Current Project Information:
 |
Jonsson,
C.B., L. Allen, Y. Chu, R. Owens,
D.G.
Goodin, J.M.S.
Hutchinson, E. Pontelli, D. Ranjan, S. Tran, M. Almiron.
2004. The Impact
of Rapid Anthropogenic Land Cover Change in the Chaco and
Interior
Atlantic
Forest
in
Paraguay
on Hantavirus Ecology. National
Institutes of Health, $1,857,996. |
Project Overview
From a landscape perspective, addressing
the goal of understanding the spatial and temporal patterns on microbial
diversity requires an understanding of the spatial ecology of the rodent
populations that serve as a reservoir for the viruses.
We envision a landscape data collection program that approaches
this problem using a variety of landscape metrics at multiple spatial
scales, analyzed within a Geographic Information System (GIS) framework to
characterize and model the landscape factors associated with spread of
disease vectors. Collection of
ecological data will combine remote and in-situ observation.
Although not widely applied, remote sensor data have been used to
detect environmental factors associated with vector habitats and human
transmission risks for a number of infectious diseases.
(Beck et al. 2000). These
applications have included detecting ecological parameters associated with
Rift Valley Fever in east Africa (Linthicum et al., 1987) and detecting
and mapping conditions associated with the Sin Nombre hantavirus in the
southwestern United States (Boone et al., 2000, Glass et al., 2000).
Data and Methods
Landscape
ecology emphasizes the effect of spatial heterogeneity and patterning on
ecosystems, and especially on biotic and abiotic process. Theoretical and
observational studies have shown to important effect of landscape spatial
structure on the viability and activity of various organisms, including
rodents (With 1994; With et al.,1997, 1999 ). Remote detecting and
mapping of the landscape characteristics associated with
disease-bearing rodents is challenging because the characteristic spatial
scale of the habitat is finer than the spatial resolution of most
operational remote sensing systems. To overcome this problem, we will
create synthetic high spatial resolution multispectral datasets by fusing
two data types; Landsat ETM+ and monochromatic digital orthophotography.
Landsat ETM+ data have spatial resolution of 30 meters and spectral
coverage in the VNIR and MIR spectral regions. These data are
optimized for observing vegetation in the spectral domain, but their
spatial resolution is too coarse to detect much of the spatial ‘fine
structure’ that we believe will differentiate the various habitats.
We therefore propose to fuse these data to one meter panchromatic images
collected by the IKONOS sensor, using a combination of
Intensity-Hue-Saturation (IHS) transformation coupled with redundant
wavelet decomposition (Chibani and Houacine, 2002). The resulting
images will retain the superior spectral properties of the ETM+ data along
with the fine spatial resolution of the IKONOS data.
Unlike
earlier studies linking remotely sensed data to viral activity (e.g.
Linthicum et al., 1987; Glass et al. 2000), we propose to go beyond simple
correlation of spectral reflectance (derived from images) to viral
activity. Instead, we propose to extract three types of information
from these fused remote sensing image data; 1) habitat type, 2) landscape
structure, and 3) canopy biophysical characteristics. These remote
sensing indices represent ecological information about the surface (which
simple spectral reflectance does not), thus our efforts to model the
spatial and temporal aspects of microbial behavior can be carried out
within the framework of realistic and relevant landscape variables.
Habitat mapping will be done using an unsupervised spectral-domain
clustering technique such as ISODATA, with maximum likelihood decision
rule (Roberts, 1987). The classification scheme will be
developed in close cooperation with mammologist, epidemiologists, and
modelers on our research team, insuring that the resulting cover map is of
optimal use in studying the spatial ecology of disease-bearing rodents.
Habitat maps will be constructed for both study sites for each of the
first four years of the project. Landscape spatial structure will be
quantified using a number of standard landscape structural metrics; for
example patch shape and area index, patch connectivity (McGarigal and
Marks, 1995), lacunarity (Henebry and Kux, 1995), fractal dimension
(Emerson et al., 1999), image texture (Musick and Grover, 1991), and
geostatistical range (Cohen, 1990). Biophysical canopy information
will be extracted from the remote sensor by exploiting the well-known
relationships between vegetation indices such as the Normalized Difference
Vegetation Index (NDVI, calculated as the difference between reflectivity
in the red and near infrared bands divided by the sum of the two
bands) and canopy biophysical variables such as leaf area index, biomass,
and canopy moisture (Price and Bausch, 1995).
Maps
of habitat, landscape structure, and canopy biophysical properties will be
combined with in-situ data on location and seropositivity status of rodent
populations, allowing correlative analysis and modeling of the habitat
preferences and spatial activity spaces of the disease-bearing rodents
associated with hantaviruses. These data sets will also be useful
for assessing the ecological conditions conducive to the presence of
viruses, potentially leading to the development of remote sensing based
tools for predicting conditions favorable for viral outbreak and
transmission to humans (Beck et al., 2000).
Remote sensing analysis will be supported
by on-site environmental observations to determine the microscale habitat
characteristics of the study area. Transects will be established
through the study areas; and biophysical characteristics such as community
composition, temperature, humidity, leaf moisture content, canopy
density/leaf area index (via ceptometer), canopy structure (via Decagon
First Growth camera), leaf physiology (via chlorophyll flourometry) soil
moisture (via short probe TDR), soil temperature, and soil type (and
others). These data will provide further information about the
characteristics of rodent habitats, as well as assisting in the
interpretation of the remote sensing data.
References
Chibani,
Y. and Houacine, C. 2002. The joint use of IHS transform and redundant
wavelet decomposition for fusing multispectral and panchromatic images. Remote
Sensing of Environment. 23:3821-3834.
With,
K.A. 1994. Using fractal analysis to assess how species perceive landscape
structure. Landscape Ecology 9: 25-36.
With,
K.A., S.J. Cadaret, and C. Davis. 1999. Movement responses to patch
structure in experimental fractal landscapes. Ecology 80:
1340-1353.
With,
K.A., R.H. Gardner, and M.G. Turner. 1997. Landscape connectivity and
population distributions in heterogeneous environments. Oikos 78:
151-169.
Cohen,
W. B., T.A. Spies, And G.A. Bradshaw. 1990.
Semivariograms of Digital Imagery for Analysis of Conifer Canopy
Structure. Remote Sensing of Environment.
34: 167-178.
Emerson,
C.W., N.S-N. Lam, and D.A. Quattrochi, D.A. 1999. Multi-scale
fractal analysis of image texture and pattern. Photogrammetric
Engineering and Remote Sensing 65:51-61.
Henebry,
G.M. and H.J.H. Kux. 1995. Lacunarity as a texture measure for
SAR imagery. International Journal of Remote Sensing
16:565-571.
Musick,
H.B. and H.D. Grover. 1991. Image textural measures as indices
of landscape pattern. In: Quantitative Methods in Landscape
Ecology, Turner, M.G. and Gardner, R.H., eds. New York:
Springer-Verlag.
|
