Duck Nesting in Retired Croplands at Horicon National Wildlife Refuge, Wisconsin

Duck nesting was studied in 85 retired agricultural fields at Horicon National Wildlife Refuge, Wisconsin, during 1976-78, as a basis for management recommendations to increase upland duck production. Blue-winged teal (Anas discors) nests comprised 94% of the 691 nests found. Numbers of mallard (A. platyrhynchos) nests in uplands appear to have declined since 1954. Nest initiation was earliest in hayfields, intermediate in tracts retired for 5+ years, and latest in 1and 2-year-old cover plantings. Bluegrass (Poa spp.) was the most frequent cover species at nests, especially at blue-winged teal nests. Mallard nests were generally in taller, denser cover, and farther from water than nests of blue-winged teal. Ducks used tall grasses and alfalfa heavily for nesting in May, but thereafter nested mostly in short grasses. The mean nesting density, as indicated by flushed hens, was 1.5/ha during 1977-78; found nests averaged 1.2/ha. Nesting densities were highest in unmown hayfields and long-retired tracts in 1977-78, and in fields dominated by tall grasses and alfalfa in 1977. Short-grass fields had high nesting densities in 1978, perhaps because of heavy spring rains. Overall nest success, corrected for stages of incubation, was 9%; predators, primarily striped skunks (Mephitis mephitis) and raccoons (Procyon lotor), destroyed 537 (81%) nests. Generally, successful nests were initiated earlier in the season, in taller and denser cover, and farther from water than nests destroyed by predators. Management recommendations include reduction of cover-reducing land-use practices in refuge uplands, and predator control in areas with high nest densities and low nest success. J. WILDL. MANAGE. 45(1):27-37 Descriptions of nest sites are common throughout the waterfowl literature, but most detailed studies of the relationships of land use to duck production are recent. Removal of cover by mowing, grazing, applications of herbicides, field burning, and annual cropping was found to be detrimental to duck production in numerous studies (e.g., Martz 1967, Kirsch 1969, Dwernychuk and Boag 1973, Fritzell 1975, Higgins 1977), but practices recommended for maintenance of cover differed. Also, possible effects of more subtle differences in land use, such as ages of fields, were not well studied, although differences in plant succession may affect duck production in uplands (Stoudt 1969, Miller 1971, Nelson and Duebbert 1974). As a part of a management plan to reduce autumnal concentrations of Canada geese (Branta candensis) at Horicon National Wildlife Refuge (HNWR), Wisconsin, cooperative farming of refuge uplands to produce food for geese was terminated after 1976 (Klepinger and Ellis 1975). Also, 19 refuge fields totalling 68 ha were seeded during 1976-78 to provide cover for nesting ducks. This paper describes a study of duck nesting in uplands at HNWR designed to develop managerial recommendations to increase duck production. Specifically, this project attempted to: (1) assess the impacts of land-use histories, cover types, and locations of recently retired croplands on species composition, initiation, sites, densities, and success of duck nests; (2) compare duck-nesting parameters of recent cover plantings with those of other field types; and (3) identify the primary factors limiting duck production in uplands at HNWR. This research was supported by the U.S. Fish and Wildlife Service, McGraw Wildlife Foundation, and the College of Agriculture and Life Sciences and the 1 Present address: Museum of Natural History, University of Kansas, Lawrence, KS 66045. J. Wildl. Manage. 45(1):1981 27 This content downloaded from 157.55.39.104 on Sun, 19 Jun 2016 06:43:47 UTC All use subject to http://about.jstor.org/terms 28 DUCK NESTING IN RETIRED CROPLANDS * Livezey Graduate School of the University of Wisconsin, Madison. I am grateful for the guidance of R. A. McCabe, and for the help of 12 field assistants. I also thank refuge personnel, especially J. E. Toll, R. L. Drieslein, H. R. Bushweiller, and A. Pachan, for their help and suggestions; J. R. Cary for advice on statistical analyses; and R. J. Blohm, J. J. Hickey, and D. H. Rusch for reviewing this paper. STUDY AREA AND METHODS Data on duck nesting were gathered in 85 retired agricultural fields at HNWR totalling 235 ha, between 20 April and 31 July 1976-78. Only 1,526 ha (18%) of the total area of the refuge are upland. The fields fringe the north, east, and west sides of a 6,834-ha marsh dominated by common cattail (Typha latifolia) and softstem bulrush (Scirpus validus) (HNWR, unpubl. data). Fields were selected and classified by mowing and cropping histories determined from refuge records; distances from fields to water were estimated on Soil Conservation Service aerial photographs. In this paper, the term long-retired tracts refers to fields untilled for 5+ years. Cover plantings seeded in 1976 were mixtures of intermediate wheatgrass (Agropyron intermedium) and switchgrass (Panicum virgatum), and in 1977 were combinations of big bluestem (Andropogon gerardi), little bluestem (A. scoparius), yellow indiangrass (Sorghastrum nutans), and switchgrass. In addition, fields other than cover plantings were grouped by a cluster analysis (Jardine and Sibson 1971) of random samples, and classed as shortgrass (primarily bluegrass) or tallgrass/alfalfa (primarily timothy [Phelum pratense], quackgrass [Agropyron repens], smooth brome [Bromus inermis], and reed canarygrass [Phalaris arundinacea]) associations. Cover in fields was measured at random sites approximately biweekly from late April to mid-July in 1977 and 1978. Cover species were assigned to 1 of 5 relative-abundance classes, and cover density was measured by counts of visible squares, after Jones (1968). Standing and prostrate cover from 0.125-m2 quadrats was clipped, dried, and weighed on 3 occasions between early May and late June. Nest searches were made twice annually, between early May and mid-July, by cable-chain dragging (Higgins et al. 1969) or by rope dragging (Labisky 1957). A nest was defined as a scrape or bowl known to have contained at least i viable egg when found. Stage of incubation and initiation dates were determined by candling (Weller 1956) and back-dating (Sowls 1955). Height, density, and composition of cover were determined as at random sites, and nest sites were classified by cluster analyses into cover types. Each nest was marked with a wooden lath placed about 5 m away, and nest fate was determined on a subsequent visit. To minimize disturbance by investigators, few nests were visited more than once during laying and incubation (Livezey 1980). Nests in which at least 1 egg hatched were considered successful, and predation was classed by the criteria of Rearden (1951). Distances from nests to selected landmarks were measured with a Leitz 500-mm range finder or with the aid of aerial photographs and field maps. Equal-area subunits of fields, classified by field type and distance to water, were used in analyses of 1977-78 nesting densities. Half-hectare subunits were used to allow inclusion of most searched areas in comparisons. Hens flushed from fields during searches were regarded as in the process of nesting, and therefore were used as an index of nest densities in the