The excessive use of nitrogenous fertilizers has thrown the nitrogen cycle out of balance. As a result, nitrate contamination in our water resources is pervasive, creating severe environmental and health consequences. The electrocatalytic reduction of nitrate to ammonia is an attractive strategy for addressing this issue, as it would treat nitrate contamination while simultaneously recovering the active nitrogen component. Several examples in the literature demonstrate the electrocatalytic reduction of nitrate, yet the water treatment industry has not adopted catalysis as a general strategy for nitrate removal and resource recovery. Catalyst selectivity has been a key roadblock to the adoption of electrocatalytic nitrate reduction technology, as the nitrate reduction mechanism is complex and highly sensitive to catalyst composition, structure, and operating environment. In this work, we apply density functional theory (DFT) to understand key factors that influence the nitrate reduction mechanism. We find that elementary reactions on the catalyst surface are highly sensitive to surface coverage. For example, metals that exhibit high coverage of adsorbed NO* promote N-N coupling and the production of dinitrogen over ammonia. Our results also show that nitrate adsorption and dissociation is sensitive to the structure of the exposed catalyst facet, and thus the morphology of the catalyst particle heavily influences reaction selectivity. In particular, we find that the triangular arrangement of atoms on (111) facets yield a lower nitrate dissociation barrier compared to the square arrangement of atoms on (100) facets. This behavior explains trends observed by our experimentalist collaborators, who found enhanced nitrate activation rates on octahedron particles that predominantly expose the (111) facet. These two examples illustrate how mechanistic insight from DFT can be used to guide strategies for tailoring catalyst composition and structure to achieve active and selective nitrate reduction to ammonia.