Solar buildings as one type of decentralized renewable energy systems have been widely adopted to reduce carbon
emissions. Related policy making faces two questions: how much total solar energy can be produced in a city and
what proportion of building energy use can be supplied by the solar power? These questions remain hard to answer
because of the lack of appropriate modeling systems, due to the data inconsistency and the limitation of current
building energy and solar potential modeling methods in accounting for the urban context influences. This study tries
to fill this gap by developing a GIS-based energy balance modeling system for urban solar buildings. This modeling
system extends the system boundary from a single building to the urban building system, uses urban-scale data
instead of costly survey, adopts widely used GIS-platform, and makes reasonable trade-offs between speed and
accuracy. It consists of four major models: the Data Integration model, Urban Building Energy model, Urban Roof
Solar Energy model and Energy Balance model. This modeling system is applied to Manhattan as a case study. The
results show the spatial and temporal variations of building energy uses, the solar power potentials in the usable roof
areas, and the self-supply and surplus ratio of buildings in Manhattan in 2012.