In order to evaluate the performance of emerging wireless systems such as relay, sensor, and mobile ad hoc networks with multihop coverage extensions, spatial-temporal network channel models are required. These models should include the directional characteristics of network information flows in order to exploit the spatial domain due to the deployment of advanced antenna systems. Furthermore, the models should be capable of handling non-stationary scenarios with dynamic evolution of network nodes. In an attempt to solve these two problems which have not been fully addressed in the existing literature, and motivated by the useful analogy between classical propagation channels and wireless networks, we propose a novel spatial-temporal network modeling framework, where the relevant figure-of-merit (FOM) may be received signal power, channel capacity, event detection error exponent, etc., depending on the network's type. Our methodology would be most useful for the design and analysis of future networks featured by decentralized, random, and dense placement of nodes.