Developing aerial robotic systems for measuring wind velocity near the Earth's surface is critical for understanding the surface-atmosphere interactions driving the dynamic state of the atmospheric boundary layer (ABL). How the ABL evolves in space and time impacts how effective crop health management practices are, determines how far air pollutants travel, and influences how fast wildfire spreads. However, most conventional sensors only reach tens of meters above the ground when the height of the daytime ABL is in the order of hundreds of meters and are often cost-prohibitive and time-consuming to deploy over water and complex terrain. This talk will present a model-based framework for on-demand and targeted measurements of atmospheric wind flows using aerial robotic systems. The first part of the talk will present a model-based wind sensing framework for aerial robotic systems that leverages flight data-driven modeling and state estimation techniques. The second part will discuss how integrating model-based wind sensing frameworks and low-cost sensor technology can help characterize the wind-induced drift of irregularly shaped objects in aquatic environments, providing a new tool for improving the lead time of ocean search and rescue operations.