Application of GIS to
Rapid Inventory for Unit
Management Planning
Year 4 Summary Report
November 2007
Stacy McNulty,
Stephen Signell,
Benjamin
Zuckerberg, and
William Porter
Adirondack
Ecological Center
of SUNY ESF
on behalf of the
UMP-GIS
Consortium
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
2
Background
The Adirondack Park consists of a patchwork of publicly- and privately-owned land. The New
York State Department of Environmental Conservation (DEC) is responsible for stewardship of units of
publicly-owned land collectively called the Forest Preserve. Stewardship of the units is guided by a Unit
Management Plan (UMP), which conforms to guidelines set forth in the Adirondack Park State Land
Master Plan (APSLMP). An inventory of the natural resources and physical characteristics of a unit is
required to provide an understanding of the significant biological resources the DEC is charged with
managing, and to ensure optimal siting of proposed facilities such as trails and campsites. Only after
inventory has been completed can DEC planners identify management objectives to protect the resources
and allow public use consistent with the unit’s carrying capacity. The UMP-GIS Consortium arose from
the need to assemble existing digital data into a Geographic Information System (GIS) and develop
datasets and tools to facilitate the inventory portion of the UMP process in the Adirondack Park.
Overview
We report here on the fourth year of the five-year cooperative UMP-GIS project. The objectives were to:
1. Assemble the GIS database describing the ecological content of the units and adjacent lands.
2. Interpret the context of the unit within the surrounding landscape.
3. Provide training to DEC planners to enable future interpretation of GIS data.
4. Ensure protection and archival of the data.
The focus during year four shifted to park-wide issues, with less emphasis on working with planners
who now possess significant ability to access and interpret natural resource information. The year was
characterized by accelerated GIS modeling and maps of the entire Adirondacks. An ecosystems map was
derived from several GIS data layers, giving an overview of the areas of the park with higher potential
richness and biodiversity. These contextual analyses provided a means to begin looking at park-wide
issues of land management, such as connecting snowmobile corridors.
Discussion with DEC and other UMP-GIS partners culminated in a list of major priorities for year five:
1. Refine contextual analyses for park-wide assessment of natural and recreational resources
2. Refine Gap Analysis Program (GAP) data viewer tool
3. Continue to provide GIS training, analysis & maps for planners
4. Archive data and complete five-year project wrap-up and assessment
Year five is the last year of the UMP-GIS project as originally conceived. Since discussions began in 2001,
much has changed in the Adirondacks, in terms of land ownership, technology, data availability, and the
socio-economic picture for the region. There are still significant questions about how to balance
protection of natural resources, recreational opportunities, and maintenance of vibrant towns and
communities. Conservation easements represent a challenge for land managers and both public and
private land use; UMP-GIS can provide information useful to the dialogue. We will discuss with DEC
and other partners the current needs for GIS and ecological data, to best assist land planning.
The GIS database is a tool that provides a “first cut” at a comprehensive natural resource inventory. The
database allows planners to characterize the natural resources and biodiversity within a unit and in the
surrounding area, something that the DEC has recognized as integral to good Forest Preserve
stewardship. Although existing GIS data will not address every data requirement for comprehensive
inventory, the UMP-GIS database will allow DEC planners to focus on filling in information gaps and
access an objective information base to support the decision-making portion of the UMP process.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
3
Objective 1. Assemble a GIS database describing the ecological content
of the units and significant ecological attributes of immediately adjacent
public and private lands.
During year four we completed the Adirondack Ecosystems Model, which was three years in
the making. The following is an excerpt from an article recently submitted for publication in
the Adirondack Journal of Environmental Studies (AJES).
Introduction
Comprising over six million acres, with 2.5 million acres of public land, the Adirondack
Park is the largest protected wilderness east of the Mississippi River. Documenting and
maintaining biodiversity within the park is one of the major goals of those tasked with
managing public park lands (APSMLP 2001). The Adirondack Park contains many of the most
exemplary, contiguous and best-protected lands in the Northeast. However, in order to manage
the park’s ecosystems, planners need to know where ecosystems occur on the landscape, how
they are arranged in relation to one another, and to what extent the lands are protected and
healthy.
For the purposes of this project, ecosystems represent recurring groups of biological
communities that are found in similar physical environments and are influenced by similar
dynamic ecological processes. Over long periods of time, a region may support a succession of
ecosystem types as environmental variables change, e.g. as glaciers expand and contract, or as
mountains are formed and worn
Figure 1. Ecosystem and vegetation types in relation to
down.
In the Adirondacks, for
landscape
position
and
disturbance
regime.
example,
a
boreal,
coniferousdominated
ecosystem
formed
following glaciation might now
support a temperate hardwood
(deciduous) forest.
Over shorter periods of time
(e.g., a human lifetime), climatic
variations are less noticeable and
ecosystems seem relatively stable.
Various combinations of bedrock, soil,
elevation, and landform position
create unique environments that are
amenable to certain organisms and
communities. Variability in landscape
position or elevation, for example, can
give rise to different ecosystem types
Diagrammatic illustration of three landscape ecosystem
(Fig. 1). Note, however, that a single
types differentiated by landform position. Following
ecosystem type can be expressed as
clear cutting of part of ecosystem 2, two forest types (2a,
multiple ecosystem types 2a and 2b,
old growth beech-maple forest, and 2b, early successional
aspen/birch forest) are distinguished, illustrating the fact
often as a result of disturbance. Wind
that different cover types are not necessarily different
storms, light surface fires, insect
ecosystem types. Modified from Barnes, et al. 1997.
outbreaks and timber operations
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
4
usually do not alter the bedrock material or climatic conditions of an ecosystem.
Species assemblages such as early successional aspen forests or mature hardwood
forests may come and go, but the basic underlying processes and site conditions remain fairly
constant and so the ecosystem type does not change. For this reason, factors such as parent
material, physiographic position and moisture conditions must be considered along with land
cover type when mapping ecosystems. Many existing measures of ecosystem health rely on
current characteristics such as land cover which may reflect recovery from disturbance (e.g.,
agriculture, logging) and are therefore ephemeral. A better ecosystem map would model
potential conditions regardless of disturbance.
Toward this end, the UMP-GIS Consortium initiated the Adirondack ecosystems model
project in 2003 with the purpose of producing a GIS-derived ecosystem map of the Park. Our
objective was to develop a model that would map the distribution and locations of “potential”
ecosystems in the Adirondacks. The term “potential” is meant to describe the process of
identifying ecosystems based on GIS data. As such, these ecosystems represent the potential of
the landscape to support unique communities and are not necessarily a reflection of current land
cover and land use practices. The outcome of this project was a spatial dataset and
corresponding map of potential ecosystem types designed for use by land managers.
Methods
The backbone of the ecosystems map was a preexisting GIS layer of Ecological Land
Units (ELUs), originally developed by The Nature Conservancy (Anderson et al., 1999). The
ELU map was derived from several GIS layers including elevation, bedrock geology, parent
material, moisture availability, and landform and combined to create ELUs.
To convert the ELUs into an ecosystem map, we had to determine which single or
groups of ELUs form distinctive ecological systems or ecosystem types. Regional experts in
community ecology attended two workshops (June 2-3, 2004, and March 21-22, 2005) with the
aims of: 1) deciding on a classification system and 2) classifying ELUs within that framework.
The classification system we chose was derived from NatureServe's Ecological Systems
of
the
United
States
(see
Appendix
1
or
visit
http://www.natureserve.org/getData/USecologyData.jsp for more information on these
systems). This represents the first version of a mid-scale ecological classification developed by
NatureServe for use in conservation and environmental planning (Edinger et al., 2002). We
focused primarily on upland, forested areas rather than wetland communities, as the
Adirondack Park Agency is finalizing a high-quality, park-wide Wetland Cover Type map. The
workshops produced an ELU-ecosystem crosswalk spreadsheet which assigned an ecosystem
type to each ELU (Table 1; Appendix 1).
Table 1. Sample set of three ecological land unit codes from the ELU value attribute table for the
Northern Appalachians/Boreal Forest (NAP) ecoregion. Formula for calculating ELU codes:
ELU_Code =ELEVZONE+SUBSTRATE+LANDFORM30.
ELU Code
ELEVZONE
ELEVZONE_DESC
SUBSTRATE
SUBSTR_DESC
LANDFORM30
LF30_DESC
5113
5000
>4000ft
100
acidic sedimentary/
metasedimentary
13
Slope crest
3423
3000
1700-2500ft
400
moderately calcareous
sed/metased
23
Sideslope NW-facing
1831
1000
0-800ft
800
coarse sediments
31
Wet flats
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
5
However, we felt that the model could be improved by incorporating other data layers
such as the APA “mesosoils” layer, and a moisture index layer (SUNY-ESF, unpublished data).
These layers were used to make decisions in cases where experts believed that two different
ecosystems might occur within a single ELU code. For example, pixels with an ELU code of
2131 were classified as Lowland Spruce-Fir (LSF) if the moisture index score was > 90;
otherwise they were classified as Northern Hardwoods (NH).
Results
The final map contained 30 ecosystem types (Table 2; Fig. 2). Northern Hardwoods
(NH) and Lowland Spruce-Fir (LSF) were the most common ecosystems, together comprising
over 65% of the study area. Other relatively abundant types were Alkaline Hardwoods (AHF),
Lowland Alkaline Hardwoods (LAK), Alkaline Hemlock-Hardwood (AHH), Montane SpruceFir (MSF), Pine Hemlock-Hardwood and COVE communities, which together represent 21% of
the land area in the park. The alkaline ecosystems (AHH, AHF, and LAK) were concentrated in
the Champlain Valley in the eastern part of the park, and in some areas in the northwestern
portions of the park underlain by alkaline bedrock.
Table 2: Ecosystem types, codes and acreages (see Appendix 1 for descriptions).
Ecosystem
Code
21
17
28
24
6
19
7
16
13
4
27
25
30
23
18
15
1
26
11
3
5
10
29
2
8
14
12
22
20
9
Ecosystem
Abbreviation
NH
LSF
WATER
PHH
AHF
MSF
AHH
LAK
COVE
ACRO-WPRP
SWB
SAND
dry flats
NH-S
MAK
DOF
ACCT
SAND-PB
AKRO-MAK
ACRO-DOF
ACS
AKRO-AHF
WMSM
ACRO
AKCT
CSF
ALPB
NH-N
MSF-MAK
AKF
Ecosystem Name
Northern Hardwood
Lowland Spruce-Fir
Water Body
Pine-Hemlock-Hardwood
Alkaline Hardwood
Montane Spruce-Fir
Alkaline Hemlock-Hardwood
Lowland Alkaline
Cove Community
Acidic Rocky Outcrop-White/Red Pine
Subalpine Woody Barren
Sand Plain
Dry Flats
Dry Northern Hardwoods
Montane Alkaline
Dry Oak Forest
Acidic Cliff & Talus
Sand plain/pine barren
Alkaline Rocky Outcrop/Montane Alkaline
Acidic Rocky Outcrop-White/Dry Oak Forest
Acidic Swamp
Alkaline Rocky Outcrop/Alkaline Hardwood Forest
Wet Meadow/Shrub Marsh
Acidic Rocky Outcrop
Alkaline Cliff & Talus
Conifer Seepage Forest
Alpine Barren
Wet Northern Hardwoods
Montane Spruce-Fir/Montane Alkaline
Alkaline Forest
Total
*Ecosystems comprising < 0.1% of the area.
Acres
2,764,240
1,552,135
443,275
364,459
281,143
245,184
203,763
156,420
131,931
109,944
91,681
58,322
52,153
33,342
33,192
19,685
18,039
8,198
7,267
2,179
1,555
641
337
325
174
30
9
5
4
1
6,579,632
Percent of
Total Acres
42.0%
23.6%
6.7%
5.5%
4.3%
3.7%
3.1%
2.4%
2.0%
1.7%
1.4%
0.9%
0.8%
0.5%
0.5%
0.3%
0.3%
0.1%
0.1%
*
*
*
*
*
*
*
*
*
*
*
*
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
6
Figure 2: Ecosystems Map of the Adirondack Park, NY*.
*Less common types were combined for display purposes: NH-S and NH-N were grouped in with NH; ACROWPRP and ACRO-DOF were grouped with ACRO; MSF-MAK was combined with MSF, and AKRO-MAK and
AKRO-AHF were combined into a single AKRO category.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
7
From the model, we created an ecosystem richness map for the Adirondack Park,
defined as the number of ecosystems found within a 1 km radius of each pixel center (Fig. 3a) or
by Forest Preserve unit (Fig. 3b). There is great spatial variability in ecosystem richness, with
the highest scores in large river corridors and the eastern Adirondacks.
Ecosystem rarity was calculated by assigning higher value to ecosystems with fewer
pixels and then summing within a 1 km radius of each pixel center (Fig. 4a) or by unit (Fig. 4b).
While many of the rarest ecosystems occur in the Champlain Valley, this is also one of the most
heavily disturbed areas of the park, so remaining intact parcels in this area may be of value.
Figure 3: Ecosystem richness by pixel (a) and mean ecosystem richness by Forest Preserve unit (b).
a)
b)
Figure 4: Ecosystem rarity by pixel (a) and mean ecosystem richness by Forest Preserve unit (b).
a)
b)
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
8
Accuracy Assessment
How does one evaluate the accuracy of a map designed to show “potential” ecosystems?
Using land cover or other vegetation data to evaluate the accuracy of an ecosystem model is not
ideal because land cover can vary within an ecosystem, as discussed in the introduction. To
illustrate this problem, consider the fact that logging operations of the late 19th and early 20th
century selectively removed many conifers from Adirondack forests (McMartin 1994).
Consequently, conifers are greatly underrepresented in some modern forests. Similarly,
agricultural activities change the current vegetative expression on the ground. These
disturbances present a dilemma. If the model classifies an area as “Lowland Spruce Fir” and
the current vegetation data from the area contains mostly deciduous trees and few conifers, one
might conclude that the model was wrong. Yet the model may in fact have predicted the
ecosystem correctly, despite the current vegetation differing from its original configuration due
to logging. Without detailed spatial information on disturbance history, there is no way to
know which case is true.
Despite a lack of good alternatives, we sought to assess how closely the ecosystems
model approached current vegetative conditions in the Adirondack Park. We used groundbased vegetation data to evaluate the model with 874 forested plots or stands from four sources:
1) USDA Forest Service provided Forest Inventory and Analysis (FIA) data for 347 plots
distributed across the Adirondacks (http://fia.fs.fed.us/).
2) SUNY ESF’s Huntington Wildlife Forest Continuous Forest Inventory (CFI): 280 plots
located on 15,000 acres in the center of the park (http://forest.esf.edu/).
3) SUNY ESF NASA FoREST project: 167 plots, also located in the central part of the park on
state Forest Preserve land (http://forest.esf.edu/).
4) New York State Office of Real Property Services (ORPS) provided data for 80 upland
forest stands from across the Adirondacks.
The final data matrix contained 874 rows (plots) and 47 columns (46 species + eco_code).
We selected the 9 ecosystems with > 10 ground-based plots for further analysis.
To test how well the ground data matched the model, importance values were subjected
to discriminant analysis, a multivariate technique that classifies observations into two or more
groups based on pre-defined categories (in this case, ecosystem code). Discriminant analysis
provides a measure of how well groups can be distinguished based on the input data
(importance values).
According to the ground data, plots classified as NH were dominated by sugar maple,
beech and yellow birch. These species were present as minor components in LSF plots where
red maple, red spruce and balsam fir had greater importance. MSF and PHH composition were
similar to LSF, but MSF had a higher proportion of paper birch, and PHH had more pine and
hemlock. These findings are consistent with the NatureServe descriptions of these types
(Appendix 1). Composition of AHH was also generally consistent with the type description,
although these areas might better be termed Alkaline Pine-Hardwood forests due to more pine
than hemlock.
Data from the COVE, AHF and ACRO plots did not match NatureServe descriptions as
closely. This was reflected in the inability of discriminant analysis to distinguish these from
other types (Table 3). For example, most COVE plots, with a high percentage of sugar maple,
beech and yellow birch, were classified as NH. In general, the ability of the model to
discriminate correctly increased with sample size. However, on average, discriminant analysis
of predicted the correct ecosystem 56% of the time, with correct classification rates ranging from
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
9
7% (COVE, n=13) to 60% (LSF, n=188) (Table 3). This was substantially better than random—
when ecosystem codes were randomized, only 30% of the plots were classified correctly.
Table 3: Discriminant analysis of ground-based importance values (true group) and ecosystem code.
True Group
_
Ecosystem Code
ACRO
AHF
AHH
COVE
LAK
LSF
MSF
NH
PHH
ACRO
4
1
1
0
1
9
2
38
1
AHF
0
5
1
2
0
4
1
19
2
AHH
0
0
5
0
0
0
0
1
1
COVE
1
1
0
1
0
2
0
7
1
LAK
0
0
1
1
10
10
0
9
1
LSF
2
4
0
0
5
112
1
88
11
MSF
1
0
0
0
1
2
5
14
1
NH
4
5
1
7
3
36
2
306
6
PHH
1
0
2
2
0
13
0
41
20
Number of Plots
13
16
11
13
20
188
11
523
44
N correct
4
5
5
1
10
112
5
306
20
Proportion
0.308 0.313 0.455 0.077 0.500 0.596 0.455 0.585 0.455
Total Proportion Correct = 0.558
Total N = 839
Total N Correct = 468
Discussion
The ecosystems model is in substantial agreement with measurements of present-day
vegetation, indicating that the model performs reasonably well. The ecosystems model
represents an improvement over the National Land Cover Data set (NLCD) which is lessdetailed and based on current vegetation and classified satellite images. Some of the ecosystem
types in our model need refining with other GIS data layers or additional ground plots to
increase sample size, but overall the map represents the major Adirondack upland forest types.
As with any model, the ecosystems map is scale-dependent, useful only at spatial scales
of 1:62,500 or greater (for reference, 7.5 minute topographic maps are at a higher-resolution
1:24,000 scale). The bedrock geology and soil depth layers used to create the model were
derived from the APA mesosoils layer. The mesosoils layer has a minimum mapping threshold
of 40 to 100 acres (16-40 ha), and was meant to be used at a scale of at least 1:62,500. The 30m
pixel resolution of the output ecosystems map is potentially misleading and it would be
inappropriate to use the model for small to intermediate scale (< 800 ha) decisions such as
locating new trails or placing campsites. Therefore, the power of the ecosystems map lies
within its application to park-wide planning.
The ecosystems model can provide a number of benefits for unit management planning.
For instance, the model can be used to identify areas that are more biologically significant or
that might host species of interest. The model also identifies areas that are least protected or
most affected by current land use. Units with lower ecosystem richness or rarity could be
suitable for higher-intensity recreational activities, while units containing a high proportion of
rare ecosystems could be managed more protectively (cf. APSMLP 2001).
The Adirondack ecosystems model makes possible an ecosystem-based approach to
managing the processes that sustain species and their interactions within ecosystems (Poiani et
al. 2000). Most decision-making occurs at smaller scales (e.g., individual wetlands, deer
wintering yards, old growth patches) and local species assemblages. The ability to summarize
ecosystem richness and rarity by Forest Preserve unit provides a larger context by which
planners can assess the importance or contribution of a unit to park-wide biodiversity. Because
the Adirondack Park makes up 20% of the Northern Appalachian/Boreal ecoregion and
contains some of its most contiguous blocks of forest (Anderson et al. 1999), natural resource
management should incorporate larger spatial scales, based in part on the ecosystems model.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
10
Objective 2. Interpret the context of the unit within the surrounding
landscape.
Park-wide Contextual Analyses
In state land planning, as directed by the State Land Master Plan (APSLMP 2001), DEC
staff are to consider the unit within the context of the immediate public and private lands and
should consider the unit compared to the park as a whole. Abundance of important natural
features, availability of recreational facilities and degree of human impact vary widely across
units. Knowledge of a unit’s park-wide rank in terms of these factors will help the DEC make
more informed management decisions. For example, units with highly sensitive ecological
systems might be managed with biodiversity protection as top priority, whereas units heavily
impacted by camping or other human activity are managed differently.
During year four, we focused on this objective to a greater extent. We created several
GIS models designed to meet two objectives:
1. Identify park-wide spatial/temporal patterns of natural resources and of human use
(recreation and impact) to identify unique areas and similar areas.
2. Evaluate grouping of Forest Preserve units into complexes to improve management
coordination/efficiency.
There are many park-wide GIS data layers relating to natural features, human impact
and recreation. However, considering each data layer separately does not produce a clear
picture of how units rank in comparison to others. Therefore, we combined data layers into
simple indices and created three preliminary GIS models designed to provide an overview of
how resources are arranged spatially within the park:
1. Natural Features Index. Where are the most valuable/numerous natural features?
2. Recreational Facilities Index. How are recreational opportunities distributed
3. Human Impact Index. Which areas have experienced the most human impact?
These new data layers can be combined in various ways to illuminate the unique aspects
of each unit and how it relates to the park as a whole.
Table 1: Input data layers used to create the park-wide models.
Index
Natural Features
Recreational Facilities
Human Impact
Data Set
Primary Aquatic communities
State Land Acquisitions
Charismatic Megawetlands
Natural Heritage Points
Ecosystem Richness
Ecosystem Rarity
Recreation Points
Public W ater Bodies
Terrain Ruggedness
Aquatic Invasive Infestations
Traffic Density
Roads
Sulfate Deposition Model
Index of Biotic Integrity (IBI)
Data Type
polyline
polygon
polygon
point
grid
grid
point
polygon
grid
polygon
polyline
polyline
grid
polygon
Data Source
TNC
APA
APA
NYS Heritage Program
SUNY-ESF
SUNY-ESF
APA
APA
SUNY-ESF
TNC
NYS DOT
APA
SUNY-ESF
SUNY-ESF
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
11
Methods
Each index represents an amalgamation of several existing GIS point, line, and polygon
(vector) and grid (pixel-based) data sets (Table 1). As it is difficult to compare these disparate
data types, we first needed to convert all data to grids so that they could be mathematically
combined to create a single index data layer. One way to create grids from vector data is to
calculate a “distance” layer showing how far all points are from an feature of interest. This
approach seemed especially appropriate for these models, because proximity to features of
interest can be important. For example, lakes and streams close to existing aquatic invasive
plant infestations are more vulnerable to invasion than remote water bodies. Accessibility of
scenic vistas decreases with distance from mountaintops. Upland communities surrounding
important wetland complexes or pristine aquatic communities may be particularly sensitive to
human impact. In addition, some data sets are mapped as points but should actually be
polygons that have not yet been digitized. These data sets, specifically the NY Natural Heritage
Program Element Occurrences (EOs) of ecological communities such as a cedar swamp, may
not be representative of the spatial extent of natural features. Therefore a distance grid is a
better approximation than a single EO point. For these reasons, we used the distance approach
extensively in the creation of the three indices.
We converted all vector data to grids by calculating distances, e.g., “distance from
roads” or “distance from pre-1900 land acquisitions.” Each of these distance grids was then
reclassified on a 1-10 scale using a geometric mean (Fig. 1). The process was repeated for each
vector data layer. Data layers already in grid format were also reclassified on a 1-10 scale. Once
all data layers were standardized, calculating the indices consisted of adding all the input data
layers together and reclassifying the combined grid back to a 1-10 scale (Fig. 2). Figure 3 shows
the three output models in panels A-C. One advantage of working with grids is that they may
be transformed mathematically, for instance by adding or subtracting. Figure 3D shows a layer
created by subtracting the Human Impact Index from the Natural Features Index. This
identifies areas (shown in dark green) with important natural features and little human impact.
Index scores were summarized for each unit, and mean scores obtained. These scores were then
used to create a table ranking each unit for each of the three indices (Table 2).
Figure 1: Example of how polygon-, line- or point-based data layers were converted to grids. A) Primary aquatic
communities layer (Source: The Nature Conservancy). B) Grid of distance to primary aquatic communities. C)
Reclassified grid, with scores ranging from 1 (furthest) to 10 (closest).
A)
B)
C)
Application
of GIS toofRapid
Inventory
forunits
Unit Management
- Year
4 Report
12
Table
2: Rankings
Forest
Preserve
according toPlanning
model index
scores.
(1=highest score, 50=lowest).
Cells are color-coded to facilitate interpretation. For Natural Features and Recreational Facilities, units with the
highest scores are colored green, and lowest scores are red. The color scheme for Human Impact ranks is
reversed, as high scores in this category are not desirable, and thus colored red.
UNIT
Hudson Gorge
Raquette Jordan Boreal
Hammond Pond
Whiteface Mt
Fulton Chain
Hammond Pond West
St Regis
Jessup River
Saranac Lakes
Raquette River
Pigeon Lake
Giant Mt
Dix
Vanderwhacker
McKenzie Mt
Hoffman Notch
High Peaks
Siamese Ponds
Debar Mt
Silver Lake
Sentinel Range
West Canada Lake
Lake George
William Whitney
Wilcox Lake
Moose River Plains
Sargent Ponds
Pharoah Lake
Split Rock
Hurricane Mt
Five Ponds
Blue Ridge
Jay Mt
Bog River
Round Lake
Wilmington
Grasse River
Blue Mt
Cranberry Lake
Ferris Lake
Aldrich Pond
Shaker Mt
Watson East Triangle
Ha De Ron Dah
Taylor Pond
Black River
Pepperbox
Independence River
Chazy Highlands
Whitehill
Natural
Features
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Human
Impact Rank
31
48
10
9
1
4
29
11
6
47
32
22
30
18
17
27
45
37
33
19
23
42
2
46
14
15
13
20
3
26
49
16
36
35
40
8
43
24
38
21
41
5
50
25
7
28
44
34
12
39
Recreational
Facilities
rank
30
43
7
1
3
17
2
12
13
40
27
15
21
23
6
35
33
25
20
24
10
29
11
39
31
14
5
4
42
8
37
16
38
19
34
18
49
22
9
36
47
26
50
32
28
41
44
45
48
46
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
13
Figure 2: Six GIS data layers were used to create the Natural Features Index. All distance grids were first reclassified
to a 1-10 scale, and then added together to create a grid with values ranging from 12-55. This was reclassified to a 110 scale to produce the final “Natural Features Index” grid shown below, center.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
14
Figure 3: Park-wide indices, shown with Forest Preserve unit boundaries (black polygons). Red indicates low, yellow
medium and green high scores. A) Natural Features Index: higher scores indicate areas with more natural features.
B) Recreational Facilities Index: higher scores indicate areas with more recreational facilities. C) Human Impact
Index: higher scores (red) indicate areas of greater human impact. D) Natural Features Index minus Human
Impact Index: high scores (green) indicate areas with important natural features and little human impact. Low
scores (red) are areas of low natural interest and high impact.
A)
B)
C)
D)
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
15
These preliminary models were intended as a “first cut” towards park-wide modeling. The
intent is to modify the inputs and parameters of these models based on the input of DEC
professionals and others with expertise in the Adirondack Park.
Below are the
recommendations made during our first park-wide model workshop (see Appendix 2).
Recommendations for Improving Park-wide Analyses:
• All three indices:
o Develop confidence rating for each input data layer
• Natural Features Index
o NYNHP data:
ƒ
Digitize EO polygons where needed
ƒ
Weight EOs by global/state ranking
• Recreational Opportunities Index
o Datasets to include:
ƒ
Snowmobile trails
ƒ
Destinations - distance to a scenic view/vista – including waterfalls
ƒ
Northern Forest Canoe trail, other routes (North Country Hiking Trail)
ƒ
Railroad tourism train corridors
ƒ
Publicly accessible water bodies including ponds > 5 acres
ƒ
Motorized access lakes and floatplane allowed lakes
ƒ
Fish-stocked lakes
ƒ
Hunting data by town
ƒ
Conservation easements
ƒ
Private campgrounds and summer youth/other group camps
o Trails – negative impact of human use (for now use registers and campground
occupancy). Use DEC data on hiker perceptions/behavior.
ƒ
Use trail density as input
ƒ
Need: hiking impact on summits
ƒ
Need: comprehensive trail database, and park-wide survey of users of the
Forest Preserve. Continue to coordinate with Chad Dawson, SUNY-ESF.
o Roads - valued as both positive for access (e.g., hunting and disabled access), and
negative (e.g., noise, impact on wilderness experience/solitude)
ƒ
Need: distance from road and threshold or values for noise levels
• Human Impact Index
o Roads – also see above
ƒ
Need: buffer size for impact of road on wildlife mortality/collisions
ƒ
Need: finished Statewide Transportation database
ƒ
Need: county, local roads data – by county?
ƒ
Three road dataset options. Roads used in models were A) state DOT
database. B) ALIS road dataset has trails included and would need to be
sorted out. C) Adk_roads from APA shared CD may be best for non-state
roads.
ƒ
Weight roads by class or other feature?
o Land use?
ƒ
Use ORPS tax map parcel land use codes
o Mines and other natural resource impact areas
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
o
o
16
Pollution – Atmospheric pollution data are based on a few weather stations and
interpolated/krigged maps, therefore they don’t show variation between sites
well.
ƒ
Use ALSC acidified lakes data
ƒ
Need: mercury data
Tourism/other economic data by town
Automated Modeling
The suite of GIS tools continued to grow in year 4, and we made steps toward
improving their user-friendliness so planners can run the ArcGIS models to generate alternative
management plans with minimal assistance. In year 4, we purchased Visual Basic software and
sought advice on creating a simple “front end” graphical user interface. The next step is to
automate the ArcGIS Model Builder models, as well as create a Gap Analysis Program data
viewer. This viewer will make information on wildlife-habitat associations more useful (e.g.,
planners can summarize the probable wildlife species and habitat present in a unit).
Objective 3. Provide training to DEC planners to enable future
interpretation of GIS data.
Training during year four consisted primarily of one-on-one sessions with planners as
requested. The change in leadership at the state and DEC regional levels resulted in a reduction
of UMP activity during year four. It is anticipated that year 5 will see more UMP activity and
increasing need for UMP-GIS analyses, training and assistance.
Objective 4. Ensure protection and archival of data.
We continue to work with DEC to incorporate datasets into the Master Habitat Data
Bank, accessible to all DEC planners. By the end of 2008, we will transfer all remaining
datasets, maps, and metadata not already in the MHDB to DEC. We will also hold a meeting of
UMP-GIS partners to determine which, if any, datasets should be made public and which
should be accessible by permission only.
The ecosystems map and contextual analyses can be integrated into the MHDB,
although we caution that usage of the park-wide summary data layers is contingent upon
proper adherence to scale and other limitations of the original data. In other words, the parkwide maps of ecosystem rarity, richness, etc. should not be “zoomed in” for use within a single
Forest Preserve unit.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
17
Summary and Future Plans – Building on Year 4
We are pleased that major efforts such as the Adirondack ecosystems model and parkwide contextual analyses came to fruition in year 4. Discussion with DEC and other UMP-GIS
partners culminated in a list of four major priorities for year five, as described earlier:
1. Refine contextual analyses for park-wide assessment of natural and recreational
resources
2. Refine Gap Analysis Program (GAP) data viewer tool
3. Continue to provide GIS training, analysis and maps for planners
4. Archive data and complete five-year project wrap-up and assessment
The park-wide analyses will be refined over the fifth year of the project, having already
garnered significant attention from state, local, and nongovernmental organizations interested
in land planning and the future of the park. Work will continue on the three indices for parkwide planning, both in improvement of the models in concert with integration of additional
data sets into the models.
Another focus of year 5 will be to make the suite of GIS tools user-friendly so planners
can run the ArcGIS models to generate alternative management plans with minimal assistance.
Planners will input values into a short series of windows with pull-down menus and run the
models to quickly produce and compare alternatives. We will also create a GAP data viewer
tool. Concomitant with these user-friendly interfaces will be training to familiarize planners
with running the tools themselves. Both tools will be built into the MHDB.
The UMP-GIS project received positive feedback from the many meetings at which we
presented (Appendix 3), including two awards. Steve Signell received an award in the North
East Map Organization Map Design Competition, winning the “Best Design in Small Format
Map” category for his portrayal of invasive species around the Saranac Lakes. Ben Zuckerberg
won best student paper at the 2007 International Association of Landscape Ecologists for his
work on Breeding Bird Atlas data and bird population change. We are honored to receive these
awards from GIS and scientific colleagues as positive reflections of the UMP-GIS staff and
cooperative project.
Year five is the last year of the UMP-GIS project as originally conceived. Since
discussions began in 2001, much has changed in the Adirondacks, in terms of land ownership,
technology, data availability, and the socio-economic picture for the region. There are still
significant questions about how to balance protection of natural resources, recreational
opportunities, and maintenance of vibrant towns and communities. The many, varied
conservation easements represent a challenge for land managers and both public and private
land use; UMP-GIS can provide information useful to the dialogue. We will discuss with DEC
and other partners the future needs for GIS analyses and interpretation of spatially-based
ecological data, to best assist land planning in Adirondack Park. The UMP-GIS Consortium
continues to provide important syntheses of information to help DEC in decision-making at
both the unit and the park level, as laid out at the beginning of the project.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
18
Bibliography
Adirondack Park State Land Master Plan (APSLMP). 2001. New York State Adirondack Park
Agency, Ray Brook, New York.
Anderson, M., P. Comer, D. Grossman, C. Groves, K Poiani, M. Redi, R. Schneider, B. Vickery,
and A. Weakley. 1999. Guidelines for representing ecological communities in ecoregional
conservation plans. The Nature Conservancy.
Barnes, B. V., D. R. Zak, S.R. Denton, and S.H. Spurr. 1997. Forest Ecology. New York, USA,
John Wiley & Sons, Inc.
Edinger, G. J., D. J. Evans, S. Gebauer, T. G. Howard, D. M. Hunt, and A. M. Olivero, editors.
2002. Ecological communities of New York state. Second edition. A revised and expanded
edition of Carol Reschke's ecological communities of New York state. (Draft for review). New
York Natural Heritage Program, New York State Department of Environmental Conservation,
Albany, NY.
Poiani, K.A., B.D. Richter, M.A. Anderson, and H.E. Richter 2000. Biodiversity at multiple
scales: functional sites, landscapes and networks. Bioscience 52:133-146.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
Appendix 1. Ecosystem code definitions.
Ecosystem
Code
Ecosystem Name
Short Description
ACCT
Acidic cliff and
talus
Cliff or talus with acidic bedrock (other than shale) on and at the base of steep slopes
(>45 degrees) at low to mid elevation, with physiognomy ranging from exposed rock to
forest, conifer to deciduous.
ACRO
Acidic rocky
outcrop
Shallow soil, acidic bedrock hilltops at low to mid elevation, with open canopy
physiognomy ranging from exposed rock to woodland, conifer to deciduous.
ACS
Acidic Swamp
Acidic, forested basin wetlands with saturated to very poorly drained mineral soils, with
physiognomy ranging from conifer to deciduous.
AHF
Alkaline
Hardwood Forest
Alkaline to circumneutral, dry (to wet), forested uplands at low to mid elevation
covering a wide range of landform settings (depending on the variant), often on rich
loamy till or colluvium, with deciduous physiognomy.
AHH
Appalachian
HemlockHardwood Forest
Acidic, moderately dry, well drained to moderately well drained, low- to mid-elevation
matrix forested uplands covering a wide range of landform settings, with a mix of oaks,
northern hardwoods, white pine and hemlock and with physiognomy ranging locally
from deciduous to conifer. Indicative in the Adirondack Park of LNE.
AKCT
Alkaline Cliff and
Talus
Cliff or talus with alkaline to circumneutral bedrock (other than shale) on and at the
base of steep slopes (>45 degrees) at low to mid elevation, with physiognomy ranging
from bare rock to forest, conifer to deciduous.
AKF
Alkaline Fen
Open canopy basin wetlands over alkaline to circumneutral bedrock with inundated to
very poorly drained fibrous to sphagnous peat soils, with physiognomy ranging from
grassland to shrubland (to woodland).
AKRO
Alkaline rocky
outcrop
Shallow soil, alkaline to circumneutral bedrock hilltops at low to mid elevation, with
open canopy physiognomy ranging from exposed rock to woodland.
AKS
Alkaline Swamp
Alkaline to circumneutral, forested (to woodland) basin wetlands with saturated to very
poorly drained mineral to peat soils, with physiognomy ranging from conifer to
deciduous.
ALPB
Alpine barrens
Alpine, lands >4,000 feet or high exposed dry to wet hilltops with essentially acidic
bedrock, with open canopy physiognomy ranging from dwarf shrubland to exposed
rock. Indicative of NAP.
CPF
Clayplain Forest
Alkaline to circumneutral, forested uplands on wet to moderately wet, poorly drained
flats with relatively impermeable clay soils (deep fine sediments) at low elevation with
deciduous physiognomy. Indicative of STL.
CSF
Conifer seepage
forest
Forested seeps with calcareous bedrock; toe of slope positions physiognomy
exclusively conifer, dominated by northern white cedar.
DOF
Dry Oak Forest
Acidic to circumneutral, forested uplands with dry to moderately dry, well drained
shallow soils over bedrock on mid-elevation hilltops, slope crests, and upper
sideslopes, with a mix of oaks, hickories, and northern hardwoods and with deciduous
physiognomy. Indicative in the Adirondack Park of LNE.
LAK
Lowland Alkaline
Forest
Alkaline to circumneutral, conifer-dominated forested lowlands typically on dry to wet
flats and the lowest parts of sideslopes on somewhat moist and somewhat poorly
drained, moderately deep loamy soil; northern white cedar is the dominant conifer.
LSF
Lowland Spruce
Fir Forest
Acidic, wet to somewhat moist, somewhat poorly drained, low-elevation matrix forested
uplands on till flats and the lowest parts of sideslopes, with physiognomy ranging from
conifer to mixed and dominated by a mix of red spruce and fir.
MAK
Montane Alkaline
Forest
Convex, alkaline to circumneutral, conifer-dominated forested uplands typically on
hilltops and slope crests on somewhat dry and somewhat excessively well drained,
shallow, rocky soil; northern white cedar is the dominant conifer.
19
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
Appendix 1. Continued.
MSF
Montane sprucefir hardwood
forest
Acidic, dry to somewhat poorly drained, high-elevation (subalpine) matrix forested
uplands covering a wide range of convex landform settings, with primarily conifer
physiognomy (except when fire altered) and dominated by a mix of red spruce and fir.
Indicative of NAP.
NH
Northern
Hardwood Forest
Acidic (to circumneutral) mid- (to low-) elevation matrix forested uplands, typically on
deep mesic, well drained loamy till soils on variable landform settings, dominated by a
mix of northern hardwoods and red spruce with physiognomy ranging locally from
deciduous to conifer.
NH-N
Northern
Hardwood Forest
(North-facing
slopes)
Northern
hardwood-pinehemlockhardwood forest
Northern
Hardwood Forest
(South-facing
slopes)
Pine Barrens
NH with dominance of spruce-northern hardwood forest (estimate 60 to 100%) over
beech-maple mesic forest (0 to 40%), thus with greater proportion of conifer and mixed
cover types; in theory: typically on steep to moderately steep N-facing slopes, cooler
microclimates and moister soils.
PHH
Pine-HemlockHardwood Forest
Acidic, forested uplands with moderately dry, well drained to moderately well drained,
moderately deep, somewhat coarse till soils (typically gravelly to sandy loam) at mid
(to low) elevations on variable (but discrete) landforms (hilltops, slope crests, upper
sideslopes and stream ravine coves, the latter especially in the SE part of Adirondack
NAP), with physiognomy ranging from conifer to mixed and with a mix of white pine
and/or hemlock codominant.
COVE
Cove Forest
Matrix forest associated with cove landforms
SAND
Sandplain
Mid- to low-elevation acidic flats, typically on glacial outwash plains, with deep, well
drained to excessively well drained sandy soil, with physiognomy ranging from
exposed substrate to forest, the latter primarily conifer, and canopy dominants varying
depending on regional variants, and with a various mix of community types.
SWB
Subalpine
Woodland and
Barrens
Acidic, high-elevation (alpine to subalpine) woodland to barrens on hilltops and slope
crests, dominated by a mix of red spruce and fir with physiognomy ranging from
exposed rock to conifer-dominated woodland (krummholz). Indicative of NAP.
WMSM
Wet MeadowShrub Swamp
and Marsh
Basin wetlands on flats over acidic (to circumneutral, possibly even to alkaline)
bedrock with inundated or saturated (to poorly drained) mineral soils (and fibrous
peat), with open canopy physiognomy ranging from deep emergent/submergent
vegetation to shrubland.
WPRP
White Pine-Red
Pine Forest
Acidic, forested uplands with dry to moderately dry, well drained to somewhat well
drained, relatively shallow soils, in variable (but discrete) landform settings at low to
mid (to high) elevations (shallow soils over bedrock on mid- (to high-) elevation hilltops,
sandy soils on kames/eskers, and sandy soils on outwash sand and gravel flats), with
conifer to mixed physiognomy and with a mix of white pine and red pine. Indicative of
NAP.
NH-PHH
NH-S
PB
Low-elevation matrix forest (conifer and mixed dominated matrix forest)
NH with dominance of beech-maple mesic forest (estimate 60 to 100%) over sprucenorthern hardwood forest (0 to 40%), thus with greater proportion of deciduous cover
type; in theory: typically on steep to moderately steep S-facing slopes, warmer
microclimates and drier soils.
Low-elevation sandplain flats dominated by pitch pine, possibly oaks, and heaths with
physiognomy ranging from exposed sand to forest, the latter conifer to mixed.
Indicative of STL & LNE.
20
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
21
Appendix 2. Meetings with DEC staff and Consortium Members.
Jan 11, 2007
Steve met with Rob Daley, DEC planner in Ray Brook, to help him with several GIS projects.
April 6, 2007
Steve met with Sean Reynolds, DEC planner in Ray Brook, regarding impoundment modeling
in Debar Mt. Wild Forest.
April 30, 2007
Steve met with Stuart Messenger of Adirondack Mountain Club regarding development of GIS
datasets for use in park-wide planning.
September 18, 2007 – Park-wide Analysis and Models Meeting, Newcomb
Present:
John Barge, APA GIS
Colin Beier, AEC
Rick Fenton, DEC Region 5 Northville
Peter Frank, DEC Forest Preserve Bureau, Albany
Ed Frantz, DOT Adirondack Park and Forest Preserve Manager, Utica
John Gibbs, DEC Region 6 Potsdam
Stacy McNulty, AEC
Steve Radminski, DOT Albany GIS Coordinator
Steve Signell, AEC
Elizabeth Spencer, NY Natural Heritage Program
Rick Weber, APA State Lands Planning
Notes:
Discussed park-wide models and indices, future project goals and timeline
Delivered:
Presentation on park-wide analysis
Received:
Tasks:
See body of Year 4 Report
October 10, 2007
Steve met with Liz Arabdjis of Fountains Spatial in Lake George for consulting on developing
ArcGIS tool for viewing NY state GAP data.
Application of GIS to Rapid Inventory for Unit Management Planning - Year 4 Report
22
Appendix 3. Presentations at Professional Conferences and Meetings.
Zuckerberg, B. How to determine songbird abundance and diversity on your property.
Society of American Foresters, Liverpool, New York, February 07, 2007.
Zuckerberg, B., W. F. Porter, and K. Corwin. Implications of the Abundance-Occupancy Rule:
Can Atlas Data be Used to Monitor Avian Population Change? 22nd Annual Conference of the
US International Association of Landscape Ecologists, Tucson, AZ, April 9-13, 2007.
Signell, S.A., S. A. McNulty, and W.F. Porter. Using GIS to facilitate a shift toward a park-wide
approach to Adirondack Preserve planning. Adirondack Park Agency, Forest Preserve
Advisory Committee (FPAC) Meeting. Albany, New York, June 19, 2007
Signell, S.A., S. A. McNulty, and W.F. Porter. Using GIS to facilitate a shift toward a park-wide
approach to Adirondack Preserve planning. Adirondack Research Consortium (ARC), Tupper
Lake, NY, May 22, 2007
Signell, S.A., S. A. McNulty, and W.F. Porter. Using GIS to facilitate a shift toward a park-wide
approach to Adirondack Preserve planning. Adirondack Park Agency, Ray Brook, New York,
July 12, 2007
Zuckerberg, B., A. Woods, and W.F. Porter. Repeating Patterns: Birds Ranges Moving
Northward in New York State. Ecological Society of America, San Jose, CA, August 6-10, 2007.
Zuckerberg, B., A. Woods, and W.F. Porter. Repeating Patterns: Birds Ranges Moving
Northward in New York State in Response to Climate Change. Connecticut College, Biology
Seminar, October 17th, 2007.
Signell, S.A., S. A. McNulty, and W.F. Porter. ArcGIS tools for recreation and conservation
planning: examples from the Adirondack Park, NY. 14th Annual Conference of The Wildlife
Society, Tucson, AZ, September 22-27, 2007.
Zuckerberg, B., W.F. Porter, and K. Corwin. Implications of the Abundance-Occupancy Rule:
Can Atlas Data be Used to Monitor Avian Population Change? 14th Annual Conference of The
Wildlife Society, Tucson, AZ, September 22-27, 2007.
Zuckerberg, B., A. Woods, and W.F. Porter. Repeating Patterns: Birds Ranges Moving
Northward in New York State. 14th Annual Conference of The Wildlife Society, Tucson, AZ,
September 22-27, 2007.