**4. Leishmaniasis in the New World**

Bhunia et al 2010 [23] in India, through satellite imagery complemented with a GIS database, estimated parameters such as altitude, temperature, humidity, rainfall and the normalized difference vegetation index (NDVI) for correlation with the distribution of Kala-azar. They observed that the highest prevalence was below 150 m of altitude with very few cases located above the 300 m level and a low NDVI value ranges correlated with a high occurrence of the disease. They also showed, that most of the cases occurred in non-vegetative areas or low density vegetation zones highlighting that the low density vegetation zones were significant

Khanal et al 2010 [24] in Nepal, merged results from a serological test made in humans and domestic animals with GIS technology to evaluate the exposure to *L. donovani* on two popu‐ lations in a recent focus of visceral leishmaniasis (VL). They used a Poisson regression model to evaluate the risk of infection in humans associated with seropositive animals in the prox‐ imities of the household. It was also demonstrated that seropositive animals and humans were spatially clustered and the presence of positive goats, past VL cases and the proximity to a forest island increased the risk of occurrence of seropositivity in humans. The authors also suggested that goats might play some role in the distribution of *L. donovani*, in the VL focus

Bhattarai et al 2010 [25] also in Nepal, with the purpose of determining possible reasons for persistence of VL during inter-epidemic periods, they mapped cases *Leishmania* infections among apparently healthy persons and animals in an area of active VL transmission. The results of a bivariate K-function analysis showed the occurrence of spatial clustering of *Leishmania* spp.–positive persons and domestic animals, addition the investigation through classification tree, determined that the proximity of *Leishmania* spp.–positive goats ranked as

Salahi-Moghaddam et al 2010 [26] in Iran performed a serological study on a population of

No significant correlation between topographic conditions and the prevalence of positive cases was observed after regression analysis. Nevertheless, positive correlations were found in relation to the amount of rainfall, between infected dogs with high titers (≥1/640) and the number of days with temperatures below 0 °C during one year. The same correlation was observed when they were considered past meteorological records, conversely the humidity

The authors suggested that in mapped areas the prevailing low temperatures could represent

More recently, Bhunia et al 2013 [27] in India, assumed that the utilization of GIS and RS technologies on the control of VL dates back to the late 2000s and those control programs have mostly focused on mapping prevalence and association of *Phlebotomus argentipes* habitats,

Besides, the authors proposed that the multiplicity of satellite and sensors technics offer relevant data to assembly spatial, spectral and temporal scales. They also argued about the

for the *P. argentipes* vector distribution in the disturbed areas.

134 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

the first risk factor for *Leishmania* infection among persons.

showed an inversely correlated with the *Leishmania* infections.

an important factor influencing the distribution of leishmaniasis.

predicting transmission risk in relation to ecological transformation.

studied.

dogs from an endemic area.

One of the first works, carried out in the New World that have exploited SR- satellite imagery technology on an epidemiological survey with American Cutaneous Leishmaniasis, was presented by Miranda et al 1996 [28] in Brazil. In that study, the data were plotted on a TM-LANDSAT image a color composition of bands 3, 4 and 5 (see supplementary information on table 3,4 and 5) that were considered useful to identify the relevant vegetation (shrubs and trees) within the boundaries of the studied areas and in their neighborhood about 250 meters from the perimeter of each area. It was suggested, the use of means qualified as presenting a larger view of a geographical area, composed the advantages of remote satellite sensing to study this endemic foci.

Lima et al 2002 [29] also in Brazil, studied the geographical distribution of notified human TL cases and correlated with the occurrence of the remaining vegetation and water streams, through satellite monitoring (LANDSAT level 4).

They observed that the geographical distribution of cases displayed a higher concentration in the northern and western regions of the studied area and a close relationship between TL and modified native forest areas, gallery forest areas or the remnants of both.


**Table 3.** Parameters utilized on Landsat 4-5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM+) methodologies (based on the data obtained from the website http://landsat.usgs.gov).


**Table 4.** Parameters utilized on Landsat 8 Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS) methodologies (based on the data obtained from the website http://landsat.usgs.gov).


**Table 5.** Parameters utilized on Landsat Multi Spectral Scanner (MSS) method (based on the data obtained from the website http://landsat.usgs.gov).

Peterson et al 2004 [30] investigates the potential of ecological niche modeling techniques for interpolating into unsampled areas in order to understand the geographic distributions of vector species. They used multiple subsamples from accessible distributional points to analyze the question of how much sampling is needed to assemble a suitable distributional interpre‐ tation for vector species.

**Landsat 8 Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS)**

Bathymetric mapping, characterizing soil from vegetation and deciduous from coniferous vegetation

functional for plant vigor assessing

vegetation; infiltrates thin clouds

Enriched wetness content of soil and vegetation and thin cloud infiltration

predictable soil wetness

100 meter resolution, enhanced thermal mapping and predictable soil wetness

Sediment-laden water, delimits areas of shallow

Infiltrates atmospheric cloud over best, highlights vegetation, boundary between land and water, and

0.7-0.8 Vegetation boundary between land and water, and natural features of landscape

natural features of landscape

1.57-1.65 Distinguishes wetness content of soil and

**Band Wavelength Attributes**

**Band 2 – blue** 0.45-0.51

136 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

**Band 11 – TIRS 2** 11.5-12.51

**Band 4 - green Band 1 - green** 0.5-0.6

**Band 6 - Short-wave Infrared**

**Band 7 - Short-wave Infrared**

**(SWIR) 1**

**(SWIR) 2**

**Landsat MSS 1, 2,3 Spectral Bands**

**Band 6 - Near Infrared**

**Band 7 - Near Infrared**

website http://landsat.usgs.gov).

**Band 1 – coastal aerosol** 0.43-0.45 coastal and aerosol analyzes

**Band 3 - green** 0.53-0.59 Highlights peak vegetation, which is

**Band 4 - red** 0.64-0.67 Distinguishes vegetation slopes **Band 5 - Near Infrared (NIR)** 085.-0.88 Highlights biomass and coastlines

2.11-2.29

**Table 4.** Parameters utilized on Landsat 8 Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS)

0.8-1.1

**Table 5.** Parameters utilized on Landsat Multi Spectral Scanner (MSS) method (based on the data obtained from the

**Landsat Multi Spectral Scanner (MSS)**

**Spectral Bands Wavelength Attributes**

water

methodologies (based on the data obtained from the website http://landsat.usgs.gov).

**Landsat MSS 4,5**

**Band 3 - Near Infrared**

**Band 4 - Near Infrared**

**Band 5 - red Band 2 - red** 0.6-0.7 Cultural features

**Band 8 - Panchromatic** .50-.68 15 meter resolution, intense image definition **Band 9 – Cirrus** 1.36 -1.38 Increased detection of cirrus cloud pollution **Band 10 – TIRS 1** 10.60 – 11.19 100 meter resolution, thermal mapping and

The Genetic algorithm for rule-set prediction (GARP) was utilized for modeling the ecological niches. The authors inferred that GARP associates ecological characteristics of known occur‐ rence points to those randomly sampled from the rest of the study region, pursuing the development of a series of decision rules that can best summarize those factors related with the presence of species.

They also demonstrated that moderate sampling densities at sample sizes that possibly could characterize many epidemiological studies, including the distributions of vector or reservoir were sufficient to produce excellent briefs of the geographic distributions of species permits development of geographic predictions for poorly known species to promote the knowledge about geographic aspects of disease systems.

Carneiro et al 2004 [31] in Brazil, used geo-technologies including satellite images, as normal‐ ized difference vegetation index (NDVI), in the collection and analysis of epidemiological data from an LV endemic area. It was observed that, the power of specific variable such as: demographic density, age, occurrence of sand flies, contaminated dogs, and human living in specific area, as well as the practical value of using NDVI values to identify risk areas.

Salomón et al 2006 [32] in Argentine, utilized the RS to study the spatial distribution of Phlebotominae associated with a focus of tegumentary leishmaniasis. Satellite images were used to estimate the influence of the maximal and minimal flow of a river present on the area of study, on the transmission of the disease. The probable correlation with the gallery forest was also rated.

The images were obtained from LANDSAT 5 TM and 7 ETM, they were georreferenced using satellite ephemeris and the nearest-neighbor method. The Band 5 was also used to discriminate areas covered by the river, and the neighboring the land uncovered of vegetation trough visual identification.

The authors concluded that the fishing spots were significantly overflowed during the transmission peak because the spatial restricted flood could concentrate vectors, reservoirs, and humans in high places.

They also suggested through both spatial distribution of vectors and remote sensing data the higher transmission risk in the area it is still related with the gallery forest, despite of the urban influence.

Margonari et al 2006 [5] in Brazil, applied the GIS methodology integrated with demographic, socio-economic and environmental data to study some aspects of the epidemiology of a visceral leishmaniasis focus.

It was observed that among biogeographic parameters such as: altitude, area of vegetation influence, hydrographic, and areas of poverty, only altitude showed to influence emergence of leishmaniasis because most canine and human cases of leishmaniasis cases were localized between 780 and 880 m above the sea level and at these same altitudes, a large number of phlebotomine sand flies were collected.

Nieto et al 2006 [33] also in Brazil, used models developed within a GIS employing Genetic Algorithm Rule-Set Prediction (GARP) and the growing degree day (GDD)-water budget (WB) concept to predict the distribution and potential risk of visceral leishmaniasis (VL).

It was described a high and moderate prevalence sites for VL were significantly related to areas of high and moderate risk prediction. Indeed the area expected by the GARP model, hinged on the number of pixels that overlapped among eleven annual model years and the quantity of potential generations per year that could be completed by the *Lu. longipalpis-L. chagasi* system by GDD-WB analysis.

In both the GARP and the GDD-WB prediction models suggested that the highest VL risk was characterized by a semi-arid and hot climate (Caatinga), but the risk in the interior forest and the Cerrado was lower and the coastal forest was predicted as a low-risk area due to the unsuitable conditions for the vector and VL transmission.

Neto et al 2009 [34] in Brazil, applied GIS and SR to examine factors associated with the incidence of urban VL. They observed that the annual incidence rates were related to socioe‐ conomic and demographic indicators as well as the vegetation index.

The highest incidence occurred in the peripheral areas of the city and areas with high popu‐ lation growth and abundant vegetation. On the other hand the percentage of households with piped water was inversely associated with the disease incidence.

The authors conclude that spatial distribution of the disease in the area was heterogeneous, and the incidence was associated with the peripheral neighborhoods fullest vegetation cover, considered subject to anthropic action.

Shimabukuro et al 2010 [35] in Brazil, utilized GIS and SR to study the geographical distribution of American cutaneous leishmaniasis and its phlebotomine vectors and generate risk maps. They observed that generally, the sand fly vector species evaluated have presented unique and heterogeneous distributions, although often overlapped. Numerous sand fly species were highly localized, while the others were much more largely spread.

The authors emphasized the complexity and geographical pattern of ACL transmission in the region.

Valderrama-Ardila et al 2010 [36] in Colombia, evaluate through spatial analysis, the envi‐ ronmental risk factors for CL. The applicant predictor variables were land use, elevation, and climatic (mean temperature and precipitation).

They observed that incidence of the disease was higher in townships with mean temperatures in the middle of the county's range. The frequency was independently associated with forest or shrubs and lower population density. The coverage of forest or shrub have not presented main changes over time.

The results confirmed the effect of weather and land use in leishmaniasis transmission.

Silva et al 2011 [14] in Brazil, studied a dog population from an endemic focus of LV. Through GIS and SR and applying kernel density estimator with Gaussian function and smooth kernel of 100 m radius, they observed local variations related to infection the incidence and distribu‐ tion of serological titers, i.e. high titers were noted close to areas with preserved vegetation, while low titers were more frequent in areas where people kept chickens.

of leishmaniasis because most canine and human cases of leishmaniasis cases were localized between 780 and 880 m above the sea level and at these same altitudes, a large number of

Nieto et al 2006 [33] also in Brazil, used models developed within a GIS employing Genetic Algorithm Rule-Set Prediction (GARP) and the growing degree day (GDD)-water budget (WB)

It was described a high and moderate prevalence sites for VL were significantly related to areas of high and moderate risk prediction. Indeed the area expected by the GARP model, hinged on the number of pixels that overlapped among eleven annual model years and the quantity of potential generations per year that could be completed by the *Lu. longipalpis-L. chagasi* system

In both the GARP and the GDD-WB prediction models suggested that the highest VL risk was characterized by a semi-arid and hot climate (Caatinga), but the risk in the interior forest and the Cerrado was lower and the coastal forest was predicted as a low-risk area due to the

Neto et al 2009 [34] in Brazil, applied GIS and SR to examine factors associated with the incidence of urban VL. They observed that the annual incidence rates were related to socioe‐

The highest incidence occurred in the peripheral areas of the city and areas with high popu‐ lation growth and abundant vegetation. On the other hand the percentage of households with

The authors conclude that spatial distribution of the disease in the area was heterogeneous, and the incidence was associated with the peripheral neighborhoods fullest vegetation cover,

Shimabukuro et al 2010 [35] in Brazil, utilized GIS and SR to study the geographical distribution of American cutaneous leishmaniasis and its phlebotomine vectors and generate risk maps. They observed that generally, the sand fly vector species evaluated have presented unique and heterogeneous distributions, although often overlapped. Numerous sand fly species were

The authors emphasized the complexity and geographical pattern of ACL transmission in the

Valderrama-Ardila et al 2010 [36] in Colombia, evaluate through spatial analysis, the envi‐ ronmental risk factors for CL. The applicant predictor variables were land use, elevation, and

They observed that incidence of the disease was higher in townships with mean temperatures in the middle of the county's range. The frequency was independently associated with forest or shrubs and lower population density. The coverage of forest or shrub have not presented

The results confirmed the effect of weather and land use in leishmaniasis transmission.

concept to predict the distribution and potential risk of visceral leishmaniasis (VL).

phlebotomine sand flies were collected.

138 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

considered subject to anthropic action.

climatic (mean temperature and precipitation).

region.

main changes over time.

unsuitable conditions for the vector and VL transmission.

conomic and demographic indicators as well as the vegetation index.

piped water was inversely associated with the disease incidence.

highly localized, while the others were much more largely spread.

by GDD-WB analysis.

The authors conclude that the environment plays an important role in generating relatively protected areas within larger endemic regions, but it could also contribute to the creation of hotspots with clusters of comparatively high serological titers indicating a high level of transmission compared with neighboring areas.

Cluster analysis of the serological titers in dogs in the study area showed a non-random distribution, demonstrating that the patterns of transmission of canine VL can undergo local alterations, producing hotspots where the risk of infection was very high compared to neighboring areas.

It was suggested the possibility to predict the specific places of high-risk VL transmission within an endemic area through the mapping of canine serological titers.

Almeida et al 2011 [37] in Brazil, used spatial analysis to identify regions at highest risk of VL in an urban area. They showed from kernel ratios results, that peripheral census tracts were the most heavily affected. The spatial analysis showed that local clusters of high incidence of VL could change their locations depending on the time, suggesting that the pattern of VL is not static, and the disease may occasionally spread to other areas.

The authors also observed a spatial correlation between VL rates and all socioeconomic and demographic indicators evaluated, such as: 1) illiteracy rate; 2) children less than five years of age as a percentage of the total population; 3) mean income of heads-of-households; 4) percentage of permanent private households connected to the water supply; 5) percentage of households with regular garbage collection; and 6) percentage of permanent private house‐ holds connected to the sewage system.

Foley et al 2012 [38], created a very useful tool that comprises a new map service within VectorMap (www.vectormap.org). Using the words of the authors, "It allows free public online access to global sand fly, tick and mosquito collection records and habitat suitability models, given the short home range of sand flies, combining remote sensing and collection point data, offering a powerful insight into the environmental determinants of sand fly distribution.

Sand fly Map uses Microsoft Silverlight and ESRI's ArcGIS Server 10 software platform to present disease vector data and relevant remote sensing layers in an online geographical information system format. Users can view the locations of past vector collections and the results of models that predict the geographic extent of individual species. Collection records are searchable and downloadable, and Excel collection forms with drop down lists, and Excel charts to country, are available for data contributors to map and quality control their data.

Sand fly Map makes accessible, and adds value to, the results of past sand fly collecting efforts. It is detailed the workflow for entering occurrence data from the literature to Sand fly Map, using an example for sand flies from South America.

The proper use of a global positioning system (GPS) device and a detailed text description of the locality are encouraged to minimize this uncertainty [39]. The calculation of spatial uncertainty, for example for Martins et al [40], allows data to be matched to appropriate resolution remote sensing data, for modeling or other spatial analyses".

Saraiva et al 2012 [41] in Brazil, utilizing GIS methodology associated with serological tests, studied a VL focus. They described the occurrence of serologically positive dogs was spread out throughout all geographical area. The places of concentration of serologically positive dogs appeared both in risk areas and outside them.

Overlaying the map of the human and canines cases with factors traditionally related to VL as vegetation, hydrography, and areas of poverty, it was not possible identify a spatial correlation between them, which demonstrates that in urban areas there are still unknown parameters.

Souza et al 2012 [42] in Brazil, carried out a space-time analysis of AVL cases in humans. They conclude by the time series analysis, a positive tendency over the period analyzed, completing that the disease was clustered in the Southwest side of area of study, suggesting it could require special attention with regard to control and prevention measures.

Finally, González et al 2013 [43] in Colombia, have surveyed the spatial distribution of two vector species of *L. infantum*, after predicting its future dispersal into climate change situations to establish the potential dissemination of the disease. They used ecological niche models through the Maxent software and 13 Worldclim bioclimatic coverages. Through predictions for the pessimistic CSIRO A2 scenario, was calculated the higher increase in temperature in function of non-emission reduction, and by the optimistic Hadley B2 Scenario, was predicted the minimum increase in temperature.

Concerning the climate change projections, they observed an overall reduction in the spatial distribution of the two vector species, progressing a shift in the vertical distribution for one species and restricting the other to certain regions at the sea level.

The authors predicted an outcome for VL vectors in Colombia and suggested that Changes in spatial distribution patterns could be affecting local abundances due to climatic pressures on vector populations thus reducing the incidence of human cases.
