Objectives

1. Methods Used to Monitor the Distribution of Migrating and Wintering Waterbirds on Lake MichiganGLC Pelagic Bird Survey Workshop (30 May 2012)

2. Objectives Great Lakes Restoration Initiative: Avian Botulism in Distressed Great Lakes Environments • Document the fall distribution and foraging patterns of waterbirds at-risk to type-E botulism in the Great Lakes via low-level, systematic aerial surveys of selected areas of Lake Michigan Distribution and abundance of migrating and wintering waterbirds on Lake Michigan (FWS Region 3 Migratory Bird Conservation and USGS Base Funds) • Characterize the distribution, abundance, and temporal use patterns of waterbirds during fall through spring for impact assessment of near-shore and off-shore wind turbine placement and to characterize sea duck wintering distribution and abundance

3. Aerial waterbirdsurveys • Transect based • Conducted at about 120 knots (138 mph, 222 km) at 200’ (61 m) AGL • Occurrence and numbers of waterbirds within 200 m on each side of plane recorded along with GPS waypoint. • Species include loons, mergansers, grebes, ducks, geese, swans, cormorants, gulls

4. Equipment Partenavia P68 Observer Equipment used- plane, GPS-linked recording device, recording software • Brian Lubinski • Pilot certified - 29 years • USFWS pilot – 13 years • Conducting bird surveys – 11 years • Low-level flying experience – 1,000 hrs

5. GPS-linked observation recording

6. Observers …training, training materials (Wildlife Counts software http://wildlifecounts.com; Aerial Observers Guide by Tim Bowman, USFWS, under development), crew experience

7. Aviation Safety …aviation safety training, flight plan (FAA) & flight following (MIFC), dry-immersion survival suits, pilot-crew briefing, aviation hazard maps Great Lakes Waterbird Survey Project Aviation Plan • Hazard analysis • Operational Plan • Flight Following • Aircraft Mishap Response Chart

8. North End Summary of Lake Michigan aerial waterbird surveys 2009-20126,182 km ( 3,841 miles) within 11 survey areas Green Bay Sleeping Bear dunes National Lakeshore Door County- Garden Peninsula Ludington Bay Milwaukee- Kewaunee, WI Oceana-Ottawa Counties MI Mid-Lake Plateau Milwaukee, WI- Evanston, IL Allegan-Berrien Counties, MI South End

9. Survey Protocol …establishing survey area • Transects are within the boundaries of 11 survey areas (up to ~ 25 miles offshore) • Fixed-width transects spaced every 4.8 km (3 miles) • Transects established using snapPLAN software (TRACK'AIR Aerial Survey Systems, The Netherlands) • Flight lines oriented generally parallel to the adjacent shoreline

10. Survey Protocol …altitude, speed, survey conditions, documenting observation conditions • Targets: • 200’ @ 120 knots (61 m @ 222 km/hr) • Tally all waterbirds within 200 m on each side of plane (distance established with clinometer, e.g., 17o off horizon at 200’) • Calm days preferred. Surveys not undertaken when water surface conditions are greater than about 20 km/hr (13 mph) and waves greater than about 1 m to enhance detectability • Observation conditions documented during survey • Rated as poor, fair, good, very good, excellent • Based on sun glare, lighting conditions, water surface/color, observer fatigue

11. Survey Protocol …documenting observation conditions

12. Distribution and density of common loons and carcasses - 7-8 Oct 2010

13. Distribution and density of waterbirds susceptible to BotE – Oct 2010 & 2011

14. Waterbird abundance model Modeling the association of selected species to environmental conditions • GIS application developed by Tim Fox to create spatial layer of sequential 6-ha observation polygons (rectangles) on each side of flight-line of plane • Both observation and null polygons established • Lookup table created to associate observation data with environmental covariate data

15. Covariate Data Current Data • Polygon Type (Observation or Null) • Observation conditions • Distance to Shore • Angle to shore • Water Surface Temperature (MODIS) • Chlorophyll Density (MODIS) • Vector Ruggedness Model (slope detrended topographic roughness) • Water Depth • Topographic Slope • Substrate type (National Cooperative Soil Survey) • Distance to Know Spawning Locations (Goodyear Spawning Atlas) • Distance to Know Zebra Mussel Locations • Wind data • Prey fish distribution (USGS Hydroacoustic survey) Data possibly to be Added • Plankton - zooplankton data • Chladophora distribution (MI Tech Research Institute)

16. Hierarchiacal Bayesian spatial count models Patrick McKann and Wayne Thogmartin • Waterbird counts within observation/null polygons (birds per 6 ha) is the response. • The associated environmental covariate information for each rectangle will be tested for association with the response. • We will probably develop a suite of models with different combinations of covariates to run through the model, and compare them by information criteria to see which model or set of models fit the data the best. • These models will be hierarchical spatial count models, with random effects for within and across years, observer, side of the plane, transect, and maybe other effects. We are also going to possibly include an autoregressive term for each rectangle, which is modeling the fact that rectangles neighboring each other are very likely to be correlated (spatial correlation). • Maps produced of species relative abundance.

17. Brian Lubinski, USFWS-Region 3 Steve Houdek, Larry Robionson, Tim Fox, Patrick McKann, Wayne Thogmartin, USGS-Upper Midwest Environmental Sciences Center Acknowledgements