Decentralized detection in a network of wireless sensor nodes involves the fusion of information about a phenomenon of interest (POI) from geographically dispersed nodes. In this paper, we investigate a binary decentralized detection problem in a dense wireless sensor network. We consider a scenario in which sensor observations, conditioned on the alternate hypothesis, are independent but not identically distributed across the sensor nodes. We consider two different fusion architectures, namely, the parallel fusion architecture and the cooperate fusion architecture. For each architecture, we derive expressions for the probability of decision error at the fusion center. We analyze the impact of a realistic sensing model, and flooding protocol on the system reliability and energy consumption for both architectures. We demonstrate that a trade-off exists among spatial diversity gain, average energy consumption, node density, and POI intensity in the cooperative architecture.