Although non-point source (NPS) pathogen pollution is a leading cause of stream impairment in the United States, the sources of NPS pollution are often difficult to ascertain. While previous studies have employed land use regression methods to develop a greater understanding of the sources and dynamics of microbial NPS pollution, little work has explicitly considered the effects of local, state, and federal stormwater management policies on water quality across multiple watersheds or at larger spatial scales. How do land use and stormwater management efforts collectively influence fecal coliform (FC) levels at a regional or multiple-watershed scale? We construct a unique spatial regression model of stream FC pollution (n=327 monitoring stations) throughout the state of North Carolina (USA), incorporating both land cover and urban development variables. We then use a subset of our data (n=80 monitoring stations) to incorporate local stormwater control measures and stormwater management policies. Results demonstrate that the inclusion of policy and management variables improves the explanatory capacity for FC levels (R=0.4412 versus R=0.5323). Locally, this model can be used to better target stream restoration and water quality mitigation actions and investments, as well as help to predict FC levels at unmonitored locations throughout North Carolina's stream network. More generally, the novel structure of this model can also help examine the large-scale effects of stormwater regulations on surface water pathogen levels, helping researchers and planners better predict water quality in the absence of extensive monitoring station data.