Finger Lakes National Forest Ecological Mapping Study

1
Finger Lakes National Forest Ecological Mapping
Study Stephen D. DeGloria1 and Diane Harlow
Burbank2 1 Institute for Resource Information
Sciences, Cornell University, Ithaca, New York
(sdd4_at_cornell.edu) 2 Green Mountain Finger
Lakes National Forests, USDA Forest Service,
Rutland, Vermont (dburbank_at_fs.fed.us)
Part 2 Land Type Associations
Part I Ecological Land Types
Results Discussion
Finger Lakes Region Subsections 212Fa
Cattaraugus Highlands 212Fb Central Allegheny
Plateau 222Ib Erie-Ontario Till Plain 222Ic
Eastern Ontario Till Plain 222Id
Cattaraugus-Finger Lakes Moraine and Hills 222Ie
Eastern Ontario Lake Plain 222Pa Lake
Ontario, Bays, and Island M212Fa Tug Hill
Plateau M222Fb Tug Hill Transition
Abstract
Abstract
Results Discussion
An ecological land type classification system was
developed based on established USDA Forest
Service practices. The land types predict the
potential forest composition and productivity of
major landscape units in the absence of natural
or human-induced disturbance. Land types and
phases were mapped by combining selected
attributes of soil map units and terrain
variables derived from a digital elevation model.
Ecological land types were modified based on
effective rooting depth and soil drainage.
Ecological land type phases were derived by
partitioning modified ELT's by slope gradient
class and dominant soil series. Landscape units
within the forest were evaluated for four
management objectives (1) perpetuation of
oak-hickory forest communities, (2) re-allocation
of grassland resources, (3) potential extraction
of aggregates, and (4) expansion of blueberry
production areas. Suitability maps and ratings
were generated by reclassifying and merging
individual factor maps of soil properties,
terrain variables, and land use. Color-coded
interpretive maps, soil monoliths, and inventory
data were produced in various formats for use by
Forest Service resource specialists, land
managers, and other environmental professionals.
Land Type Associations (LTA) were developed for
the Finger Lakes National Forest (FLNF) and
environs. Land Type Associations are used for
Forest-wide planning and are required for the
FLNF forest plan revisions currently in progress.
Terrain and soil variables were integrated to
represent dominant landscape-level processes that
directly impact management of forest resources on
the FLNF. Six Land Type Associations are defined
which occur in variable proportions in the
Central Allegheny Plateau, Cattaraugus-Finger
Lakes Moraine and Hills, and Eastern Ontario Till
Plain Subsections. Further refinement and
characterization of these Land Type Associations
are required in concert with possible
re-definition of the transition zone between the
Central Allegheny Plateau and Eastern Ontario
Till Plain Subsections.
Key Spatial Environmental Variables
Figure 1. Ecological units of the eastern United
States for the Finger Lakes region. The boundary
of the FLNF is shown in all illustrations.
Background
Ecosystem management requires spatially-referenced
data and information to make informed decisions
regarding the sustainable development and use of
environmental resources. In every case,
information in a form suitable for collaborative
spatial decision-making is required. Ecological
land classification systems are required to help
organize our knowledge of the composition,
structure, and function of ecosystems at various
scales of space, time, and complexity. The USDA
Forest Service has developed an ecological land
classification and mapping framework for the
multiple-use management and planning of our
national forests (USFS 1993). The framework
describes a spatially nested, hierarchical system
for classifying units of land by their ecological
capabilities and tendencies, and recognizes that
at different spatial scales, different ecological
processes become dominant in driving the
composition, structure, and function of
ecosystems. For the eastern United States, an
ecological classification system has been
developed which expands the hierarchical
eco-region classification of Bailey, et. al.
(1994), and serves to define landscape units on
the basis of their geomorphic, pedological,
ecological, and hydrological characteristics
(Keys, et. al. 1995). These landscape units are
defined as sections and subsections, and are
nested within broader-scale provinces as defined
by the eco-region classification (Figure 1).
Subsections are further divided at finer spatial
scales into Land Type Associations (LTAs),
Ecological Land Types (ELTs), and, ultimately,
phases of Ecological Land Types (ELTp). Land
Type Associations represent an intermediate level
in the hierarchy, and are perceived to be useful
in Forest- and area-wide planning, analysis, and
management. The challenge facing land management
organizations is to define, quantify, and
visualize important ecological relationships to
advance resource planning and decision-making at
landscape scale.
Figure 3. 1988 land cover map used for change
analysis (1938-1988) and blueberry suitability
assessment (Fig. 15).
Figure 2. Digital Elevation Model (DEM) used to
generate terrain variables for modeling ELT and
resource suitability modeling.
Figure 1. Soil map units for FLNF and look-up
table for technical groupings of soil properties
for ELT modeling.
Background
Figure 2. Shaded relief map derived from 3-arc
second DEM.
Figure 3. Slope gradient classes derived
from 3-arc second DEM.
Management of public lands for multiple uses
requires timely and accurate resource
inventories. Data derived from such inventories
are used to develop land and resource management
plans based on ecological land type units mapped
at variable spatial scales. The Forest Service
defines ecological units as a function of
biophysical processes occurring from local to
landscape scales. These units are used for
forest management, planning, monitoring, and
evaluation at various spatial scales depending on
the administrative unit under consideration.
Ecological Land Types (ELT) and Land Type
Associations (LTA) are used for Forest-wide
planning, and are required for the FLNF forest
plan revisions currently in progress.
Spatial Modeling of Ecological Land Types
Figure 4. Soil parent material map derived
from State Soil Geographic Database (STATSGO).
Figure 5. Soil acidity classes derived from
STATSGO map unit reclassification.
Project Objectives
(1) evaluate the quality of existing soil
resource inventories, (2) determine the nature
and distribution of ecological land types, (3)
determine the suitability of landscape units for
meeting specific resource management objectives,
and (4) produce interpretive maps and statistical
data for land management, public information, and
environmental education and research programs in
the Finger Lakes National Forest (FLNF), New
York.
Figure 6. Soil acidity map.
Figure 7. Effective rooting depth map.
Figure 4. Soil parent material map.
Figure 5. Soil texture map.
Materials and Methods
?Evaluation of soil resource information for the
national forest included remapping,
re-correlating, and geo-referencing soil survey
data in southern Seneca and northern Schuyler
counties. Field transects were used to determine
free carbonate distribution and solum depth
relationships as a function of landscape position
and glacial till lithologies. Soil pedons were
sampled to derive soil physical, chemical, and
mineralogical data by soil horizon. Two
permanent transects were established to estimate
and monitor annual soil temperatures for
classification purposes. ?Interpretive maps for
soil parent material, soil texture class, soil
acidity, effective rooting depth, and soil
drainage class were derived by a series of
reclassification steps for each soil map unit
from the re-correlated soil survey for the
Forest. ?Terrain variables of slope aspect and
slope gradient were derived from the 3-arc second
DEM for the Finger Lakes region (USGS
1990). ?Derivative soil and terrain variables
were modeled spatially using the ERDAS MATRIX
command. The following is an example of the
spatial modeling process where soil parent
material class is combined with soil texture
class on a grid-cell basis DATA Input parent
File e/flnf/giss/par_material Input
texture File e/flnf/giss/texture Output
matrix1 File e/flnf/giss/gismo/par_tex START
Matrix1 either 0 if (parent is 0 or texture
is 0) or (parent (3 (texture-1)))
Otherwise END ?Any resulting matrix class,
including the final ELT model output, occupying
less than 1 of the total FLNF land area was
reclassified into the most closely related class.
This minimum size delineation was used as the
threshold value based on the range of ELT size
from tens to hundreds of hectares (Russell and
Jordan 1991). ?Maps for Red Oak, Pasture, and
Aggregate Extraction suitabilities were derived
from woodland suitability, trefoil-grass hay
productivity, and construction materials tables,
respectively, from the Schuyler and Seneca County
soil survey interpretation tables. The Blueberry
Suitability map was produced using a rule-based
spatial model combining soil drainage class,
slope aspect, land use/land cover conditions, and
proximity to roads. ?All spatial data were
geo-referenced to the Universal Transverse
Mercator coordinate system and projection, North
American Datum of 1927 (NAD27). The grid cell
size for the study was 31.8 meters square, or 0.1
hectare (0.25 acre) per grid cell. The FLNF
database was prepared during 1987-1990 and was
made available to this phase of the project for
ELT classification and mapping. ?The land cover
map for 1988 was derived from manual
interpretation of medium-scale (158,000)
color-infrared aerial photos acquired on 29 April
1985 by the National High Altitude Photography
(NHAP) program and updated by field observations
during summer 1988. The resulting map was used as
one spatial variable to map suitability of FLNF
landscapes for expanding areas devoted to
blueberry production.
Figure 7. Regional map of LTA02.
Figure 6. Regional map of LTA01.
Figure 8. Regional map of LTA03.
Objectives
Our objectives were to (1) define the dominant
landscape-level processes that characterize the
central Finger Lakes region in New York, (2)
identify and provide the spatial context for the
biophysical components of these processes, and
(3) define, construct, and map Land Type
Associations for the FLNF.
Materials Methods
The FLNF is located in a transitional zone
between the eastern broadleaf forest and
Laurentian mixed forest provinces of the eastern
USA. Most of the FLNF occurs in one Subsection
that represents a biophysical transition between
the northern glaciated Allegheny plateau and the
Erie-Ontario lake plain. Significant social,
economic, and biophysical characteristics define
this region, and directly influence Forest-level
resource management, planning, and monitoring
programs of the Forest Service. Smaller portions
of the FLNF occur in the Eastern Ontario Till
Plain to the north, and in the Central Allegheny
Plateau to the south. The methodology for this
project was guided by the principles of
ecological classification as outlined in the
National Hierarchical Framework of Ecological
Units (USFS 1993), the principles and basic
approach of LTA mapping in the Lake States
(Jordan, et. al. 1996), and spatial modeling
approach of Host, et. al. (1996), and previous
experience in ecological land classification in
the mid-west and northeastern USA (Russell and
Jordan 1991). There are two major approaches for
defining LTAs aggregation and regionalization.
Aggregation is the process whereby LTAs are
defined through the grouping of attributes used
to define Ecological Land Types at finer spatial
scales. This is commonly referred to as the
bottom up approach. Regionalization is the
process whereby LTAs are defined through the
integration of coarse-scale biophysical data
representing landscape-level patterns and
processes that may be more visible or obvious
over larger landscape areas. This is commonly
referred to as the top down approach.
Generally, a combination of the two approaches
incorporating both remote sensing and GIS
technologies has proved to be most successful in
other eco-regions of the USA. Priority was
given to elevation, slope gradient, and soil pH.
We stratified the range of elevation into two
zones lt 500m and gt 500 meters (amsl). We
stratified slope gradient into three classes
0-3, 4-8, and gt8. Soil pH values computed for
each STATSGO map unit were stratified into three
classes (NRCS 1994) extremely - very strongly
acid (pH lt 5.0), strongly acid - moderately acid
(pH 5.1-6.0), and slightly acid - neutral (pH
6.1-7.3). Geo-referenced grid themes of terrain
and soil properties were analyzed using the Map
Calculator Tool in ArcView to map six Land Type
Associations. We merged individual LTA models
using ArcView Map Calculator conditional
statements where each gn represented an
individual LTA grid (n1,6) (( g1
1.AsGrid) .con(1.AsGrid,(g2
1.AsGrid) .con(2.AsGrid,(g3
1.AsGrid) .con(3.AsGrid,(g4
1.AsGrid) .con(4.AsGrid,(g5
1.AsGrid) .con(5.AsGrid,(g6
1.AsGrid) .con(6.AsGrid,0.AsGrid)))))))
Attribute tables and narrative descriptions of
LTA units were defined and delineated in order to
provide the map user an understanding of the
ecological properties and differentiating
characteristics of each map unit.
Figure 8. Soil drainage map.
Figure 9. Slope gradient map.
Figure 10. Slope aspect map.
Figure 11. Ecological land type map.
Figure 9. Regional map of LTA04.
Figure 10. Regional map of LTA05.
Figure 11. Regional map of LTA06.
Spatial Modeling of Resource Suitability

• Land Type Associations Mapping Criteria
• LTA Name
  Elevation Slope
  Gradient Soil pH Class
• Allegheny High Plateau
  gt 500m lt 3
  gtExt. Acid
• Transitional Allegheny Sideslopes
  gt 500m gt 3
  gtExt. Acid
• Transitional Allegheny Till Plains
  lt500m lt 3
  ltMod. Acid
• Transitional Allegheny Moraine Hills
  lt 500m gt 3
  ltMod. Acid
• Finger Lakes Till Lake Plains, Bottoms
  Hollows lt500m lt 8
  gtMod. Acid
• Finger Lakes Sideslopes Ravines
  lt500m gt 8
  gtMod. Acid

Figure 12. LTA units in the Finger Lakes region.
Summary and Recommendations
Further refinement and characterization of these
Land Type Associations are required in concert
with possible re-definition of the transition
zone between the Central Allegheny Plateau and
Eastern Ontario Till Plain Subsections. Further
study is required to better understand how shifts
in natural and human-induced disturbance regimes
at landscape scale have impacted the nature,
extent, and pattern of ecological communities in
the region. The influence of nearly 200 years of
human settlement and agricultural
intensification, poor logging practices, and
chronic soil erosion have depleted the young,
infertile soils of the region. Removal of
natural and human-induced disturbances such as
fire, farmland through abandonment, establishment
of continuous cover crops (pastures and forest
plantations) over a relatively short time scale
make difficult the prediction and mapping of the
ecological potential and services available from
these landscapes.
Figure 15. Blueberry suitability.
Figure 13. Pasture suitability.
Figure 14. Aggregate suitability
Figure 12. Red Oak suitability.
Given this limited spatial analysis, the pastures
in the northern portion of the Forest appear to
be more productive, require less maintenance, and
have improved access and proximity to members of
the Hector Grazing Association. Very limited
areas exist for aggregate extraction in the
northern portions of the Forest given the
prevalence of large till plains and the absence
of large till moraines, outwash plains, and
alluvium. Potential areas suitable for
expanding blueberry production in the Forest are
located in the southern portions of the Forest.
This results from an optimum combination of
terrain, land use, soils, and access favorable to
blueberry culture and access by the public.
References
Bailey, R.G., P.E. Avery, T. King, W.H. McNab
(eds.) 1994. Ecoregions and subregions of the
United States (17,500,000 scale colored map and
unit descriptions). U.S. Geological Survey.
Washington, DC.
Summary Conclusions
Host, G.E., P.L. Polzer, D.J. Mladenoff, M.A.
White, and T.R. Crow. 1996. A quantitative
approach to developing regional ecosystems
classifications. Ecol. Applications 6(2)608-618.
A GIS-based Ecological Land Type (ELT) mapping
system was developed using environmental
databases for the Finger Lakes National Forest
(FLNF), located in the Finger Lakes Region of
central New York State. Environmental data
important in determining the productive capacity
of the Finger Lakes National Forest was generated
from various sources. These data were combined
using GIS operations to produce a map of the
Ecological Land Types of the FLNF and suitability
maps of resources important to forest planners
and stakeholders. Fourteen ELT's were found to
exist in the FLNF and were aggregated to five
major ecological communities. Each ELT has a
unique productive capacity and requires specific
management practices to realize this capacity.
The ELT map provides a tool to help evaluate
current management practices and adjust them
accordingly. This classification and mapping
system may serve as a tool for making
multiple-resource management decisions on the
Forest. The system represents the natural
productive capacity of the land in the absence of
human intervention and disturbance. With
additional knowledge of land use characteristics
of a site in addition to the information that
ELT's provide, land use management can be
adjusted to reduce soil erosion, locate
recreational uses to minimize conflicts with
other land allocations, and improve wildlife
habitat and timber management. Management
alternatives to improve productivity, manage
lands for multiple use, and allocate land
resources, while protecting the environment can
then be implemented based on these
characteristics. Additional applied research is
required to refine the ELT mapping process on the
FLNF. This should occur with strong
collaboration between plant community ecologists,
soil scientists, and experts in spatial analysis
to appropriately apply such models for forest
land management and planning.
Jordan, J.K., E. Padley, D.T. Cleland, J.A.
Gates, D.J. Hoppe, L.S. Kempf, B. Leuelling, K.L.
Seleen, D.A. Shadis, and J. Sibernagel. 1996.
Land Type Associations - Origins, concepts,
mapping, and applications in the Lake States
National Forests. U.S. Department of
Agriculture, Forest Service. 149 p (Draft).
Keys, Jr., J., C. Carpenter, S. Hooks, F. Koenig,
W.H. McNab, W. Russell, and M.L. Smith. 1995.
Ecological units of the eastern United States
first approximation (map unit tables and booklet
of map unit tables). U.S. Department of
Agriculture, Forest Service. Atlanta, Georgia.
83 p.
NRCS. 1994. State soil geographic (STATSGO)
data base. Misc. Publ. 1492. National Soil
Survey Center. Natural Resources Conservation
Service, U.S. Department of Agriculture.
Washington DC.
Russell, W. E. and J.K. Jordan. 1991. Ecological
classification system for classifying land
capability in mid-western and northeastern U.S.
national forests. In Proc. Symp. Ecological
Land Classification Applications to Identify
the Productive Potential of Southern Forests.
Charlotte, NC. USDA Forest Service General
Technical Report SE-68.
USFS. 1993. National hierarchical framework of
ecological units. ECOMAP, U.S. Department of
Agriculture, Forest Service. Washington, DC.
(October). 21p.
USGS. 1990. Digital elevation models. Data
Users Guide 5. U.S. Geological Survey, Reston,
Virginia. 51p.
This research was supported by the Green Mountain
National Forest, USDA Forest Service, 231 North
Main Street, P.O. Box 519, Rutland, VT 05701,
under Agreements 40-1484-7-0434
43-1681-8-0179 43-1681-9-0204 and MID-98-28.
Presented at the 1st Annual Finger Lakes Research
Conference, 08 October 2005, Finger Lakes
Institute at Hobart and William Smith Colleges,
Geneva, New York http//fli.hws.edu
Corresponding Author S.D. DeGloria,
607-255-5459 voice 607-255-2644 fax
sdd4_at_cornell.edu e-mail IRIS, Department of Crop
and Soil Sciences, 232 Emerson Hall, Cornell
University, Ithaca, NY 14853 http//www.css.corne
ll.edu/faculty/degloria.html