Monica Brelsford, Weed Ecology Lab
Land Resources and Environmental Science, MSU
The methodology that follows is for determining land use change at the sub-county
level. The strength of such an analysis is that land managers can focus on
specific “hot spot” locations that appear to be changing with
respect to land cover or land use based on countywide data from the Census
of Agriculture or other indicator. Based on the analysis the effects on wildlife,
humans, or recreation can be assessed.
Additionally, fine scale analysis can help determine where and to what extent
local farmland or important habitat may be lost to rural residential development,
the impacts on conservation or public access opportunities, or if change is
taking place in the way the public or land managers perceive.
The process consists of four steps:
- Selection of the study area
- Acquire, scan, and georectify air photos
- Interpret photos
- Analyze change over time
The Study Areas:
The study area may be chosen based on several criteria. These may include:
the need to investigate an area of sensitive habitat (i.e. a specific river
corridor), an area of a county undergoing rapid development (land on the forest
edge along a national forest boundary), or an area of increased recreation
or agricultural intensity. The size of the study area is limited by time and
available resources but parts of counties are most appropriate given the number
of photos that must be acquired and studied. At least two time periods are
typically studied – for most areas a decade long interval will provide
insight into the nature of land use change.
Acquire, Scan, and Georectify Air Photos
Aerial photos of the study area for the two time periods are required. These
are available as Digital ortho quarter-quads (DOQQ) available for free for
most areas via the Internet (for Montana: www.nris.state.mt.us/), or “hard
copy” photos ordered from USGS in Salt Lake City, UT. DOQQs are black
and white aerial photos that are ready to use in ArcView or other digital
mapping program. Hard copy photos will need to be scanned into a computer.
Some years are either not available or quality is not as high as other years
so some flexibility is required for the analysis.
DOQQs are combined into a mosaic of the study area and then georectified
to be ready for use in ArcView. Rugged topography takes longer to georectify
then flat topography. Georectification and building the mosaic of the landscape
is done by professional technicians and involves registering landmarks on
the photomosaic to know points on the landscape – major roads, railroad
crossing, bridges, etc. Technicians attempt to georectify within 20 meters
of specific locations on the DOQQs.
Interpret photos
Polygons are created and classified on the digital photomosaic using the
classification scheme developed by Harrison and Potter (2001) (see table below).
Each section of photo is studied to determine the type of land use/cover for
each time period and a polygon is drawn around the land use area. Each polygon
is given a value based on the dominate land use. For example, in the figure
below Land Use T1 consists of three polygons of two different land uses –
a and b.
Houses or residential areas can be mapped as point features overlaying digitized
photos. Home sites can be verified by driving around the study area. Barns
and outbuildings are mapped and can be useful for determining present use
of the land. They are typically separated from the housing database before
further analysis. The final product at this stage is at least two photomosaics
of the study area with a layer of polygons depicting land use/cover, a data
layer of points depicting homes, and a final layer of other buildings. Additional
data layers could consist of management jurisdictions, conservation easements,
recreation access points, or specific species habitat depending on the area
of interest.
Analyze change over time
Land use polygons for different years can be intersected and the resulting
polygon coverage depicts the type of change, the area of change, and the location
of change that has occurred. This information can be displayed visually
and spatially.
In the above example where a = river, b = riparian c = agriculture, in time
period 1 (T1), the river is through riparian vegetation. By T2, the river
has changed course, but still has riparian vegetation and by T3 a portion
of the riparian vegetation has become agricultural land. The change is represented
by the hyphen letters, so a transitioned to b (a-b), b went to a (b-a), and
finally b to c (b-c). So the actual amount of change is only the area of
the hyphen letters. Relevant questions might be how much riparian is turning
into agriculture land, how much does the river change, is the total acreage
of riparian land remaining the same and simply shifting, are residential structures
at risk given the migration of the river corridor.
As part of an effort to understand the impact of homes in an agricultural
setting on land use and the environment in general, each home site point can
be assumed a spatial signature equivalent to other land use/cover types; a
sort of ecological footprint on the land. Emergent research has shown that
the zone of influence is largely based on dispersal distances of animals that
accompany humans (pets) and their impacts on native animal species (birds
and mammals) (Hansen et al, 2002; Odell et al, 2003; Theobald et al, 1997).
The considerations for how to assign an area to each house is dependent on
the land use/cover type where the house is placed, what elements of the landscape
or environment may be impacted, and dispersal rates of organisms that may
cause an impact. For example, if a house is placed in agricultural land it
is likely to be a source of weeds, which can impact crop production. Dispersal
distances and rates vary tremendously with different weed species and mechanisms
of dispersal (wind, water, animals, etc.) (Cousens and Mortimer, 1995).
The photos below are an example of the output. The example adopts a conservative
zone of influence - 100 m diameter circle (7.76 acres) for the influence zone
on home sites points. The location is in the upper Paradise Valley, Park County,
MT. Habitat maps, conservation easements, or other data layers of interest
to land managers could be overlaid on the photo output.
Aerial photos indicating 100 m radius zones
of influence around house site points (left) and 100 m radius house site zones
of influence colored to show intersection with different land use/cover types
(right).
Conclusion
The analysis described here may be particularly useful to land managers who
wish to focus on perceived hot spots of change at the subcounty level. The
analysis requires GIS capability and technical expertise available to most
public lands agencies and is not overly data intensive except for the labor
involved in air photo interpretation and preparation of the polygon layers.
The investment in time and data may have significant payback with respect
to highlighting fine scale changes and resultant management opportunities
in those areas undergoing localized land use change.
Land Use / Cover Classification Scheme
I: Urban
1101—Farm (house with out buildings)
1103—Commercial Areas or Structures
1106—Airport
II: Agriculture
2020—Center Pivots
2021—Irrigated Hay
2022—Dry land Annual Crop
2023—Dry land Hay
2024—Irrigated Annual Crop
III: Grasslands
3000—Moist Grassland
3150—Grassland/ Dry land Pasture
3170—Grassland with Sagebrush
IV: Shrub Lands
3400—Juniper with deciduous shrubs and minor component pine
3350—Sagebrush (>20% canopy)
V: Forest Lands
4000—Low Density Coniferous Forest
4200—High Density Coniferous Forest
4300—Streams with mixed Deciduous/ Coniferous Forest
4500—Cut over forest
VI: Water
5000—Water (Yellowstone river and gravel bars, lakes and ponds)
VII: Riparian
6120—Deciduous Trees (cottonwood/willow juniper under-story possible)
6130—Shrub Riparian (young cottonwoods, willow)
6160—Riparian meadow/ bottomland pasture
6190—Broadleaf Trees on Ditch Lines
VIII: Barren Lands
7300—Rock, Rock Outcrop
7500—Gravel Pits, Disturbed
IX: Alpine
8100—Alpine (treeless)
Literature Cited
Cousens, R. and M. Mortimer. 1995. Dynamics of Weed Populations. Cambridge
University Press.
Hansen, A., R. Rasker, B. Maxwell, J. Rotella, J. Johnson, A. Wright, U.
Langner, W. Cohen, R. Lawrence, M. Kraska. 2002. Ecological causes and consequences
of demographic change in the new west. Bioscience 52:151-162
Odell, E.A., D. Theobald and R. L. Knight. 2003. Incorporating ecology into
land use planning. APA Journal 69:72-81.
Theobald, D.M., J.R. Miller and N. Thompson Hobbs. 1997. Estimating the cumulative
effects of development on wildlife habitat. Landscape and Urban Planning 39:25-36.
