PE&RS February 2018 Full - page 101

Removing Rural Roads from the National Land
Cover Database to Create Improved Urban Maps
for the United States, 1992 to 2011
Christopher E. Soulard, William Acevedo, and Stephen V. Stehman
Abstract
Quantifying change in urban land provides important infor-
mation to create empirical models examining the effects of
human land use. Maps of developed land from the National
Land Cover Database (
NLCD
) of the conterminous United
States include rural roads in the developed land class and
therefore overestimate the amount of urban land. To bet-
ter map the urban class and understand how urban lands
change over time, we removed rural roads and small patches
of rural development from the
NLCD
developed class and cre-
ated four wall-to-wall maps (1992, 2001, 2006, and 2011) of
urban land. Removing rural roads from the
NLCD
developed
class involved a multi-step filtering process, data fusion using
geospatial road and developed land data, and manual edit-
ing. Reference data classified as urban or not urban from a
stratified random sample was used to assess the accuracy of
the 2001 and 2006 urban and
NLCD
maps. The newly created
urban maps had higher overall accuracy (98.7 percent) than
the
NLCD
maps (96.2 percent). More importantly, the urban
maps resulted in lower commission error of the urban class
(23 percent versus 57 percent for the
NLCD
in 2006) with the
trade-off of slightly inflated omission error (20 percent for
the urban map, 16 percent for
NLCD
in 2006). The removal
of approximately 230,000 km
2
of rural roads from the
NLCD
developed class resulted in maps that better characterize the
urban footprint. These urban maps are more suited to model-
ing applications and policy decisions that rely on quantitative
and spatially explicit information regarding urban lands.
Introduction
Urban land is defined in a variety of ways, but is generally dis-
tinguished from other forms of developed land based on high-
er population density, higher building density, higher land use
intensity, and/or more impervious cover. For the purposes of
this article, urban is defined as densely developed clusters of
land including residential, commercial, and industrial land
uses. We define “rural roads” as all roads located outside of ur-
ban areas, as well as small patches of rural development such
as mixed-use agricultural communities, low-density outposts,
mining drill pads, railroads, and highway rest stops. We sub-
sume all non-urban land types under the label “rural roads”
to simplify terminology because rural roads represent the vast
majority of area represented by these types of non-urban de-
velopment. Monitoring the status, trends, and spatial patterns
of urban lands is essential to understanding the causes and
consequences of human land use practices on the landscape.
By linking urban changes to spatiotemporal factors and policy
drivers (Bengston
et al
., 2004; Wang
et al
., 2012), scientists,
planners, and land managers may be able to anticipate future
changes on the landscape. Similarly, understanding how ur-
ban changes directly impact wildlife habitat (including habitat
loss and fragmentation) (Fahrig, 2003; Swenson and Franklin,
2000), surface and ground-water hydrology (Göbel
et al
., 2004;
Strauch
et al
., 2008), and surface albedo (Taha
et al
., 1988), as
well as how changes indirectly alter species biodiversity and
air and water pollution (Fahrig, 2003; Strauch
et al.
, 2008)
may help land-use planners mitigate against further impacts
by establishing urban growth limits or boundaries. Effective
land management planning relies on dependable, consistent
data representing changes in the urban footprint over time.
While many tabular data and spatially-explicit maps cur-
rently provide national-scale information on developed lands
in the United States (US Department of Agriculture, 2011; Fal-
cone, 2015; Loveland
et al
., 2002; Mitchell, 1977; Nusser and
Goebel, 1997; Price
et al
., 2006; Sleeter
et al
., 2013; Soulard
et
al
., 2014; Theobald, 2014; US Census Bureau, 2016a, US Cen-
sus Bureau, 2016b, US Geological Survey, 2016), few focus on
the urban component of developed lands. Developed lands
often include both rural and urban areas, and are commonly
defined using a modified Anderson classification scheme (An-
derson, 1976). Anderson defines developed lands as:
Areas of intensive use with much of the land covered by
structures. Included in this category are cities, towns, vil-
lages, strip developments along highways, transportation,
power, and communications facilities, and areas such as
those occupied by mills, shopping centers, industrial and
commercial complexes, and institutions that may, in some
instances, be isolated from urban areas.
Perhaps the most commonly used resource for maps of de-
veloped lands over time is the National Land Cover Database
(
NLCD
).
NLCD
provides wall-to-wall maps of land cover and
land-cover change for the conterminous United States (
CONUS
)
across four map dates (1992, 2001, 2006, and 2011) (Fry
et al
.,
2009; Homer
et al
., 2004; Jin
et al
., 2013; Vogelmann
et al
.,
2001; Xian et al., 2009). The 2001, 2006, and 2011
NLCD
clas-
sification employed classification and regression trees (
CART
),
impervious surface area estimates, and ancillary data to clas-
sify land-use/land-cover (
LULC
) types from Landsat imagery
based on a modified Anderson Level II (Anderson, 1976)
Christopher E. Soulard is with the US Geological Survey,
Western Geographic Science Center, Menlo Park, CA, orcid.
org/0000-0002-5777-9516 (
).
William Acevedo is Emeritus, US Geological Survey, Western
Geographic Science Center, Menlo Park, CA.
Stephen V. Stehman is with the Department of Forest and
Natural Resources Management, State University of New
York, Syracuse, NY, orcid.org/0000-0001-5234-2027.
Photogrammetric Engineering & Remote Sensing
Vol. 84, No. 2, February 2018, pp. 101–109.
0099-1112/17/101–109
© 2018 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.84.2.101
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
February 2018
101
51...,91,92,93,94,95,96,97,98,99,100 102,103,104,105,106,107,108,109,110,111,...114
Powered by FlippingBook