Decentralized detection in a network of wireless sensor nodes involves the fusion of information about a phenomenon of interest (POI) from geographically dispersed sensor nodes. In this paper, we investigate a binary decentralized detection problem in a dense wireless sensor network (WSN). We assume that the strength of the POI varies spatially. We consider a fusion architecture which allows cooperation among the nodes. In order to allow the nodes to cooperate, we propose a consensus flooding protocol and analyze the average energy consumption when cooperation is present. Unlike similar works, we analyze the effect of a realistic sensing model, flooding protocol and network connectivity on the system reliability and energy consumption. We demonstrate that a trade-off exists among spatial diversity gain, average energy consumption, delivery ratio of the flooding protocol, network connectivity, node density, and POI intensity in the cooperative architecture. Consequently, we are able to determine when the additional performance gain provided by cooperation outweighs the increase in average energy consumption required to cooperate.