Indigenous Peoples and local communities report ongoing and widespread climate change impacts on local social-ecological systems

The study addresses how climate change impacts are governed by local biophysical and sociocultural conditions, which complicates efforts to produce a comprehensive global understanding that includes local experiences. Instrumental measurements and models provide essential global-scale evidence but often lack resolution and relevance for areas with sparse stations, especially territories inhabited by Indigenous Peoples and local communities, and may overlook locally salient variables. Meanwhile, place-based studies centering Indigenous and local knowledge richly detail localized impacts but are difficult to compare across sites due to heterogeneous methods. To bridge these gaps, the authors conducted a globally coordinated, cross-culturally comparable study documenting locally experienced climate change indicators and impacts across 48 sites and assessed how reports vary by climate zone and livelihood, aiming to integrate Indigenous and local knowledge into global climate research and policy.

The paper synthesizes two major streams of prior work. First, research based on instrumental measurements and climate modeling clarifies global patterns of change (e.g., heterogeneous temperature and precipitation shifts) but loses accuracy when downscaled in data-sparse regions and tends to focus on a limited set of variables defined by the scientific community, potentially neglecting locally critical phenomena (e.g., non-major crops) and the complex ways people experience environmental change. Second, an expanding body of work documents Indigenous and local knowledge of climate change, providing nuanced, site-specific insights into observed changes, drivers, and responses, yet lacking standardization needed for global comparison and synthesis. The IPCC has called for strengthening evidence on local knowledge contributions, and recent scholarship emphasizes integrating diverse knowledge systems, recognizing the intertwined climate and biodiversity crises, and highlighting the importance of locally perceived impacts for adaptation and risk planning.

Design and ethics: The research was conducted within a network designed to standardize the collection and coding of locally relevant yet cross-culturally comparable data on climate change indicators and impacts. The protocol was approved by the Ethics Committee of Universitat Autònoma de Barcelona (CEEAH 4781). Permissions from local authorities and Free Prior Informed Consent were obtained at each site, with national ethics approvals where required. The study protocol and training materials are publicly available. Site selection and participants: Data were collected during 2019–2022 in 48 field sites worldwide characterized by predominantly nature-dependent livelihoods. A site comprised a group of villages or households with relatively homogeneous environmental and sociocultural features. Sites spanned four Köppen-Geiger climate zones (tropical, arid, temperate, snow/polar) and a range of livelihoods (agropastoralism, agriculture, fishing, pastoralism, other nature-dependent livelihoods). Data collection: In each site, researchers conducted semi-structured interviews with 15–25 key participants focusing on environmental changes observed over decades, eliciting direction and perceived drivers, starting from open prompts and probing specifically for temperature, precipitation, wind, seasons, soils, water, and wild/domestic biota. Narratives often included multiple drivers; only observations at least partly attributed to changes in temperature, precipitation, air masses, or seasons were retained. Indicator development and agreement: Observations were consolidated into local indicators of climate change and its impacts (LICCIs) through a hierarchical system: (1) system (atmospheric, physical, life), (2) subsystem (e.g., precipitation, temperature), (3) impacted element (e.g., mean temperature, seasonal temperature), and (4) indicator. Verbatim observations of the same change were grouped into the same indicator. Each site held 3–5 focus groups to resolve contradictions and validate indicators; due to COVID-19, some sites used additional interviews. Indicators were considered "site-agreed" if free of contradictions in interviews or explicitly agreed upon in focus groups. Contextual classifications: Sites were assigned to climate zones using their location and the Köppen-Geiger classification. Site-level average decadal temperature change (1901–2018) was derived from IPCC Atlas estimates based on CRU-TS data and categorized as high positive (>0.3 °C/decade), medium positive (0.15–0.3), mild positive (0–0.15), or negative (<0). The main livelihood activity per site (agropastoralism, agriculture, fishing, pastoralism, other nature-dependent livelihoods) was identified from interviews, recognizing that households often employ mixed strategies. Statistical analysis: Site was the unit of analysis. Generalized linear models (Poisson) compared the site average number of LICCIs within each subsystem across (i) climate zones (reference: snow/polar), (ii) decadal temperature change (reference: high positive), and (iii) main livelihood (reference: agriculture), followed by Tukey post-hoc tests where applicable. A multivariate analysis assessed how site LICCI composition (number and type of impacts in the physical and life systems) varied with climate zone, decadal temperature change, and livelihood using a Bray-Curtis dissimilarity matrix and non-parametric MANOVA (NPMANOVA) implemented in vegan (R 4.2.1).

- Coverage: 48 sites across all inhabited continents and four Köppen-Geiger climate zones (tropical n=16; arid n=12; temperate n=13; snow/polar n=7). Most sites exhibited increasing decadal average temperature (20 sites at medium increase: 0.15–0.3 °C/decade). Predominant livelihoods: agropastoralism (n=16), agriculture (n=13), fishing (n=12), pastoralism (n=4), other nature-dependent livelihoods (n=3). - Volume of evidence: 1,661 site-agreed observations consolidated into 369 indicators (94 indicators of climate change in the atmospheric system; 275 indicators of impacts in physical and life systems). - Distribution by system/subsystem: 46.4% of observations concerned the atmospheric system, especially precipitation (20.4%) and temperature (13.0). Physical system impacts accounted for 19.6% (notably freshwater 9.1%). Life system impacts constituted 33.9%, with plant cultivation (9.9%) and terrestrial flora (7.9%) most frequent. - Most-cited indicators: Changes in mean temperature (45 citations; 2.7%) and in mean seasonal temperature (34; 2.0%). Most-cited impacts: Changes in crop productivity (52; 3.1%) and changes in abundance/density of wild plant or fungi species (33; 2.0%). - Variation across climate zones: Tropical sites had higher average numbers of freshwater-related impacts than snow/polar sites. Arid and snow/polar sites reported significantly more impacts in pastures and grasslands than tropical sites and fewer impacts in plant cultivation than temperate sites. Snow/polar and temperate sites reported more impacts in ice and snow than arid/tropical (where rare) and more impacts on terrestrial fauna than tropical sites. - Variation with decadal temperature change: Sites with high decadal warming (>0.3 °C/decade) reported higher average impacts in pastures and grasslands than sites with medium or mild warming, and higher land cover change/land degradation than sites with medium warming. - Variation across livelihoods: Agricultural sites reported higher averages of precipitation indicators than sites with other NRD livelihoods and higher temperature indicators than pastoralist sites; they also reported more impacts on freshwater and soils than agropastoralist sites. Fishing sites reported more air mass indicators than agriculture, agropastoralism, and pastoralism, more oceans and seas impacts than agropastoralism and other NRD, and more marine ecosystem impacts than agriculture. Pastoralist sites reported more impacts in pastures/grasslands and land cover change/degradation than all other livelihoods. Sites with other NRD livelihoods tended to report more ice/snow impacts (not statistically significant; driven largely by Inuit reports). - Multivariate composition: LICCI composition was significantly associated with climate zone (NPMANOVA F3,47=1.8, p<0.01) and main livelihood (F4,47=1.2, p=0.04), but not with decadal temperature change, indicating that both ecological context and resource use shape the profile of reported impacts.

The study provides strong, cross-site quantitative evidence that Indigenous Peoples and local communities observe ongoing, tangible, and widespread climate change indicators and impacts. Despite site-specific nuances, consistent patterns emerge, particularly the preponderance of precipitation-related observations and cascading freshwater, soil, and biotic impacts, reflecting the centrality of water cycle changes in local social-ecological systems. Importantly, the composition and frequency of reported indicators and impacts vary systematically with climate zone and predominant livelihood activities, supporting the hypothesis that the way people interact with their environments (e.g., agriculture, fishing, pastoralism) is a key predictor of observed changes. The authors note critical caveats: attribution of observed impacts solely to climate change is difficult due to multiple, interacting drivers (e.g., land-use change, extraction), and the site sample is relatively small and geographically biased with overlaps between climate zones and livelihoods. Furthermore, translating localized, culturally grounded knowledge into standardized scientific categories risks overgeneralization and misinterpretation. Nevertheless, the standardized protocol and cross-cultural comparability enable identification of global patterns while preserving locally meaningful insights through the linked qualitative work. Implications include the need to integrate Indigenous and local knowledge into adaptation planning and risk analyses, tailored to both climatic and livelihood contexts. The methodology offers a pathway to systematically identify local economic and non-economic loss and damage, potentially supporting just compensation mechanisms. Protecting and mobilizing Indigenous and local knowledge systems remains essential for effective adaptation and mitigation, particularly given accelerating disruptions to these knowledge bases.

This globally coordinated study demonstrates that locally experienced climate change indicators and impacts among Indigenous Peoples and local communities are pervasive and diverse, with systematic differences linked to climate zones and predominant livelihoods. By standardizing the documentation of 369 indicators across 48 sites and coupling quantitative analyses with rich qualitative insights, the work advances integration of Indigenous and local knowledge into global climate research and policy. Policy recommendations include: (1) incorporating local knowledge to design context-relevant adaptation strategies; (2) applying the protocol in tools (e.g., OpenTEK, Oblo) to assess economic and non-economic loss and damage and inform compensatory measures; and (3) supporting the continuity and mobilization of Indigenous and local knowledge systems for adaptation and mitigation. Future research should expand geographic coverage, deepen mixed-methods integration for attribution and causal inference, and enhance multi-scalar coordination to translate local observations into actionable national and global policies.

- Attribution: Reported impacts cannot always be solely attributed to climate change due to complex, interacting drivers (e.g., land-use change, overextraction), though such drivers often amplify climate-related vulnerabilities. - Sampling and generalizability: The number of sites is limited, geographically biased, and shows overlap between climate zones and livelihoods, potentially affecting balance and external validity. - Epistemic translation: Standardizing locally meaningful observations into scientific categories risks overgeneralization and misinterpretation, given ontological and cultural differences in how changes are perceived and valued. - Data constraints: Some sites lacked focus groups due to COVID-19, relying on additional interviews to resolve contradictions, which may influence consensus assessment. - Downscaling context: While decadal temperature change categories were assigned from gridded datasets, uncertainties remain in data-sparse regions, potentially affecting exposure classifications.