Hybrid Free Space Optical (FSO) and Radio Frequency (RF) networks promise highly available wireless broadband connectivity and quality of service (QoS), particularly suitable for emerging network applications involving extremely high data rate transmissions such as high quality video-on-demand and real-time surveillance. FSO links are prone to atmospheric obscuration (fog, clouds, snow, etc) and are difficult to align over long distances due the use of narrow laser beams and the effect of atmospheric turbulence. These problems can be mitigated by using adjunct directional RF links, which provide backup connectivity. In this paper, methodologies for modeling and simulation of hybrid FSO/RF networks are described. Individual link propagation models are derived using scattering theory, as well as experimental measurements. MATLAB is used to generate realistic atmospheric obscuration scenarios, including moving cloud layers at different altitudes. These scenarios are then imported into a network simulator (OPNET) to emulate mobile hybrid FSO/RF networks. This framework allows accurate analysis of the effects of node mobility, atmospheric obscuration and traffic demands on network performance, and precise evaluation of topology reconfiguration algorithms as they react to dynamic changes in the network. Results show how topology reconfiguration algorithms, together with enhancements to TCP/IP protocols which reduce the network response time, enable the network to rapidly detect and act upon link state changes in highly dynamic environments, ensuring optimized network performance and availability.