Date of this Version
Characteristics of stormwater-management ponds that contribute to avian hazards to aviation at airports have not been quantified. We selected 30 stormwater-management ponds (average 0.1 ha), approximately 50km from Seattle-Tacoma International Airport, as surrogates to on-airport facilities. We conducted 46 weeks of avian surveys (between 14 February 2005 and 17 February 2006) and evaluated model fit of 6 a priori models relative to pond use by an avian group via Kullback–Leibler information. Our full model, composed of pond surface area (sa), ratio of area of open water to area of emergent and woody vegetation (ow:ew), perimeter irregularity, and geographic isolation, was among 3 best approximating models for pond use by 9 of 13 groups (within Anatidae, Ardeidae, Charadriidae, Columbidae, Accipitridae, Laridae, and Rallidae) considered. The full model and models lacking sa or ow:ew were indistinguishable in fit for a group composed of avian species considered hazardous to aviation. For models selected, Akaike weights (i.e., relative likelihoods) ranged from 0.869 to 0.994. In contrast, relative likelihood for a mean model (i.e., a model including only an intercept) was <10−4 for all groups. We suggest that designs of stormwater management ponds at airports in the Pacific Northwest should minimize the pond perimeter via circular or linear designs. Also, ponds should be located so as to reduce the number and proximity of other water resources within 1 km. For existing stormwater-management ponds at airports, we suggest reducing the availability of open water via covering or drawdown.