- Summary 1. Functional diversity (FD) has become a principal concept for revealing mechanisms driving community assembly and ecosystem function. Multiple assembly processes, including abiotic filtering, competition and multi-trophic relationships, operate simultaneously to structure FD. In water-limited plant communities, FD is likely to reflect trade-offs between drought resistance vs. disturbance resistance and competitive ability. 2. We propose a mathematical mechanistic model for understanding the organization and function of water-limited plant communities. The approach captures the interplay between abiotic filtering, below- and above-ground competition and disturbance. We exploit this powerful model to uncover mechanisms underlying changes in functional diversity along stress gradients. 3. Our approach links biomass production and FD to environmental conditions through plant resource capture ability. Functional groups are defined along a single trade-off axis according to investment in capturing light (shoot) vs. water (root). Species growth rate is determined dynamically by the species traits, water availability and grazing stress. We derive biomass production, functional diversity and composition along precipitation and grazing gradients. 4. Model’s results revealed several regimes structuring FD along the precipitation gradient: ‘Struggle for water’ at low precipitation, ‘competition for water’ at intermediate precipitation and ‘competition for light’ at high precipitation. 5. We observed a shift in grazing effect on FD from negative at very low precipitation, to positive at higher precipitation. Unimodal FD–grazing intensity relationship was observed under high precipitation, while under low precipitation, FD decreased moderately with increasing grazing intensity. 6. Synthesis. Our model showcases how fundamental tradeoffs in plant traits may drive functional diversity and ecosystem function along environmental gradients. It offers a mechanism through which novel understandings can be obtained regarding the interplay between water stress, belowand above-ground competition and disturbance intensity and history. We discuss further model testing possibilities as well as required empirical work.