Trade-offs between water use and food production in Egyptian agriculture under an uncertain future

Abstract

Egypt’s agricultural system is facing increasing pressure from limited water resources – sourced predominantly from the Nile River – rising demand, and climate change. Higher temperatures and evapotranspiration will intensify water stress, though CO₂ fertilisation may partially offset impacts by boosting yields and reducing water use. Meanwhile, upstream damming and extreme weather further threaten water availability and crop productivity. These challenges demand adaptive, integrated land and water management to safeguard food security for a rapidly growing population.

Despite these constraints, Egypt maintains agricultural potential due to its climate suitability for year-round crop production, extensive irrigation infrastructure, and underutilised groundwater resources. Policy efforts have focused on irrigation modernisation, improved seed varieties, and large-scale agricultural expansion. Agriculture accounts for 17% of employment and 13% of the economy, while most of the farmers are smallholders. Additionally, food subsidies have been used to help maintain food security, although they have also deepened reliance on imports, increasing vulnerability to global market fluctuations.

The interconnected nature of these challenges necessitates analytical tools that are capable of integrating biophysical and socioeconomic dimensions. Existing biophysical models effectively simulate climate impacts but treat farmer decision-making as static, whilst agent based models capture human responses but simplify biophysical dynamics. This thesis addresses these limitations by analysing current characteristics and recent trends in farming households, and then developing a coupled agent-based land-use change and ecohydrological model that integrates farmer decision-making with biophysical processes to evaluate climate change and socioeconomic impacts on agricultural production and resources.

The research begins with a typology analysis of nationally representative labour market surveys from 2012 and 2018, revealing four distinct agricultural household clusters: specialised farming households focused on intensive production; village households reliant on subsistence agriculture; diversified income households combining agricultural and nonfarm activities; and landlord households with minimal direct involvement. The analysis iv demonstrates trends towards income diversification away from agricultural income as well as agricultural abandonment.

Building on these findings, this study develops a coupled modelling framework that integrates the agent-based land-use change model CRAFTY (Competition for Resources between Agent Functional Types) with the ecohydrological model SWAT+ (Soil and Water Assessment Tool). Both models operate on the same 900 m resolution grid: SWAT+ simulates daily crop growth and irrigation demand, while CRAFTY allocates land use annually by modelling competition between 56 agent functional types based on socioeconomic resources and agent characteristics.

Agents represent distinct cropping strategies and production systems, broadly classified as either commercial or subsistence types. Commercial agents reflect capital-intensive farmers who use deficit irrigation and depend on higher levels of financial and manufactured resources. In contrast, subsistence agents represent smallholders who rely on traditional flood irrigation and are more likely to persist under lower profitability conditions. These categories capture key differences in access to resources, irrigation practices, and labour dynamics, enabling the model to simulate transitions between farming systems across a range of future scenarios.

To evaluate future outcomes, the model applies a set of integrated socioeconomic and climate scenarios combining assumptions about diets, trade, and farmer behaviour, based on the Shared Socioeconomic Pathways (SSPs) and Representative Concentration Pathways (RCPs). Under high-emission scenarios, yields for key crops increase by up to 25%, with water use efficiency improving by 69%, while low-emission scenarios show modest gains (4%). Production shortfalls vary widely: the inequality pathway results in the smallest gap (22%), while the sustainability scenario still leaves a 52% shortfall despite improved efficiency. No scenario fully meets demand for domestic production, with shortfalls shaped by population growth, dietary patterns, and trade dynamics. Notably, the regional rivalry scenario, which targets agricultural self-sufficiency but also drives high exports, faces the largest gap, with unmet demand for domestic production reaching 78%.

This research shows that neither agricultural expansion nor dietary change alone can close the food production gap under Egypt’s limited water and land resources. The coupled model v highlights the trade-offs that emerge when policies focus on single objectives, underscoring the need for integrated strategies that account for socioeconomic diversity, spatial variability, and environmental constraints. By capturing how different farming households respond to change, the model supports more adaptive and targeted planning, emphasising the combined role of population trends, dietary shifts, trade policy, strategic agricultural expansion, and irrigation efficiency in ensuring future food security.