Diversity differs among three variations of the stress gradients hypothesis in two representations of niche space

• 3 modes of interaction for the stress gradient hypothesis (SGH) are simulated.
• Hierarchical and separated niche responses on environmental gradients are represented.
• Positive interactions increase coexistence through greater source-sink dynamics.
• SGH effects increase coexistence by the separation of facilitation and competition.
• Among the 3 SGH modes coexistence decreases variably with environmental change.

How does the stress-gradient hypothesis affect coexistence in relation to established theory? For two orthogonal stress gradients, a spatially explicit agent based simulation is used to project diversity for simple competitive and facilitative interactions and for three variations of the stress-gradient hypothesis: intraspecific and interspecific competitive and facilitative interactions are a function of the abiotic environment; interactions are relative to species-specific fitness along gradients; or interaction is fixed by species regardless of the abiotic environment. Simulations are run with two orthogonal environmental gradients for two representations of niche. Facilitation can increase diversity by maintaining larger source populations and thus higher establishment rates and sink populations. With species hierarchically related in niche space, the simulations show that positive interactions and changing interactions along a stress gradient maintain greater diversity through intraspecific competition that is effective where dominance would occur and through facilitation where stress is high. A changing environment that favors some species and harms others decreases diversity in the hierarchical cases, where poor competitors most likely subject to interspecific interaction respond most strongly. Diversity outcomes differ among the three stress gradient variations because the intensity of interactions differs across the environmental gradients, not because of change in the environment.