MOSART-urban: Integrated regional-scale urban flood modeling
Numerical modeling of hydrological processes in urban areas plays a key role in urban planning and disaster management. Urban infrastructures intertwine hydrologic and hydraulic processes even more strongly and as a result, accurate large-scale modeling of urban watersheds becomes a challenging task. In this study, a conceptual model called MOSART-urban is developed that takes into account all the dominant urban processes in a computationally efficient framework. The model is developed based on the channel routing module of E3SM (Energy Exascale Earth System Model), MOSART. Currently, MOSART performs the channel routing through three mechanisms; hillslope, tributaries and main channel. MOSART-urban replaces the hillslope and tributaries components of MOSART with sewershed and surface storage components to account for the effect of urban infrastructures. A sewershed describes water pathways through a built environment that direct a portion of surface runoff on the impervious surface to the sewer system. The sewer system has a limited capacity and may overflow. The surface storage component adds effects of elements such as retention ponds and wetlands, and is driven by the head difference between the storage and the main channel. MOSART-urban conveys the portion of the surface runoff that is not stored in the surface storage through the sewershed based on a diffusive wave approach to account for the backwater effect. The routing in the sewershed is done via two mechanisms; sewer network which is based on an instantaneous network unit hydrograph, and sewer overflow. In this approach the actual sewer network is replaced with a synthetically generated network that provides a network width and a travel time distribution for routing the flow through the sewer network. Houston in Texas is used as a case study to measure performance of MOSART-urban. The results show significant improvement in regional-scale modeling of urban areas while being computationally efficient.