Birds fly within a moving atmosphere, but possess inertia relative to the Earth. As a bird is subjected to changes in the wind, its inertia acts as an anchor, holding it in place while the air washes past it, accelerating it relative to that air. Such wind variations (gradients), if exploited effectively, thus provide valuable opportunities for kinetic energy gain; a concept known as gradient soaring. Seabirds such as the albatross are famous for utilising this effect, exploiting wind gradients above ocean waves and in the atmospheric boundary layer to fly for thousands of miles with minimal energy expenditure. While gradient soaring has been most widely studied for these predictable and large-scale wind gradients, the same mechanisms permit energy gain from small, local and stochastic temporal variations in the wind. Urban environments in particular provide a rich landscape for atmospheric energy gain, where obstacles such as buildings and trees create complex flows with updrafts, wind shear and turbulence; previous research suggests that birds can almost half their energetic costs by utilising the wind in urban environments. Based on the dynamics of a gull-sized aircraft, flights through urban wind flow models have been simulated within an energy-aware path planning framework to elicit energy-gain manoeuvres and quantify the flight cost savings available. This work seeks both to explain the flight strategies of urban birds and to enhance the energy efficiency of small aerial robots.