Monitoring Land Cover and Land Use Change for Great Lakes Network Parks

1
Monitoring Land Cover and Land Use Change for
Great Lakes Network Parks
2
Land Cover/Land Use Change Monitoring Questions

• What is the magnitude and direction of change in
  spatial patterns of LULC in and adjacent to parks
  over time? Spatial patterns to include such
  measures as land cover class, corridors,
  fragmentation, juxtaposition, edge, road density,
  number of homes, etc.
• What are the changes in area and shape in urban,
  agricultural, and other areas dominated by human
  landuse within a defined monitoring region for
  each park.
• What are the changes in forest harvest amounts
  and patterns in each park monitoring region?
• How has the human occupation density, measured
  either by population or building density, changed
  in each monitoring region?

3
Acreage of Parks and Land Cover Area of Interest
4
Three Distinct Protocols

• Differences are based on
• Size of parks
• Urban vs Forested Adjacent Landscapes
• MISS, Lower SACN, INDU Minnesota Land Cover
  Classification System (MLCCS), which incorporates
  NVCS classification with additional urban classes
• PIRO, SLBE, APIS, GRPO Fine scale imagery
  and metrics, using NVCS classification to the
  formation level
• Upper SACN, VOYA, ISRO Fine scale imagery,
  but either broader metrics or sampling rather
  than complete census
• Protocols for all parks will also include
  delineation of roads, buildings and developed
  areas as well as disturbance events (fire,
  windthrow, etc.), for the within park area of
  interest (Park 300 meter buffer)

5
We are currently working to incorporate two
classification approaches National Vegetation
Classification System (NVCS) Minnesota Land
Cover Classification System (MLCCS)
Example of MLCCS land cover mapping on urban
landscape
6
Fine Scale Metrics

• Fine-scale (Park and immediately adjacent lands)
  Metrics
• Land Use/Land Cover
• Land cover classes mapped/measured
• Closed-canopy forest area
• Total agricultural area
• Row-crop agriculture
• Urban/suburban area
• Forested, non-forested wetland area
• Built-structure density
• Road location and type

• Derived metrics from measured/mapped data above
• Size
• Average forest patch size
• Median Forest patch size
• Average urban/agricultural patch size
• Shape
• Average forest perimeter/area ratio

7

• Derived metrics - Continued
• Edge
• Forest Edge density km/km2 or similar metric
• Mean distance to agriculture/urban edge
• Road density
• Average distance to roads
• Area greater than 80 meters to a road
• Fragmentation
• Wetlands
• Net change in total wetlands area within each
  NVCS physiognomic class
• Area of per-class transition for NVCS subclass
  wetlands.
• Dynamic Environments These components will most
  likely be sampled locations rather than complete
  census
• Average change in bluff base and top position
• Change in sand bar, point, and beach size, shape,
  and location
• Change in stream channel location

8
NVCS Hierarchical Classification
Physiognomic Categories Category . . . .Example
Class . . . .Open Tree Canopy Subclass . . .
.Evergreen Open Tree Canopy Group . . . .
.Temperate or Subpolar Needle-leaved Evergreen
Open Tree
Canopy
Subgroup . . . .Natural/Seminatural
Formation . . . . Rounded-crowned
temperate or subpolar needle-leaved
evergreen open
tree canopy. Floristic Categories
Alliance . . . .Juniperus occidentalis
Woodland Alliance
Association . . . . Juniperus occidentalis
/Artemesia tridentata

9
MLCCS has adapted the NVCS, using similar
hierarchical approach, but has developed more
detailed classes for cultural landscapes
10
Coarse Scale Land Cover Metrics

• Coarse scale metrics will be very similar, but
  will include a reduced suite of classes
• The Network will likely use existing datasets
  rather than manual delineation of land cover,
  roads, buildings, or other features
• Land Cover Analysis using National Land Cover
  Dataset (Currently developed with LANDSAT Imagery
  30 meter satellite imagery)
• Analysis of Road Density using Roads Network from
  U.S. Census TIGER data
• Analysis of Human Population Change and Building
  Density using U.S. Census data

11
Frequency of Land Cover Analysis

• Schedule will follow the Terrestrial Vegetation
  Protocol, which currently involves a 5 year
  return interval
• In addition to complete coverage of land cover
  classification to the NVCS Formation level, the
  LCLU protocol will likely include updating the
  NVCS Vegetation Map at 1 kilometer blocks
  centered on permanent plots established in the
  Terrestrial Vegetation Program
• Aerial photography will be acquired on this
  return interval, either through direct
  contracting with vendors, or cooperation with
  other programs.
• Current plans are to standardize on 112,000
  scale imagery, and 0.15-0.2 meter resolution
  ortho-imagery

12
Comparison of Satellite Imagery and Aerial
Photography
13
Remote Sensing Priorities and Products Acquired
to Date
14
Current Projects in Land Cover/Land Use

• Land Cover/Land Use Protocol Development Paul
  Bolstad, Professor in Forest Ecology
• Land Cover Mapping and Analysis for MISS Lower
  SACN using the Minnesota Land Cover
  Classification System Bart Richardson, MN DNR
• Early Generation Aerial Photography to Provide
  Historical Context to Current Land Cover within
  and adjacent to Parks
• INDU Project with Volker Radeloff at
  University of Wisconsin using 1938, 1980s, 2005
  imagery
• VOYA Project with Bill Befort at Minnesota
  DNR using 1927 imagery
• APIS Project with Paul Bolstad at University
  of Minnesota using 1938 imagery
• GRPO MISS Project with Mike Hyslop at
  Michigan Technological University using 1940
  (GRPO) and 1937 (MISS) imagery

15
Housing and Road Density Growth in and around
Indiana Dunes National Lakeshore 1938-2005
Shelley Schmidt1, Volker C. Radeloff1, Peter
Budde2, and Ulf Gafvert3

Results
2005 Orthophoto of Study Area
Introduction
2005 orthophotography of Indiana Dunes National
Lakeshore

• National Parks serve as areas of ecological
  conservation
• by protecting habitat and preventing landcover
  change.
• Housing and road development near national parks
  can
• isolate them and limit their ecological
  function by
• Fragmenting wildlife habitat
• Altering hydrology
• Spreading exotic species
• Polluting natural areas
• We focused on one park, Indiana Dunes National
• Lakeshore (INDU), established in 1966, for this
  study.

Building density has increased 7 fold between
1938 and 2005.
Road density has nearly doubled between 1938 and
2005.
Road density outside of INDU has seen a
pronounced increase, while the increase within
INDU has been more subtle.
Objective
The percentage of the study area occupied by
roads (25 meter buffer) is highest outside the
park in 2005.
Road Housing Density Change, 1938 to 2005
1938
2005
Quantify the change in road and building density
in and around Indiana Dunes National Lakeshore
between 1938 and 2005 using aerial photography
and satellite imagery.
Patch density has increased both within and
outside of INDU, although to a much greater
degree outside the park.
Study Area
Landscape fragmentation is occurring, with
decreasing mean area per patch both within and
outside of INDU.
ONTARIO
Conclusions
WISCONSIN

• The current landscape is highly fragmented and
  few
• areas are free from human disturbance.
• Patch density and road density have increased
• significantly, resulting in decreased mean
  patch size.
• Road and housing growth outside the park have
• increased at more than twice the rate as growth
  inside
• the park.
• Housing and road development are creating
• fragmentation patterns on the landscape that
  are not
• easily reversed.
• Future fragmentation could be limited by
• Protecting existing large forest patches
• Clustering new development
• Decommissioning unnecessary roads

MICHIGAN
Indiana Dunes National Lakeshore, containing
6,000 hectares, runs for nearly 40.2 km along
southern Lake Michigan. The park is home to
beaches, sand dunes, bogs, wetlands, woodland
forests, an 1830s French Canadian homestead, and
a working 1900 era farm. Indiana Dunes is ranked
7th among national parks in native plant
diversity. Ninety of the 1,418 plants that have
been found in the park are on Indianas
threatened or endangered list
Buildings
ILLINOIS
OHIO
INDIANA
1938
2005
Methods
Roads and buildings within the boundaries of INDU
and within a 3.2 km buffer outside of the park
were digitized for 1938 and 2005. 25 meter
buffers around roads was applied to account for
edge effects. Change in road and building
density and the resulting landscape fragmentation
was calculated.
Roads
Future Work
Upcoming steps in this project will include
expanding to the rest of INDU, incorporating a
1980 time point, classifying landcover, and
expanding our methods to other Great Lakes
network National parks, beginning with Pictured
Rocks National Lakeshore.
1938
2005
1Department of Forest Ecology and Management,
University of Wisconsin-Madison 2National Park
Service Midwest Region GIS Technical Service
Center 3National Park Service Great Lakes
Network Inventory and Monitoring Program
We thank the National Park Service and the US
Forest Service North Central Research Station for
funding this project
For further information, contact Shelley Schmidt
at slschmid_at_wisc.edu
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
Great Lakes Network Vegetation Mapping Schedule