Objectives: To describe the risk and details of injuries associated with ground hardness in community level Australian football (AF). Design: Prospective injury surveillance with periodic objective ground hardness measurement. Methods: 112 ground hardness assessments were undertaken using a Clegg hammer at nine locations across 20 grounds, over the 2007 and 2008 AF seasons. Details of 352 injuries sustained by community level players on those grounds were prospectively collected as part of a large randomised controlled trial. The ground location of the injury was matched to the nearest corresponding ground hardness Clegg hammer readings, in gravities (g), which were classified from unacceptably low (<30. g) to unacceptably high hardness (>120. g). Results: Clegg hammer readings ranged from 25 to 301. g. Clegg hammer hardness categories from low/normal to high/normal were associated with the majority of injuries, with only 3.7% (13 injuries) on unacceptably high hardness and 0.3% (1 injury) on the unacceptably low hardness locations. Relative to the preferred range of hardness, the risk of sustaining an injury on low/normal hardness locations was 1.31 (95%CI: 1.06-1.62) times higher and 1.82 (95%CI: 1.17-2.85) times higher on locations with unacceptably high hardness. The more severe injuries occurred with low/normal ground hardness. Conclusions: Despite the low number of injuries, the risk of sustaining an injury on low/normal and unacceptably hard grounds was significantly greater than on the preferred range of hardness. Notably, the severity of the injuries sustained on unacceptably hard grounds was lower than for other categories of hardness.