Modeling Waterfowl Migration Using Radar Imagery 1 March 2008 through 28 February 2009

United States Fish and Wildlife Service and Illinois Department of Natural Resources Grant/Contract No: W-148-R-2Research from the last two decades has elucidated the importance of migration and stopover in the annual life-cycle of waterfowl.However, conservation goals and objectives for migrant waterfowl in the mid-continent have often been based on antiquated data or conjecture, and significantinformation needs remain with respect to the migration ecologyof ducks. Length of time spent at stopover sites and factors that affect departure probability are among the key uncertainties that have eluded us due to the innate difficulties of studying movement events that most often occur at night over large-scales. Weather surveillance radar is one tool with the potential to overcome these challenges and reveal new insights into migration. Although the technique has been useful in studying bird movements at broad spatial and taxonomic levels, it has not yet been applied to quantitative questions related to waterfowl due to the challenges of target classification and the calibration of radar echo to bird density. In the fall of 2008, we used a thermalinfrared camera (TIR) to ground-truth duck movements on weather radar. Our field estimate of the mean radar cross section of all dabbling ducks (112.5 cm2) aligned well with the one published estimate for mallards (120.0 cm2). We also found a strong relationship between TIR estimates of duck density and corresponding measures of echo on radar (R2=0.91, n= 8). Joint Ventures (JVs)relyon mid-continent focus areas such as those in the Illinois River valley (IRV) to provide the energetic requirementsof ducks throughout fall migration. However, ducks arrive and depart thesewetlands throughout fall, resulting in considerable turnover in fall populations. Site surveys can reveal the overall magnitude of use, but theactual number of birds using a siteand length of time they spendat these stopovers is largely unknown. 3We examined 8 years of data (1996-1998, 2003,and 2005-2008)from the KILX radar in Lincoln, Illinois to identify and enumerate ducks emigrating from a known stopover site. To estimatethe mean stopover duration for each year(days/duck), we dividedaerial inventory estimates of total use-daysfor the site bythe estimatednumber of ducks seen departing on radar. The overall mean (10.1±2.8 days [SD]) differed greatly from thecurrent value used by JVs in energetic carrying capacity models (28 days). Because our estimates of stopover duration varied substantially between years, we explored the role of habitat quality as a potential driver of these fluctuations. A regression analysis revealed a positive, linear trend and a strong relationship (R2=0.58).We also developed a comprehensive database of duck departures throughout these seasons toexamine the relationship between departure timing and exogenous weather conditions. This approach substantially improves those ofprevious studies of this question by identifyingexplicit cases of departureand non-departure. We used biologically-justifiable covariatesrelated to wind direction and speed, visibility, cloud height, temperature, and pressure toconstruct a discrete set of competing modelsintended to explain variationin daily emigration. We will employan information theoretic approachandevaluate competing modelsaccording to complexity and parsimony.INHS Technical Report Prepared for United States Fish and Wildlife Service and Illinois Department of Natural Resource