PERS_September_2018_Flipping_86E2 - page 559

A Long-Term, Consistent Land Cover History of
the Southeastern United States
AAG Remote Sensing Specialty Group 2018 Award Winner 1
Matthew P. Dannenberg, Conghe Song, and Christopher R. Hakkenberg
Abstract
Land-Cover/Land-Use (LCLU ) change is a critical aspect of
global environmental change, with profound social and eco-
logical consequences. The southeastern U.S. in particular is
changing rapidly, but a long-term, consistent LCLU history at
fine spatial resolution does not exist for the region. Here, we
present a new LCLU history of the southeastern U.S. based on
temporal extension of the 2011 National Land Cover Database
(NLCD) back to 1986. We used Automatic Adaptive Signature
Generalization ( AASG) to generate this product from Landsat
TM/ETM+ imagery and ancillary topographic information. AASG
identifies stable sites between two images and uses these
stable sites to generate a new training dataset for updating a
classification from one date to the next. Our long-term LCLU
classifications are broadly consistent with the NLCD while
providing a much longer historical record for characterizing
recent changes in the southeastern U.S. and contextualizing
their consequences for ecosystem services in the region.
Introduction
The amount of land dedicated to human use has increased
dramatically over the past several centuries (Ellis et al., 2010;
Foley et al., 2005), and both urban and agricultural land uses
will likely continue increasing in the near future (Seto et al.,
2012; Tilman et al., 2011). Land cover and land use change al-
ter Earth’s biogeochemical cycles, biophysical processes, and
ecological systems. Conversion of forests and other natural
cover types to agricultural, commercial, or residential land
uses results in a net release of carbon to the atmosphere (Can-
adell et al., 2007; Le Quéré et al., 2009), while forest regrowth
on abandoned agricultural land and afforestation on previ-
ously unforested land results in a net uptake of carbon by the
land surface (Birdsey et al., 2006; Pan et al., 2011). Conver-
sion of land cover alters the biophysical environment through
changes in albedo, surface hydrology, evapotranspiration and
precipitation recycling, and partitioning of net radiation to
sensible versus latent heat flux (Bonan, 2008; Bounoua et al.,
2002; Bronstert et al., 2002; Diffenbaugh, 2009; Kaufmann et
al., 2007; Pongratz et al., 2010). The spatial area and configu-
ration of different land cover types also have important con-
sequences for biodiversity and for the movement of animals
across landscapes (Fischer and Lindenmayer, 2007; Nagendra
et al., 2004; Newbold et al., 2015; Sala et al. , 2000; Tylianakis
et al., 2008). Land cover and land use change therefore alters
the global carbon cycle, local ecosystem services, and human
health, e.g., through exposure to heat-related illness (Jenerette
et al., 2016; Kovach et al., 2015) and vector-borne diseases
(Wimberly et al., 2008).
The southeastern United States (hereafter, “the Southeast”)
is one of the most rapidly changing regions in the country
(Fry et al., 2011; Sleeter et al., 2013; Wear and Greis, 2013),
with gross forest cover losses that are among the highest rates
observed globally during the 21st century (Hansen et al., 2010
and 2013). During the 19th and early 20th centuries, secondary
forest cover in the Southeast increased following agricultural
abandonment, but in recent decades this trend has reversed,
with significant forest cover losses resulting primarily from
timber harvesting and rapid urban expansion (Drummond and
Loveland, 2010; Napton et al., 2010). By the mid-21st century,
the amount of land devoted to urban and suburban use in the
Southeast is projected to be two to three times greater than at
present (Terando et al., 2014). However, long-term, consistent,
high-resolution maps of land cover over the Southeast are
needed in order to characterize past land cover and land use
change, to assess its biogeophysical, biogeochemical, and eco-
logical consequences, and to contextualize likely future urban
and suburban growth across the region.
With the recent opening of the Landsat archive and ad-
vances in computing technology, it is now more feasible than
ever to develop innovative algorithms for long-term, mod-
erate-resolution land cover classifications over large areas.
Several methods have been proposed and implemented for
large-area mapping (e.g., Franklin et al., 2015; He et al., 2017;
Sheffield et al., 2015; Shih et al., 2016; Xian et al., 2009; Yin
et al., 2014). Here, we use one such method, automatic adap-
tive signature generalization (AASG) (Dannenberg et al., 2016;
Gray and Song, 2013), to extend the National Land Cover
Database (NLCD) backwards in time for the Southeast. Our ob-
jectives in this research are: (1) to operationalize AASG, which
had previously only been tested on small scales, for large-area
mapping; (2) to provide a new long-term, consistent land
cover dataset for the rapidly changing and dynamic Southeast
region; and (3) to assess the accuracy of the new Southeast
land cover dataset, thus providing new evidence regarding the
Matthew P. Dannenberg is with the Dept. of Geographical and
Sustainability Sciences, Univ. of Iowa, Iowa City, IA 52242,
and formerly with the School of Natural Resources and
the Environment, University of Arizona, Tucson AZ 85721
(matthew-dannenberg@uiowa.edu).
Conghe Song is with the Department of Geography, University
of North Carolina, Chapel Hill NC 27599.
Christopher R. Hakkenberg is with the Department of
Statistics, Rice University, Houston TX 77005.
Photogrammetric Engineering & Remote Sensing
Vol. 84, No. 9, September 2018, pp. 559–568.
0099-1112/18/559–568
© 2018 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.84.9.559
1. In recognition of the 100th Anniversary of the Association of Ameri-
can Geographers (AAG) in 2004, the AAG Remote Sensing Specialty
Group (RSSG) established a competition to recognize exemplary
research scholarship in remote sensing by early career scholars in
Geography and allied fields. Matthew P. Dannenberg submitted this
paper which was selected as the 2018 winner.
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING September 2018 559