HABITAT SELECTION AND NESTING RESPONSES TO SPRING FLOODING BY EASTERN WILD TURKEY HENS IN LOUISIANA FREDERICK G. KIMMEL, School of Forestry, Wildlife and Fisheries, Louisiana State University, Baton Rouge, LA 70803 PHILLIP J. ZWANK, Louisiana Cooperative Wildlife Research Unit, School of Foresty, Wildlife and Fisheries, Louisiana State University, Baton Rouge, LA 70803 Abstract: Observations were made on the movements of 14 radio-tagged eastern wild turkey (Meleagris gallopavo silvestris) hens during 1983 spring flooding of the Mississippi River. The study area, Point Lookout, East Carroll Parish, Louisiana, was flooded for 64 days, from 18 April to 20 June 1983. One hen died during this period. Monitored turkeys remained in trees throughout the day during flooding. High ground in fields was used as it became available. Habitats were ranked by relative usage and preference. Flooded riverfront hardwoods had the greatest use during the day as well as for roosting. Wheat fields were the preferred habitat at the lower water levels. Uncultivated fields were preferred at higher water levels when wheat was inundated. Mean minimum home ranges were similar for juvenile (308 ha) and adult (313 ha) hens. No young turkeys were sighted on the area, indicating reproduction probably did not occur. In the southeastern United States, some of the highest wild turkey populations are found in the flood plain of the Mississippi River (Lewis 1967). Much of this bottomland type has been converted to row crops in Louisiana so these wild turkey populations are currently found in the batture lands which is the narrow strip of land between the Mississippi River and the flood protection levee. Flooding may limit wild turkey populations along the Mississippi River. Few studies document responses of turkeys to severe flooding. Dalke et al. (1946) stated that floods along the Mississippi River may force turkeys to levees and tree tops for days at a time. Hollis (1950) reported turkeys driven permanently from habitat during flooding. He believed, however, that flood damage to turkeys is slight. Schorger (1966) stated that floods are detrimental only to turkey nests and 155 156 young. This paper presents information on wild turkey nesting success, habitat use, and preference during 1983 flooding of the Mississippi River, which was the highest since 1973 and the fifth highest of the century, lasting for 64 days (18 April - 17 June). We would like to thank International Paper Co. and Anderson Tully Co. on whose land this study was conducted. We appreciate the cooperation and assistance of the members of Point Lookout Hunting Club and Big Eight Hunting Club. Our thanks to the Louisiana Department of Wildlife and Fisheries, which provided radio transmitters, and to D. Dennett, C. Parker, and C. Williams for their assistance. STUDY AREA The 2,500 ha study area is located between the Mississippi River and the flood protection levee in East Carroll Parish, Louisiana. Flooding of the lower elevations occurs annually. The river begins to flood a significant portion of the study area when the U.S. Army Corps of Engineer's gauge at Lake Providence, Louisiana, measures 10.6 m, which measurement corresponds to 30.68 m on U.S. Geological Survey topographic maps of the study area. During major floods a majority of the study area is inundated. The study area was stocked in 1966 with 31 wild turkeys. Twenty- three were killed in 1972 when the first hunting season was held. Spring gobbler hunting seasons have been held every year since except 1974 when the season was closed due to severe flooding in spring 1973. Peak harvest on the study area occurred in 1980 when 55 gobblers were killed. The 1983 wild turkey population was estimated at 500-600 turkeys (C. Williams, pers. commun., La. Dept. of Wildl. and Fish., Monroe, LA.). Approximately 1,200 ha of the study area is riverfront hardwoods (Fig. 1). Sugarberry (Celtis laevigata), box elder (Acer negundo), sweet pecan (Carya illinoensis), and green ash (Fraxinus pennsylvanica) are the primary overstory species. Black willow (Salix nigra) and cottonwood (Populus deltoides) are the predominate overstory species on approximately 312 ha of the willow-cottonwood habitat type. Pure stands of black willow and cottonwood are found in the lower elevations, with sugarberry and box elder interspersed at higher elevations. Approximately 240 ha were in agricultural production. Cotton fields, occupying 237 ha, were fallow when data were collected. Wheat was planted annually for wildlife on about 100 ha. Four American sycamore (Plantanus occidentalis) plantations occupied 196 ha. Uncultivated fields in perennial and annual grasses covered 243 ha. Pipelines, roads, and permanent water made up the remaining area. Fig. 1. Habitat types on Point Lookout, East Carroll Parish, Louisiana, 1983. 158 METHODS In February 1983, 5 adult and 9 subadult wild turkey hens were captured with a cannon net and fitted with radio transmitters (Teleonics Telemetry Consultants, Mesa, AZ ; and Telemetry Systems, Mequon, WI) and released at the capture site. and yagi antenna were used for tracking. A portable receiver, scanner hens were determined by triangulation. Locations of the transmittered Attempts were made to locate each hen 4 days each week. Each day was divided into 4 time periods: early morning (0.5 hr before sunrise to 0900), midday (0901-1400), late afternoon (1401-sunset), and night (0.5 hr after sunset). not taken from sunset to 0.5 hr after sunset. Readings were locate Attempts were made to all turkeys within a single time period each day. Data were also collected from observations of unmarked hens. The extent of flooding over the study area was determined by calculating the elevation of flood waters and plotting flooded areas on a U. S. Geological Survey topographic map. The daily river stage from the Lake Providence gauge was added to the elevation of gauge zero. The average slope of the river was calculated and the reading corrected for slope because the gauge was approximately 16 km north of the study area. Locations of monitored hens were plotted on a topographic map. Each location was classified by habitat type. identified: Eight habitat types were riverfront forest, flooded riverfront forest, willow- cottonwood, sycamore plantation, flooded sycamore, fallow field, uncultivated field, and wheat field. A preference index was developed to rank habitat preference. The amount of available area was estimated for each habitat type. Available habitat was determined for water levels at low flood stage (30.48-32.00 m), mid-flood stage (32.00-33.52 m), and high flood stage (33.53-35.05 m ) . These levels corresponded with the 1.52 m contour intervals on the topographic map. We could not accurately interpolate elevations between contour intervals; therefore, only maximum and minimum values of available habitat for each water level/habitat type were determined. The minimum available habitat value was determined by assuming that all land within the contour interval was flooded. The maximum available habitat value was determined by assuming no land within the contour interval was flooded. The index value was then calculated as: observed locations within a habitat type/water level expected locations within a habitat type/water level. Expected locations were derived based on the assumption that the habitat types were used in proportion to their availability. Flooded, unforested habitat types were considered unavailable and were not included in determination of expected locations. The index was calculated for each minimum and maximum value. Because the actual preference index value lies between the maximum and minimum values, the midpoint was used as the index value. to obtain a larger sample size. All daytime locations were pooled separately. Night locations were analyzed This method was used primarily to rank relative preferences for habitat types. No inferences were made on the degree of preference 159 as indicated by the index values. Chi-square analysis was used to determine whether significant differences (P<0.05) existed in habitat use between age classes. If no significant difference existed, the data from both age classes were pooled; otherwise, the data were analyzed separately. Home ranges were determined by the minimum home range method (Mohr 19479. A t-test was used to determine differences in mean home range of adult and subadult hens. RESULTS AND DISCUSSION Nesting During flooding, three nests of unmarked hens were found. Two were located in clumps of Johnson grass (Sorghum halepense) in an open field along the Mississippi River. This was the only large area (42 ha) that was not inundated. The third nest, also in a Johnson grass clump, was located at the edge of a woods road. All appeared to have been destroyed. A nest of a radio-marked hen was found in a sycamore plantation on 13 May. The nest, which contained 9 eggs, was later covered by flood waters. Levees provided potential nest sites but were not used by radio marked hens. Vehicular traffic, other human disturbances, and grazing may have limited their use. Thirteen turkey eggs were found scattered over the study area during the flood period. These did not appear to have been associated with a nesting attempt and may have been eggs "dumped" by hens. Usually, a single egg was found in an open field with little surrounding cover or on woods roads. In one instance, 3 eggs were found approximately 15 m apart in a sandy woods road. Lack of suitable nesting sites, stress, or predator harassment may have triggered the "dumping" of eggs. Savage (1977) reported finding 2 "dumped" turkey eggs on elevated roads during flooding. Wheeler (1948) attributed egg "dumping" to interruption of laying by predators or inability of the hens to reach their nests on time. An active nest containing 8 eggs was found on 4 August in a brush pile along a woods road. The flood waters had been off the nest site for 34 days. The hen was flushed by tree trimming crews and did not return. Four eggs were removed from the nest, 2 of which were infertile. The remaining 2 eggs contained embryos estimated to be 10 days old. Deer hunters were asked to report sightings of turkeys during the first 2 days of the deer hunting season. None of 50-60 hunters 160 interviewed reported sighting young turkeys. The 1984 spring gobbler harvest data also indicated a minimal turkey hatch. Six adult gobblers were killed, no subadult gobblers were killed or seen. The previous year, 9 adult gobblers and 28 subadult gobblers had been killed. Limited reproduction may have resulted from lack of nesting habitat, predation, and nutritional deficiencies. During the peak of flooding, approximately 42 ha of the 2,500 ha study area were not inundated. Predators, such as striped skunks (Mephitis mephitis) and coyotes (Canis latrans), both abundant on the study area, became concentrated as flood waters rose. A radio-marked hen shared an island of barren ground approximately 25 m 2 with 6 striped skunks. Additionally, the quality of the foods available may have been inadequate for reproduction. Pattee (1978) found that ovary development was correlated with protein, phosphorus, and insects in the crop. Gardner and Arner (1968) observed that inadequate diets resulted in lowered egg production. Nesting that occurs after the normal nesting season may maintain turkey populations when nesting during the normal nesting period is interrupted. Renesting following a late snowfall in Montana produced sufficient poults to replace annual mortality (Jonas 1968). Nesting, although reduced, was delayed 5 weeks due to heavy flooding in April 1975 on Huntington Point, Mississippi (Johnson et al. 1977). Floods of prolonged duration, however, reduce chances of successful reproduction. Spermatazoa survive up to 70 days in the domestic turkey though the commonly known fertile period is usually shorter (Guilbert 1974). The proportion of infertile eggs then increases the longer nesting is delayed (Schorger 19669. Habitat Use During flooding, 170 daylight locations and 68 night locations were determined for 14 hens (x=17 per hen, s.d.=5.6). Daylight habitat use by adult and subadult hens was not significantly different (P<0.05) at low flood stage (x =10.22, df=7) or mid-flood stage (x =5.64, df=4). Habitat use between age classes was significantly different (P<0.05) at high flood stage (x =5.64, df=2). No difference (P<0.05) occurred between adult and subadult roosting habitat. Figures 2-5 illustrate changing habitat availability and use for 3 flood levels. Use of flooded riverfront forest and flooded sycamore habitat increased with rising water levels. Use of riverfront forest, sycamore, and wheat habitat decreased with rising water. Use of fallow fields was highest at mid-flood stage. Fallow fields contained islands of high ground that were available until the water reached the high flood stage. Uncultivated fields represented the highest ground on the study area, and at high flood stage were the only large areas (42 ha) of dry ground available. Use of these fields by turkeys increased with 161 Fig. 2. Habitat availability estimates and distribution of locations of radio-marked hens during low flood stage, Point Lookout, East Carroll Parish, Louisiana, 1983. 162 Fig. 3. Habitat availability estimates and distribution of locations of radio-marked hens during mid-flood stage, Point Lookout, East Carroll Parish, Louisiana, 1983. 1 6 3 Fig. 4. Habitat availability estimates and distribution of locations of adult, radio-marked hens during high flood stage, Point Lookout, East Carroll Parish, Louisiana, 1983. 164 Fig. 5. Habitat availability estimates and distribution of locations of subadult, radio-marked hens during high flood stage, Point Lookout, East Carroll Parish, Louisiana, 1983. 165 rising water and peaked during high flood stage. At night, radio-marked hens roosted most frequently in flooded riverfront hardwoods. Use ranged from 50% of the locations at the low flood stage to 85% at the high flood stage. Flooded riverfront hardwood forest was the preferred roosting habitat at low flood stage (UI=2.90) and high flood stage (UI=1.61) (Table l), but unflooded riverfront hardwoods were preferred at mid-flood stage (UI=1.91). Although the willow-cottonwood type was preferred by adult hens during daylight at high flood stage, this habitat type is not known to have been used for roosting. Radio-marked hens preferred unflooded habitat during the day. Wheat field was the most preferred habitat type at the low water level (UI=41.67), followed by sycamore (UI=3.13), fallow field (UI=1.90), and riverfront hardwood (UI=1.52) (Table 2). Wheat, fallow field, and riverfront hardwood types continued to be preferred at the mid-flood stage; however, sycamore was utilized less. Uncultivated fields were the preferred habitat type of subadult hens (UI=33.33) (Table 3) at high flood stage when this was the only habitat type containing a large area not inundated. Adult hens preferred willow-cottonwood (UI=2.08) at high flood stage. Preference for wheat fields, fallow fields and sycamore plantations may be due to a preference for green vegetation in spring. Kimmel (1984) found that during the period preceding the flood (1 March - 18 April 1984), the same radio-marked hens used sycamore plantations, fallow fields, and wheat fields in proportion to their availability. Savage (1977) observed use of wheat fields in Tensas Parish, Louisiana, during spring for feeding and nesting. Fallow fields and sycamore plantations, which lacked dense grass, received high use, possibly due to the abundance of early spring forbs. Low use of uncultivated fields during low and mid-flood stages may have been due to the absence of high quality foods in these fields in early spring. Competition from dense perennial grasses and compacted soil probably limited forb production. Despite the preference for unflooded habitat types, flooded riverfront forest was the most frequently utilized daylight habitat during the flood period (26% of all locations). Turkeys were observed in trees throughout the day while flooding occurred. Even though the willow-cottonwood habitat type was located on the lowest elevation on the study area, and was flooded longer than any other habitat type, it had the second highest rate of use at mid-flood stage (23%) and high flood stage (35%). These forested habitat types, based on abundance and degree of usage, appear to be the most important turkey habitat on the area during severe flooding when tree tops are the only available habitat. The forested habitat types provide roosting areas as well as food, in the form of fruit, leaves, and buds, during flooding. Kimmel (1984) found that sugarberry comprises approximately 37% of the overstory in the riverfront hardwoods. Sugarberry buds are utilized by turkeys when the 166 Table 1. Ranking of available roosting habitat types used by eastern wild turkey hens during flooding on Point Lookout, East Carroll Parish Louisiana, based on preference indicies calculated as observed locations within a habitat type/water level divided by expected locations within a habitat type/water level. Low Flood Stage (30.48-32.00 m) Mid-Flood Stage (32.00-33.52 m) High Flood Stage (33.52-35.05 m) Flooded riverfront hardwoods (2.90) 1 Riverfront hardwoods Sycamore plantations (1.88) Riverfront hardwoods (0.91) Willow-cottonwood (0.80) Flooded sycamore plantations (0) (1.91) Flooded riverfront hardwoods (1.47) Sycamore plantations (0.51) Willow-cottonwood (0.48) Flooded sycamore plantations (0) Flooded riverfront hardwoods (1.61) Riverfront hardwoods (1.42) Flooded sycamore plantations (0) Willow-cottonwood (0) 1 indices shown in parenthesis 167 Table 2. Ranking of available daylight habitat used by wild turkey hens during low and mid-flood stage on Point Lookout, East Carroll Parish, Louisiana, based on preference indicies calculated as observed locations within a habitat type/water level divided by expected locations within a habitat type/water level. Low Flood Stage Mid-Flood Stage (30.48-32.00 m) (32.00-33.52 m) Wheat fields (41.67) 1 Sycamore plantations (3.13) Fallow fields (1.92) Riverfront hardwoods (1.52) Flooded riverfront hardwoods (1.03) Flooded sycamore plantations (0.34) Willow-cottonwood (0.25) Uncultivated fields (0) Wheat fields (6.41) Fallow fields (1.68) Uncultivated fields (1.03) Flooded sycamore plantations (0.97) Riverfront hardwoods (0.96) Flooded riverfront hardwoods (0.90) Willow-cottonwood (0.72) Sycamore plantations (0) 1 indices shown in parenthesis 168 Table 3. Ranking of available daylight habitat used by wild turkey hens during high flood stage (33.52-35.05 m) on Point Lookout, East Carroll Parish, Louisiana, based on preference indicies calculated as observed locations within a habitat type/water level divided by expected locations within a habitat type/water level. Adult Hens Subadult Hens Willow-cottonwood (2.08) 1 Uncultivated fields (33.33) Flooded sycamore plantations (0.93) Flooded riverfront hardwoods (1.02) Flooded riverfront hardwoods (0.42) Riverfront hardwoods (0.90) Riverfront hardwoods (0) Flooded sycamore plantations (0.76) Uncultivated fields (0) Willow-cottonwood (0.28) 1 indices shown in parenthesis 169 birds are confined to tree tops (Schorger 1966). On several occasions turkeys were observed in red mulberry (Morus rubra) trees. This species is not abundant on the study area, indicating that mulberry may have been sought out. Red mulberry fruit occasionally is utilized by wild turkeys in spring (Moore and Thomas 1977). Insects and insect galls in the overstory may also be utilized as food (Korschgen 1967). The omnivorous and opportunistic nature of the wild turkey (Schorger 1966, Korschgen 1967, Williams 1981) appears to contribute to its ability to survive under flood conditions. Hens returned to dry ground as soon as it became available. On several occasions, flocks of 5-10 hens were observed on small islands (0.10 ha) recently exposed by falling water. On only 1 occasion was a radio-marked hen located on a levee. Radio-marked hens were not observed to have used land on the opposite side of the levee. One radio-marked hen died during the flood period. The cause of death could not be determined. Mean minimum home range of 9 subadult hens was 308 ha. The 5 adult hens had a mean minimum home range of 313 ha. There was no difference (P<0.05) in home range of adult and subadult hens. Average home range of radio-marked hens during flooding (310 ha) was consistent with the findings of other researchers. A prenesting home range of 396 ha was found by Johnson et al. (1977) on a batture study area in Mississippi. Savage (1971) observed a home range of 584 ha for hens on a similar study area in Mississippi. Since only one radio-marked hen died during the study period, a maintenance diet of food must have been available during the flood. Because of the impossibility of having a controled environment, the study was unable to determine whether habitat use at low water levels was due to seasonal preferences, flooding, or a combination of factors. Habitat use during low flood stage may be due, in part, to breeding behavior and dispersal to suitable nesting areas. At mid-flood and high flood stages, however water levels played an increasingly important role in habitat selection. At high flood stage, water level was probably the primary factor determining habitat use. The primary effect of flooding appears to be reduction in successful nesting. Mortality of adult and subadult hens was low even with concentrations of predators. Turkeys were able to survive by foraging in the overstory for extended periods, and used high ground when it was available. The wild turkey can live to 10 years of age (Mosby 1967) so it can withstand the loss of production for 1 year but successive years of severe flooding, however, could deplete a local wild turkey population. Pages 55-56 in Lowell K. Halls, ed. by wildlife. Southern fruit producing plants used U. S. Dep. Agric. For. Serv. Gen. Tech. Rep. SO-16. South. For. Exp. Sta., New Orleans, LA. 170 LITERATURE CITED DALKE, P. D., A. S. LEOPOLD, and D. L. SPENCER. 1946. The ecology and management of the wild turkey in Missouri. Missouri Conserv. Comm. Tech. Bull. No. 1. 86 pp. DICKSON, J. G., C. D. ADAMS, and S. H. HANLEY. 1978. Responses of turkey populations to habitat variables in Louisiana. Wildl. Soc. Bull. 6:163-166. GARDNER, D. T., and D. H. ARNER. 1968. Food supplements and wild turkey reproduction. Trans. N. A. Wildl. and Nat. Res. Conf. 33:250-258. GUILBERT, A. B. 1974. Poultry. Pages 301-323 in E.S.I. Hafez, ed. Reproduction in farm animals. Lee and Febiger, Philadelphia, PA. HOLLIS, F. D. 1950. The present status of the wild turkey in Louisiana. La. P. R. Proj. 3-R. 78 pp. JOHNSON, R. L., D. E. WESLEY, C. J. PERKINS, and A. D. SULLIVAN. 1977. Effects of even-aged cottonwood culture on wildlife populations and habitats in the Mississippi Delta. Final Rep., U. S. Forest Serv. Agreement No. 19-177. 214 pp. JONAS, R. J. 1968. Adverse weather affects Merriam's turkey reproduction in Montana. J. Wildl. Manage. 32:987-989. KIMMEL, F. G. 1984. Habitat use by wild turkey hens in the Mississippi River floodplain. M.S. Thesis, La. State Univ., Baton Rouge. 115 pp. KORSCHGEN, L. J. 1967. Feeding habits and foods. Pages 137-198 in O. H. Hewitt, ed. The wild turkey and its management. The Wildl. Soc., Washington, DC. 589 pp. LEWIS, J. B. 1967. Management of the Eastern wild turkey in the Ozarks and bottomland hardwoods. Pages 371-407 in O. H. Hewitt, ed. The wild turkey and its management. The Wildl Soc., Washington, DC. MOHR, C. O. 1943. A summary of North American small animal censuses. Am. Midl. Nat. 29:545-587. MOSBY, H. S. 1967. Population dynamics. Pages 113-136 in O. H. Hewitt, ed. The wild turkey and its management. The Wildl. Soc., Washington, DC. MOORE, D. M. and W. P. THOMAS. 1977. Red mulberry/Morus rubra. 171 PATTEE, O. H. 1978. Effects of nutrition on wild turkey reproduction in South Texas. Dis. Abstr. Int. B. Sci. Eng. 38(8):3489. SAVAGE, J. L. 1977. Status and habitat selection of two introduced turkey populations in bottomland hardwoods of Louisiana. M.S. Thesis., La. State Univ., Baton Rouge. 122 pp. SCHORGER, A. W. 1966. The wild turkey: its history and domestication. Univ. Okla. Press, Norman. 625 pp. SPEAKE, D. W., T. E. LYNCH, W. J. FLEMING, G. A. WRIGHT, and W. J. HAMRICK. 1975. Habitat use and seasonal movements of wild turkeys in the southeast. Pages 122-130 in L. K. Halls, ed. Proceedings of the third national wild turkey symposium. Texas Chapter, The Wildl. Soc. WHEELER, R. J. 1948. The wild turkey in Alabama. Ala. Dept. Cons. 92 pp. WILLIAMS, L. E. 1981. The book of the wild turkey. Winchester Press, Tulsa, OK. 181 pp.