This study examined the impact of road networks on the distribution of dengue fever cases in Trinidad, West Indies. All confirmed cases of dengue hemorrhagic fever (DHF) observed during 1998 were georef- erenced and spatially located on a road map of Trinidad using Geographic Information Systems software. A new digital geographic layer representing these cases was created and the distances from these cases to the nearest classified road category (5 classifications based on a functional utility system) were examined. The distance from each spatially located DHF case to the nearest road in each of the 5 road subsets was determined and then subjected to an ANOVA and t-test to determine levels of association between minor road networks (especially 3rd and 4th class roads) and DHF cases and found DHF cases were located away from forests, especially 5th class roads). The frequency of DHF cases to different road classes was: 0% (1st class roads), 7% (2nd class roads), 32% (3rd class roads), 57% (4th class roads) and 4% (5th class road). The data clearly demonstrated that both class 3 and class 4 roads account for 89% of nearby dengue cases. These results represent the first evidence of dengue cases being found restricted between forested areas and major highways and would be useful when planning and implementing control strategies for dengue and Aedes aegypti mosquitoes.
2. R.S. Mahabir et al. / Acta Tropica 123 (2012) 178–183 179
The behavioral factors which influence the dispersal of Ae.
aegypti mosquitoes include the flight associated with securing
blood meals, like flying indoors, while oviposition flight is influ-
enced by the availability of breeding sites like finding artificial
water holding containers. Utilizing the need by female mosquitoes
to forage for blood meals (Dunn, 1927) and the need to seek out
oviposition sites (Chadee, 1997), a mark-release-recapture study
reported that Ae. aegypti in Queensland, Australia readily crossed
small quiet roads but avoided crossing a major highway near the
release point and concluded that busy roads may impede dispersal
(Russell et al., 2005). In a similar study, Laurance et al. (2009) inves-
tigated the impacts of roads and linear clearings in tropical forests
and found various ‘beetles, flies, ants, bees, butterflies, amphibians,
reptiles, birds, bats and small and large mammals tend to avoid
even narrow (<30 m wide) clearings or forest edges’. These results
were confirmed by a study which examined the influence of a major
highway in Trinidad on the population dispersal or dynamics of
Ae. aegypti using a panel of microsatellites, two SNPs and a 710 bp
sequence of mtDNA cytochrome oxidase and found strong evidence
of limited gene flow across the highway, which effectively frag-
mented the population on the east and west side of the highway
(Hemme et al., 2010).
These previous studies have examined the impact of roads on
the dispersal of the vector of dengue fever but the impact of roads
or landscape barriers on the transmission of dengue fever has never
been reported and studies of this kind are long overdue. The results
of these studies may well explain the clustering of dengue cases
and identify the impediments faced by infected mosquitoes to dis-
tribute the infection to other geographical localities. The present
study was conducted to investigate the impact of the different road
networks on the distribution or dispersal of dengue fever cases in
Trinidad, West Indies.
2. Materials and methods
2.1. Study area
Trinidad (10.5◦N; 61.5◦W) is the southernmost island within
the Caribbean archipelago located northeast of Venezuela and sep-
arated by a distance of 15 km (Bradbury et al., 1981; Flenley,
1993) (Fig. 1). It is estimated that Trinidad became separated
from the South American mainland some 11,000–15,000 years
ago (Bradbury et al., 1981; Flenley, 1993) and the geology, flora
and fauna on the island is basically continental (Mohan et al.,
2009). Trinidad is rectangular in shape covering 4828 km2 (TIDCO,
2006), with an estimated population of 1.2 million inhabitants
(PROCICARIBE, 2006). One-third of the island is considered to be
cultivable land (PROCICARIBE, 2006) while the remaining land sur-
face is dominated by naturally occurring native vegetation making
up an estimated 61.2% of the island (MALMR, 1995).
The island experiences a tropical climate with average daily
temperatures ranging between 22 ◦C and 32 ◦C (Hoag et al., 2001)
with two distinct seasons: a dry season from December to May, and
a wet season from May to November (Mohan et al., 2009). Due to the
prevailing northeast trade winds and orographic effects, the high-
land areas of northeast Trinidad receive up to 3800 mm per year of
rainfall with precipitation ranging from 1250 mm in the northwest
to 3000 mm in the southwest part of the island respectively (Hoag
et al., 2001).
2.2. Case data
A definition of a confirmed DHF case was provided by the Min-
istry of Health, Insect and Vector Control Division. This definition
was framed as persons (child or adult) having the symptoms;
Fig. 1. Distribution of DHF cases in Trinidad (1998).
temperature of 38 ◦C or higher for 5 days, accompanied by
headache, myalgia, and other non specific clinical presentations
(Chadee, 2009; WHO, 2010). These persons were closely monitored
for signs of hemorrhage. All suspected cases of DHF were labora-
tory confirmed by virus isolation, detection of specific IgM antibody
and/or seroconversion (Chadee, 2009; WHO, 2010).
Records representing confirmed DHF cases for 1998 were
collected from the Ministry of Health. These contained street
addresses of persons confirmed to be infected by the disease. The
year 1998 was specifically chosen because it represented a major
outbreak of the disease in Trinidad.
2.3. Address geocoding
The street addresses of confirmed DHF cases were located on a
Geographic Information System (GIS) road layer of Trinidad using
the ArcGIS software. Successfully located cases were then plotted
as point features; creating a new GIS layer representing point loca-
tions of DHF cases in Trinidad. Due to insufficient information in the
DHF case records (e.g. missing addresses or much generalized loca-
tions such as city or county level) only 76.6% (157) of cases were
geocoded. This had the net effect of making it almost impossible
to get accurate home or work place addresses due mainly to poor
standardization of data collection forms used in primary, secondary
and tertiary care institutions.
2.4. Cluster analysis
In order to validate or confirm the data analysis, the Average
Nearest Neighbor (ANN) method was used (Getis and Franklin,
1987) to determine whether the spatial distribution of dengue cases
was due in part to chance or non-random spatial clustering. This
method calculates a nearest neighbor index based on the average
distance from each case to its nearest neighboring case. If the aver-
age distance is less than the average for a hypothetical random
distribution, the distribution of the cases being analyzed are con-
sidered clustered, if not they are considered dispersed. The index
is represented as the ratio of the observed distance divided by the
expected distance (expected distance is based on a hypothetical
random distribution with the same number of features covering
3. 180 R.S. Mahabir et al. / Acta Tropica 123 (2012) 178–183
the same total area) (Getis and Franklin, 1987) and is calculated
using the formula
ANN =
¯DO
¯DE
where ¯DO is the observed mean distance between each feature and
the nearest neighbor and ¯DE is the expected mean distance for the
features given in a random pattern.
2.5. Distance to roads
Roads are classified based on their administrative and util-
ity roles. The administrative function demarks road ownership
while utility looks at the technical requirements and maintenance
practices, which influence the administrative classification and
financing (Talvitie, 1996). In Trinidad, a functional approach has
been adopted with the hierarchy of roads based on size, use for
mobility and accessibility parameters. Mobility refers to the actual
ability of the road to move traffic while accessibility refers to the
ease of entering or exiting a roadway to or from adjacent properties
(land access). Trinidad categorizes roads into five classes, starting at
class 1 representing arterials which allowed high mobility and low
accessibility, and increasing in class order by a value of 1 to class
5 representing local roads and indicating low mobility and high
accessibility. This functional classification system provides a surro-
gate measure for the volume of traffic that may be expected from
the various road classes in Trinidad with traffic volume decreasing
with an increase in class number or mobility.
In Trinidad, class 1 and 2 roads are the longest found, stretching
along the length and breadth of the island, with over 60% of the
population living along and within close proximity to these road
networks. In contrast, class 3 and 4 roads are less than 20 km long
and serve mainly minor towns and villages which house less than
30% of the population. While, class 5 roads usually vary in their
length from 10 to a few kilometers and serve less than 10% of the
Trinidad population. These roads are found in more urbanized land-
scapes throughout Trinidad compared to class 3 and class 4 roads
which are found in more rural settings (Fig. 2).
The geographic road layer acquired in this study contained
attribute information expressing the names and the various
Fig. 2. Dengue incidents and distance to roads occurring during the 1998 outbreak
in Trinidad, West Indies.
classifications into which each road in Trinidad was assigned. The
road layer together with the layer identifying the point locations
of confirmed DHF cases were then used in tangent with the Near
tool in ArcGIS to identify the type of road or road class nearest
to each dengue case. The Near tool determines the Euclidean or
straight line distance between features in one layer and the near-
est feature in another. Its use emphasizes that Ae. aegypti is not
usually obstructed during flight by natural or manmade features.
All distances calculated for analysis were expressed in kilometers.
Next, the road layer was used to create five other GIS layers
based on the attached road class information. These layers were all
subsets of the original road layer. Each new layer represented a set
of roads belonging to an individual category or class of road. Using
the ANN tool, the distance from each DHF case to the nearest road
in each of the five layers was calculated. Distances calculated for
each layer were then placed into 1 km bins. A threshold of 6 km was
used as a cutoff point as distances beyond this were far to disperse
to warrant further investigation.
2.6. Distance to forests
A GIS land cover layer for Trinidad for the year 1998 was
obtained and used to extract a layer representing forested areas
across the island. Next, the forest map was used together with the
Near tool in ArcGIS once more to identify the distance from each
case of DHF to the nearest forested area. These distances were sub-
divided into 1 km bins. Each bin was further subdivided into the
frequency of the nearest road class to DHF cases.
2.7. Software and digital datasets
Data were analyzed using Geographic Information System soft-
ware package, ArcGIS version 9.3 (Earth Science Resource Institute,
Redlands, California) (ESRI, 2009) and the Data Analysis Tools sta-
tistical extension for Microsoft Excel 2007. All digital geographic
datasets collected or created were represented as thematic lay-
ers and converted to a common geographic coordinate system,
WGS 1984 Zone 20N to support uniform analysis of the data. Geo-
graphic datasets used were collected in 1994 from a variety of
sources, principally, the Ministry of Health, Trinidad and Tobago
and the Department of Geomatics Engineering and Land Manage-
ment, University of the West Indies, St. Augustine, Trinidad (based
on Anderson Level 1). An Analysis of Variance (ANOVA) and t-test
were used for all statistical tests and P values <0.05 were considered
to be statistically significant (Petrie and Sabin, 2000).
3. Results
3.1. Spatial distribution of dengue
The spatial distribution of DHF cases recorded in Trinidad in
1998 is shown in Fig. 1. Most cases were clustered around the two
major cities; Port of Spain and San Fernando located on the west-
ern side of the island and with large numbers of DHF cases found
in either north or south Trinidad. The distribution of DHF cases is
concomitant with human settlements which are primarily located
on the western part of the island.
3.2. Clustering of dengue cases
The calculated value for the Average Nearest Neighbor (ANN)
index for DHF cases was 0.54 at a significance level of 0.01. Together,
these values suggest that DHF cases were spatially clustered and
there is less than a 1% likelihood that the observed cluster pattern
could be attributed to random chance thereby rejecting the null
4. R.S. Mahabir et al. / Acta Tropica 123 (2012) 178–183 181
Table 1
The distance of DHF cases (home addresses) and their proximity to road classes in
Trinidad, West Indies (1998).
Roads classes Distance (km)
Less than 1 1–2 2–3 3–4 4–5 5–6
1 20 21 9 14 9 4
2 97 18 18 9 5 2
3 110 29 8 5 0 2
4 148 9 0 0 0 0
5 100 21 14 8 5 2
Table 2
The significance values found for distances away from road classes at which DHF
cases were observed (t-test values).
Distances Within 1 km 1–2 km 2–3 km 3–4 km 4–5 km 5–6 km
P-value 0.006 0.004 0.047 0.071 0.342 0.161
hypothesis (with 99% confidence). These results provide sufficient
evidence to support further ecological analysis of this data.
3.3. Distance to roads
Fig. 2 shows the spatial distribution of DHF cases superimposed
onto road classes across Trinidad. These results show a greater
number of dengue case clusters around class 3 and class 4 roads
compared to any other road classes. The frequency of DHF cases
to different road classes was: 0% (1st class roads), 7% (2nd class
roads), 32% (3rd class roads), 57% (4th class roads) and 4% (5th class
road). The data clearly demonstrated that both class 3 and class
4 roads account for 89% of nearby dengue cases. The relationship
between road distances to dengue cases was further investigated
using an Analysis of Variance (ANOVA). The results show a sig-
nificant (F = 2.621; P < 0.0000) relationship between distance from
roads and dengue cases (Table 1). The results of the t-test (assum-
ing unequal means) confirmed this feature with more dengue cases
being found within 1–3 km away from the various road classes
(Table 2). That is, as road classes increase from arterial (class 1)
to local (class 5) there is a general increase in the prevalence of
dengue for roads within the first 3 km of cases. These results are
indicated by significance values less than 0.05 for distances within
the first 3 km and values greater than this threshold for distances
beyond 3 km (Fig. 2).
3.4. Distance to forests
Table 3 shows the number of DHF cases found near forested
areas and different road classes in Trinidad. In general the number
of DHF cases increased from arterial (class 1) to local roads (class
5), that is, within the first 1 km of forested areas but most cases
(P > 0.006) of DHF were clustered around class 3 and class 4 roads.
In addition, the relationship between DHF cases to forest locations
was further investigated using an Analysis of Variance (ANOVA)
and Near tool (ArcGIS) which indicated more cases of DHF occurred
Table 3
The distance of DHF cases (home addresses) and their proximity to forested areas
and class 5 roads in Trinidad, West Indies (1998).
Roads classes Distance from forest (km)
Less than 1 1–2 2–3 3–4 4–5
1 0 0 0 0 0
2 10 2 0 0 0
3 48 3 0 0 0
4 70 11 1 2 1
5 4 2 0 0 0
Table 4
The significance values found for distances away from forested areas and class 5
roads classes at which DHF cases were observed (t-test values).
Distances Within 1 km 1–2 km 2–3 km 3–4 km 4–5 km
P-value 0.083 0.39 0.006 0.009 0.006
away from forested areas (F = 2.866; P > 0.031). The results of the t-
test (assuming unequal means) confirmed this feature that is DHF
cases occurred away from the forest and class 5 roads (Table 4).
4. Discussion
The results show a positive correlation between dengue cases
and road networks in Trinidad, West Indies, with significant num-
bers of dengue cases geographically distributed close to minor
motorways (especially 3rd and 4th class roads) than with major
motorways (1st and 2nd class roads). This finding is supported
by the mark-release-recapture studies conducted in Queensland,
Australia which reported Ae. aegypti would readily crossed small
quiet roads but significantly fewer crossed major highways near
the release point and concluded that busy roads impeded dispersal
(Chadee, 1997). Goosem (2002) also reported that wider roads and
highways strongly hindered animal movement and is supported
by work conducted on Ae. aegypti movement across the Uriah But-
ler Highway in Charlieville, Trinidad (Hemme et al., 2010) using
molecular markers which found strong evidence of limited gene
flow across the highway, which effectively fragmented the popula-
tion on the east and west side of the highway. These studies provide
the environmental factors which limit the distribution of the vec-
tor and ultimately the transmission of dengue fever in some parts
of Trinidad (Hemme et al., 2010). Similar findings were observed
in Puerto Rico when a cluster of homes with similar human and
vector populations only 27.4 m from two foci of dengue remained
free of infection for more than two months (Neff et al., 1967). These
results strongly suggested a limited flight range or impaired disper-
sal of Ae. aegypti in the natural environment. Similarly, although
dengue fever is recognized as a residential disease its dispersal
or spread may be limited: to movement of infected human pop-
ulations within and outside of their home environment and the
dispersal of infected Ae. aegypti mosquitoes based on the type of
housing patterns and the type of road network available in the
area. The present results suggest that the road network can create
“barriers” to the infected and non infected mosquito’s dispersal.
Recent ecological studies have highlighted the impact of the
‘edge effect’ on the dispersal of a wide variety of arthropods,
reptiles, birds and small and large mammals which avoid even
narrow <30 m wide clearing or forest edge (Russell et al., 2005).
The present study also indicated that DHF cases were not observed
in close proximity to forested areas but rather in locations where
minor motorways are found (Fig. 1). Studies conducted by Colton
et al. (2003) in five townships along minor roadways showed
genetic heterogeneity among the various Ae. aegypti populations in
Trinidad and provided molecular genetic (RFLPs) evidence for ‘skip
oviposition’ in the field (Corbet and Chadee, 1993). These results
suggest that gravid Ae. aegypti females were moving in and around
houses and were not evidently impaired by the minor road network
in these study sites.
These results indicate that Ae. aegypti populations are often lim-
ited in their distribution between major motorways and forested
areas. These two physical features in the environment (forest and
highway) represent the borders within which Ae. aegypti live and
dengue fever is transmitted. Houses located along minor motor-
ways or well designed urban centers with block plan housing
patterns some distances away from forested areas produce con-
ducive conditions for Ae. aegypti breeding sites, resting sites, blood
5. 182 R.S. Mahabir et al. / Acta Tropica 123 (2012) 178–183
feeding sites, oviposition sites and areas to disperse. In contrast,
the major motorways may form major ‘barriers’ to the flying
mosquitoes because the volume of traffic is greater on these roads
and especially during blood feeding periods for mosquitoes (early
morning and later evening) since these usually coincide with jour-
neys to and from work. The flight speed of mosquitoes allows
these foraging mosquitoes to succumb to the traveling velocity
of vehicles at this time. Added to this, the width of major roads
increases the probability of being caught in the ‘wind tunnel’
of vehicles. Conversely, minor motorways usually have a lower
volume of traffic with narrower widths. This general pattern is
reflected in the results with an increase in the number of dengue
cases occurring in closer proximities to more minor class road-
ways. In fact, results highlight that no cases of dengue were found
close to 1st class roads compared to other classes of roads, a find-
ing similar to that reported from Trinidad (Hemme et al., 2010)
where major roadways were identified as barriers for the dispersal
of Ae. aegypti.
Work in the Central Plain of Thailand identified high risk areas
for dengue transmission and found different types of urbaniza-
tion were linked with different intensities of dengue transmission
(Barbazan et al., 2000). In fact, Barbazan et al. (2000) using Landsat
images found large numbers of dengue cases and most epidemics
were occurring in sub-districts with medium density housing areas
away from main roads. Data from the present study showed dengue
cases were never identified in housing communities close to class 1
roads (Fig. 2). These findings are contradictory to the main descrip-
tion of DF and DHF risk areas found in the literature (Gubler and
Kuno, 1997); that is, the densest urban areas where the virus can
spread easily were reported to be located near high traffic roads
allowing the importation of virus by infected travelers and the
vectors. In the Thailand study however, no attempt was made to
understand these results further (Barbazan et al., 2000) but the
present results showed the incidence of dengue increases with
decreasing road mobility and increasing levels of land access from
major roads (class 1) to local roads (class 4) after which values
decline once more (Tables 1–4).
In India (Arunachalam et al., 2004), Peru (Hayes et al., 1996),
Thailand (Barbazan et al., 2000), and in Puerto Rico (Morrison et al.,
1998) DF/DHF outbreaks in urban and rural communities suggested
dengue transmission occurred outside their housing communities,
at citizens work sites or when they traveled outside of the housing
community visiting relatives and friends within the wider commu-
nity. In India dengue transmission was possibly due to the extensive
movement of people (Arunachalam et al., 2004). In Thailand the
high traffic roads were responsible for allowing the importation of
virus by infected travelers (Barbazan et al., 2000) while in Peru a
high prevalence of dengue and Ae. aegypti were identified in urban
and rural centers but barriers to dengue transmission were not
identified (Hayes et al., 1996).
In Puerto Rico (Morrison et al., 1998) work on the tempo-
ral and spatial distribution of reported dengue cases suggested
that natural barriers between the neighborhoods and the low sur-
vival rates for completion of the extrinsic incubation limited the
dispersal of the dengue virus to other geographically separated
neighborhoods. The present study identified one of the physical
barriers to dengue transmission as 1st and 2nd class roads and fur-
ther explains the apparent visual clustering of cases around cities
and urban centers (Fig. 1). Tables 3 and 4 show the relationship
between the forest and the distance to dengue cases and road
classes and showed the number of cases increased as the classes
of roads were less used by vehicular traffic (Table 3). The find-
ings of this research provide preliminary evidence to support a
more in depth study into the association of roads on the distri-
bution of dengue in Trinidad. In addition, these results can be used
to develop and guide surveillance for dengue cases by developing
risk maps in both time and space. This approach can also help to
focus control strategies to areas where DF transmission is ongo-
ing and where both adult and immature mosquito populations are
high.
Conflict of interest
None.
References
Arunachalam, N., et al., 2004. Studies on dengue in rural areas of Kurnool District,
Andhra Pradesh, India. Journal of the American Mosquito Control Association
20, 87–90.
Barbazan, P., et al., 2000. Dengue hemorrhagic fever (DHF) in the Central Plain of
Thailand. Remote sensing and GIS to identify factors and indicators related to
dengue transmission. In: Conference Proceedings of the Chao Phraya Delta:
Historical Development, Dynamics and Challenges of Thailand’s Rice Bowl,
Kasetsart University, Bangkok, 12th–15th December 2000.
Bradbury, J.P., et al., 1981. Late quaternary history of lake Valencia. Science 214,
1299–1305.
Chadee, D.D., 1984. Aedes aegypti aboard boats at Port of Spain, Trinidad, W.I.
(1972–1982). Mosquito News 44, 1–3.
Chadee, D.D., 1997. Effect of forced egg-retention on the oviposition patterns of
Aedes aegypti (L.) mosquitoes. Bulletin of Entomological Research 86, 539–541.
Chadee, D.D., 2003. Surveillance for the dengue vector Aedes aegypti in Tobago, West
Indies. Journal of the American Mosquito Control Association 19, 199–205.
Chadee, D.D., 2009. Dengue cases and Aedes aegypti indices in Trinidad, West Indies.
Acta Tropica 112, 174–180.
Chadee, D.D., Doon, R., Severson, D.W., 2007. Surveillance of dengue fever cases using
a novel Aedes aegypti population sampling method in Trinidad, West Indies: the
cardinal points approach. Acta Tropica 104, 1–7.
Chadee, D.D., Ritchie, S., 2010. Oviposition behaviour and parity rates of Aedes
aegypti collected in sticky traps in Trinidad, West Indies. Acta Tropica 116,
212–216.
Christopher, S.R., 1960. Aedes aegypti (L.), The Yellow Fever Mosquito. Its Life History,
Bionomics and Structure. Cambridge University Press, Cambridge, UK.
Colton, Y.M., Chadee, D.D., Severson, D.W., 2003. Natural oviposition behavior of the
mosquito Aedes aegypti as evidenced by codominant genetic markers. Medical
and Veterinary Entomology 16, 195–201.
Corbet, P.S., Chadee, D.D., 1993. An improved method for detecting substrate pref-
erences shown by mosquitoes than exhibit “skip” oviposition. Physiological
Entomology 18, 114–118.
Dunn, L.H., 1927. Tree-holes and mosquito breeding in West Africa. Bulletin of Ento-
mological Research 18, 139–144.
Dutta, P., et al., 1998. Distribution of potential dengue vectors in major townships
along the national highway and trunk roads of north east India. Southeast Asian
Journal of Tropical Medicine and Public Health 29, 173–176.
ESRI (Earth Science Resource Institute), 2009. How Average Nearest Neighbor
Distance (Spatial Statistics) Works, Available from: www.esri.com (accessed
02.10.09).
Flenley, J.R., 1993. The palaeoecology of Easter Island and its ecological disaster. In:
Fischer, S.R. (Ed.), Easter Island Studies. Contributions to the History of Rapanui
in Memory of William T. Mulloy. Oxbow, Oxford, pp. 27–45.
Gubler, D.J., Kuno, G., 1997. Dengue and Dengue Haemorrhagic Fever. University
Press, CAB International, Cambridge, p. 453.
Garcia-Rejon, et al., 2008. Dengue virus infected Aedes aegypti in the home environ-
ment. American Journal of Tropical Medicine and Hygiene 79, 940–950.
Getis, A., Franklin, J., 1987. Second-order neighborhood analysis of mapped point
patterns. Ecology 68, 473–477.
Goosem, M., 2002. Effects of tropical rainforest roads on small mammals, fragmen-
tation, edge effects and traffic disturbance. Wildlife Research 29, 1–13.
Haverfield, L.E., Hoffman, B.L., 1966. Used tires as a means of dispersal of Aedes
aegypti in Texas. Mosquito News 26, 433–435.
Hayes, C.G., et al., 1996. The epidemiology of dengue virus infection among urban,
jungle, and rural populations in the Amazon region of Peru. American Journal of
Medicine and Hygiene 55, 459–463.
Hemme, R.R., et al., 2010. Influence of urban landscapes on population dynamics in a
short-distance migrant mosquito: evidence for the dengue vector Aedes aegypti.
PLoS Neglected Tropical Diseases 4, e634.
Hoag, R.B., et al., 2001. Estimation of groundwater recharge in Trinidad using mete-
orological, Geographic Information Systems (GIS), and watershed modeling. In:
Proceedings of the 10th Caribbean Water and Wastewater Association Confer-
ence, Cayman Islands, 1st–5th October 2001.
Hughes, J.H., Porter, J.E., 1956. Dispersal of mosquitoes through transportation, with
particular reference to immature stages. Mosquito News 16, 106–111.
Laurance, W.F., Goosem, M., Laurance, S.G.W., 2009. Impact of roads and lin-
ear clearings on tropical forests. Trends in Ecology and Evolution 24,
659–669.
Le Maitre, A., Chadee, D.D., 1983. Arthropods collected from aircraft at Piarco Inter-
national Airport, Trinidad, West Indies. Mosquito News 43, 21–23.
Mackenzie, J.S., et al., 1996. Dengue in Australia Journal of Medical Microbiology 45,
159–161.
6. R.S. Mahabir et al. / Acta Tropica 123 (2012) 178–183 183
MALMR (Ministry of Agriculture, Land and Marine Resources) 1995. Trinidad and
Tobago. Country report to the FAO International Technical Conference on Plant
Genetic Resource. Government of Trinidad and Tobago, Trinidad, 77 pp.
Mohan, A.R.M., et al., 2009. Epidemiology of human leptospirosis in Trinidad
and Tobago, 1996–2007. A retrospective study. Acta Tropica 112,
260–265.
Morrison, A.C., et al., 1998. Exploratory space-time analysis of reported dengue cases
during an outbreak in Florida, Puerto Rico, 1991–1992. American Journal of
Tropical Medicine and Hygiene 58, 287–298.
Neff, J.M., et al., 1967. Dengue fever in a Puerto Rican community. American Journal
of Epidemiology 86, 162–184.
Ordonez-Gonzalez, J.G., et al., 2001. The use of sticky ovitraps to estimate dispersal of
Aedes aegypti in northeastern Mexico. Journal of the American Mosquito Control
Association 17, 93–97.
Petrie, A., Sabin, C., 2000. Medical Statistics at a Glance. Blackwell Science Lts., Osney
Mead, Oxford, UK.
PROCICARIBE, 2006. Trinidad and Tobago. Land Resources. Trinidad, PROCI-
CARIBE.2005. Available from: http://www.procicaribe.org (accessed 11.09.06).
Reiter, P., Amador, M.A., Clark, G.G., 1995. Dispersal of Aedes aegypti in an urban
area after blood feeding as demonstrated by rubidium-marked eggs. American
Journal of Tropical Medicine and Hygiene 52, 177–179.
Reiter, P., Gubler, D.J., 1997. Surveillance and control of urban dengue vectors. In:
Gubler, D.J., Kuno, G. (Eds.), Dengue and Dengue Haemorrhagic Fever. CAB Inter-
national, Fort Collins, CO, pp. 425–453.
Russell, R.C., et al., 2005. Mark-release-recapture study to measure dispersal of the
mosquito Aedes aegypti in Cairns, Queensland, Australia. Medical and Veterinary
Entomology 19, 451–457.
Talvitie, A., 1996. International experiences in restructuring the road sector. Paper
presented at Transportation Research Board Annual Meeting, 1996, Washington,
DC.
TIDCO (Tourism and Industrial Development Company of Trinidad and Tobago),
2001. About Trinidad and Tobago Geography. Government of Trinidad and
Tobago, Trinidad, Available from: http://www.visittnt.com/General/about/
geography.html (accessed 19.08.06).
WHO, 2010. Dengue: Guidelines for Diagnosis, Treatment, Prevention and Control,
New Edition 2009. WHO Press, Geneva, p. 147.
