Small wetlands project

The Oswegatchie-Black wetlands/watershed relationship project was based upon photointerpretation of NHAP 1:58000 color infrared aerial photography, with an anticipated minimum wetlands delineation and classification size of one acre. For the two test areas, the NHAP 1:58000 photos were flown May 24, 1985, and exhibited full leaf-out for broad-leaf deciduous trees. As a point of reference, the NHAP goal generally is to acquire spring pre-leaf-out imagery, but atmospheric conditions, phenological variations, and other extenuating circumstances sometimes result in later-than-planned-for flying dates. In examining the 1:20000 pre-leaf-out (1994) imagery and the NHAP 1:58000 full-leaf (1985) imagery, the following observations were noted:

1. The 1:20000 pre-leaf-out imagery (especially the Kodak 2448 Aerochrome MS) was excellent for small wetland delineation and classification. Nine of eleven 2'x 2' targets placed in representative small wetlands were identified on the photographs (two targets were masked by tree canopy) and wetland areas smaller than 0.05 acres (approximately 2000 square feet) were identified. In addition, rills (1-2 feet wide) serving as linear surface water connections between wetland polygons were easily seen on the imagery. It was observed that most rills contained some surface water on the May 2 field check, but rill characteristics (channel size and shape, vegetation, slope, and detritus material) indicate that surface flow is most likely ephemeral, with the rills becoming dry during the summer months.
2. Based upon comparison of 1:20000 and 1:58000 photography it was determined that the majority of wetland polygons smaller than one acre can be identified on the 1:58000 NHAP imagery. A total of 24 separate (not contiguous with a larger wetland system) wetland polygons less than one acre in size were identified with NHAP imagery and mapped in the test areas while only three polygons were not delineated. Although wetland polygons as small as 0.1 acre were identified on the 1:58000 imagery, the very small physical area on the photography made accurate delineation difficult. In delineating the smallest wetlands, the actual wetland area was smaller than the width of the pen line on the overlay.
3. Because of leaf-out conditions on the NHAP 1:58000 photography, a number of small linear connections between wetland polygons were not identified. Clearly seen on the 1:20000 pre-leaf-out imagery, these 1-2 foot wide surface rivulets were completely masked by leaf canopy and could not be seen on the 1:58000 imagery even when the interpreter knew of their existence. For the test areas a total of 15 linear wetland segments were delineated on the 1:58000 photography while an additional 19 linear wetlands were identified on the 1:20000 imagery.
4. Classification of NWI (National Wetlands Inventory) covertypes for the small wetlands was very accurate for both the 1:20000 and 1:58000 imagery. Only one polygon 'PFO4E' was classified incorrectly as 'PF04/PF01E' on the 1:58000 photos. Even with leaf-out conditions, it was still possible to discern covertype for the small delineated wetlands.
5. The accurate location of small wetlands can be enhanced significantly using a GPS unit. In conjunction with the Trimble Community Base Station, wetland location was determined to ±2 meters with the Trimble Pathfinder Basic handheld rover unit. Wetland boundary data, with any necessary attributes, can be transferred directly into various GIS formats, including ARC/INFO, and recorded as a polygon/linear data layer. It should be understood, however, that GPS delineation of wetland boundaries can be very labor intensive.

In summary, large scale 1:20000 color and color infrared aerial photography is superb for delineating and classifying very small wetlands. Under ideal conditions (i.e., a topographic depression, leaf-off vegetation, and some standing water or wet soil) wetlands less than 2000 square feet (0.05 acres) were identified in the Cranberry Lake and in the Inlet/Oswegatchie test areas. Simultaneous photo analyses of 1:20000 and 1:58000 imagery indicate that wetlands as small as one acre can be consistently and accurately identified and classified with the smaller scale 1:58000 NHAP photography. Except for linear wetlands under leaf-out conditions, the 1:58000 NHAP is also suitable for delineation and classification of most wetlands smaller than one acre, with some wetlands as small as 0.1 acre providing a recognizable signature. The inability to identify rill-like linear connections between wetlands on leaf-out photography remains a problem.

The NAPP 1:40000 color infrared aerial program is the new national photographic standard and preliminary analysis indicates this imagery, assuming high quality photography and proper phenological timing, should be an excellent compromise scale for reliably mapping small wetlands units. Utilizing larger scale photography does increase interpretation time and costs, however. The number of photos (with 60% forward overlap) per 7.5 minute quadrangle, for example, increases from four at 1:58000 to ten at 1:40000, but the increased accuracy level should offset the additional costs. GPS technology is an excellent tool for accurately locating wetlands in the landscape.

Medium or poor fens and ombrotrophic bogs were designated as "critical" because of their perceived rarity in the Adirondack Park (and the Northeast in general) and because of indications that anthropogenically derived acidic and nutrient inputs are increasing the rate at which successional change is occurring (Gorham et al. 1984). It is known that fens can undergo a successional evolution to bog, a more acidic and less productive wetlands habitat. Bogs, particularly ombrotrophic bogs, are considered critical because of their sensitivity to atmospheric nitrogen additions (Tickle 1992).

This study attempted to remotely sense critical wetlands (peatlands) through the use of airphoto interpretation. Results indicate that peatlands (fens and bogs) cannot be separated from other wetland maps via airphoto interpretation, essentially because depth of peat, a critical distinguishing feature, is not directly visible on the photography. Fens, wet meadows, and emergent marshes are very similar on a gross morphological level, as are some bogs, coniferous swamps and shrub swamps. All can exhibit similar species assemblages (physiognomically and floristically), water chemistries and associated topography. But, it is the substrate which separates them. Fens and bogs are, by definition, types of peatlands, having deep accumulations of decomposed or partially decomposed organic debris. Wet meadows and emergent marshes do not accumulate a thick organic layer although their predominantly mineral substrate can have fairly high fractions of organic matter. It is believed that the accumulation of peat and development of fen or bog instead of wet meadow, emergent marsh, coniferous swamp or shrub swamp is dependant on original topographic position, hydrology, and climate. The relationships between these wetlands and relative landscape position may be elucidated with further field work and additional GIS data and analyses. The techniques did not allow the consistent separation of ombrotrophic bogs from other wetlands with similar signatures such as coniferous swamp or shrub swamp due to the complexity of covertype interspersion.

It became evident during this study that aerial photography was not a suitable medium by which to identify deep organic accumulations. Although the techniques employed failed to separate fens from emergent marsh or wet meadow, they did allow us to identify all the potential habitats for peatland inhabiting protected species such as the white fringed orchid [Platanthera blephariglottis (Willd.) Lindl.]. Figure 11 is an example quadrangle showing potential critical wetlands, those with "EM" labels, which are expected to contain all of the fens as well as other wetland covertypes within that signature. Figure 12 is a map of the same quadrangle showing Adirondack Park Agency land classifications depicting public and private lands. By electronic overlay it can be determined, for instance, what proportion of any type of wetland is located on public land versus private land for any given geographic area within the Oswegatchie-Black watershed.

Once completed, the pond watershed and wetlands coverages were combined and analyzed using GIS techniques. The total watershed area for the Oswegatchie-Black drainage is 398,783 ha. The total wetland area for the entire drainage is 60,766 ha characterized by 200 unique covertypes representing approximately 15% of the total drainage area. A total of 1223 ponded watersheds were identified in the Oswegatchie-Black drainage.

Evaluation of the distribution of wetlands and of their covertypes are presented and discussed in the following tables. To demonstrate GIS capability of combining geographic data with tabular data, which was one of the main objectives of this study, a subset of the lake watershed data is also presented. For this demonstration, the 23 lake watersheds associated with the Adirondack Long Term Monitoring (ALTM) study within the Oswegatchie-Black drainage basin, were selected. These are part of 52 total ALTM waters parkwide that represent a range of different hydrologic lake types and for which the longest record of aquatic chemistry exists in the Adirondack Park. One of the research objectives of this intensive lake monitoring effort is to track lake chemistry response to acidic deposition reductions that will result from the 1990 Clean Air Act Amendments. Therefore, these are key waters for the State as well as for the Nation. The watershed and wetlands maps for each of these ponds are included in Appendix K. The wetland covertype labels appear only when the default scale (to printer map size) is smaller than 1:24000.

Relative coverage of the 200 unique wetland covertypes is presented in Table 2 arranged by relative coverage and Table 3 arranged in alphanumeric order by covertype. A summary of the major wetland classifications by covertype are provided in Table 4 and Table 5. Spruce-fir forest wetland is shown to be the predominant covertype, appearing in over 47% of the labels. Dead shrubs, rooted vascular aquatic bed and tamarack forest were the least frequent representing collectively less than 1% of the covertypes. A wetland polygon can be labeled with up to two covertypes, each exceeding 40% coverage of the polygon area. By convention, the first label is the taller vegetation, not necessarily the predominant cover.

The degree of wetness of the vegetative covertypes is indicated by the water regime label. The coverage of each water regime label is presented in Table 6 by relative wetness (driest to wettest) and Table 7 arranged by frequency of coverage. Over half of the wetlands are considered saturated. This classification, 'B', represents a different hydrologic gradient than the other classifiers. In the saturated wetlands, the surface of the wetland soil expands or rises as it gets wetter, thus they are not considered flooded.

A total of 1223 ponds were identified in the Oswegatchie-Black drainage. Watershed data are presented in Appendix L arranged numerically by pond identification number and Appendix M arranged in decreasing percent area of watershed as wetland for all 1223 ponds. Lake ANC represents acid neutralizing capacity and lake DOC represents dissolved organic carbon. Lake elevation is included. These lake chemistry data were imported from an existing ALSC database file for demonstration.

One of the project goals was to test the capability of manipulating a large geographic databases with an existing tabular database using a significant Adirondack problem. Acid deposition and lake acidification was selected as a test problem. Lake acid-base chemistry is, in part, described by acid neutralizing capacity (ANC) and influenced by dissolved organic carbon (DOC). There is discussion in the literature that wetlands influence lake DOC and that DOC, in turn, influences the concentration of certain parameters in the aquatic cycle such as mercury, which is deposited from the atmosphere (Driscoll et al. 1994). Bioaccumulation of mercury by fish has been documented in remote lakes with low ANC, and is of concern to Adirondack scientists (Simonin et al. 1994).

One project objective was to create the wetlands and watershed coverages and examine them with an existing lake chemistry database. The Adirondack Long Term Monitoring (ALTM) lakes data were used and are presented in Table 8 and Table 9. These data do not show a strong correlation between general percent watershed as wetlands and lake DOC or lake ANC. It is expected however, that these digital databases will elicit closer examination of relationships between, for instance, wetland landscape position or covertypes and lake chemistry.

It has been demonstrated that GIS can be used to relate detailed watershed features, such as wetlands at less than one hectare threshold, over a sizable (400,000 hectare) area with other large tabular databases such as the Adirondack Lakes Survey Corporation lake chemistry data. Key to the success of this demonstration was the use of the ArcInfo version 7 Regions data model to create the hierarchical watershed labelling coverage that allowed complex watersheds to be identified relatively easily.

Continue reading next section of OB1 Report -- Public Forums