New Population of the Rare Dragonfly Ophiogomphus howei (Odonata: Gomphidae) in Southern Michigan, United States.

Ophiogomphus howei Bromley is a rare North American dragonfly, given a global conservation rank of Vulnerable by NatureServe. This species inhabits localized stretches of a limited number of typically undisturbed, high-quality, forested rivers in two disjunct regions in North America. We describe a new population in between the known ranges from an impaired river in a largely urban watershed in southern Michigan, United States. We also report a previously overlooked specimen from a new location in Pennsylvania, United States, and provide current occurrence and conservation status of the species in North America.

Gomphidae is the second-largest dragonfly (Anisoptera) family in the world; there are 102 species in 17 genera in North America (Ware et al. 2017, Tennessen 2019). Nearly all species inhabit flowing waters, and their aquatic nymphs are highly modified to burrow into substrate, a niche exploited by only one other family of Odonata (Tennessen 2019). Ophiogomphus Selys is a Holarctic genus of 29 species, 20 of which are restricted to the New World (Garrison et al. 2006, Tennessen 2019, Schorr and Paulson 2020). They are small to medium-sized dragonflies that occupy clean streams or rivers, often with rocky or gravelly substrates (Paulson 2011). Ophiogomphus howei Bromley (Pygmy Snaketail) is the smallest Ophiogomphus in North America, with abdomen and hindwing lengths not exceeding 35 and 22 mm, respectively (Needham et al. 2014).

Ophiogomphus howei occurs in two disjunct geographical areas in North America (Fig. 1). A western population is located in the Great Lakes region, centered in the northern half of Wisconsin, extending into Minnesota and the western Upper Peninsula of Michigan; there is a single record in northwestern Ontario. An eastern population occurs in scattered locations from New Brunswick, Canada, to South Carolina, United States, with the westernmost records from Kentucky. Ophiogomphus howei has not been recorded from states between the western and eastern ranges, despite considerable survey efforts (Curry 2001, Olcott 2011, ODS 2019, Tennessen 2019).

Fig. 1.

Range of Ophiogomphus howei. Shaded areas include river systems where O. howei has been found, but not all states within shaded areas have records of O. howei; see Table 1 for summary of occurrence. Symbols represent: Diamond = single Ontario, Canada record; triangle = single Mississippi River record in Minnesota; star = population described in this paper; plus sign = new historical record described in this paper.

Table 1.

Details of current occurrence and conservation status of Ophiogomphus howei by state/province

State/Province	S rank	Status	SGCN?	Notes	Sources
West
Ontario	S1	Endangered	N/A	Only record is a single exuvia collected in 2007 in the Namakan River, Rainy River District. Subsequent searches in 2008, 2009, 2010, and 2016 were unsuccessful.	Hamill 2013, COSEWIC 2018
Minnesota	S3	Special concern	Yes	Confirmed breeding populations in the St. Croix River system in 2 counties along Wisconsin border. Records from nearby Kanabec County and a more northerly record from the Mississippi River in Itasca County are each based on single exuvia.	Haarstad 1994; Steffens and Smith 1999; MDNR 2013, 2016
Wisconsin	S4	N/A	No	Reported from 23 counties in the northern half of the state with nearly 1,500 flight season records.	WOS 2019
Michigan	S1	Threatened	Yes	Previously only known from exuviae and nymphs collected the Paint River, Iron County, in the Upper Peninsula, all but one from 1997. An undated record from Menominee County (Donnelly 2004, Abbott 2006-2020c) is unsupported by data and considered erroneous.	Lee 2007, Derosier et al. 2015, SCAN 2020
East
New Brunswick	S2	Special concern	N/A	First recorded in 2002, now known from 6 large river systems; considered rare.	COSEWIC 2018
Maine	S2	Special concern	Yes	Known from 22 populations in 11 rivers.	Brunelle and deMaynadier 2005, MDIFW 2015
New Hampshire	S2	None	No	Known from 3 counties with most records from the Merrimack and Contoocook River systems.	Hunt 2012, NHNHB 2018
Massachusetts	SX	None	No	The type specimen was a female collected in 1922 in Massachusetts; there have been no records since.	Bromley 1924, Nikula 2002, Buchsbaum et al. 2016
New York	S1	Special concern	Yes	Most common in the upper Hudson River.	White et al. 2010, Schlesinger 2017
Pennsylvania	SH	None	Yes	Historical records are from 1921 to 1988 from Cumberland and Susquehanna counties. Additionally, we located a previously unpublished Forest County specimen from 1924 in the Cleveland Museum of Natural History (SCAN 2019); see text for details.	Calvert 1924, Kennedy and White 1979, Shiffer et al. 2014, PGC-PFBC 2015, PNHP 2020
Maryland	S1	None	Yes	First recorded in 2011 and known from the Potomac River in two counties.	MDDNR 2016,  Orr 2019
Virginia	S1S2	None	Yes		VDGIF 2015, Roble 2020
Kentucky	S1S2	Endangered	No	Known from 6 counties in the eastern third of the state; no records since 1998.	KDFWR 2014, OKNP 2019
Tennessee	S3?	None	Yes	Known from two sites on the Tellico River in Monroe County; the first record from 1971 is from a site now inundated by the Tellico Dam project. State rank with question mark reflects uncertain status.	Louton 1982, Tennessen 1993, TSWAPT 2015
North Carolina	S1	Significantly Rare	No	Found only in extreme northwestern part of state in the New River system.	Ratcliffe 2018, LeGrand et al. 2020
South Carolina	Not ranked	None	No	Starting in 2008, multiple exuviae and at least one adult have been collected from both sides of the Chattooga River, which forms the border of Oconee County, South Carolina and Rabun County, Georgia.	Abbott 2006-2020d, Beaton and Dobbs 2010, Dobbs 2012
Georgia	Not ranked	None	No	See South Carolina.

S rank from NatureServe (2020), lower numerical ranks indicate increased risk of extinction or extirpation. SX, presumed extirpated; SH, possibly extirpated. For ranking methods, see Faber-Langendoen et al. (2012); SGCN, Species of Greatest Conservation Need as indicated by each US state’s latest Wildlife Action Plan (see AFWA 2019 for background on action plans).

The distribution of O. howei within each region is discontinuous, as this species is localized and found only along limited stretches of some river systems. Although it can be common in spots (Paulson 2007), it has been considered rare and restricted throughout its range (Tennessen 1993). This scarcity and patchy distribution has led O. howei to be given a global conservation rank of G3 (‘Vulnerable’) by NatureServe (2020), whereas the IUCN Red List categorizes the status as Least Concern using the same lines of evidence (Abbott et al. 2017). Table 1 presents the most current occurrence and conservation status of O. howei by state and province.

Ophiogomphus howei is listed as Threatened in the state of Michigan (Derosier et al. 2015). Prior to records described herein, it was known in Michigan only from 19 exuviae (shed exoskeletons of newly emerged nymphs) and 2 nymphs collected in 1997, and a single exuvia collected in 2014, all from the Paint River, Iron County, in the western Upper Peninsula (Lee 2007, Craves 2015, SCAN 2020). We describe a new population of O. howei in southern Michigan from the Grand River in Ingham and Eaton counties. This location is ~390 km from the closest western population site in Wisconsin, and about 485 km from the closest reported eastern population location in northern Kentucky.

Observations

On 22 May 2017, DM photographed a small male gomphid at Riverbend Natural Area, Delhi Township, Ingham County, Michigan (Abbott 2006-2020a). This location is along the Grand River, ~13 km south of the state capital city of Lansing. The small size and tinting on the basal half of the hindwings, unique among gomphids, confirmed its identity as O. howei. A photograph of a male was taken at the same location by R. Nirschl on 1 June 2017 (Abbott 2006-2020b). Additional searches by the authors and other observers for adults or exuviae in 2017 and in 2018 at this site, as well as other nearby locations along the Grand River, were unsuccessful.

In 2019, DM observed a teneral (recently emerged) female O. howei at Riverbend Natural Area (hereafter referred to as RNA) on 2 June. During a search by the authors at RNA on 4 June 2019, three females were netted (one collected) and at least two other individuals (sex undetermined) were observed. All were flushed from low vegetation or were seen taking very short flights (<10 m). RNA and nearby locations were searched again on 12 June 2019, but no O. howei were located.

Around this time, we were forwarded a photograph of a male gomphid that had been found moribund on 29 May 2019 on the floor inside a home in Dimondale, Eaton County, which we confirmed was O. howei. This residence is about 80 m from the Grand River and ~7.5 km downstream from RNA.

In 2020, DM searched RNA for adult O. howei without success on 31 May. JAC and DSO looked for adults and exuviae at RNA and downstream at neighboring Burchfield Park on 1 June. No adults were seen, but two exuviae of O. howei were found on the river bank at Burchfield. On 4 June, JAC and DSO searched the banks again from Burchfield to RNA and located four more exuviae.

On 13 June, DM observed a Skillet Clubtail, Gomphurus ventricosus (Walsh) (Odonata: Gomphidae) consuming a small dragonfly at RNA; the prey item was a female O. howei.

Site Description

RNA, Burchfield Park, and Dimondale are within the Lansing-East Lansing Metropolitan Statistical Area, which has a population of over 550,000 (U.S. Census Bureau 2020a). The Grand River is the second-largest watershed in Michigan. In the ~130 km between the headwaters and RNA, it flows north through the urban Jackson Metropolitan Statistical Area (population ~158,000) and the small city of Eaton Rapids (population ~5,200) (U.S. Census Bureau 2020a,b). The entire main stem of the Grand River has water quality impairments, including fish consumption advisories due to mercury and polychlorinated biphenyl (PCB) levels in fish tissue, and PCBs in the water column (MDEQ 2017). Segments of the river upstream have also failed to meet dissolved oxygen and sedimentation standards (MDEGLE 2020a).

RNA and adjoining Burchfield Park are managed by Ingham County and cover an area of over 218-ha bordering 5 km of the Grand River (Kaschinske 2015). Major land cover types in the 7.24 km2 local catchment (the drainage area for a specific stream segment) that includes RNA and Burchfield are 32.9% forest, 41.4% agriculture, and 6.7% developed (USEPA 2018). Figure 2 shows the Grand River and surrounding landscape from Eaton Rapids to Dimondale with the local catchment highlighted.

Fig. 2.

Landscape overview that includes the portion of the Grand River in southern Michigan occupied by Ophiogomphus howei. The outlined area is the local catchment surrounding sites where all exuviae and most adults were found. The Village of Dimondale, where a single adult was found, is noted. The red line is the county line dividing Eaton County (west) and Ingham County (east). The river flows from south to north then west in this region.

RNA consists of hiking and mountain bike trails through wooded floodplain and associated upland habitats bordering a large meander in the Grand River. Approximately 35–40 m from the river is a <2 ha old field where all the adult O. howei were observed, except for the Dimondale individual. In addition to heavily used trails, Burchfield has recreational amenities including canoe landings, picnic shelters, playgrounds, disc golf course, and a ball diamond. Forested areas at Burchfield have open understories, and clearings are mostly groomed or mowed. In both parks, trails are often widened and interconnected from off-trail use. The Grand River is 30- to 40-m wide, and the riverbank is highly eroded and steep in many places. The opposite bank is residential or forested, with access to the river restricted by private property.

Ophiogomphus howei exuviae were found on the riverbank adjacent to a trail that runs along the Grand River through Burchfield to RNA. All exuviae were <1 m from the water’s edge on soil or horizontal forbs or duff on sections of bank with <45 degree slopes. Each spot where they were found was on the downstream side of a canoe landing or a tree that projected into the river, resulting in a disruption of the current and a slow or calm pocket of water.

The Dimondale location is in a residential area just outside Dimondale village limits, which has a population of 1,249 (US Census Bureau 2020b). Given that the O. howei at this location was inside a home, we exclude it from the following in-stream habitat descriptions.

The Michigan Department of Environment, Great Lakes and Energy (formerly Michigan Department of Environmental Quality) periodically surveys watersheds throughout the state, evaluating both physical habitat conditions and biological communities. A station near the RNA/Burchfield border was sampled in 2006, 2011, and 2016 (Holden 2007, 2012; Parker and Rippke 2017). This station corresponds, within a few meters, of one site where we found O. howei exuviae.

Depth of the river in these surveys ranged from 0.76 to 1.8 m. In 2016, substrate in the center of the river was recorded as cobble, with margins of sand and silt. Flashiness scores declined from ‘Excellent’ in 2006 to ‘Poor’ in 2011 and 2016, indicating bank scouring and erosion. Parker and Rippke (2017) took note of high eroding banks, exacerbated by foot traffic and off-trail use (Parker and Rippke 2017), as we observed. The quality of the habitat in the river was scored ‘Excellent’ in 2006, and ‘Good’ in 2011 and 2016 (Holden 2007, 2012; Parker and Rippke 2017; for substrate composition definitions, scoring methods, and metrics, see MDEQ 2008).

The station had benthic macroinvertebrate community scores of ‘Excellent’ in each sampling period (Holden 2007, 2012; Parker and Rippke 2017; scoring criteria in Creal et al. 1996). Macroinvertebrates in these surveys are only identified to family, and this station consistently had much higher numbers of individual Gomphidae than any other stations along the upper Grand River, including those on tributaries. In 2016, gomphids were the dominant taxa at the RNA/Burchfield station, with 43 individuals making up 16.6% of the macroinvertebrate community. Five of the other 30 stations sampled in 2016 had gomphids, but none had more than two individuals (Parker and Rippke 2017).

Discussion

Our observations of Ophiogomphus howei in southern Michigan represent the first adults observed or collected in the state, and a new population between the two known population ranges. The southern Michigan location is roughly 390 km from the closest Wisconsin location (Epstein et al. 2002) and 490 km from the closest Michigan site in the western Upper Peninsula.

The closest eastern population records are from northeastern Kentucky, about 485 km from the southern Michigan location, but there have been no reports of O. howei from the state since 1998 (KDFWR 2014). An extant population in the Potomac River, Alleghany County, in northwest Maryland (Hubick and Brighton 2012–2020, Orr 2019) is ~615 km distant from the southern Michigan site.

In the course of our research, we discovered a specimen of an adult female O. howei in the Cleveland Museum of Natural History (catalog no.: CMNHENT0039305; SCAN 2019) collected 19 May 1924 at Tionesta, on the Alleghany River, Forest County, Pennsylvania, that has not been included in previous lists of Pennsylvania Odonata (Shiffer 1985, Shiffer et al. 2014). There have been no records of O. howei in Pennsylvania since 1988 and the species is considered ‘historical’ in the state (PGC-PFBC 2015). Aside from a 1921 record from south-central Pennsylvania, all records have been from the Susquehanna River in extreme northeastern Pennsylvania (Shiffer et al. 2014). The Tionesta location extends the Pennsylvania range at least 200 km west, and at 440 km from our southern Michigan location would be the closest eastern site if the population were extant.

In addition to being apparently extirpated from Pennsylvania, O. howei has not been detected during recent statewide Odonata atlas projects in West Virginia (Olcott 2011) or Ohio (Glotzhober and McShaffrey 2002, ODS 2019). The southern Michigan population is the first to be found in the gap between the western and eastern ranges.

Collectively, our observations of adult O. howei from 2017 to 2020 give a conservative estimated flight period in southern Michigan from 22 May to 13 June. The majority of records of O. howei across North America are based on exuviae, not adults (Tennessen 1993, White et al. 2010, COSEWIC 2018). Therefore, there is inadequate data from the eastern population at latitudes similar to southern Michigan for comparison of adult flight dates. Flight season in Wisconsin, where the range of O. howei is further north than southern Michigan but has a large adult sample size, is late May to early July, with 77% of observations occurring in June (WOS 2019).

Presence of adult O. howei over several years and the recent discovery of exuviae indicate that there is a breeding population at RNA/Burchfield. Exuviae are considered conclusive evidence that odonates have completed development close to the location where they are found (Tennessen 1993, Raebel et al. 2010, Bried et al. 2012, DuBois 2015, DuBois and Smith 2016). Ophigomphus howei nymphs are reported to emerge on mud banks or low vegetation, close to the water’s edge, concentrated where current slows abruptly such as below structures along erosional banks (Kennedy and White 1979, COSEWIC 2018), which coincides with our collections along the Grand River. Emergence sites may not always reflect the exact in-stream location where nymphs developed if they drifted downstream prior to emergence. Ophiogomphus nymphs are burrowers, which might make them less susceptible to drift (Hardersen and Toni 2019), although there is some evidence that O. howei nymphs come out of the substrate and drift with the current, probably late at night and prior to emergence (DuBois and Pratt 2017). Little is known about the frequency and distance involved, but limited data suggest there is not substantial downstream displacement of O. howei via drift (summarized in Gibbs et al. 2004).

Estimating abundance or population size at this southern Michigan site is hampered by difficulty in detection and limited survey effort. Adult O. howei are rarely seen even where they are known to occur (White et al. 2010, Paulson 2011, COSEWIC 2018), as they are presumed to spend most of their lives in the forest canopy (Kennedy and White 1979, Shiffer 1985, COSEWIC 2018). They have a flight period of ~6 wk that is often concentrated over ~10 d (Brunelle and deMaynadier 2005, White et al. 2010, Hunt 2012, WOS 2019). Prior to our efforts there were few records for collections or surveys of adult odonates from the RNA area during the appropriate time frame (Michigan Odonata Survey, unpublished data).

To our knowledge, we are the only people that have made an effort to survey for exuviae. Frequent searches are recommended for accurate detection of exuviae (Aliberti Lubertazzi and Ginsburg 2009, Raebel et al. 2010, Bried et al. 2012). The bulk of O. howei nymphs emerge over 4–6 d (Gibbs et al. 2004), providing an abbreviated window for surveys. Odonate exuviae do not persist long in the environment. A primary cause of exuviae loss is heavy rainfall and wind (Bried et al. 2012, DuBois 2015). We were unable to perform systematic surveys for O. howei adults or exuviae, and several of our planned searches for exuviae were prevented by weather events. Further, river morphology and more gently sloping, less-eroded banks suggest that emergence locations, and therefore exuviae deposition, may be more numerous on the opposite bank of the river from RNA and Burchfield, where we had no access due to private property.

That O. howei should be found in the Grand River in southern Michigan at all is surprising, given that some landscape and water quality characteristics reported to be important habitat indicators for this species are markedly different from most sites where O. howei has been found.

Habitat requirements are typically reported as medium to large, ‘fast-flowing’, ‘high-quality’ rivers. Landscapes are noted as being ‘forested’ and relatively ‘undisturbed’ with ‘minimal’ agriculture. Surveys of rivers that meet these criteria in states and provinces where O. howei has been found (Nikula 2002, White et al. 2010, Hamill 2013, COSEWIC 2018) or where it is presumed they could occur (Wagner and Thomas 1999, Klymko 2010, Olcott 2011, ODS 2019, Pfeiffer et al. 2019, Savard 2019) have often failed to locate any individuals. This localized and patchy distribution and the apparently narrow habitat requirements has led Ophiogomphus howei to be designated as a habitat specialist (Kennedy and White 1979, White et al. 2010, COSEWIC 2018).

A comparison of the southern Michigan site with the Paint River site in Michigan’s Upper Peninsula where nearly all O. howei exuviae have been collected (Lee 2007) and a site on Wisconsin’s St. Croix River with a large O. howei population (DuBois and Pratt 2017), utilizing local catchment data from the U.S. Environmental Protection Agency’s WATERS GeoViewer application (USEPA 2018) reveals some notable contrasts. The RNA/Burchfield local catchment occupied by O. howei has a lower estimated mean annual flow velocity (0.95 ft/s) compared with 1.27 ft/s the Paint River site or 1.83 ft/s at the St. Croix River site. The RNA/Burchfield location is also less forested (32.9%) with more agriculture (41.4%) than the Paint River (76.2% forest, no agriculture) or St. Croix River sites (53.5% forest, 27.9% agriculture). The degree of potential agricultural influence on the river is reflected in the level of crop inputs. For example, the level of pesticide usage in the RNA/Burchfield catchment (77.3 kg/km2) and synthetic nitrogen fertilizer application (17.8 kg/ha/yr) is far higher than reported for the Paint River location (negligible for both) or the St. Croix River site (4.5 kg/km2 and 7.2 kg/ha/yr, respectively). Further, nutrient levels for the upper Grand River watershed (encompassing the RNA/Burchfield sites and upstream to the headwaters) are high relative to other rivers in southern Michigan (MDEQ 2017), with agrochemicals being the primary source of nitrogen and phosphorus (Luscz et al. 2015). High nutrient levels can contribute to depressed dissolved oxygen levels in waterways. Kennedy and White (1979) described dissolved oxygen levels where O. howei nymphs were found in the New River, North Carolina, as ‘always near the saturation point’. Average July temperature for the Grand River at the upstream end of RNA is ~25°C, at which the saturation point would be expected to be >8.0 mg/liter (Hanshue and Harrington 2017). Dissolved oxygen measurements averaged 6.3 mg/liter at this location from 2010 to 2014 (HRC 2017). The state standard minimum level in warm waters is 5.0 mg/liter, and portions of the Grand River upstream from RNA have a history of failing to meet this standard (MDEGLE 2020b). More comprehensive and comparable data are not available for this water quality metric.

The extent of developed land in the RNA/Burchfield local catchment (6.7%) is roughly similar to the Paint River (2.8%) and St. Croix River (6.4%) catchments. However, <13 km upstream from RNA, the Grand River passes through Eaton Rapids, an urban area with catchments that are 69.6% developed. Another ~66 km upriver is the city of Jackson, with a local catchment that is 83.7% developed. The density of road crossings upstream from RNA is triple the density present upstream from the Paint Creek or St. Croix locations. Impervious surfaces and road runoff are major causes of stream sedimentation and pollutants in waterways (Tetra Tech 2006). Conditions upstream influence the habitat, water quality, and environmental conditions experienced by aquatic invertebrates further downstream (Strayer 2006, Collins and McIntyre 2017). Pollution, including that from pesticides and other agricultural runoff, is listed as a threat to O. howei that needs further investigation due to a lack of data (Hunt et al. 2010, Environment Canada 2013, COSEWIC 2018). While urbanization tends to affect Odonata diversity negatively (reviewed in Villalobos- Jiménez et al. 2016), studies suggest there are many mechanisms contributing to ‘urban stream syndrome’ (Paul and Meyer 2001, Walsh et al. 2005) and their impacts on Odonata may be very species-specific (Villalobos-Jiménez et al. 2016, Prescott and Eason 2018, Tippler et al. 2018). We compared the southern Michigan site with only two other locations for which we could pinpoint specific stream reaches and local catchments. These sites are representative of typical O. howei site descriptions, but careful comparisons of a range of specific habitat metrics over many sites would help clarify which qualities are most critical to O. howei reproductive success.

It is unknown if the degraded landscape and water quality at the southern Michigan location is suppressing the O. howei population, accounting for the low number of individuals recorded so far. However, the Grand River has a long history of impaired water quality (Parker and Rippke 2017); the section between Eaton Rapids and Dimondale that includes RNA and Burchfield has had fish consumption advisories dating back to at least 1998 for mercury and/or PCBs (USEPA 2016). Considering that members of the genus Ophiogomphus tend to stay close to their in-stream habitats, with dispersal distances estimated to be no greater than ~10 km (Collins and McIntyre 2017, COSEWIC 2018), it seems likely this population has been present for some time. These circumstances suggest that O. howei may be less specialized, or more adaptable, than previously believed. Given the very low encounter rates of adults even where intensive surveys have been conducted (e.g., 11/114 records in Maine and New Brunswick, COSEWIC 2018), we are optimistic that the observations of at least 10 adults reported here from a limited number of surveys over several years could indicate a robust population at the southern Michigan site.

No genetic studies have been done with O. howei comparing the eastern and western populations (COSEWIC 2018) despite variation in the morphology of nymphs in the southernmost portion of the eastern range from individuals in the western population (Beaton and Dobbs 2010, Tennessen 2019). Given that diminished capacity for dispersal may lead to low gene flow and consequently higher genetic differentiation between populations in lotic species (Hof et al. 2006, Marten et al. 2006), this type of investigation could reveal diversity in the genetic makeup of this species across North America.

Due to their more limited habitat availability and geographical range sizes, lower dispersal capabilities, and often narrower ecological requirements, lotic Odonata species are at higher risk of local or regional extirpation or extinction, with riverine Gomphidae often identified as particularly vulnerable (Clausnitzer et al. 2009, White et al. 2015, Collins and McIntyre 2017, Rocha-Ortega et al. 2020). Identifying habitat requirements, sensitivity to environmental change, and accurate knowledge of temporal and distributional data are all key to effective conservation actions (Cardoso et al. 2011).