Moderate_resolution_GEC

MARCH 2012
Remote Sensing & Classified Land Cover
Essential Land Use Decision Support Tools Using
Moderate-Resolution Imagery
GLOBAL ECOSYSTEM CENTER www.systemecology.org

Remote Sensing & Classified Land Cover: Essential Land Use Decision Support Tools
2 GLOBAL ECOSYSTEM CENTER
For the past half century, satellites have been providing information regarding Earth’s sur-
face. And, just as the satellite data collection technology has evolved, so has the ability to ac-
curately interpret and analyze the imagery. Currently, data collected by Earth-observing satellites
can be classified into discrete land cover categories, allowing for the documentation and analysis
of the landscape. Additional information regarding landscape structure and functions of the land-
scape can be determined using a Geographic Information Systems (GIS). The combination of land
cover metrics and engineering/scientific models permit accurate calculations of human-induced
impacts on the landscape.
While conflicts between humans and nature are becoming more complex, larger in scale,
and potentially leading to more severe consequences, information technology provides an un-
precedented ability to gather and assess information. The challenge is to better utilize available
resources when land use decisions are formulated and avoid potentially disastrous conflicts.
Natural disasters triggered by hurricanes are an example of the value of remote sensing and
GIS. The effectiveness of barrier islands in reducing hurricane strength and protecting the main-
land is well understood. Through landscape modeling, scientists are able to calculate buffering
effects of these islands. Satellite imagery has documented their condition for over 30 years, and
in New Orleans, decision makers were aware of the decline of the barrier islands. Unfortunately,
land use policies were not established to rebuild or protect these islands and the city suffered
irrevocable consequences and will likely never fully recover.
This document identifies data and analysis opportunities available to improve land use deci-
sion making and to avoid the disastrous consequences. It is divided into the following sections:
T
here has been a long history of conflict between humans and nature related
to land use. Historic records document catastrophic flooding, fire, disease and
drought. As these events occurred, they were often perceived as “acts of God”
and beyond our control; however, people have become aware that many of these epi-
sodes were triggered by human actions. The powerful influences of humans on natu-
ral landscapes are too often minimized as cities and towns are rebuilt after disasters.
Future disasters are assured as settlements are reestablished without changing devel-
opment patterns or making essential adjustments in land use policies. While ignorance
may be a valid excuse for past mismanagement, it is not a valid excuse today.
“Give me a lever long enough and a fulcrum on which to place it, and I shall move the world.”

-Archimedes

3MARCH 2012

1. Land Cover Metrics - Data for Land Use
Strategies
2. Landsat Imagery - The Pre-Eminent Source for
Moderate-Resolution Land Cover
3. Land Cover Classification Targeted to Local
Needs
4. Coastal Southern California Case Study
5. Landsat imagery - A Logical Choice for
Standardized Global Land Cover
1. Land Cover Metrics - Data for Land Use
Strategies
Land cover metrics are measurements of Earth’s
land surface, including vegetation, geology, hydrology,
or anthropogenic features. Land cover data is capable
of providing direct and objective indications of land use
impacts on natural conditions. These measurements
are among the most significant and detectable indica-
tors of global ecological change (Figure 1).123
Land cover
directly impacts biological diversity4
while contributing
to local, regional, and global climate change.5
Land cover measurements are acquired through
land surveys or remote sensing (RS). Historically, land
surveys were the primary source of data, however
remotely sensed data has become ascendant for land-
scape analysis. In the United States, the most ambitious
and comprehensive land survey efforts have been the
Natural Resource Inventory (NRI) conducted by the Nat-
ural Resource Conservation Service.6
This comprehen-
sive effort has provided excellent land cover statistics
for over 80 years, but it uses labor-intensive methods
for field data collection and it is inefficient compared to
remote sensing options. Additionally, remote sensing
provides information for the entire landscape, unlike
statistical sampling techniques used in field collection
for the NRI.
In the United States, remote sensing has become
the most practical, indispensable and timely method for
producing land cover classifications. The USGS National
Figure 1 - Landsat imagery and classification. A) Visible spec-
trum, B) Near and Mid-infrared (bands 4,5,3), and C) Land
cover classification (generalized categories include: 3 urban
categories (white/gray), natural vegetation (shades of greens),
other landscape features such as water, open space, agricul-
ture, pasture, wetlands, etc.
A
B
C

Land Cover Database (NLCD) and NOAA Coastal Change Analysis Program
(C-CAP) land cover classifications provide standardized land cover products
(Figure 2). These land cover classifications from remote sensing-derived
metrics may be used as a proxy for biological indicators, allow real-time per-
spectives to follow the rate of landscape change and establish base line data
for “change detection” and growth scenarios.7
While remote sensing allows the production of land cover classifications
far more efficiently than land surveys, classification schemes require stan-
dardization for appropriate analysis and comparisons. The establishment
of a standardized scheme for an area as large as the United States requires
massive efforts involving the expertise from many agencies and institutions
and the commitment to a larger purpose. In the United States, this has
been completed by the formation of the Federal Geospatial Data Commit-
tee (FGDC) to coordinate the effort. Subsequently two Federal Agencies,
the USGS and NOAA have undertaken the production of land cover products
Figure 4 GEC specializes in creating (updating or
backdating)classifiedlandcoverdatasetsfromany
archived Landsat image.
Figure 2 - By obtaining archived imag-
ery available from the the EROS Data
Center for 1985 and 2011 and conduct-
ing a change detection analysis, urban
grown can be visually displayed and
calculated. The land cover classification
for 1985 and 2011 were standardized to
the official C-CAP data set.

5MARCH 2012
(NLCD and C-CAP). The products are now produced every five years and are
available to the public free of charge. These products not only provide plan-
ning agencies with high quality data, but also provide venerable data sourc-
es for the production of the intermediate land cover classifications needed
to fill the time spans between the production of the official data sets.
2. Landsat imagery, The Pre-Eminent Source of Moderate-Resolution Land
Cover
Satellite imagery is generally grouped into three categories: high, mod-
erate and low-resolution. The resolution of the image is determined by its
pixel size, generally ranging from less than a meter to more than a kilometer.
The focus of this paper is on moderate-resolution Landsat imagery which is
Figure 3 - The EROS Data Center uses a
convenient path/row system to archive
Landsat imagery. The satellites have a
consistent orbit which allows imagery
collection of any landmass on Earth
every 16 days. The Landat series has
been obtaining 30-meter imagery since
1984 and has an archive of over 2.5
million images. The path/row system is
illustrated below.

Extending National Land Cover Classifications
Satellite imagery has permanently changed our
understanding of Earth. The data collectors on the
Landsat satellite’s are designed to identify relevant land cover
objects collecting and archiving imagery since 1984. Unfortunately only
about 1% of archived Landsat imagery of the United States has ever been classified
into land cover categories. The historical data holds tremendous potential for under-
standing and guiding future land use decisions.
Figure 4 illustrates the immense imagery archive available by exhibiting the 790 images
available for one location in Southern California (Path 41, Row 36) as of October, 2011.
The imagery is especially valuable in the United States where standardized land cover
classifications have been produced by NOAA and the USGS so private organizations like
the GEC can utilize the archive to extend land cover analysis to cover 1984 to present.

7MARCH 2012
considered a pre-eminent data source for land cover
classification.
Landsat satellites have acquired multi-band
digital imagery of Earth’s surface for over three de-
cades, enabling the examination of changes caused
by both natural processes and human practices.8
The specifications for data collection were devel-
oped by experts in land cover analysis, and in the
case of Landsat, the goal was the documentation
of significant land cover objects. Landsat imagery is
collected in seven spectral bands at a 30-meter pixel
resolution, and it is designed to capture major land
features such as roads, bridges and buildings includ-
ing larger natural features.
While the Landsat satellite series has been
providing continuous coverage of the Earth’s surface
since 1972, the collection system was upgraded with
Landsat 4 in 1982 to provide the 30-meter resolu-
tion imagery. The data is now widely used to create
moderate-resolution land cover classifications.
Imagery collected by the satellites is downloaded to
the Earth Resources Observation and Science (EROS)
data center in Sioux Falls, South Dakota where it is
archived and available for use free of charge. The
image archive uses a path/row cataloging system
that allows easy navigation and acquisition (Figure
3). The extensive image library allows temporal
comparison through the analysis of imagery during
different time periods. There is almost limitless po-
tential value of this data for documenting land cover
change and assisting land use decision-making.
While Landsat satellites collect data over land
masses globally, the potential for assisting land use
planning is most evident in the United States where
standardized land cover classification systems have
been established by the Federal Geographic Data
Committee (FGDC). This committee includes repre-
sentatives from relevant federal agencies in addition
to experts from academia and private companies.
This standardized classification system provides
technical guidelines for distinguishing land cover
types as well as documenting critical land cover clas-
sification procedural steps.
The USGS and NOAA are charged with the
task of developing and maintaining land cover data
sets under the FGDC system. The USGS has produced
National Land Cover Database (NLCD) classifications for
1992, 2001, 2006 (technical issues exist with the 1992
data) and NOAA has produced Coastal Change Analysis
Program (C-CAP) classifications for 1996, 2001, 2006,
and 2011. While both data sets use the same classifi-
cation standard, C-CAP is primarily limited to coastal
areas and provides more detailed wetland categories.
These federally created data sets provide valuable and
essential land use planning data. However, updates are
periodic; the most recent land cover classifications are
generally between 5 to 10 years old. Additional local
and regional land cover classifications are needed for
most land use decisions and local agencies can procure
necessary land cover classifications from private com-
panies with remote sensing expertise.
3. Land Cover Classification Targeted to Local Needs

The Global Ecosystem Center (GEC) specializes in
creating land cover classifications in accordance to
FGDC standards. In the United States, these classifica-
tions can be built upon existing national land cover
classifications provided by the USGS or NOAA. The
effort of the federal agencies is analogues to opening
the door to a huge library archive so that scholars can
research the data, interpret and communicate find-
ings.
With imagery available in the USGS archives, land
cover classifications can be developed for any area in
the United States and for any period between 1984 and
present. This data are especially valuable in the United
States where base classification schemes have been
developed. Additionally, the GEC has developed techni-
cal methods for connecting relevant ancillary data (soil
type, rain fall, air quality etc.) to land cover classifica-
tions, allowing for ecosystem service calculations and
better decision-making. The calculations use models
that have been peer-reviewed and are widely used by
the scientific and engineering communities. Details
about the Global Ecosystem Center and the services it
provides are available at www.systemecology.org.

8 GLOBAL ECOSYSTEM CENTER MARCH 2012
Featured Case Study - Coastal Southern California
Coastal Southern California is a large, highly populated region of Califor-
nia in the United States. The two largest cities and metropolitan areas are
Los Angeles and San Diego. The urban area stretches along the coast from
the northern suburbs of Los Angeles to the border with Mexico. Coastal
Southern California is a major economic center for the state of California
and the nation.
The landscape of this area has undergone considerable change over the
last quartercentury; much of the original natural system has been replaced
by a human network. Data collected by Landsat satellites has recorded these
changes and archived the imagery for evaluation.
The GEC obtained archived imagery from 1984 and 2011 and the result-
ing classification extended the analysis period from 10 years to 27 years.
Standardized land cover classifications have numerous practical implications
when extended over longer time frames. Some of the specific applications
of the data for land use planning, natural resource management, and vulner-
ability assessments is outlined in the sidebar that follows this one page case
study overview.
Figure 5 - Land cover change
and fire hazard areas super-
imposed on a land cover base
map. Generally, most urban
developments (turquoise) are
near high fire risk areas. Santa
Ana winds often make it very
difficult to control fires.

MARCH 2012 9
Water
Water is a valuable and scarce resource
in the Southern California region and natural
vegetation affects water supply. The land cover
this region was accurately measured using
archived Landsat imagery and the official land
cover terminology developed by the FGDC. The
analysis showed there has been a considerable
increase in impervious surfaces and consider-
able loss in vegetation. Apart from precipitation,
evapotranspiration is one of the largest outflow
components of the hydrologic cycle, particularly
in arid areas.9
As natural vegetation decreases relative to
impervious surfaces, evapotranspiration rates
are acceler-ated increasing the water needs for
the remaining plants. Impervious surfaces cou-
pled with urban drainage systems alter natural
hydrology by increasing stormwater runoff and
reducing groundwater recharge. The negative
results are more frequent flooding, higher flood
peakflow, lower base flow in streams, and lower
water table levels.10
Fire Hazards
While wild land fires are part of the natural
system in this region, the expansion of man-
made developments into fire-prone lands has
dramatically increased the number of fires and
the risk of serious damage. Most of Southern
California is at risk of damage from wild fire
in the native chaparral and sage and that risk
is increasing due to the enduring drought and
residential encroachment into wild land. Wildfire
risk will increase in southern California as well as
in the western United States in the coming years.
This risk can be reduced by using land cover
imagery to identify the least hazardous areas for
urban expansion and preventing fragmentation
of large blocks of natural areas. Figure 5 shows
fire hazard areas and their proximity to urban de-
velopments.
Trend analysis over 27 years demonstrates
the region between Los Angeles and San Diego
experienced the highest rates urban growth and
that most of these new developments are now in
areas of significant wildfire risk.
Growth Models and Projections (Scenario Model-
ing)
California is expected to grow from 35 million
to approximately 45 million residents by 2020.
Since 1990, the population of the Southern Califor-
nia region has expanded from 14.6 million to 16.5
million – an increase of 12.8%.
A scenario modeling algorithm in the Urban
Ecosystem Analysis methodology calculates the
impact of land cover change on natural systems.
The impact of past growth can be determined using
archived imagery from 1985, 1996, 2001, 2005,
and 2011. Using this historical data a trend analysis
can be constructed and future land cover and the
associated ecosystem services calculated for 2020
(Table 1).
The trend data shows that the scrub area (sage
and chaparral) will decrease by over 2.% between
2011 and 2020. An Urban Ecosystem Analysis of
the area estimates stormwater flow will increase
by over 3.5 billion cubic ft2
during this time, and
conversely water infiltration will decrease by the
same amount. Increasing stormwater and decreas-
ing infiltration is a critical water conservation issue
for the arid Southwest United States now and the
conditions are expected to intensify over the next
decade.
Detailed information on specific watersheds
can be obtained by selecting 12 digit watersheds
for analysis and using high-resolution imagery as
the land cover data source.

GLOBAL ECOSYSTEM CENTER
Figure 6 - San Diego sub-basin has experienced growth
in urbanized areas over past two decades as new homes
are scattered into scrubland chaparral fragmenting the
natural, fire-prone landscape.
Table 1 - San Deigo Watershed - An Urban Ecosystem Analysis demonstrates the impact of increasing impervious surfaces
stormwater.
10
Air Pollution Air Pollution Carbon Carbon Stormwater Runoff Stormwater Benefits
Year Removal Removal Value Stored Sequestered Reduction* @ $2 per cu.ft
(lbs/yr) ($) (tons) (tons) (cu.ft) ($)
1985 79,519,452 221,908,930 31,724,659 246,985 2,451,805,525 4,903,611,050
1996 83,217,494 232,228,778 33,200,010 258,471 2,686,347,533 5,372,695,066
2001 82,989,740 231,593,201 33,109,147 257,764 2,692,532,618 5,385,065,235
2005 82,704,033 230,795,901 32,995,163 256,876 2,719,732,078 5,439,464,155
2011 79,461,572 221,747,405 31,701,567 246,805 2,548,924,655 5,097,849,310
2020** 82,165,608 229,293,360 32,780,356 255,204 6,007,815,263 12,015,630,526
* Stormwater Runoff Reduction = If existing land cover replaced to Impervious Surfaces: Buildings/Structures
** Scenario of -2% Shrub to Urban Residential, rest of the categories remain that of 2011

MARCH 2012
Case Study - Kalimantan, Indonesia
Indonesian, Kalimantan and Malaysian, and Borneo comprise the third
largest island in the world. It’s geographic location is in the South China Sea
and ecologically it houses rich tropical forest, peatlands, and extensive biodi-
versity including threatened animal species like orangutans, elephants, and
tigers.
Extensive illegal logging has removed over half of the island’s forest cover
which often grows over peatlands 10 to 12 meter deep. Once the forests are
removed, the land is drained for farming and the peatland is burned releasing
massive amounts of CO2
into the air causing Indonesia to be the third largest
emitter of CO2
.
Below classified land cover images processed by the Global Ecosystem
Center reveal the extent of forest and peatland loss between 1985 and 2010.
The source of the imagery is the Landsat satellite series.
1989
2010
11

Case Study - Goiânia, Brazil
1985 Spectral (4,5,3) 1985 Classified 2011 Spectral (4,5,3) 2011 Classified
Goiânia is a planned city founded in 1933 and was designed for a
population of 50,000 inhabitants. Currently, it has a metropolitan area over
1.5 million people. Illegal or informal settlements have recently appeared,
with 7,000 housing units located in environmentally hazardous areas.16
These include river banks and places subject to periodic ﬂooding. Slum
settlements have been overwhelmingly built in these sensitive watershed
areas.
An analysis of Landsat satellite images between 1985 and 2011
reveal the extensive growth and development of the Goiânia metropolitan
area. The raw data was obtained from archived Landsat imagery available
through the USGS, and processed into eight land cover categories.
The land cover data was used by GeoAdaptive to assist land use
planners in the city with growth and development planning.

MARCH 2012 13
Case Study - Osa Peninsula, Costa Rica
The Osa Peninsula is located in southwestern Costa Rica surrounded by the
Pacific Ocean. It is one of the most biologically diverse places on Earth and
home to at least half of the species found in Costa Rica. Most of the area is
undeveloped tropical forests and wetlands, although a portion of the natural
wetlands has been converted to rice production.
The Global Ecosystem Center used the Landsat archive to obtain imagery
from 1985, 2000 and 2011. The imagery was processed by the image ana-
lysts. Electronic bands were combined to form a spectral image. These
non-visible wavelengths (near infrared, thermal etc) allow the landscape
to be classified into discrete land cover using spectrometry. This land cover
classification revealed 14 discrete land cover categories following guidelines
provided by the local government. Below are the land cover classes identi-
fied from the imagery in 2011.
This classification was conducted by image analysts at the Global Ecosystem
Center using Erdas Imagine and See5.
2011 Spectral 2011 Classified

4. Conclusion:
Land cover change may be the most significant
agent of global change; it has a significant influence
on climate, hydrology, and global bio-geochemical
cycles. Arguably, over the next 20 to 50 years, land
cover change will have a more direct influence on
human habitability than climate change. In ad-
dition to its importance as an input variable to
other areas of global change research, it is also an
important area of study in its own right. Land cover
is an issue with far-reaching policy implications on
international, national, national and local scales.
Land cover change is inextricably linked to policy,
sustainable development, and a wide range of
research.
Remote Sensing technology and land cover
data are an essential part of land use decisions.
Data derived from moderate resolution satellite
imagery, collected by the Landsat satellite series,
provides extensive data describing the landscape
(land cover) over the last 30 years. Furthermore, an
extensive archive of moderate resolution Landsat
satellite data is available from the USGS at no cost.
It can be obtained over the internet and converted
into standardized land cover categories for use in
geo-graphic information systems. Unfortunately,
the potential of this data to improve land use deci-
sions has barely been tapped. Two problems seem
to exist 1) decision makers are unaware of the data
and 2) lack of expertise in processing the imagery.
Over the past 27 years, the Landsat satel-
lite series has scanned hundreds of images over
every part of the world and over two and a half
million images are available for public use in the
archives. The Landsat satellite completes a cycle
of the entire globe every 22 days and downloads
digital files to facilities on the ground. These data
have tremendous potential for helping people ad-
dress some of the world’s greatest ecological and
environmental challenges. These data allow us to
accurately measure conditions in the past as well
as the present. With this data it is possible to quan-
tify ecological changes between past and present
conditions and identify trend lines that foresee the
future. The data available from moderate resolution
satellites can provide decision-makers with the data
they need to make good decisions today that can
improve the future.
Remotely sensed land cover data provides deci-
sion makers with a complete set of facts allowing
them to accurately analyze growth and develop-
ment. With the use of engineering algorithms, eco-
logical services can be calculated and future trends
forecast. Putting dollar values to ecosystem services
helps make technical data relevant to public policy
makers.
Remotely sensed land cover produces data that
allow us to accurately remember the past, evalu-
ate present, and predict the future. It allows view-
ing land cover changes and trends at various scales
from sections of a county, to entire continents or
even to global scales. The Global Ecosystem Center
specializes in land cover classification from remotely
sensed imagery and has the capacity to translate
the archived Landsat images to highly accurate land
cover data and even help decision makers quantify
the ecosystem services provided by the land.

15MARCH 2012
1 Turner, M.G., 1990. Landscape changes in nine rural counties in Georgia,
Photogrammetric Engineering and Remote Sensing 56(3): 379-386
2 Lambin EF, Baulies X, Bockstael N, Fischer G, Krug T, Leemans R, Moran
EF, Rindfuss RR, Sato Y, Skole D, Turner II BL, Vogel C. 1999. IGBP Report
No 48 and IHDP Report No 10: Land-use and Land-cover Change (LUCC):
Implementation Strategy. Stockholm, Sweden: International Geosphere-
Biosphere Programme (IGBP); Bonn, Germany: International Human
Dimensions Programme on Global Environmental Change (IHDP).
3 Di Gregorio A. 2005. Land Cover Classification System (LCCS), Version
2:Classification Concepts and User Manual. FAO Environment and
Natural Resources Service Series, No 8. Rome, Italy: Food and Agriculture
Organization
4 Sala OE, Vitousek PM. 1994. Beyond global warming: Ecology and global
change. Ecology 75:1861–1876.
5 Brovkin, V.; Claussen, M.; Driesschaert, E.; Fichefet, T.; Kicklighter, D.;
Loutre, M. F.; Matthews, H. D.; Ramankutty, N.; Schaeffer, M.; Sokolo. 2006
Biogeophysical effects of historical land cover changes simulated by six Earth
system models of intermediate complexity. Climate Dynamics, 6.
6 Harlow, J., 1994. History of Soil Conservation Service National Resource
Inventories Resources, Natural Resource Conservation Service.
7 Landscape and Urban Planning 94 (2010) 158–165
8 landsat.gsfc.nasa.gov
9 Hanson, R.L, 1991, Evapotranspiration and Droughts, in Paulson, R.W., Chase,
E.B, Roberts, R.S., and Moody, D.W, Compilers, National Water Summary
1988-1989—Hydrologic Events and Floods and Droughts: U.S. Geological
Survey Water-Supply Paper 2375, p 99-104
10 The Impacts of Impervious Surfaces on Water Resources. The New Hampshire
Estuaries Project, University of New Hampshire, Hewitt Annex 2007
11 Technical Background Report to the 2003 Safety Element, City of Glendale,
California, Earth Consultants International
12 California Fire Hazard Severity Zone Map Update Project. California
Department of Forestry and Fire Protection. http://www.fire.ca.gov/fire_
prevention/fire_prevention_wildland_statewide.php
13 California 2020 Projected Urban Growth 2010. http://koordinates.com/
layer/670-california-2020-projected-urban-growth/
14 Census Data. Southern California Association of Governments. http://
koordinates.com/layer/670-california-2020-projected-urban-growth/
15 Skole, D.L., Justice, C.Townshend, J.R.G.Janetos, A. 1997. A Land Cover
Change Monitoring Program: Strategy for an International Effort. Mitigation
and Adaptation Strategies for Global Change. Vol 2. Issue 2.
16 http://www.citiesalliance.org/ca/sites/citiesalliance.org/files/CA_Docs/
resources/cds/liveable/goiania.pdf
ENDNOTES

1607 22nd St. NW, Washington, DC 20008
Phone: 202.290.3530
Fax: 202.683.6729
http://www.systemecology.org
Gary Moll, President
Kenneth Kay, Geospatial Specialist
Binesh Maharjan, Geospatial Specialist
MARCH 2012
GLOBAL ECOSYSTEM CENTER www.systemecology.org
Remote Sensing & Classified Land Cover
Essential Land Use Decision Support Tools Using
High-Resolution Imagery
Also available:
Essential Land Use Decision
Support Tools Using High-Resolution
Imagery

Moderate_resolution_GEC

More Related Content

What's hot

Similar to Moderate_resolution_GEC

More from Kenneth Kay

Moderate_resolution_GEC