This was a first-time baseline survey, for which we devised our own methodology and had no basis for historical comparison. As such, we lacked any means other than anecdotal comparisons, which we avoided relying upon, for assessing changes of buckthorn presence.
Because we worked within a nature preserve, a paramount methodological consideration was that our methods be as unobtrusive as possible. We restricted ourselves to procedures employing no destructive testing, no physical sample collecting or removals, no excavation or other permanent habitat alteration, and minimal trampling or wildlife disturbance. We therefore sampled only at close proximity to existing trails.
We obtained data at twenty-two locations (Sites A through V), where we attached brass marker tags to ground level tree trunks for future revisitation. Although sites were near existing trails to minimize sampling disturbance, we randomly selected specific tag locations for forty meter sampling transect origins. The sites included varying soil moisture conditions (in non-wetlands), degrees of disturbance, and woodland formations (Figure 3). We excluded wetlands and water bodies as observation sites (Figure 4) because our preliminary reconnaissance indicated little or no buckthorn presence in these settings, but especially because sampling activities might adversely affect these extra-sensitive habitats.
We collected data from within random one-meter sampling quadrats in forty-by-twenty meter grids lying transverse to trails, using the site tags as grid centers. Each quadrat thus had a unique and reconstructable location. For such reasons as falling onto trails, or to avoid especially sensitive or hazardous positions (e.g., endangered plants, wasp nests, poison ivy, etc.), the number of quadrats within each 800 m2 grid varied from twenty to 140 (Figure 5). Within each quadrat we recorded 1) the location (to the nearest centimeter) of aboveground buckthorn stems, 2) the diameter of each buckthorn stem, and 3) the number of non-buckthorn woody stems.
Possible data inaccuracies may have
existed. Our principal concerns
were:
Species identification: although we observed during September and October, when buckthorn foliage is particularly obvious, occasional confusion of buckthorn with other plants (e.g., Prunus, Alnus, Betula) was possible.
Sampling paucity:
twenty-two sites is a sparse sample for projecting spatial distribution
patterns across several hundred hectares. We therefore regard our initial
depictions of buckthorn extent as preliminary, pending future sampling
verification.
Site selection: bias may exist from sampling only along trails within twenty meters of origin tags, especially if these were avenues of buckthorn entry. In this project we had no means to judge the actual mechanism of buckthorn introduction.
Our initial analysis procedure was to use site means of buckthorn dominance, stem density, and stem diameter as centroid values for creating isoline maps. Comparison of the resultant patterns enabled us to identify locations of relative concentration by abundance and size, and revealed whether or not these attributes coincide spatially. It then was possible to reexamine the study area, seeking other ecological characteristics that might promote--or inhibit--the establishment of buckthorn.
To further the latter endeavor, we also created an attribute matrix showing site letters where community parameters coincided with more (above average abundance) or less (below average abundance) buckthorn presence. The gray-shaded attributes are those which occurred at half or more of all sampling sites. Excessive wetness in low-lying locations, shading and competition, and deeper litter layers were the most common habitat factors at sites having less buckthorn. Clustering of buckthorn stems also was a frequent characteristic, especially where the stems were larger.