Background

This study examines landscape-scale patterns of plant species richness, environmental heterogeneity, community composition and diversity, and remotely sensed spectral measures of biodiversity. One of the aims of the research is to combine classic vegetation analysis with quantitative analysis of satellite imagery to test the possibility of using remote sensing as a tool to recognize biodiversity hotspots within a landscape.

Arctic rivers, due to their undisturbed nature, provide an excellent opportunity to study patterns in riparian ecosystems uncomplicated by anthropogenic influences. Arctic riparian areas occupy unique positions as sites of relatively high biodiversity (biodiversity "hotspots") in an environment generally low in biodiversity. In order to determine biodiversity patterns and establish a link between spectral diversity and biodiversity, we are testing the hypotheses that 1) species richness is a function of summer temperatures and environmental heterogeneity, and richness will be highest at points where these indices are highest, 2) the diversity of plant communities at a given site is primarily determined by the environmental heterogeneity of the site, and 3) the diversity of plant communities will be reflected in the diversity of spectral reflectance values recorded by satellite imagery.

Goals

The goal of this research is to characterize the riparian ecosystem in terms of species, community, and spectral diversity in order to develop methods of distinguishing biodiversity hotspots using satellite imagery. The objectives are to answer the following questions:

• To what extent is satellite imagery useful in distinguishing gradients of community diversity along the riparian corridor and between riparian and surrounding areas?
• What are the respective roles of climate, environmental heterogeneity, location along the riparian corridor, soil type and chemistry, and snow depth on patterns of species composition, species richness, plant community composition, and diversity of plant communities?
• How do species composition, species richness, and plant community composition differ along climatic, elevational, and latitudinal gradients in arctic riparian areas?

One outcome of this study will be a means of estimating biodiversity using remote sensing. This will be accomplished by linking species diversity data to community diversity data obtained from ground-based surveys with remotely sensed spectral data reflecting landscape diversity. The proposed research will test the idea that areas with a greater degree of environmental heterogeneity will show a greater degree of species and community diversity. This diversity will be reflected in a more heterogeneous pattern of greenness in the landscape, detectable by remotely sensed satellite imagery. A second outcome will be baseline data providing the groundwork for a multi-river study of species diversity along rivers in the Canadian, Russian, and Alaskan Arctic. It will further test hypotheses generated by the analysis of arctic riparian landscape patterns in the study proposed here.

One unique aspect of this study is the extensive use of undergraduate assistants in the remote field work and the combination of field work with field teaching. Over forty students from the United States and Canada have been involved in this research to date, spending four to six weeks in the field as students enrolled in a field course, enrolled for independent study, or as student volunteers. These undergraduates share a portion of the costs of travel to the site, costs of transporting gear, and aid in data collection and logistic support. This approach to using undergraduate help and providing them with instruction and work experience is an excellent way to increase the pool of students interested in arctic science.

Results

Results are available for the portion of this work relating species richness, temperature, and environmental heterogeneity:

Landscape patterns in plant species richness along an Arctic river

William A. Gould and Marilyn D. Walker

We examined the correlation between vascular plant species richness, mean July temperature, and components of landscape heterogeneity in order to determine the relative influence of temperature and the physical landscape on a gradient in plant richness along the Hood River in the Northwest Territories of Canada. This is the first analysis of vascular species richness along the complete length of a North American river.

The vascular flora for the study area includes 210 species. Species richness at seventeen sites along the length of the river ranged from 69 to 109 vascular plant species. Sites with the lowest species richness were those in the upper portions of the river, with diversity increasing towards the coast. The observed variation in species richness along the river is correlated with environmental heterogeneity along the river (r^2 = 0.514, P = 0.0012), the most significant component of which is an increase in soil pH range, with pH increasing as the river nears the coast. The best predictor of species richness is the average site soil pH (r^2 = 0.857, P < 0.0001) (Fig. 1). The primary cause of higher soil pH is the presence of uplifted marine sediments and tills derived from non- acidic Precambrian rock commonly occurring at elevations of less than 150 m along the river. We originally postulated that the gradient in species richness found along the river might be due to a temperature gradient associated with decreasing elevation from the headwaters to the coast (356 m). Mean July temperatures measured at sites along the Hood ranged from 9.6 to 10.3ÁC. Although the temperature gradient was correlated with decreasing elevation (r^2 = -0.378, P = 0.0147), mean July temperatures were not well correlated with species.

The vascular flora for the study area includes 213 taxa (112 forbs, 75 grasses and sedges, 19 deciduous woody shrubs, and 7 evergreen woody shrubs), 210 species, and 91 genera. 187 species were found within the species richness plots in this study and other records are from a concurrent analysis of riparian plant communities. The inclusion of these records brings the vascular plant flora for the Bathurst Inlet region to 274 species. The Hood River flora represents 78% of this regional flora. Thirty-two families are represented within the area, with over 40% of the species present belonging to three families: Cyperaceae (36 species), Poaceae (27 species), and Asteraceae (20 species). Fifty-three percent of the species are forbs, 35% are graminoid, 9% are deciduous shrubs, and 3% are evergreen shrubs. The Hood River flora is made up primarily of circumpolar, arctic-alpine species.

An ordination indicates that the four sites on the uplifted marine sediments have a distinctive species composition from the sites located upriver, above the uplifted sediments (Fig. 2). The relatively short axes (1.4 SD units) are a reflection of both the number of widespread species found at nearly all sites, and of the use of only presence/absence data rather than presence and abundance data often used in ordinations. SD units reflect the amount of dissimilarity between sites. One SD unit indicates about 50% similarity between floras. Each site flora is a composite of all the communities found at a site. The first axis in this ordination is a complex gradient related to position along the river and soil chemistry.

Figure 2. An ordination accomplished using detrended canonical correspondence analysis of the floras of seventeen sites along the Hood River. Sites found on uplifted marine sediments (below 150 m) separate out as floristically distinct from sites found above the uplifted sediments. All sites share a large percentage of species (30%) which reduces separation between site in terms of SD units. Separation along the first axis is related to gradients in soil pH and a complex gradient in elevation. Separation along the second DCA axis is unexplained.

Fall colors along the Hood.

Study area: click on the map for a larger version.