The Role of Remotely Sensed Data in Future Scenario Analyses at a Regional Scale - Fluvial Remote Sensing for Science and Management - page 270

local (Aggett et al., 2009), regional (Bolliger et al., 2007), or
hemispheric (Lawler et al., 2006). Various time scales can
also be considered, with retrospective analysis (Bolliger
et al., 2007, Brierley et al., 2009), middle term prospective
(Kepner et al., 2004, Santelmann et al., 2006), long-term
prospective (Fullerton et al., 2009, Lautenbach et al.,
2009), or very long-term prospective analysis (Virkkala
et al., 2008, Baas et al., 2010). In geomorphology, such
approaches are often used to assess the impact of human
activities (Zezere et al., 2004), land use change (Bouman
et al., 1999, Kepner et al., 2004) or the effects of natural
processes on the morphology of a river and its floodplain
(Wolski et al., 2008, Magdaleno et al., 2011). Recently,
these approaches have been increasingly used to assess the
future quality of habitats in response to human activities
(Hemstrom et al., 2001, Vache et al., 2002, Fullerton et al.,
2009), or climate change (Lawler et al., 2006).

cornerstone of its representation of the territory under
study, western Oregon's Willamette River basin.

12.1.4 Alternative futurescenarios for the
WillametteRiver,Oregonasacasestudy

This chapter is based on the research of more than 30
scientists from Oregon State University, University of
Oregon, and the US Environmental Protection Agency
(EPA) who assembled ecological and social information
about the Willamette River basin (WRB) to assess trajec-
tories of environmental change from 1850 to 2050. Past,
present, and future landscapes were depicted in maps to
represent scenarios of landscape conditions and human
actions that shape the trajectories of ecological condi-
tions in this basin. Rather than having scientists define
the assumptions and relationships for future scenarios,
we used the recommendations of groups of stakeholders
who met with us for more than two years to articulate and
review the assumptions used to create spatially explicit
characterisations of three future scenarios for 2050 (Hulse
et al., 2004). This summary of environmental trajectories
in the WRB is reported in detail in the Willamette River
Basin Planning Atlas (Hulse et al., 2002) and a special
issue of Ecological Applications (Baker et al., 2004 and
associated papers).
This chapter aims to: 1) explore many of the aquatic
and terrestrial resource responses that were informed by
remotely sensed information and 2) illustrate an approach
at nested spatial analyses within a large river basin. This
basin scale assessment of trajectories of ecological change
is based on spatially explicit representations of 1) histor-
ical land cover ca. 1850 prior to recent trends of land
conversion and resource consumption, 2) current land
cover and land use ca. 1990 based on remotely sensed
data from Landsat images and aerial photography, 3)
alternative future scenarios of land use and land cover ca.
2050 based on an interactive stakeholder process, and 4)
modeled projections of major ecological characteristics
of these landscape scenarios based on observed empirical
relationships in the WRB.
The landscape assessment in this research required inte-
gration of remotely sensed data and ecological models to
explore complex relationships between people and the
changing landscapes that shape their futures. The 30,000-
km 2 Willamette basin makes up 12% of the state of
Oregon and supports more than two-thirds of Oregon's
population, which is expected to double by 2050. The
basin provides important agricultural and forest prod-
ucts for the region but also is one of the most ecologically

12.1.3 Methodsof employingremotelysensed
information inalternative futures

Three major methods are commonly used to project
biophysical or human infrastructural characteristics of
the landscape from remotely sensed information. First,
remotely sensed information can be classified or inter-
preted to explicitly represent landscape features, such
as land cover (e.g., vegetation types, water, rock or soil)
(Steel et al., 2004, Fullerton et al., 2009). Second, remotely
sensed information can be enhanced with social informa-
tion to increase the resolution of the classification, such
as land use (e.g., crop type, residential and industrial zon-
ing, transportation data, census data) (Baker et al., 2004).
Third, algorithms or models can be used to estimate bio-
physical characteristics from remotely sensed information
(e.g., fish and wildlife community structure, abundance,
habitat quality) (VanSickle et al., 2004, Schumaker et al.,
2004, Lawler et al., 2006).
The purposes of the analysis strongly influence the
processes for applying remotely sensed information.
Processes used to develop scenarios include definition
of assumptions for scenarios and their interpretation
by scientists, stakeholder groups, decision makers and
managers.
The remainder of this chapter describes, as a case study,
work conducted between 1995 and 2002 by the Pacific
Northwest Ecosystem Research Consortium, a multi-
university/federal agency partnership with support from
the US Environmental Protection Agency. This effort con-
sisted of a river-basin extent alternative future scenario
analysis that employed remotely sensed information as a