2007 NORTHEASTERN NATURALIST 14(3):415–424 Ozone Injury to Plants within the Seney National Wildlife Refuge in Northern Michigan Donald D. Davis* Abstract - Annual field surveys were conducted from 1999–2004 within the Seney National Wildlife Refuge in northern Michigan to determine if ambient ozone levels at this remote location were great enough to injure refuge vegetation. Ozone injury was observed on sensitive bioindicator plants during each survey year; however, the incidence (percentage) of plants exhibiting symptoms was low and varied among species and years. Ozone-induced symptoms occurred on Sambucus canadensis (American elder), Prunus serotina (black cherry), Asclepias syriaca (common milkweed), and Apocynum androsaemifolium (spreading dogbane). The most sensitive species was spreading dogbane. In addition, ozone injury was observed on a viburnum species, tentatively identified as Viburnum nudum var. cassinoides (withe-rod). Ambient ozone has been monitored since 2002 at an EPA monitoring site within the refuge. Cumulative SUM60 ozone levels (ppb-hrs) by the end of August for each survey year were greatest in 2003, followed by 2002, and least in 2004. The annual incidence of ozone injury for the 3 years was not directly related to level of ambient ozone, but was likely confounded by environmental factors such as drought. Based on the 2004 survey, the threshold level of SUM60 ozone needed to induce visible symptoms on sensitive vegetation in this remote refuge is close to 5000 ppb-hrs. Introduction The Seney National Wildlife Refuge (Seney NWR), located in the Upper Peninsula (UP) of northern Michigan (Fig. 1), was established in 1935 as a refuge and breeding ground for migratory birds and wildlife. The refuge is located near the villages of Seney and Germfask, approximately 130 km northwest of the Mackinac Bridge that joins the UP with lower Michigan. The refuge contains 38,600 ha that are managed to provide quality habitat for a diversity of waterfowl and wildlife species. In 1970, Congress conferred wilderness status on part of the Seney NWR, naming it the “Seney Wilderness,” and designating it as a Class-I air-quality area, affording the refuge further protection under the Clean Air Act as amended in 1977 (US Congress 1977). In that act, the US Congress gave federal land managers of Class-I areas responsibility to protect all air-quality related values (AQRVs) in these areas. AQRVs include vegetation, wildlife, water, soils, visibility, and cultural resources. By federal law, AQRVs in Class-I areas must be protected from deterioration. However, despite this special protection, significant levels of ambient ozone impinge on many Class-I air-quality areas in the eastern US, adversely affecting AQRVs (Lefohn and Manning 1995, *Department of Plant Pathology and Penn State Institutes of the Environment, 211 Buckhout Laboratory, The Pennsylvania State University, University Park, PA 16802; ddd2@psu.edu. 416 Northeastern Naturalist Vol. 14, No. 3 Manning et al. 1996). This paper is the third in a series dealing with ozone injury on vegetation within our national wildlife refuges (Davis 2007, Davis and Orendovici 2006). Refuge characteristics (adapted from refuge brochures) The Seney NWR, located in the Great Manistique Swamp, contains natural wetlands, grasslands, and forests. Pinus strobus L. (white pine) once dominated the forested uplands of the refuge, but these were logged during the late 1800s and the area burned many times. Controlled burning, a habitat management treatment to encourage those plant species and communities favored by fire, is practiced today within the refuge. The upland portions of the refuge are now covered with second- and third-growth stands of Populus spp. (aspen), Pinus banksiana Lamb. (jack pine), Pinus resinosa Ait. (red pine), and white pine. Wet areas are dominated by thick stands of Alnus rugosa (Du Roi) Spren. (speckled alder) and Picea mariana (Miller) BSP (black spruce), with Typha latifolia L. (common cattail) growing in the open areas. Strangmoor bogs, or “string bogs,” sub-arctic formations rare at this latitude, characterize more than half of the Wilderness Area. The string bogs are dominated by sparse stands of aspen, jack pine, and red pine, fringed with thick clumps of speckled alder, black spruce, and occasional Larix laricina (Du Roi) K. Koch (tamarack). A series of manmade dikes, resulting in 21 artificial pools and 750 islands, were built within the refuge to encourage use by waterfowl. Because of this varied habitat, the refuge is now home to more than 400 species of plants. The most predominant, ozone-sensitive bioindicator species (US DOI 2003) found within the Seney NWR is common Asclepias syriaca L. (milkweed); many thousands grow on the open, exposed, sandy dikes of the refuge. Populus species were also common in the refuge, but those growing near the survey sites were often infected with leafspots, confounding symptom evaluation. Other bioindicators in the refuge include Sambucus canadensis L. (American elder), Prunus serotina Ehrh. (black cherry), and Apocynum androsaemifolium L. (spreading dogbane). These sensitive species generally respond to ozone by producing a characteristic adaxial “stipple,” a symptom first described by Richards et al. (1958). Stipple on the leaves of ozone-sensitive plants is a reliable diagnostic symptom that can be used to evaluate the incidence of ozone injury during field surveys within wildlife refuges (Davis 2007, Davis and Orendovici 2006). Ambient ozone levels Ground-level ozone is the most important plant-damaging air pollutant in eastern North America, where elevated ozone concentrations occur annually (Comrie 1994, Coulston et al. 2003). These elevated ozone levels can travel long distances and are capable of injuring native plants in remote locations such as wilderness areas (Lefohn and Manning 1995, Manning et al. 1996) and wildlife refuges (Davis 2007, Davis and Orendovici 2006). Prior to the first survey in 1999, the 1997–1998 ozone data from the nearest EPA 2007 D.D. Davis 417 monitoring site (EPA AIRS site #55-029-0004) were examined. Data from this monitor, located approximately 120 km southwest of the refuge at Ellison Bay, WI, revealed that elevated levels of ambient ozone had occurred in both 1997 and 1998. Assuming ozone levels monitored at Ellison Bay were comparable to those occurring at the Seney NWR, located 120 km from the monitor, ozone injury was likely to occur to sensitive vegetation within the refuge. The ambient ozone levels monitored at Ellison Bay, along with the presence of ozone-sensitive bioindicators (US DOI 2003) within the refuge, provided impetus for conducting this survey. In 2002, an ozone monitor (EPA AIRS site #26-153-0001) was established within the boundaries of the Seney NWR. Comparison of 2002– 2004 ambient ozone data from the EPA site at Ellison Bay with data monitored within the Seney NWR revealed that the levels and patterns of ozone were not similar between the two monitoring sites (data not shown). Therefore, only ozone data monitored within the Seney NWR are presented in this paper. The main objective of this study was to determine if ozone injury occurred on vegetation growing in the remote Seney NWR, and if so, to determine the incidence (percentage) of the various species of bioindicator plants exhibiting stipple. A secondary objective was to determine if the incidence of ozone injury was related to ambient ozone levels and soil-moisture stress. This second objective was limited since only 3 years of monitored ozone data were available. To meet these objectives, annual surveys were conducted in the refuge during the 6-year period from 1999–2004. Methods General survey methods were similar to those used within the Brigantine NWR in New Jersey (Davis and Orendovici 2006) and Moosehorn NWR in Maine (Davis 2007). All fieldwork was conducted solely by the author. In 1998, maps of the Seney NWR were examined to select tentative survey sites that had unrestricted air movement and sunlight, criteria for suitable sampling sites (Anderson et al. 1989). Based on the maps, 25 tentative survey sites were selected and visited. Eleven sites did not contain suitable bioindicators and were eliminated, resulting in 14 final sampling sites (Fig. 1). With some modifications, the general location of these 14 sites formed the basis for the 1999–2004 field surveys. Data were not taken each year at each site, depending on the foliar condition of the bioindicator plants. The refuge was surveyed once a year during August of 1999–2004 on the dates shown in Table 1. American elder, black cherry, common milkweed, and spreading dogbane were examined at each site (Table 1). The number of plants examined within a species varied from year to year, mainly due to insect infestations that rendered foliage unusable in some years. The total number of individual plants within each bioindicator species, as well as the number of plants exhibiting stipple, were counted. Ozone-induced stipple 418 Northeastern Naturalist Vol. 14, No. 3 was simply recorded as present or absent on individual bioindicator plants; severity of injury was not evaluated. Incidence (percentage) was calculated as the number of symptomatic plants/number of plants examined for each bioindicator species. A general linear model (GLM) was used to determine if significant (p = 0.05) differences in incidence of ozone injury occurred among species across all 6 years (Minitab 2003). Since species (US DOI 2003), ozone level (Hildebrand et al. 1996), and drought stress (Showman 1991, Yuska et al. 2003) influence the level of ozone injury, we investigated the relationship between these factors using binomial logistic regression (Davis and Orendovici 2006). Annual ambient ozone levels were expressed as SUM60 (ppb-hrs), the cumulative sum of hourly ozone concentrations equaling or exceeding 60 ppb during the growing-season. Annual soilmoisture stress as of July 31 of each survey year was expressed as the Palmer Drought Severity Index (PDSI) (Palmer 1965), using data from the web at http://climvis.ncdc.noaa.gov/cgi-bin/ginterface. Predictive abilities of binomial regression analyses were severely limited since only 3 years of ambient ozone data were available. To illustrate the pattern of ambient ozone levels that impinge upon the refuge during the growing season, SUM60 ozone levels for May through September of each survey year were graphed (Fig. 2). Figure 1. Location of 14 survey sites (circles with site numbers) in the Seney National Wildlife Refuge within the Upper Peninsula of northern Michigan (map courtesy of the US Fish and Wildlife Service). 2007 D.D. Davis 419 Results and Discussion Symptom description and incidence Although not all bioindicator species were present at all survey sites, American elder, black cherry, common milkweed, and spreading dogbane were the most common bioindicators in the Seney NWR, and all exhibited classic dark, adaxial stipple as first described by Richards et al. (1958) and illustrated by Skelly (2000). These symptoms were similar to those observed on ozone-sensitive bioindicator plants in wildlife refuges in New Jersey and Maine (Davis 2007, Davis and Orendovici 2006). Chlorosis, reddening, and premature defoliation also occurred occasionally on bioindicators, but were not recorded since such symptoms can be elicited by other stresses such as heat, moisture stress, and early onset of autumnal coloration/senescence (Orendovici et al. 2003). Overall, the incidence of stipple within this remote refuge in northern Michigan was very light, except on spreading dogbane. The incidence of ozone injury on black cherry and common milkweed was much less than that reported in the more polluted areas of southwestern Michigan (Bennett et al. Table 1. Summary of observations made during the 1999–2004 surveys at the Seney National Wildlife Refuge. Numbers in table refer to number of plants exhibiting ozone-induced stipple as compared to the total number of plants evaluated for that bioindicator; data also expressed as percentages. Black American Common Spreading SUM and Year (survey date) cherry elder milkweed dogbane average 1999 (Aug 11–15) Number plants examined 49 20 7068 47 7184 Number plants injured 1 0 18 14 33 Percentage 2.04% 0.00% 0.25% 29.79% 0.46% 2000 (Aug 24–27) Number plants examined 22 35 7060 325 7442 Number plants injured 0 0 39 99 138 Percentage 0.00% 0.00% 0.55% 30.46% 1.85% 2001 (Aug 20–23) Number plants examined 62 42 5416 480 6000 Number plants injured 1 5 64 111 181 Percentage 1.61% 11.90% 1.18% 23.13% 3.02% 2002 (Aug 18–22) Number plants examined 36 20 3514 108 3678 Number plants injured 0 0 51 15 66 Percentage 0.00% 0.00% 1.45% 13.89% 1.79% 2003 (Aug 20–23) Number plants examined 90 48 2228 50 2416 Number plants injured 1 2 44 3 50 Percentage 1.11% 4.17% 1.97% 6.00% 2.07% 2004 (Aug 15–17) Number plants examined 50 21 2055 51 2177 Number plants injured 0 0 33 3 0 Percentage 0.00% 0.00% 1.61% 5.88% 0.00% Average 0.97% 3.76% 0.91% 23.09% 1.62% 420 Northeastern Naturalist Vol. 14, No. 3 2006). The overall ozone-sensitivity ranking of the bioindicators, based on mean percentage of individuals exhibiting stipple (in parentheses) across all years was: spreading dogbane (23.1%) > American elder (3.8%) > black cherry (1.0%) > common milkweed (0.9%) (Table 1). However, only spreading dogbane exhibited significantly (p = 0.05) greater incidence values, as compared to the other species, which had statistically similar incidence ratings. The incidence of ozone injury on spreading dogbane ranged from approximately 6% in 2004 to 30% in 1999 and 2000. However, these high values were mainly due to one very sensitive clone growing at one site. By 2003 and 2004, this clone was being overgrown by competing vegetation, complicating evaluation of temporal patterns of ozone injury on this species. Nevertheless, the high sensitivity of spreading dogbane also has been reported in Maine within both the Acadia National Park (Kohut et al. 2000) and the Moosehorn NWR (Davis 2007). In addition, Bergweiler and Manning (1999) reported that flowering and reproduction of spreading dogbane was inhibited by ambient ozone in Massachusetts. These observations indicate that spreading dogbane is a sensitive bioindicator and useful for detecting elevated levels of ozone. However, both Davis (2007) and Kohut et al. (2000) observed that the leaves of spreading dogbane became chlorotic and spotted, and began to senesce by late summer. These foliar disorders limit the usefulness of spreading dogbane as a bioindicator. Figure 2. Sum of hourly ozone concentrations equaling or exceeding 60 ppb (SUM60, ppb-hrs) recorded from May 1 to September 31, 2002–2004, at a monitoring location (EPA AIRS site #26-153-0001) within the Seney National Wildlife Refuge. 2007 D.D. Davis 421 American elder, black cherry, and common milkweed were also injured by ambient ozone within the Seney NRW. These species are known ozone bioindicators (US DOI 2003), but American elder and black cherry plants were found only at widely scattered locations in the refuge, making them less useful. Also, although black cherry has been reported to be sensitive to ozone (Bennett et al. 2006, Davis and Skelly 1992, Davis et al. 1981), this species exhibited a very low incidence (less than 1%) of ozone injury within the Seney refuge. This low level of ozone injury was similar to the incidence of ozone injury observed on black cherry reported for wildlife refuges in New Jersey (Davis and Orendovici 2006) and Maine (Davis 2007). The reason for low level of injury on black cherry observed during these surveys is unknown, but may be due to genetic differences in ozone sensitivity among different populations of black cherry. Common milkweed is known to be very sensitive to ozone (US DOI 2003) and occurred in very high numbers on dikes throughout the Seney NWR. More than 27,000 milkweed plants were examined during these surveys, illustrating that milkweed was the most common bioindicator in the refuge. Unlike spreading dogbane, injured milkweed plants appeared to be widely scattered among the population. One species of viburnum, tentatively identified as Viburnum nudum L. var. cassinoides (L.) Torr. & Gray (witherod), also exhibited classic stipple, but occurred infrequently. This species also exhibited classic ozone stipple at the Moosehorn NWR in Maine (Davis 2007). If stipple can be induced on withe-rod under controlled, ambient ozone dosages to prove cause and effect, withe-rod may prove useful as an ozone bioindicator in eastern North America. Ambient ozone levels The cumulative SUM60 ozone levels (ppb-hrs) monitored at the EPA site within the Seney NWR were: 22,200 (2003) > 12,685 (2002) > 5171 (2004). Ozone accumulation during the growing season followed fairly similar patterns from year to year, gradually increasing from May to the end of the summer, and becoming fairly constant in September (Fig. 2). However, the pattern was slightly different in 2002, when the ozone levels rose quickly early in the growing season. In comparison to other wildlife refuges in eastern and central US, ambient ozone levels monitored within the remote, northern Seney refuge were very low. There are few industrial or urban areas upwind from the Seney NWR, and these areas are relatively unpolluted and produce few ozone precursors. Ozone levels at Seney were less than at the Moosehorn NWR in northeastern Maine, as estimated by the nearest EPA monitor (EPA AIRS site #23-009- 0102) at Acadia National Park, ME. At Acadia, the cumulative SUM60 ozone levels ranged from approximately 26,000–33,000 ppb-hrs by late summer in 1999–2003 (Davis 2007). It is likely that the higher ozone levels at Acadia were related to ozone precursors forming in the megalopolis along the eastern seaboard southwest of Maine. In New Jersey, the cumulative SUM60 ozone levels within the Forsythe NWR (EPA AIRS site #34-001-0005), often 422 Northeastern Naturalist Vol. 14, No. 3 exceeded 40,000 ppb-hrs by the end of the summer, and actually exceeded 70,000 ppb-hrs during a high-ozone year (Davis and Orendovici 2006). These high levels were also likely influenced by ozone precursors forming in the megalopolis along the eastern seaboard. High SUM60 ozone levels, also exceeding 70,000 ppb-hrs by late summer, have been reported as well near the Mingo NWR in southeastern Missouri, as estimated from the nearest ozone monitor (EPA AIRS Site #29-186-0005). Relationship of incidence to ambient ozone and drought levels Although the incidence of ozone injury was not related to ambient ozone levels for 2002–2004, the analyses were severely limited by having only 3 years of ambient ozone data. However, Eckert et al. (1999) also reported the lack of a relationship between ambient ozone levels in Acadia National Park (Maine) and ozone injury, and attributed the lack of correlation to the confounding effects of moisture stress on stomatal functioning and resultant gas uptake. In the current study, the incidence of ozone injury on spreading dogbane, but not other species, was weakly, but significantly, related to the drought index (PDSI) according to the Pearson goodness-of-fit statistic (Minitab 2003). However, this relationship was too weak to be used for predictive purposes. Nevertheless, the threshold SUM60 ozone level needed to induce stipple on sensitive plants within the Seney refuge is likely 5000 ppb-hrs under the environmental conditions of these surveys. If this threshold level is correct, undetected ozone injury might have occurred within the refuge on plants in a sensitive stage of development as early as June in 2002 and 2003 (Fig. 2). Since most ozone injury surveys are conducted in mid- to late August, such early-season injury might no longer be apparent at survey time, due to abscission of symptomatic leaves and reduced visibility resulting from new growth. Also, high ozone levels in June may cause injury on those plant species that emerge and complete their life cycles early in the growing season, such as late-spring or early-summer ephemerals (Davis and Orendovici 2006). Ozone-induced stipple occurred on at least one bioindicator species within the refuge during each survey year, revealing that bioindicators in this remote area can be used as predictive tools as indicators of environmental deterioration. Ambient ozone levels at this remote refuge are likely great enough during most years to cause injury on refuge plants, including those growing in the Seney Class-I wilderness area. The US Fish and Wildlife Service can utilize the results of these surveys when making air-quality management decisions, including those related to review of Prevention of Significant Deterioration permits, and such data can be used to strengthen our National Ambient Air Quality Standards for ozone (US EPA 1996). Acknowledgments The author gratefully acknowledges receipt of financial support and ambient-ozone datasets from the US Fish and Wildlife Service, Air Quality Branch, Denver, CO. 2007 D.D. Davis 423 Literature Cited Anderson, R.L., C.M. Huber, R.P. Belanger, J. Knighten, T. McCartney, and B. Book. 1989. Recommended survey procedures for assessing on bioindicator plants in Region-8 Class-1 Wilderness areas. USDA Forest Service, Forest Pest Management, Asheville Field Office, Asheville, NC. Report 89-1-36. 6 pp. Bennett, J.P., E.A. Jepson, and J.A. Roth. 2006. Field responses of Prunus serotina and Asclepias syriaca to ozone around southern Lake Michigan. Environmental Pollution 142:354–366. Bergweiler, C.J., and W.J. Manning. 1999. Inhibition of flowering and reproductive success in spreading dogbane, Apocynum androsaemifolium (L), by exposure to ambient ozone. Environmental Pollution 105:333–339. Comrie, A.C. 1994. A synoptic climatology of surface ozone in rural ozone pollution at three forest sites in Pennsylvania. Atmospheric Environment 28:1601–1614. Coulston, J.W., G.C. Smith, and W.D. Smith. 2003. Regional assessment of ozone sensitive tree species using bioindicator plants. Environmental Monitoring and Assessment 83:117–127. Davis, D.D. 2007. Ozone-induced symptoms on vegetation within the Moosehorn National Wildlife Refuge in Maine. Northeastern Naturalist 14:403–414. Davis, D.D., and T. Orendovici. 2006. Incidence of ozone symptoms on vegetation within a National Wildlife Refuge in New Jersey, USA. Environmental Pollution 143:555–564. Davis, D.D., and J.M. Skelly. 1992. Foliar sensitivity of eight eastern hardwood tree species to ozone. Water, Air, and Soil Pollution 62:269–277. Davis, D.D., D.M. Umbach, and J.B. Coppolino. 1981. Susceptibility of tree and shrub species and response of black cherry foliage to ozone. Plant Disease 65:904–907. Eckert, R., R. Kohut, T. Lee, and K. Stapelfeldt. 1999. Foliar ozone injury on native vegetation at Acadia National Park: Results from a six-year (1992–1997) field survey. National Park Service , Boston, MA. 41 pp. Technical Report NPS/BSO– RNR/NRTR/00–12. Hildebrand, E., J.M. Skelly, and T.S. Fredericksen. 1996. Foliar response of ozone sensitive hardwood tree species from 1991 to 1993 in the Shenandoah National Park, VA. Canadian Journal for Forestry Research 26:658–669. Kohut, R., J. Laurence, P. King, and R. Raba. 2000. Identification of bioindicator species for ozone and assessment of the responses to ozone of native vegetation at Acadia National Park. National Park Service , Boston, MA. Technical Report NPS/BSO–RNR/NRTR/00–13. 126 pp. Lefohn, A.S., and W.J. Manning. 1995. Ozone exposures near Class-I wilderness areas in New Hampshire and Vermont. Atmospheric Environment 29:601–606. Manning, W.J., S.V. Krupa, C.J. Bergweiler, and K.I. Nelson. 1996. Ambient ozone (O3) in three Class-I wilderness areas in northeastern USA: Measurements with Ogawa passive samplers. Environmental Pollution 91:399–403. Minitab Inc. 2003. Quality Plaza, 1829 Pine Hall Road, State College, PA. Orendovici, T., J.M. Skelly, J.A. Ferdinand, J.E. Savage, M.-J. Sanz, and G.C. Smith. 2003. Response of native plants of northeastern United States and southern Spain to ozone exposures: Determining exposure/response relationships. Environmental Pollution 125:31–40 Palmer, W.C. 1965. Meteorological Drought. Research Paper Number 45. US Department of Commerce, Washington, DC. 58 pp. 424 Northeastern Naturalist Vol. 14, No. 3 Richards, B.L., J.T. Middleton, and W.B. Hewitt. 1958. Air pollution with relation to agronomic crops. V. Oxidant stipple of grape. Agronomy Journal 50:559–561. Showman, R.E. 1991. A comparison of ozone injury to vegetation during moist and drought years. Journal Air and Waste Management Association 41:63–64. Skelly, J.M. 2000. Tropospheric ozone and its importance to forests and natural plant communities of the northeastern United States. Northeastern Naturalist 7:221–236. United States Congress (US Congress). 1977. The Clean Air Act as amended August 1977. P.L. 95–95. US Government Printing Office, Washington, DC. United States Department of the Interior (US DOI). 2003. Ozone sensitive plant species on National Park Service and US Fish and Wildlife Service lands: Results of a June 24–25, 2003 workshop, Baltimore, MD. US DOI, Denver, CO. Natural Resource Report NPS/NRARD/NRR-2003/01. 21 pp. United States Environmental Protection Agency (US EPA). 1996. Air-quality criteria for ozone and related photochemical oxidants. Vol. 1 of 3 reports. EPA/600/P- 93/004aF. Research Triangle Park, NC. Yuska, D.E., J.M. Skelly, J.A. Ferdinand, R.E. Stevenson, J.E. Savage, J.D. Mulik, and A. Hines. 2003. Use of bioindicators and passive sampling devices to evaluate ambient ozone concentrations in north central Pennsylvania. Environmental Pollution 125:71–80.