What is the added value of decadal predictions compared to traditional long-term climate projections when it comes to forecasting wintertime precipitation in the Euro-Mediterranean region? A new paper, conducted by CMCC researchers, reveals the benefits of this approach for sectors like water resource management, agriculture, and infrastructure planning.

The Mediterranean region is highly vulnerable to climate change, with rising temperatures and decreasing rainfall posing risks to agriculture, water management, and disaster preparedness. In this context, accurately predicting climate trends 5 to 10 years in advance is essential so that decision-makers and stakeholders can adequately prepare.

A new study, conducted entirely by CMCC researchers including lead author Dario Nicolì, reveals how decadal climate predictions (therefore focusing on the next decade) provide a more reliable and accurate tool for forecasting winter rainfall in Southern Europe and North Africa compared to traditional long-term climate predictions.

Decadal predictions also allow for enhancing regional climate assessments through dynamical and statistical downscaling, as well as machine learning and artificial intelligence methods, which could further improve climate risk evaluation, inform policy decisions, and support the development of targeted adaptation and mitigation strategies. The promising results of this study therefore emphasize the need to integrate decadal predictions into these efforts.

“By incorporating real-time observations and accounting for natural climate variability, our decadal forecasts offer more reliable climate information for decision-making over the next 5 to 10 years on adaptation strategies, infrastructure planning, and climate services in Mediterranean countries,” says Nicolì. “Additionally, our hybrid model further refines these predictions, improving winter rainfall forecasts and providing essential insights for sectors like water resource management, agriculture, and infrastructure planning in the region.”

Figure 2. Anomaly correlation coefficient (ACC) for the multi-model Init mean (a), (b) and the difference with the respective ACC for the NoInit (c), (d) for precipitation field, assessed for forecast years 1–5 (a), (c) and forecast years 1–10 (b), (d) for the DJFM season. The stippling denotes points where the ACC values are found to be statistically significant. Source: Nicolì et al 2025

The study compares an ensemble of eight decadal–prediction systems with the corresponding uninitialized long-term climate projections. Key findings include:

• Superior performance: Decadal predictions consistently outperform long-term projections for both 1–5 and 1–10-year forecasts in the Euro-Mediterranean region (figure 2c and d), demonstrating higher predictive skill, particularly in winter precipitation across Southern Europe and North-Western coast of Africa.
• Better representation of ocean and atmospheric patterns: Decadal predictions capture critical climate factors, such as the North Atlantic Oscillation (NAO) and subpolar North Atlantic sea surface temperatures (SPNA-SST), which influence downstream precipitation over the Euro–mediterranean area.
• Hybrid model innovation: The study introduces a hybrid model combining dynamical and statistical approaches, improving forecast accuracy for precipitation across the Euro-Mediterranean area based on the aforementioned indices.

The study demonstrates that operational decadal predictions, which incorporate internal variability through yearly initialization, are more suitable than climate projections for forecasting Euro-Mediterranean precipitation on a decadal timescale.

“The potential of decadal predictions adds value by supporting climate projections – such as constraining or subsampling climate projections – and offering decision-makers a more accurate perspective on both near-term and long-term climate evolution,” says Nicolì.

More information:

Dario Nicolì, Silvio Gualdi, Panos J Athanasiadis; Decadal predictions outperform climate projections in forecasting Mediterranean wintertime precipitation, 2025,  Environ. Res. Lett. 20 034034 DOI 10.1088/1748-9326/adb59e 

CMCC is actively involved in several projects in this field, including the ASPECT project and the CMIP6 Decadal Climate Prediction Project (DCPP).