Dissecting the vital role of dietary changes in food security assessment under climate change

Impacts of the climate crisis on food security, economic growth, and living conditions comprise one of the greatest wicked problems facing humanity in the 21st century. Both climatic trends and weather extremes negatively impact the sustainability and consistency of plant-based food production. The United Nations defines food security as the state in which people, at all times, have physical, social, and economic access to sufficient, safe, and nutritious food that meets their dietary needs. Despite calls for inclusion of social and economic assessments, primary food metrics informing food security demographically persist in biophysical dimensions, such as requiring sufficient supply and nutrient intake. Even though food security is an emergent property of the balance between food production (supply side) and food needs (demand side), most studies assessing climate change impacts have focused on supply, in part because demand-side measures are more difficult to quantify reliably. Climate change and weather variability strongly affect seasonal and annual food supply, alongside land use management, genotypes, environment, and their interactions. On the demand side, variation in food demand (per capita demand for different food groups) is governed by population growth and trends in dietary choices, among other factors. Previous studies have shown that climate change and variability negatively impact food supply; historically, for grains, weather impacts were outweighed by productivity gains from technology, practices, and machinery. However, such observations warrant revisiting because anthropogenic dietary preferences were not well accounted for. Dietary patterns—the types and quantities of food consumed—vary among countries and among geographic and socioeconomic groups. Globalization and expanding middle-class affluence have increased consumption of carbohydrates and animal-based proteins, with potential short-term health benefits. In the long term, shifting toward plant-based diets may reduce detrimental health outcomes, such as overweight and obesity.

Study scope and data: The analysis integrates supply- and demand-side drivers of food security across 155 countries from 1990–2018, and projects outcomes for 2031–2060. Ten food categories were analyzed: fruits, non-starchy vegetables, starchy roots, nuts/seeds, pulses, grains/cereals, meat, fish, eggs, and dairy. Supply-side data came from FAOSTAT Food Balances, with trade indicators from the World Integrated Trade Solution (WITS), agricultural water input and land data from AQUASTAT, and historical climate (daily to annual aggregates of mean temperature and precipitation) from NASA POWER. Future climate projections for 2031–2060 used six CMIP6 models under SSP2-4.5 and SSP5-8.5. Demand-side dietary intake data (grams per person per day by sex and 22 age groups) were sourced from the Global Dietary Database 2018 (GDD2018). Historical and projected population data were from the United Nations. Cereal-equivalent (CE) metric: To compare diverse foods on a resource-use basis, the study converted both supply and demand into cereal-equivalents (CE), reflecting land and feed resource requirements (following Rask et al.). Weighted mean CE coefficients for meat subcategories (beef, mutton/goat, pig, poultry) were computed for 1990 and 2018 using FAOSTAT production data. Historical demand estimation: National CE food demand was computed by multiplying per-capita dietary intake by CE coefficients, days per year, and population for each country. To isolate the effect of dietary change from population growth, changes in CE demand between 1990 and 2018 were calculated holding population constant. Historical supply estimation and climate attribution: National CE food supply (sum over categories of supply times CE) was log-transformed and modeled with a country-level panel regression to estimate effects of interannual climate variability and trade on CE supply. The model included quadratic terms for annual mean temperature and precipitation; interactions with agricultural water input to capture compensation effects; export and import product shares; and country/year fixed effects, weighted by agricultural land area. Model performance was assessed with 1000-iteration bootstrap cross-validation (80% train, 20% test). Climate change effects on CE supply were computed by comparing predicted CE supply using observed historical climate versus a counterfactual with non-climate variables fixed to 1990. Future demand scenarios (2031–2060): To remove population growth effects, the study used 2018 dietary patterns across 22 age groups with projected populations to estimate baseline demand. Eight dietary change scenarios were analyzed relative to the 2018 baseline diet: Business-as-Usual (BAU), Chinese Food Guide Pyramid, Mediterranean, Global Dietary Guideline, National Recommended Diet by region, National Recommended Diet by income, Flexitarian, and EAT-Lancet. Reference intakes for the ten categories were compiled from published guidelines and scaled across age groups. Future supply projections (2031–2060): The historical climate–supply panel model was driven by CMIP6 projections under SSP2-4.5 and SSP5-8.5 to estimate changes in global CE supply relative to 2018. Integrated food security assessment: Six scenarios describe combinations of increases/decreases in CE supply (from climate) and CE demand (from dietary change), classifying national outcomes as positive or negative for food security. Historical (1990–2018) and future (2031–2060) country outcomes were mapped by region and income group.

Historical period (1990–2018): - Climate change detrimentally impacted CE food supply in 92% of countries. - Dietary changes increased country-level CE food demand with a global median of 18%, largest in Asia (median 19%) and smallest in North America (3%) and Europe (4%). Changes were smaller in high-income countries than in lower-income groups. - Integrating climate-driven supply change with diet-driven demand change, food security was negatively impacted in 70% of countries. Situation I (negative climate impact on supply compounded by increasing demand) occurred in almost 60% of countries, affecting about 4.8 billion people (59% of global population), including ~1.4 billion in China, ~0.3 billion in Indonesia, and ~0.1 billion in Mexico. Situation V (negative supply, decreasing demand) occurred in Kenya, Latvia, Madagascar, Malaysia, and Uruguay (~0.1 billion people). Situation IV (positive supply outstripping increasing demand) occurred in two countries (Russia and Tianan), affecting ~0.2 billion people (~3% of global population). Situation VI (negative supply offset by decreasing demand) occurred in 28% of countries, affecting 36% of the global population. Food security was positively affected in the remaining 40% of countries. Diet trends (1990–2018): - Per-capita demand increased for all ten major categories globally; nuts rose most (+176%), driven by +447% in Latin America & the Caribbean. In Asia, per-capita demand rose for vegetables (+93%), meat (+83%), fish (+61%), eggs (+133%), and milk (+56%). Starchy roots demand per capita declined in Europe (−30%) and North America (−12%). Future projections (2031–2060): - Under BAU diet, global CE demand increases by about 74% relative to 2018 (ignoring population dynamics); other diet scenarios raise demand by 1–41%, with Flexitarian <5%. EAT-Lancet reduces total demand by about 15% versus 2018. - Average model projections indicate global CE supply reductions of ~3% under SSP2-4.5 and ~4% under SSP5-8.5 relative to 2018 (with multi-model ranges reported up to ~8% under SSP5-8.5). - Under SSP2-4.5, negative outcomes dominate with BAU diets (Situation I, II, V in 82%, 2%, and 6% of countries, respectively). Shifting diets reduces the share of countries facing negative outcomes; under sustainable diets, negative effects occur in less than 50% of countries (31% for Flexitarian; 38% for EAT-Lancet). Sustainable diets (Flexitarian, EAT-Lancet) could mitigate food security threats by about 42% and 35%, respectively.

Considering the impact of climate change on food production alone oversimplifies food security assessment. Because food security arises from the balance of supply and demand, changes in dietary preferences and resultant per-capita demand can offset or exacerbate climate-driven supply shocks. Historically (1990–2018), climate change reduced CE food supply in most countries (92%). Yet, in 28% of countries (36% of the global population), decreases in demand (or slower growth) offset supply losses such that food security was not threatened. Conversely, much of the global population did not benefit from positive supply changes because dietary shifts increased demand more rapidly, underscoring the decisive role of diets in determining national food security under climate change. Looking ahead, more frequent and intense extremes will likely aggravate supply volatility, with high- and low-emission futures potentially pushing an additional substantial share of the population into hunger during rare but severe events. Projections indicate modest average global CE supply reductions by 2031–2060 (about 3–4%) due to climate change, but population growth alone could increase demand by roughly 21%, and BAU dietary change could raise demand by about 74% versus 2018, compounding pressures. Adopting nutritionally adequate, lower-impact dietary patterns—especially Flexitarian and EAT-Lancet diets dominated by plant-based foods—substantially reduces demand growth and the share of countries and people exposed to negative food security outcomes. Thus, policies targeting both production resilience and demand-side dietary transitions are critical to improve food security under climate change.

This study provides a global, integrated assessment of food security that explicitly combines climate-driven changes in food supply with diet-driven changes in food demand using a resource-based cereal-equivalent framework. It shows that climate change has already reduced food supply in most countries and that dietary shifts have been a decisive determinant of whether food security deteriorates or improves. Future projections indicate that, under business-as-usual diets, demand growth will far outpace projected supply declines, worsening food security risks; by contrast, sustainable, plant-forward dietary patterns (e.g., Flexitarian and EAT-Lancet) can substantially mitigate these risks. Policy and research should jointly target: (1) accelerating resilient, low-emission production systems; (2) enabling and incentivizing healthy, sustainable dietary shifts; and (3) integrating demand management (including reducing food loss and waste) into food security planning. Future work should incorporate price and market feedbacks, explicitly consider food loss and waste, improve consistency between intake and production/economic data, and refine behavioral adherence and regional heterogeneity in diet transitions.

- Food loss and waste were not modeled; the study notes food waste accounts for approximately 29% of the total and substantially affects food demand. Addressing loss and waste could improve food security while reducing GHG emissions and resource pressures. - Prices and market dynamics (including trade responses and affordability) were not explicitly represented; changes in the supply–demand balance can influence prices, trade, and access, thereby affecting food security. - Supply and demand data come from different measurement systems: GDD reflects intake, whereas FAOSTAT reflects economic output. To bridge this, the study uses cereal-equivalent (CE) rather than calories or monetary values, but mismatches may remain. - The panel model evaluates relative effects of climate change on supply and dietary change on demand, rather than estimating absolute levels; unmodeled factors or structural changes could affect estimates. - National characteristics (e.g., climate-driven preferences) and adherence to dietary scenarios were simplified; actual transitions may be gradual and heterogeneous.