Geoscience Reference
In-Depth Information
Reserve (Nigeria) and in Gambia. A detailed review of satellite-derived variables
and their use in public health is given by Hay (
2000
) and Goetz et al. (
2000
).
More complex methods involve using statistical models that relate various
components of the transmission of parasites to satellite data for the description,
explanation and prediction of vector-borne diseases. Remote sensing data are used
as inputs to drive (i) spatial models of transmission risk (Danson et al.
2006
),
(ii) hydrological models that aim to describe or predict disease transmission in
catchment basins (Spear et al.
1998
), and (iii) climate based parasite forecast models
that reveal hotspots of disease occurrences and predict the risk area and intensity of
diseases (Zhou et al.
1998
,
1999
).
Epidemiological studies have taken advantage of the steady improvement of
satellite sensors. In some instances, investigators used aerial color-infrared pho-
tography (Welch et al.
1989
); however, passive satellite sensor data have been
used most commonly for epidemiological studies (Rogers et al.
2002
): Landsat
MSS (Barnes and Cibula
1979
); AVHRR (Baylis et al.
1999
; Brooker et al.
2002
;
Estrada-Peña and Venzal
2006
; Mendelsohn and Dawson
2008
); Landsat TM (Pope
et al.
1992
; Hugh-Jones et al.
1992
; Brown et al.
2008
); Landsat ETM (Brooker
et al.
2004
; Goossens et al.
2006
); ASTER (Tatem et al.
2004
); MODIS (Raso et al.
2006
; Gilbert et al.
2007
); METEOSAT (Hay et al.
1998
; Hay et al.
2006
); SPOT
(Tran et al.
2002
; De La Rocque et al.
2004
); IKONOS (Mushinzimana et al.
2006
);
QuickBird (Ratana et al.
2005
). While earth observing satellites have a higher spa-
tial resolution (e.g. 1-4 m for IKONOS, 60-70 cm for panchromatic, 2-4 m and
2-8 m for multispectral QuickBird; 10-20 m for SPOT), meteorological satellites,
despite their lower resolution, have the major advantage of a high temporal resolu-
tion (two images are collected every day by AVHRR and one image every 30 min
by the geostationary satellite, Meteosat). Recent studies have focused on the use
of active remote sensing for epidemiology applications. Kaya et al. (
2004
)used
RADARSAT-1 and RADARSAT 2 for vector-borne disease risk mapping.
Remote sensing methods have been employed for a wide range of diseases over
various regions worldwide: e.g. Dengue fever in Argentina (Carbajo et al.
2001
;
Rotela et al.
2007
); onchocerciasis in West and Central Africa (Thomson et al.
2000
); Rift Valley fever in Kenya (Linthicum et al.
1987
), Yemen (Abdo-Salem
et al.
2006
), Senegal (Lacaux et al.
2007
); schistosomiasis in Africa (Brooker
2007
)
and China (Zhou et al.
1998
; Guo et al.
2005
); cholera in Bangladesh (Lobitz et al.
2000
); tick-borne encephalitis in the Baltic States (Šumilo et al.
2006
); visceral
leishmaniasis in Sudan (Elnaiem et al.
1998
) in India (Sudhakar et al.
2006
) and in
Southwestern Asia (Cross et al.
1996
); human trypanosomiasis in Africa (Rogers
and Randolph
1994
; Rogers et al.
1996
; Kitron et al.
1996
; Rogers
2000
); helminth
infections in Cameroun (Brooker et al.
2002
); human tick-borne encephalitis infec-
tion in the Czech Republic (Daniel et al.
2006
); lyme disease in Wisconsin (Kitron
and Kazmierczak
1997
). A massive body of remote sensing-based studies focus
on malaria in various parts of the world, including the Western Kenyan highlands
(Mushinzimana et al.
2006
), Mexico (Beck et al.
1997
), Southern France (Tran
et al.
2008
), China (Liu and Chen
2006
); Belize (Achee et al.
2006
); Philippines
(Leonardo et al.
2005
); India (Srivatsava et al.
2001
); Bangladesh (Rahman et al.
