2011 NORTHEASTERN NATURALIST 18(4):489–496 Can Turtle Mortality be Reduced in Managed Fields? Lori Erb1,* and Michael T. Jones1,2 Abstract - Early-successional habitats, including agricultural fields, appear to provide important foraging and nesting resource areas for Glyptemys insculpta (Wood Turtle) and Terrapene carolina carolina (Eastern Box Turtle) in the northeastern US. Mowing and agricultural activities can elevate turtle mortality rates. We performed two experiments to evaluate the risk of turtle mortality associated with 1) style of mower, 2) mower blade height, and 3) tractor tires. Mower blade height did not affect mortality rates when set to ≤15 cm. Different types of mowers appear to exert differential effects on mortality, with sickle bar mowers resulting in 50% lower mortality rates than rotary mowers and other models. However, mortality due to crushing by tractor tires may be as high as 46%, independent of blade type and height. Introduction Early-successional habitat and agricultural fields are important foraging and nesting resource areas for Glyptemys insculpta Le Conte (Wood Turtle) and Terrapene carolina carolina L. (Eastern Box Turtle) (Compton et al. 2002, Dodd et al. 2006, Nazdrowicz et al. 2008, Saumure 2004). Recent studies suggest that population densities of Wood and Eastern Box Turtles may be higher where early-successional habitat exists in close proximity to other important habitat features (Jones 2009, Nazdrowicz et al. 2008). Natural succession of earlysuccessional and fallow agricultural fields will in time reduce the availability of habitat types critical for nesting, foraging, and thermoregulation. This local scarcity may force turtles to travel longer distances to find similar habitat elsewhere, which in some cases will put them at greater risk of mortality caused by vehicular traffic. This increase in risk may be especially true of gravid females seeking appropriate nesting habitat (Ashley and Robinson 1996, Fowle 1996, Gibbs and Steen 2005, Haxton 2000). Therefore, maintaining these habitat types is of great importance. Rates of turtle mortality due to mowing of early-successional fields and agricultural activities may be high. Two recent studies of Wood Turtle, for example, have reported substantially higher rates of mortality resulting from agricultural activities than from automobiles (Jones 2009, Saumure 2004, Saumure et al. 2007). Nazdrowicz et al. (2008) and Tingley et al. (2009) reported that mowing is the major source of human-induced mortality in Eastern Box Turtles in Delaware and Wood Turtles in Nova Scotia, respectively. These reports stress the importance of determining ways to eliminate or reduce turtle mortality while 1Natural Heritage and Endangered Species Program, Massachusetts Division of Fisheries and Wildlife, 1 Rabbit Hill Road, Westborough, MA 01581. 2Current address - Department of Environmental Conservation, University of Massachusetts, Amherst, MA 01003. *Corresponding author - lori.erb@state.ma.us. 490 Northeastern Naturalist Vol. 18, No. 4 actively creating and maintaining early-successional habitat, particularly within key conservation areas for Wood and Eastern Box Turtles. In a series of controlled field experiments, we examined whether mortality rates differ among mowing equipment types and blade heights, and calculated the proportion of agriculture-related mortalities resulting from crushing by tractor tires. Methods Study site Experiments were conducted within an early-successional field in Hampshire County, MA. The site is relatively flat and comprises a fallow hayfield currently managed for grassland wildlife. No mowing occurred during the growing season prior to our experiments. Experiment 1 Exploratory experiments were performed to determine which mowing equipment has the greatest impact on turtle mortality. As a means of estimating potential turtle mortality, we deployed acorn squash halves as non-living turtle models. To capture the full potential of mortality risk, we cut each acorn squash to 7.9 ± 0.5 cm in height, the largest known height of Wood Turtles in Massachusetts (Jones 2009), the taller of the two focal species. Squash weights ranged from 368–538 g and were similar to the average weights of Eastern Box Turtles in Massachusetts, the lighter of the two turtle species (Willey 2010). As the lighter species, Eastern Box Turtles are more susceptible to the possible suction generated by rotary mowers. Before each trial, we measured ground to blade height at several locations along the width of the blade, and these values were averaged. Models were placed arbitrarily within approximately a 1- by 2-m area within the mower blade path. We placed turtle models in front of the mower, and between the tractor tires, and the mower was subsequently driven over the models. The area varied slightly in accordance with variation in the mower head width, which was 1.2–2.1 m. Models were retrieved and examined for damage. We tested four types of mowers: 1) 2-m sickle bar mower of unknown make and model, 2) John Deere 605 rotary mower, 3) Fecon BH99H-1 mulching head mower, and 4) Bobcat FC200 flail mower. A sickle bar mower is a long, thin, straight rectangular cutterbar that operates on the same principle as hair clippers, with both stationary and oscillating blades. Modern sickle bars have guide bars to avoid hitting hard objects. The mulching head mower is a rotating cylinder with large teeth. Rotary mowers have a shaft with two rotating cutting blades. The flail mower is a long horizontal tube with rows of rotating or freeswinging cutting blades. The sickle bar mower was tested at a 10-cm height with ten models. A mulching head mower was tested at 10- and 15-cm blade heights with four and eight models, respectively. A rotary mower was tested at 10 and 15 cm with four models each, and 19-cm blade height with eight models. The flail mower was tested at 15-cm blade height with five models. The flail mower has a fixed blade, with a 2011 L. Erb and M.T. Jones 491 non-adjustable guide bar that rests on the ground to protect the cutting unit from hitting the substrate and protruding rocks. Jones (2009) and Saumure and Bider (1998) reported that Wood Turtles occasionally survive minor injuries sustained from mowers. Turtles with healed dents, scrapes, gouges, and puncture wounds on the carapace consistent with mower injuries are frequently observed in the wild in Massachusetts (Jones 2009; M.T. Jones, unpubl. data). A 1-cm-deep wound on the dorsal surface of the carapace could feasibly damage the lungs, depending on the width of the cut and placement. For this reason, we divided turtle models hit by the mower blades into two categories: 1) injured and presumed to have lived, if a model had a shallow cut <1 cm deep or tire-related damage that was within 2.5 cm of the outer edge of the shell and 2) dead, if cut ≥1 cm deep, or if indented or compressed >2 cm in the central portion of the model (≥2.5 cm from edge). Mower tires, rather than tractor tires, may cause indentation or compression but not crush the turtle models, because less pressure is exerted on the model. While we do not know the psi required to crush a turtle shell, we assume that if a live turtle is run over by the tractor, which weighs between 4775–11,965 lbs. (www.TractorData.com), it would not survive. Alternatively, if the model was only nicked by the tire, we assumed it would survive. Experiment 2 We examined the relationship between blade height and mortality of turtles in more detail using a rotary mower, because experiment 1 indicated a relatively high mortality risk to turtles and it is a popular mower among farmers. Rotary mowers were also found to cause the majority of agricultural-related mortalities for a population of Wood Turtles in Canada (Saumure et al. 2007). We tested 14 turtle models in each of three treatments: 1) 10-cm blade height, 2) 15-cm blade height, and 3) no mower. We also tested the effects of the tractor tires in treatment three. In this experiment, we used a 1986 Ford 540 tractor—a typical representative tractor—and the rotary mower from experiment 1. To estimate potential turtle mortality, we deployed cabbage halves as non-living turtle models. We assumed that both cabbages and acorn squash are adequate model material to indicate whether the model was run over by a tire, nicked by a tire, or cut by a mower blade. Model heights ranged from 6.9–8.0 cm (mean = 7.8 cm), with most >7.6 cm, and widths ranged from 13–15 cm; these values are on the high end of the average range of the height and width for Wood Turtles in Massachusetts (Jones 2009). Turtle model weights ranged from 453–500 g, within the typical range of weight for an Eastern Box Turtle in Massachusetts (Willey 2010). We flagged off three 20-m by 20-m treatment areas and used random GPS coordinates, generated by the True Random Number Service (Haahr 1998), to place 14 cabbage halves in each of the three plots. We measured blade heights from the ground at several locations along the width of the blade and averaged the values. In treatment areas 1 and 2, we mowed an overgrown fallow hayfield, with the blade set to an average height of 10 cm and 15 cm, respectively. We drove the tractor through treatment area 3 with no mower head attached, 492 Northeastern Naturalist Vol. 18, No. 4 as a control to test for damage due to the effect of tractor tires alone. In each treatment area, we mowed from the outer edge toward the center in a circular fashion, as is typical practice for most farmers and land managers. Treatment area 2 was mowed first, followed by treatment area 1, and then treatment area 3. Turtle models were retrieved after all treatment areas were mowed. Damaged turtle models were divided into two categories following the same criteria as described above in Experiment 1. We calculated the percent turtle mortality and used the VassarStats software program (Lowry 2008) to run a 2-by-3 contingency table test to compare expected to the observed values. Results Experiment 1 Sickle bar mowers caused substantially less mock mortality than other mower types, and raising the blade height to ≥15 cm decreased the risk of mortality with the mulching head mower and to a lesser degree with the rotary mower (Table 1). No mortalities were observed during any of the trials with the sicklebar mower, all at a fixed, 10-cm blade height. Mortality from the rotary mower was ≥37.5% higher at 10-cm than at 15- and 19-cm blade heights. Mortality from the mulching head mower was higher at 10-cm than 15-cm blade height. Mortality from the flail mower set at a 15-cm height was 100%. No non-lethal injuries were observed in any of the treatment categories. Experiment 2 Results of treatment area 1 were 5 mortalities, 2 non-lethal injuries, and 7 uninjured. In treatment 2 there were 7 mortalities, 3 non-lethal injuries, and 4 uninjured. For treatment 3 there were 5 mortalities, 2 non-lethal injuries, 6 uninjured, and one was not recovered (Table 1). Mortality was high in all treatments, and no differences were observed between treatments (χ2 = 0.62, P = 0.733, and Cramer’s V = 0.123). Tires posed the greatest risk, with 80 and 100% of mortalities and 33 and 50% of non-lethal injuries in treatment 1 and 2, respectively, attributed to tires. All 5 mortalities and 2 non-lethal injuries in treatment 3 were due to tractor tires since no mower was used. Table 1. Percent mortality with different mowing equipment and at different blade heights. Sample sizes are in parentheses. Blade height Mower equipment 10 cm 15 cm 19 cm Tires-no mower Experiment 1 Sicklebar mower 0 (10) Rotary mower (brush hog) 50 (4) 0 (4) 12.5 (8) Mulching head mower (bull hog) 100 (4) 0 (8) Flail mower 100 (5) Experiment 2 Rotary mower (brush hog) 50 (14) 36 (14) 38 (13) 2011 L. Erb and M.T. Jones 493 Discussion Our results remind us of the significant risk that turtles face from tires of heavy equipment, regardless of the type of mower and blade height used. Results showed an overall 46% mortality rate due to the tractor tires alone. The rear tires of the tractor used in these experiments account for approximately 37% of the width of the tractor, which is defined here as the distance between outer edges of the tires. Furthermore, the typical path taken when mowing a field includes some overlapping of area covered in consecutive passes across a field. Taken together the area covered by tires corresponds well with the mortality risk observed in this study. However, this may be an overestimation since turtle behavior has not been addressed. Some turtles will likely flee an area as a tractor approaches or mowing commences (Saumure et al. 2007), while others will retract into their shell. Survivorship may also be higher in moist soils where turtles can sink into the mud and be cushioned from crushing. Additionally, turtles are incredibly resilient, and it may be possible that our threshold for non-lethal injuries needs to be refined (i.e., ≥1 cm cut, or >2 cm dent). Regardless, our results indicate that the most effective way to reduce the risk of turtle mortality in early-successional habitat and agricultural fields is to keep heavy equipment out of these areas, to the extent possible, during the turtle active season. Similar patterns were observed in Wood and Eastern Box Turtle populations in two other studies (Jones 2009, Nazdrowicz et al. 2008). The Wood and Eastern Box Turtle activity season in Massachusetts runs from approximately April 1 through November 1, with a peak in use of earlysuccessional habitat from May 1 through September 15 (Jones 2009, Willey 2010). Saumure et al. (2007) observed agricultural mortalities during June– August in a Wood Turtle population in Canada, and found no significant difference in mortalities between sexes. However, female Eastern Box Turtles may disproportionately use early-successional habitat in May and June. This time period also encompasses the majority of telemetry locations in early successional habitat for Eastern Box Turtle studies in Connecticut and Massachusetts (Quinn 2008, Willey 2010) and is likely the most critical time to avoid mowing in Eastern Box Turtle habitat. The critical avoidance dates may differ by geographic location, with more northern latitudes having shorter turtle seasons, and they can also differ from year to year, depending upon both air and substrate temperatures (Dodd 2001). Conservation land managers should seriously consider alternatives to mowing where Wood or Eastern Box Turtles have been identified as priority species. In some cases hand-cutting and herbicide applications may be the best alternative to control woody plants and avoid impacts to turtles, and can be used in concert with off-season mowing. Another method of controlling succession is to use grazing animals. Sheep, goats, and even cattle at low densities are likely to be more compatible, than mowing, with turtles and other wildlife species that use early-successional habitat, although care should be taken that livestock do not degrade bank habitat or water quality. Grazing is already becoming a common 494 Northeastern Naturalist Vol. 18, No. 4 tool used for the control of succession and invasive plants for other turtle species, such as Glyptemys muhlenbergii (Schoepff) (Bog Turtle) (Tesauro and Ehrenfeld 2007, USFWS 2006). In situations where mowing is necessary during the turtle active season, such as for agricultural fields, mower style and blade height should be considered carefully. Mower type may affect the mortality risk to Wood and Eastern Box Turtles. Our data show that certain mower types are likely to cause disproportionately higher rates of turtle mortality. The use of rotary and flail mowers results in increased risk to turtles compared with older-style sickle bar mowers. The flail mower has a guide bar that rolls along the ground and carries the entire weight of the mower head (1630 kg according to Bobcat equipment specifications). As a consequence, with use of the flail mower, our data suggests nearly 100% mortality. Our recommendation is that flail mowers not be used during the turtle active season on conservation lands in Wood or Eastern Box Turtle habitat unless the guide bar is raised and the safety switch disconnected. Additional types of mowers should be tested, such as rotary mowers with a corkscrew-shaped blade. In experiment 1, we observed a decreasing trend in mortality from the 10- to 15-cm blade height with the rotary mower. However, experiment 2 indicated that raising the rotary mower blade height to 15 cm did not significantly decrease turtle mortality. Treatments 1 and 2 in experiment 2 indicated an 8–17% mortality risk due to mower blades alone, excluding tire-related mortality. Our sample sizes were relatively small for detecting small trends, and the effect may have been diluted by the large number of mortalities due to tires. It may be that with a larger sample size this trend would prove to be significant. Alternatively, raising the blade to greater than 15 cm may be required to reduce mortalities due to mower blades. Experiment 1 also indicated a decreasing trend in mortality (100% to 0%) for the mulching head mower when raised above 10 cm. Raised blade heights can have benefits for farmers by reducing blade wear (Rider and Barr 1987), reducing soil erosion, and promoting soil moisture retention, which can increase yield in subsequent harvests (Sharp et al. 1995, Smith 1978). Mowing to maintain early succession fields for conservation reasons should only require multi-year rotations (e.g., mowing once every 2–3 years), although this will not prevent the invasion of forbs. If mowing is combined with another maintenance method such as chemical control of invading woody plants, mowing during the turtle active season may not be necessary. If periodic mowing is the sole method used for maintenance, woody plant cover on the site will likely increase over the long term, and mowing during the active season will be necessary to maintain the area as early-successional habitat. In some years, very frequent mowing may be required to reduce woody plant abundance. Wood and Eastern Box Turtles heavily use the field-forest ecotone, and are usually located within 30–45 m of the field edge (e.g., Tingley et al. 2009, Willey 2010). Therefore, it may be beneficial to leave an unmowed strip of early-successional vegetation along the edge of the field during the turtle active season, particularly along edges near overwintering habitat (i.e., streams for Wood Turtle and forest for Eastern Box Turtle). This edge area could be mowed only during 2011 L. Erb and M.T. Jones 495 the turtle inactive season. Saumure et al. (2007) observed a disproportionately high number of injuries on the right sides of Wood Turtle shells. They speculated that the cause was due to the general practice of mowing field margins first and in a counter-clockwise direction. Turtles are hit as they try to flee the field. They further suggested that turtles may learn from previous experience to escape these areas during mowing events. If this phenomenon is widespread, mowing from the center portion of a field outward would likely decrease mortality risk for fleeing animals. Future research should seek to determine the effect on turtle mortality risk of leaving an unmowed edge, mowing directionality (mowing from the center of the field outward), time of day, and mower speed. Acknowledgments We would especially like to thank D. Fuller and J. Wright for their habitat management expertise and use of mowing equipment. We thank J. Chayes, G. Murphy, L. Willey, and S. Haggerty for their field assistance. Special thanks to P. Sievert of the US Geological Survey Cooperative Fish and Wildlife Research Unit at the University of Massachusetts for statistical advice. Thank you to B. Lohr for review of an early draft. This manuscript was also substantially improved by four anonymous reviewers. Literature Cited Ashley, P.E., and J.T. Robinson. 1996. Road mortality of amphibians, reptiles and other wildlife on the Long Point Causeway, Lake Erie, Ontario. Canadian Field Naturalist 110:403–412. Compton, B.W., J.M. Rhymer, and M. McCollough. 2002. Habitat selection by Wood Turtles (Clemmys insculpta): An application of paired logistic regression. Ecology 83:833–843. Dodd, C.K. 2001. North American Box Turtles: A Natural History. University of Oklahoma Press, Norman, OK. 231 pp. Dodd, C.K., A. Ozgul, and M.K. Oli. 2006. The influence of disturbance events on survival and dispersal rates of Florida Box Turtles. Ecological Applications 16:1936–1944. Fowle, S.C. 1996. Effects of roadkill mortality on the Western Painted Turtle (Chrysemys picta belli) in the Mission Valley, western Montana. Pp. 205–223, In E.L. Evink, P. Garrett, D. Zeigler, and J. Berry (Eds.). Trends in Addressing Transportation-Related Wildlife Mortality, Proceedings of the Transportation Related Wildlife Morality Seminar. FL-ER-58-96, Florida Department of Transportation, Tallahassee, FL. Gibbs, J., and D.A. Steen. 2005. Historic trends in sex ratios of turtle populations in the United States. Conservation Biology 19:552–556. Haahr, M. 1998–2011. The true random number service. Available online at http://www. random.org. Accessed June 2008. Haxton, T. 2000. Road mortality of Snapping Turtles, Chelydra serpentina, in central Ontario during their nesting period. Canadian Field Naturalist 114:106–110. Jones, M.T. 2009. Spatial ecology, population structure, and conservation of the Wood Turtle (Glyptemys insculpta) in central New England. Ph.D. Dissertation. University of Massachusetts, Amherst, MA. 233 pp. Lowry, R. 2008. VassarStats. Available online at http://faculty.vassar.edu/lowry/newcs. html. Accessed 22 March 2009. 496 Northeastern Naturalist Vol. 18, No. 4 Nazdrowicz, N.H., J.L. Bowman, and R.R. Roth. 2008. Population ecology of the Eastern Box Turtle in a fragmented landscape. Journal of Wildlife Management 72:745–753. Quinn, D.P. 2008. A radio–telemetric study of the Eastern Box Turtle (Terrapene carolina carolina) home range, habitat use, and hibernacula selection in Connecticut. M.Sc. Thesis. Central Connecticut State University, New Britain, CT. 84 pp. Rider, A.R., and S.D. Barr. 1987. Hay and Forage Harvesting: Fundamentals of Machine Operation. Deere and Co., Molin, IL. 213 pp. Saumure, R.A. 2004. Spatial ecology and conservation of the North American Wood Turtle (Glyptemys insculpta) in a fragmented agri-forested landscape. Ph.D. Dissertation. McGill University, Montréal, QC, Canada. 134 pp. Saumure, R.A., and J.R. Bider. 1998. Impact of agricultural development on a population of Wood Turtles (Clemmys insculpta) in southern Québec, Canada. Chelonian Conservation and Biology 3:37–45. Saumure, R.A., R.D. Titman, and T.B. Herman. 2007. Effects of haying and agricultural practices on a declining species: The North American Wood Turtle, Glyptemys insculpta. Biological Conservation 135:581–591. Sharp, W.C., D.L. Schertz, and J.R. Carlson, Jr. 1995. Harvesting and storage. Pp. 243–262, In R.F. Barnes, D.A. Miller, and C.J. Nelson. (Eds.), Forages, the Science of Grassland Agriculture. Vol. II. Iowa State University Press, Ames, IA. 357 pp. Smith, D., 1978. Forage Management in the North. Kendall/Hunt Publishing, Dubuque, IA. 237 pp. Tesauro, J., and D. Ehrenfeld. 2007. The effects of livestock grazing on the Bog Turtle (Glypemys [= Clemmys] muhlenbergii). Herpetologica 63(3):293–300. Tingley, R., D.G. McCurdy, M.D. Pulsifer, and T.B. Herman. 2009. Spatio-temporal differences in the use of agricultural fields by male and female Wood Turtles (Glyptemys insculpta) inhabiting an agri-forest mosaic. Herpetological Conservation and Biology 4:185–190. US Fish and Wildlife Service (USFWS). 2006. Effects of the implementation of habitat restoration projects on the Northern population of Bog Turtle. US Fish and Wildlife Service Technical Report, Hadley, MA. 39 pp. Willey, L.L. 2010. Ecology of Eastern Box Turtles in central Massachusetts. Ph.D. Dissertation. University of Massachusetts, Amherst, MA. 200 pp.