurs. Health Sci. 15, 398–405 (2013).Article Google Scholar Zhang, L., Tochen, R. M., Hibbard, M. & Tang, Z. H. The role of local leaders in environmental concerns in master plans: an empirical study of China’s eighty large municipalities. J. Plan. Educ. Res. 38, 301–313 (2018).Article Google Scholar Bai, N., Luo, R., Nourian, P. & Pereira Roders, A. WHOSe heritage: classification of UNESCO world heritage statements of ‘outstanding universal value’ with soft labels. In Findings of the Association for Computational Linguistics: EMNLP 2021 (eds Moens, M.-F. et al.) 366–384 (Association for Computational Linguistics, 2021).Huber, M., Zollner, D., Pecher, S. & Wolf, L. Mid-Term Programme Evaluation: World Heritage Leadership—Final Evaluation Report (IUCN, 2023); https://www.iucn.org/sites/default/files/2022-05/world-heritage-leadership-mid-term-evaluation-2020.pdfWannewitz, M. et al. Progress and gaps in climate change adaptation in coastal cities across the globe. Nat. Cities 1, 610–619 (2024).Article Google Scholar Deng, J. & Lin, Y. The benefits and challenges of ChatGPT: an overview. Front. Comput. Intel. Syst. 2, 81–83 (2023).Google Scholar Sandmann, S., Riepenhausen, S., Plagwitz, L. & Varghese, J. Systematic analysis of ChatGPT, Google search and Llama 2 for clinical decision support tasks. Nat. Commun. 15, 2050 (2024).Article CAS Google Scholar Niu, Y. M., Wu, J., Jiang, S. Y. & Jiang, Z. B. The Bullwhip Effect in Servitized Manufacturers (SSRN, 2024).World Heritage List (UNESCO WHC, 1978); https://whc.unesco.org/en/list/Peel, M.C., Finlayson, B.L. & McMahon, T.A. Updated world map of the Köppen–Geiger climate classification. Hydrol. Earth Syst. Sci. 11, 1633–1644 (2007).Advisory Bodies (UNESCO WHC, 1996); https://whc.unesco.org/en/advisorybodies/Woodruff, S. C. & Stults, M. Numerous strategies but limited implementation guidance in US local adaptation plans. Nat. Clim. Change 6, 796–802 (2016).Article Google Scholar Niraula, D. et al. Intricacies of human–AI interaction in dynamic decision-making for precision oncology. Nat. Commun. 16, 1138 (2025).Article CAS Google Scholar Chicco, D. & Jurman, G. The Matthews correlation coefficient (MCC) should replace the ROC AUC as the standard metric for assessing binary classification. BioData Min. 16, 4 (2023).Article Google Scholar Yang, C. et al. World Heritage documents reveal persistent gaps between climate awareness and local action. figshare https://doi.org/10.6084/m9.figshare.28823297 (2025).Download referencesAcknowledgementsThis work was supported by the National Natural Science Foundation of China (grant no. 52394222, C.S.), the National Key R&D Program-Strategic Scientific and Technological Innovation Cooperation (grant no. 2022YFE0208600, Q.D.), the Heilongjiang Touyan Innovation Team Program (C.S.) and the China Postdoctoral Science Foundation (grant no. 2024M754190, L.Z.). We are also grateful to X. Hou and Y. Liu for providing the artwork and to X. Xiao and G. Zhang for contributing the photography.Author informationAuthors and AffiliationsSchool of Architecture and Design, Harbin Institute of Technology, Harbin, ChinaYang Chen, Dayang Wang, Luchen Zhang, Cheng Sun & Qi DongKey Laboratory of Cold Region Urban and Rural Human Settlement Environment Science and Technology, Ministry of Industry and Information Technology, Harbin, ChinaYang Chen, Dayang Wang, Luchen Zhang, Cheng Sun & Qi DongCollege of Landscape Architecture, Nanjing Forestry University, Nanjing, ChinaChongxiao WangAuthorsYang ChenView author publicationsSearch author on:PubMed Google ScholarDayang WangView author publicationsSearch author on:PubMed Google ScholarLuchen ZhangView author publicationsSearch author on:PubMed Google ScholarChongxiao WangView author publicationsSearch author on:PubMed Google ScholarCheng SunView author publicationsSearch author on:PubMed Google ScholarQi DongView author publicationsSearch author on:PubMed Google ScholarContributionsY.C. together with D.W., C.S. and Q.D. designed the work. Y.C. led the paper, wrote the draft and, together with D.W. and C.W., analysed the data. L.Z., Q.D. and C.S. substantively revised the draft. All other authors acquired and interpreted the data and edited the drafts.Corresponding authorsCorrespondence to Cheng Sun or Qi Dong.Ethics declarationsCompeting interestsThe authors declare no competing interests.Peer reviewPeer review informationNature Climate Change thanks Chris Ballard, Attila Buzási, Travis Coan and John Hughes for their contribution to the peer review of this work.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Extended dataExtended Data Fig. 1 Distribution of WHS and their climate awareness.a Geographic distribution across different themes of WHS (n = 562, including repeated values). Among the 535 WHS with valid climate awareness scores, 27 sites belong to multiple themes. b Distribution of climate awareness scores across different climate zone (n = 535). The different colors in the figure represent various regions, with the actual values corresponding to the average climate awareness scores for each region.Supplementary informationSupplementary InformationSupplementary Figs. 1 and 2, Notes 1–8 and Tables 1–11.Rights and permissionsSpringer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.Reprints and permissionsAbout this article