Due to the presence of location dependent channel errors on wireless links, many wireless scheduling algorithms have been proposed which aim to provide efficient distribution of resources and fair allocation of available bandwidth to flows. Traditionally, these algorithms have adopted a two state model to represent the bandwidth of individual wireless channels, i.e. channels can be in a ?good? state or in a ?bad? state. In actual fact, the bandwidth of each channel depends on the adaptive techniques employed by the wireless system (e.g. coding and modulation). This paper considers different levels of rainfall and examines the impact of these levels in conjunction with suitable adaptive mechanisms on the available system bandwidth. A wireless link bandwidth model is presented which shows an accurate depiction of bandwidth variations to be used for design, optimization and benchmarking of wireless network schedulers. Knowledge of the link bandwidth is crucial for an intelligent scheduling algorithm since the scheduler must know how much bandwidth is available in order to perform fair reallocation.