Atlas of diatoms (Bacillariophyta) from diverse habitats in remote regions of western Canada

Abstract High-resolution LM images of diatoms from remote regions of western Canada are presented as a contribution to our knowledge of diatom floristics, ecology and biogeography in North America. Approximately 600 taxa are imaged in 132 plates. Genera with the most taxa are Cymbella (19 taxa), Cymbopleura (29), Encyonema (23), Encyonopsis (15), Eunotia (77), Gomphonema (42), Navicula (47), Neidium (20), Nitzschia (35), Pinnularia (50) and Stauroneis (34). Diatoms were collected from diverse habitats in four of North America’s major biomes: Arctic tundra, taiga, Rocky Mountains and Pacific rainforest. Many of the photographed taxa could not be identified to species and are likely new to science. Other taxa may represent new records for North America or Canada. Images of voucher specimens are keyed to individual collection sites. Detailed descriptions of the collection sites include GPS coordinates, colour photographs, vegetation, algal substrates, elevations, pH, temperature and conductivity. Samples were collected from natural substrates in fresh to brackish, flowing and standing waters. Voucher slides are deposited in the Montana Diatom Collection (Helena) and the University of Montana Herbarium (Missoula). Cleaned diatom frustules have been deposited in the Diatom Herbarium of the Academy of Natural Sciences of Philadelphia.


Introduction
From 2009 to 2017, the authors collected 96 diatom samples from a variety of freshand brackish-, standing-and flowing-water habitats in remote regions of western Canada ( Fig. 1). Collection sites were located in four of North America's major biomes: Arctic tundra, taiga, Rocky Mountains and Pacific rainforest. Representative specimens from those collections are presented here in a series of high-resolution photographic plates as a contribution to our knowledge of diatom floristics, ecology and biogeography in North America.
This atlas is intended as a resource for the study of diatom floristics, diatom biogeography and diatom ecology. It is not intended as a taxonomic resource nor as a definitive account of taxa richness. Priority was given to imaging unknown, unusual, uncommon and visually distinctive taxa and, of the ones that were photographed, only clean, good quality images were used. If a good quality image of a taxon was not available, we did not include that taxon in the atlas. For this reason, we expect that most of the included taxa can be identified from the images alone. For practical reasons, images are presented without written descriptions.
This atlas of diatom images is comparable to Schmidt's Atlas (A. Schmidt 1874(A. Schmidt -1959, but with three significant differences: (1) this atlas addresses only diatoms from western Canada; (2) illustrations are high-resolution (600 dpi) digital images rather than line drawings; and (3) detailed descriptions are provided for many of the collection sites, including GPS coordinates, colour photographs, terrestrial and aquatic vegetation, algal substrata, elevations, pH, water temperature and conductivity. This atlas might also be thought of as a preliminary illustrated checklist of diatoms from western Canada that are included in the Montana Diatom Collection.

Methods
Samples were collected from rivers, streams, lakes, pools, bogs, fens, beaver ponds and wet meadows in six regions of western Canada (Fig. 1). Most of the samples are from remote locations accessible only by hiking or by kayak or canoe. Elevations at collection sites range from sea level to over 2000 m a.s.l. Many of the sampled water bodies are not named on topographic maps. In these cases, locally descriptive place names are provided by the sampler. Temperature, pH and conductivity were measured in the field with an YSI 556 Multi Probe System or a Eutech Instruments Oakton pH Tester, Model 30. Colour photographs were taken of most collection sites. The Oakton pH meter was calibrated once annually before each field season against standard solutions of pH 4.01 and pH 7.01. However, sampling trips were long and conducted under challenging conditions. The meter may have gone out of calibration and readers should be cautioned about readings other than those taken in Waterton Park.
At each site, diatoms were collected from all available near-shore substrata, including aquatic macrophytes, mosses, rocks, fine sediments and woody debris. Subsamples were combined with ambient water in a single container and preserved with iodine before transport to the laboratory, where they were treated with 30% hydrogen peroxide (H 2 O 2 ) to remove organic matter. After several rinses in distilled water, cleaned diatom material was dried on cover slips and mounted permanently on glass slides using Naphrax.
Slides were examined under LM with differential interference contrast optics and images were captured using a Leica DM LB2 research microscope and a Spot Insight monochrome digital camera (Model 14.0). Slides examined for this study are deposited in the Montana Diatom Collection (MDC) in Helena and the University of Montana Herbarium in Missoula (MONTU). Vials of cleaned and dried diatom frustules have been deposited in the Diatom Herbarium of the Academy of Natural Sciences (Drexel University) in Philadelphia.
Imaged diatoms were identified by the first author to the lowest practical taxonomic unit using available identification resources. None of the species is described as new to science, but many are designated as unknowns (e.g. Navicula sp.) or as comparable to another taxon (cf.). Alternative identifications by reviewers (Acknowledgements) are given in brackets in the figure captions along with the reviewer's initials, thereby giving readers other options for identification.

Format
In keeping with the intended use of this atlas as a resource for the study of diatom floristics and biogeography, images of taxa are presented separately for each of the six regions ( Fig. 1). Regions (site descriptions and diatom plates) are presented in the order in which they were sampled, first Waterton Lakes National Park (2009), followed by Haida Gwaii (2013Gwaii ( , 2017, Clearwater River (2014), Coppermine River (2015), Baillie-Back Rivers (2016) and Hood River (2017). A master list of taxa and index to the plates is provided in a table at the end of the atlas (Appendix 2).
Taxa are presented in general phylogenetic order following the classification scheme of Round et al. (1990). Centric diatoms are presented first, then araphids, then monoraphids and finally biraphids. The order of genera in the biraphid group may vary from region to region. An effort was made to group species of the same or similar genera on the same plate, but sometimes species of dissimilar genera are presented on the same plate in order to conserve space.
When possible, multiple specimens of a taxon are presented in size-reduction series. However, this was not always possible for rare and uncommon taxa. In general, the number of specimens displayed on a plate is proportional to the relative abundance of that taxon in that region.

Descriptions of collecting sites
The following sections describe the geographic regions that were sampled and the collecting sites in each region.
In the spring and summer of 2009, 41 samples of benthic diatoms were collected from waters in Waterton Lakes National Park, the Canadian component of the Waterton-Glacier International Peace Park (Table 1, Figures 2-14). The samples were collected in the course of scheduled pond and stream monitoring and assessment projects. All samples consisted of some surface water and scraping of a submerged object. One objective of benthic diatom sampling was to determine the presence and extent in the Park of Didymosphenia geminata (Lyngbye) M. Schmidt ("Didymo" or "rock snot"), which at the time was reported in large numbers from many streams in adjacent Glacier National Park (Bahls 2007, Schweiger et al. 2011. Didymo was detected in 14 samples collected from Waterton Lakes National Park in 2009; all of the Waterton samples that contained Didymo were from flowing waters, including the Waterton River, Belly River and several smaller streams. The Haida Gwaii Archipelago lies about 500 km northwest of Vancouver Island and is separated from the British Columbia mainland by the 70 to 100 km-wide Hecate Strait (Fig. 1). The Archipelago consists of a 300 km-long, north-south trending group of islands in the shape of a "V". Along the western branch of the "V" is a mountain range with summits over 1,100 m elevation. Higher elevations on the archipelago support mountain hemlock and alpine tundra vegetation; lower elevations are dominated by coastal red cedar, pine, western hemlock and Sitka spruce (Banner et al. 1989). The coastline along the eastern edge of Haida Gwaii has fluctuated dramatically since 12,000 years BP (Josenhans et al. 1995(Josenhans et al. , 1997. At that time, sea level was about 150 m lower than it is today. By 9,000 years BP, the sea level had risen sharply to 15 m higher than today and remained at that level until about 5,000 years BP, falling to current levels by about 2,000 years BP. These fluctuations are the result of interplay between isostatic, eustatic and tectonic forces in the area. As a result of these fluctuations, Holocene archaeological sites (9,000-5,000 years BP) are stranded in the forest well above their original coastal locations. Diatom remains in sediments of coastal fresh-water ponds include evidence of past salt-water intrusions (Pienitz et al. 2003).
Six samples were collected from freshwater habitats in July 2013 and another three samples were collected from fresh and brackish waters in May 2017 ( Table 2). The three samples collected in 2017 were originally numbered 1, 2 and 3 but are renumbered 7, 8 and 9 here to avoid confusion with the 2013 samples. Sites sampled in 2013 were accessed from the coast by kayak; sites sampled in 2017 were accessed by inland routes. The following descriptions of the sample sites are taken from the field notes of Beverly Boynton.

Fresh-water samples, July 2013
All samples included squeezed vegetation, a scraped rock or stick, surface water and a few ml of iodine added. Samples were taken well above any tidal influence. Sites would all have infrequent human visitation (maybe less than yearly, or even never), especially since I (BB) walked upstream further than needed if people were getting drinking water. Forest of red alder, western hemlock, western red cedar, salal, mosses, ferns. Squeezed brown moss (which was longer and easier to squeeze than previous samples), scraped a rock, surface water. Probably no one has been to this particular place.

Fresh-and brackish-water samples, May 2017
All samples included squeezed vegetation, a scraped rock or stick, surface water and a few ml of iodine added. According to locals, May was more rainy and cooler than usual.  The following account is taken from the field notes of Beverly Boynton: Sampling was done in June 2014 while on a 20-day canoe trip from the headwaters of the Clearwater River in Saskatchewan (Forrest Lake) to its confluence with the Athabasca River in Alberta (Fig. 1), estimated to be about 480 river kilometres. Moser et al. (2004) reported on the ecology of diatoms in lakes of Wood Buffalo National Park, which is located about 200 km northwest of the Clearwater/ Athabasca River confluence.
The first six samples are from the boreal forest of the Canadian Shield. This area is largely an open-canopy jack pine, black spruce and white spruce forest, with extensive areas of lichens. There are also numerous shrubs, including willows, alders, birch and some forbs. The soil is shallow, i.e. often just an inch (2.5 cm) of forest detritus, on top of deep sand. The area has almost no rocks of any size that we saw (except in rapids) until we were almost off the Shield. The Shield itself is Precambrian granitic rock.
The last three samples are from the Western Canada Sedimentary Basin (depositions from inland seas advancing and retreating). This area has numerous outcrops and gorges of limestone and dolomite. The soil is much deeper, organic soil with clay, supporting a mixed forest of closed-canopy paper birch, white spruce, balsam poplar and a great many shrubs and forbs.
Both areas are topographically fairly flat (especially on the Shield), with an enormous number of shallow gouges from the Laurentide Ice Sheet, which are filled with water. One does not need to walk far to encounter a bog, fen, marsh, swamp, pond, lake or creek. Drainage is generally poor because of the flatness. The hydrology is still quite young and sorting itself out and still influenced by post-glacial rebound.
No samples came from the Clearwater River itself or from areas that seemed to be receiving flow from the River. Because of the low number of people on most sections of the river, along with most people's general distaste for spending time in bogs, fens, marshes and swamps, the specific sampling sites are probably rarely if ever visited and did not seem to be disturbed. Mosquitoes and blackflies were moderately bad in general on this trip, but were not worse at the sampling sites and posed little problem for the Slime Crew (my husband assisted with some collecting). As always, seeking out these microhabitats added immensely to the interest of the trip; botany and geology were highlights.
All samples consisted of some surface water, a vegetation squeeze and scraping of a submerged object. There were no rocks to scrape and the sticks (when found) had only a minimal slimy feel. If sediment was present on the surface, it was sampled.
Coordinates are WGS84, latitude-longitude in degrees and decimal minutes, elevation in metres. Unfortunately, my camera drowned after the first week; pictures are then from my Itouch; however, the last nine days were quite overcast and I barely was able to keep my Itouch recharged from my Solio. All samples are from the watershed of the Clearwater River.
Sample #1, bog near Naomi Lake (Fig. 29)  The following account is taken from the field notes of Beverly Boynton: Sampling was done while on a 28-day canoe trip on the Coppermine River ( Fig. 1), from Point Lake near the headwaters, to its mouth in Coronation Gulf of the Arctic Ocean, about 450 river kilometres. All samples were from fresh water, none closer to the Arctic Ocean than about 10 km; no samples were taken from the Coppermine River itself. Only sample #1 was from Northwest Territories; the rest are from Nunavut, Canada. The trip began in the upper Coppermine, which is a system of lakes on Arctic tundra, above the treeline. At Redrock Lake, the Coppermine leaves the tundra to enter white spruce forest with willow and birch shrubs. Treeline and its transition zone follow the protected river valley up to the Coppermine Mountains, though tundra predominates in places beyond the river valley. At Big Bend, the trees thin and become shorter. Past the Coppermine Mountains, the terrain is predominantly tundra vegetation of grasses and sedges, lichens, willows and smaller birch shrubs.
The entire trip was on the Canadian Shield, though this Precambrian granitic rock was not always visible due to postglacial till and sediment deposition. A huge glacial lake, Lake Coppermine, formed during deglaciation when a lobe of the ice sheet blocked the Coppermine's outlet to the sea. This lake extended from Fairy River to Rocky Defile and has left behind lake sediments of marl. Further downriver, the Coppermine cuts through sandstone, limestone and dolomite, forming gorges. The river then goes through the Muskox Intrusion, which was formed during the Proterozoic from mantle plumes. This is one of the globe's largest basalt flood plains and consists of rock types such as gabbro, which is the rock type from Escape Rapids to Coronation Gulf (Dredge 2001).
Place names were non-existent, so descriptive terms are used here.  The following account is taken from the field notes of Beverly Boynton: Sampling was done while on a 26-day canoe trip that went from a lake on the Baillie River (a main tributary of the Back River), to the Baillie's confluence with the Back River, then down the Back River, passing through Pelly Lake and ending on Mission Island in Upper Garry Lake, about 445 river km from our starting point (Fig. 1). Elevation at our put-in on the Baillie was 284 m, about 168 m near the confluence of the Baillie and Back Rivers (a 120 m drop over about 185 km), then the Back essentially becomes flatwater, with an elevation at our take out on Mission Island of 148 m (a drop of only 20 m over the final 260 km). All samples are fresh water from Nunavut.
The sampling area is subarctic tundra on the Canadian Shield, with continuous permafrost and a thin active layer of reportedly acidic soil. There are some areas of granitic outcrops, but most of the river corridor is covered with extensive deposits from de-glaciation of the Laurentide Ice Sheet, including sandy eskers, large sand flats and areas of mostly sorted till and sediments. In addition, there are extensive areas of peat, with sphagnum moss and other mosses. Lichen species were fairly ubiquitous. The area is well above the treeline, except for one small area on the Baillie that has a relic stand of white spruce. The shrubs include dwarf birch, small willow species, red alder and various Ericaceae species. Other plants included grasses, sedges and the expected flora for subarctic bogs, fens and uplands.
Annual precipitation is low, making this area a polar desert, but during summer, snow melt and thawing of the active layer results in waterlogged soil and a network of lakes, streams, rivers and wetlands. Throughout the collection area were numerous ponds, pools of standing water, wet peatlands and wet to moist areas of hummocks and patterned ground, in addition to dry uplands.
From local reports, May and June were rainy months, at least in Yellowknife, about 480 km to the west of our put-in. The Back clearly had high waters, as most dry riverbeds noted on maps were covered with water. The Back drains a huge area and, being a lowlands river on permafrost, the accumulated water is slow to discharge into the Arctic Ocean. We had a number of high wind days that kept us from paddling and a few periods of rain.
Place names are my own descriptive terms, sometimes adding nearby names from Canadian maps.  The following account is taken from the field notes of Beverly Boynton: Sampling was done while on a 27-day, 300-km canoe trip in Nunavut, from the headwaters of the Hood River to the north end of the peninsula dividing the Hood River mouth in Arctic Sound from Baillie Bay (both are in Bathurst Inlet of Coronation Gulf of the Arctic Ocean). Elevation at our put-in is 414 m and the mouth of the Hood is at sea level.
The Hood River is on the Central Continental Arctic portion of North America, on the Precambrian Canadian Shield. It lies between the Contwoyto Plateau to the south (a 450 m high plateau of gently rolling drift) and the Tree River uplands to the north and west (a lower, dissected granite plateau of smooth rock-knob hills with deep valleys) and flows through isolated, rugged tundra. For most of its length, the Hood is less than 150 km south of Coronation Gulf of the Arctic Ocean as it runs west to east. There are many areas of Precambrian granitic outcrops, but much of the surface along the river corridor is covered with extensive deposits from de-glaciation of the Laurentide Ice Sheet, with areas of sorted and unsorted till and sediment, including sandy eskers, sand, mud and clay flats.
The lower half of the Hood has some metamorphic rocks of quartzite and slate, with clay tills. The Wilberforce Hills to the east are the dissected edge of the Contwoyto Plateau. About 50 km from its mouth, the Hood turns abruptly to the north in its run to the coast. The river then lies in a broad flood plain with evidence of previous salt-water incursions from when sea levels were higher.
Bathurst Inlet is a physiographic division of the Shield, with a major NNE to SSW fault forming the boundary between the uplands and the Coronation Gulf Lowlands. It is a complex submerged valley, a 200 km-long extension of the Coronation Gulf lowlands penetrating the Shield, with west-dipping diabase and basalt sills, often overlying basalt. The Queen Maud Lowlands lie to the east of the Inlet.
The river is all above the Arctic Circle, well above the treeline, with continuous permafrost and a thin active layer of soil. Lichen species were ubiquitous as were Ericaceae spp., dwarf birch, willow, alder, sedges and grasses. Mosses seemed less extensive than seen on other barren grounds trips; we identified 50 species of arctic wildflowers.
Unfortunately, strong headwinds prevented us from paddling to the actual estuary of the Hood in Arctic Sound and, even more disappointing, a hike to the tip of the peninsula ended on tall undercut bluffs that prevented a descent to the ocean at the northernmost point. For this reason, the final Hood River sample was about 8 km upstream from Arctic Sound and the first Bathurst Inlet sample (on east side of the peninsula) was perhaps half a mile to the south of the headlands of the peninsula.
The Hood River was very low this season, presumably due to low winter snow, an early snow and ice meltoff and lack of rain. Compared to a personal account by friends who paddled the river in 2013 and found it to be low water, this year the river was much lower.
There is considerable research being done on the arctic freshwater system in the face of climate change. Significant changes include rising surface air temperatures, warming permafrost and shrub encroachment on the northern tundra. Storage and cycling of fresh water on land has changed along with precipitation, river discharge, lake abundance and size and soil moisture.
In total, 16 samples of benthic diatoms were collected from water bodies along the Hood River corridor. (There is no sample #10.) The following field notes describe the sampling sites. Datum is WGS84, coordinates are latitude-longitude in degrees and decimal minutes; elevation is in metres. Place names are my own (BB) descriptive terms, sometimes adding nearby names from Canadian maps.   Fig. 6 N. longiceps. Fig. 7 N