Multi-decadal climate services help farmers assess and manage future risks

The paper examines how multi-decadal climate projections (20+ years) delivered through climate services can inform farmers’ risk management decisions. Despite growth in climate services since the Global Framework for Climate Services (GFCS) was established, benefits to end users remain uncertain due to inequitable access, insufficient user–provider interaction, a product-centric focus, assumed rather than elicited demand, and evaluation centered on economics over real-world outcomes. Agriculture is a priority sector facing substantial climate change impacts, yet most farmer-facing services emphasize short-term weather (1–14 days) or seasonal forecasts (3–6 months). There is limited evidence on how multi-decadal projections are used to manage long-term risk, and climate change often feels psychologically distant to farmers. The study poses the research question: how would multi-decadal projections affect farmers’ risk management decisions pertaining to future climate risks? Using Australia’s My Climate View (MyCV)—which provides localized and commodity-specific projections for the 2030s, 2050s and 2070s—the authors conducted 24 qualitative semi-structured interviews to explore changes in risk perception and identify management responses.

Prior work has highlighted challenges for climate services: inequitable access, limited co-production, a focus on products over processes and outcomes, and evaluations not grounded in real-world impacts. Studies show farmers commonly use short-term and seasonal information, while evidence on multi-decadal projections in agriculture is scarce. Psychological distance hampers climate action; localizing and contextualizing information can reduce this distance. Existing literature also documents skepticism toward climate projections linked to experiences with short-term forecast accuracy and general climate skepticism. There is emerging work on co-design/co-development of services, the role of trusted intermediaries (e.g., extension and advisers), and the need to translate useful information into usable decisions. This study addresses gaps by empirically examining how localized, commodity-specific, multi-decadal projections influence farmers’ risk perceptions and management planning.

The study used a participatory qualitative risk analysis framework grounded in action research principles to engage farmers in identifying, assessing, and managing future climate risks. The framework involved four steps: (1) identify short-term risks currently experienced; (2) identify long-term risks pre-MyCV, introduce MyCV, and re-identify long-term risks post-MyCV; (3) assess impacts of post-MyCV long-term risks on farm businesses (hazards, severity, consequences); and (4) identify risk management strategies to address those risks. MyCV is a national-scale tool providing local, agriculturally relevant historical climate data and future projections for the 2030s, 2050s, and 2070s. Participants: 24 Australian farmers across multiple commodities and climatic zones were recruited via an outreach agency, professional contacts, and snowball sampling. Sampling continued until thematic saturation. Interviews were conducted online, semi-structured, averaged ~1 hour, recorded with consent, professionally transcribed, de-identified, and analyzed. Analysis: Thematic analysis was conducted using R (RQDA package). Transcripts were iteratively coded into themes for risk identification, assessment, and management. Multiple researchers reviewed codes to resolve discrepancies, with five coding rounds reducing subthemes from 33 to 18. Visualization relied on counts of theme mentions (not for prominence claims). Codes and visualization scripts are publicly available via CSIRO’s Data Access Portal. Ethics approval was obtained from CSIRO’s Human Research Ethics Committee (ID: 001/21).

• Sample and intervention: 24 farmers engaged with My Climate View (MyCV), a localized, commodity-specific multi-decadal projection tool (2030s, 2050s, 2070s).
• Shift in perceived risk complexity: Before MyCV, long-term risks were often ambiguous or unknown; after MyCV, identified future risks narrowed to known and uncertain categories. This indicates reduced perceived complexity and potentially reduced psychological distance.
• Risk assessment themes (post-MyCV): five main themes emerged—(1) confidence in data (mixed; trust in providers like the Bureau of Meteorology and alignment with personal observations increased confidence), (2) impacts on commodities (varied by location, scale of change, and risk tolerance), (3) existing capacity (some felt adequately prepared; others reported constraints, e.g., irrigation access), (4) location suitability (evaluating current and alternative regions for commodities), and (5) water requirements (concerns over irrigation sources, allocations, dam capacity).
• Risk management themes (post-MyCV): farmers identified strategic adaptations—(1) change farming practices (e.g., irrigation efficiency, canopy management, soil moisture retention, adjusting livestock numbers), (2) inform business investments (board-level planning, reconsidering future investments), (3) add machinery and infrastructure (e.g., overhead sprinklers, shade structures, new dams), (4) change commodities (consider switching crops where projections suggest declining suitability), and (5) continue existing practices (where current strategies suffice or confidence in projections is low).
• Data confidence as a driver: Confidence in projections strongly influenced the extent to which farmers translated information into identified risks and actions; trusted service providers and perceived alignment with on-farm experience increased confidence.

The study demonstrates that localized, commodity-specific multi-decadal projections delivered via MyCV can help farmers interpret future climate risks more concretely, reducing ambiguity and psychological distance and enabling strategic planning rather than only operational responses. Confidence in the data emerged as pivotal; skepticism stemming from experiences with short-term forecasts and general climate skepticism can impede use of projections, but trust in credible providers and alignment with observed trends help. Despite varying confidence levels, all participants engaged in risk assessment and management discussions, suggesting that interactive, contextualized engagement can foster meaningful use. The findings contribute empirical evidence for operationalizing multi-decadal projections in farm decisions and underscore the importance of co-produced, trusted, iterative climate services to translate awareness into action.

Multi-decadal climate services like My Climate View can support farmers in identifying and assessing long-term climate risks by reducing perceived complexity and psychological distance, thereby informing strategic adaptation planning. However, to convert awareness into concrete actions, services must foster confidence through trusted providers, contextualized information, and recurring, interactive engagement. The work contributes empirical insights into how long-term projections can be operationalized in agricultural decision-making and highlights the need for ongoing co-production and engagement to sustain impact. Future research should follow up on implementation of identified actions and use larger, possibly mixed-methods designs to statistically validate patterns across commodities, regions, and business types.

• Attribution: While participants’ perceptions shifted post-MyCV, the study cannot conclusively isolate the interface’s specific contribution versus the facilitated discussion process.
• Implementation follow-up: The study did not return to assess whether identified risk management actions were implemented.
• Scope and generalizability: Qualitative design with 24 participants aimed for depth and saturation, not representativeness; statistical validation of cross-business, climatic zone, and commodity patterns is warranted.
• Confidence dynamics: Findings on confidence in projections reflect perceptions at the time; longitudinal changes were not assessed.