The number of studies examining how species diversity influences the productivity of ecosystems has increased dramatically in the past decade as concern about global loss of biodiversity has intensified. Research to date has greatly improved our understanding of how, when, and why species loss alters primary production in ecosystems. However, because experiments have been performed at rather small spatial and short temporal scales, it is unclear whether conclusions can be readily extrapolated to the broader scales at which natural communities are most likely to influence ecosystem functioning. Here we develop a simple patch-dynamics model to examine some of the scale-dependent and independent qualities of the diversity-productivity relationship. We first simulate a typical diversity-productivity experiment and show that the influence of species richness on productivity is temporally dynamic, growing stronger through successional time. This holds true irrespective of whether resource partitioning or a sampling effect is the underlying mechanism. We then increase the spatial scale of the simulation from individual patches to a region consisting of many patch types. Results suggest that the diversity-productivity relationship is not influenced by spatial scale per se, but that the mechanism producing the relationship can change from sampling effects within individual patches to resource partitioning across patch types composing the region. This change occurs even though model dynamics are the same at both scales, suggesting that sampling effects and resource partitioning can represent different descriptions of the same biological processes operating concurrently at differing scales of observation. Lastly, we incorporate regional processes of dispersal and disturbance into the model and show that these processes can amplify the effect of species richness on productivity, resulting in patterns not easily anticipated from experiments. We conclude that the relative control of community structure by local versus regional processes may be a primary determinant of the diversity-productivity relationship in natural ecosystems. Therefore, past experiments having focused only on local processes might not reflect patterns and processes underlying diversity-productivity relationships in communities where disturbance and dispersal regulate species biomasses.