de Oliveira et al.

Malaria Journal 2011, 10:177

http://www.malariajournal.com/content/10/1/177

RESEARCH Open Access

Spatial patterns of malaria in a land reform

colonization project, Juruena municipality,
Mato Grosso, Brazil
Elaine Cristina de Oliveira1, Emerson Soares dos Santos2, Peter Zeilhofer3, Reinaldo Souza-Santos4 and
Marina Atanaka-Santos5*

Abstract
Background: In Brazil, 99% of malaria cases are concentrated in the Amazon, and malaria’s spatial distribution is
commonly associated with socio-environmental conditions on a fine landscape scale. In this study, the spatial
patterns of malaria and its determinants in a rural settlement of the Brazilian agricultural reform programme called
“Vale do Amanhecer” in the northern Mato Grosso state were analysed.
Methods: In a fine-scaled, exploratory ecological study, geocoded notification forms corresponding to malaria
cases from 2005 were compared with spectral indices, such as the Normalized Difference Vegetation Index (NDVI)
and the third component of the Tasseled Cap Transformation (TC_3) and thematic layers, derived from the visual
interpretation of multispectral TM-Landsat 5 imagery and the application of GIS distance operators.
Results: Of a total of 336 malaria cases, 102 (30.36%) were caused by Plasmodium falciparum and 174 (51.79%) by
Plasmodium vivax. Of all the cases, 37.6% (133 cases) were from residents of a unique road. In total, 276 cases were
reported for the southern part of the settlement, where the population density is higher, with notification rates
higher than 10 cases per household. The local landscape mostly consists of open areas (38.79 km²). Training forest
occupied 27.34 km² and midsize vegetation 7.01 km². Most domiciles with more than five notified malaria cases
were located near areas with high NDVI values. Most domiciles (41.78%) and malaria cases (44.94%) were
concentrated in areas with intermediate values of the TC_3, a spectral index representing surface and vegetation
humidity.
Conclusions: Environmental factors and their alteration are associated with the occurrence and spatial distribution
of malaria cases in rural settlements.

Background land use and dramatic environmental transformations in

In Brazil, and more specifically in the state of Mato the Amazon landscapes [2].
Grosso, the population in rural land reform settlements The informal gold extraction techniques that predomi-
and occupations increased dramatically in the 1980s and nate in these mining areas favor the creation of environ-
1990s. Additionally, thousands of people migrated to the ments appropriate for malaria vector reproduction, with
northern regions of the state during this period, numerous bodies of standing water and small areas with
attracted by the discovery of gold there. Unregulated high population densities [3].
mining became a principal economic activity, increasing As a consequence of the complex network of factors
the occupation of the region [1] and leading to intense involved in the transmission of malaria, one technique
for the analysis of the disease has utilized Geographical
Information Systems (GIS). GIS allow information from
different sources to be stored and integrated, enabling
* Correspondence: [email protected] an analysis of the interactions between relevant variables
5
Institute of Public Health, Federal University of Mato Grosso, Av. Fernando
Corrêa, Cuiabá, Mato Grosso State, 78.060-900, Brazil in their spatio-temporal context, the development and
Full list of author information is available at the end of the article

© 2011 de Oliveira et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
de Oliveira et al. Malaria Journal 2011, 10:177 Page 2 of 9
http://www.malariajournal.com/content/10/1/177

testing of hypotheses, supervision and evaluation of through the visual interpretation of a #3#4#5 color com-
interventions and the construction of predictive models posite, where the following three thematic classes were
as part of preventive operations [4-6]. differentiated: (a) Forests, (b) Agricultural area, and (c)
As the spatial patterns of malaria in a territory reflect Secondary vegetation.
complex interactions between parasites, vectors, and Two spectral indices were derived from the Landsat
human hosts, studies of the influences of environmental TM imagery: the Normalized Difference Vegetation
factors on malaria allow the simulation of epidemiologi- Index (NDVI), a widely used measure of the amount of
cal risk situations as a tool for health care decision-mak- green biomass, and the third component of the Tasseled
ing and the definition of priority actions for malaria Cap Transformation [12], an index representing the
control [7]. Many large-scale studies exist, as shown by humidity of the first reflectance layer (vegetation cover
EISEN and WRIGHT [8], but studies at different spatial or soil surface). The NDVI was calculated in the
scales are required to understand fully the relationships SPRING 4.3 software by the combination of the red (#3)
between landscape features, parasite distribution, and and near infrared (#4) bands ([#3-#4]/[#3+#4]) scaled
malaria infection. As the risk of malaria infection is into 256 grey levels. Then, three classes were defined:
related to multiple factors, the interactions between low (0 to 100), intermediate (101 to 200) and high (>
which have rarely been explored at local scales in the 200). The Tasseled Cap Transformation (hereafter
southern Amazon, the present study identifies and ana- TC_3) was performed by the respective module in ENVI
lyses the influences of multiple socio-environmental 4.1 and then differentiated into the following classes:
characteristics on the spatial distribution of malaria low (0 to 76.5), intermediate (76.6 to 165.8) and high (>
cases in the “Vale do Amanhecer” settlement of the Bra- 165.8).
zilian land reform programme in northern Mato Grosso. The following distance layers were created using the
buffer command of the TerraView 3.1.4 software: One
Methods for the distance to potential procreation habitats, with
Study area three 100-m zones (100, 200 and 300 m); one for the
The “Vale do Amanhecer” settlement is located in distance to streams, with 100-m zones between 100 m
northern Mato Grosso, about 6.2 km from the city of and 1,500 m; and one for the distance to mining areas,
Juruena, at 10°22’53.84” S latitude and 58°25’27.35” W in 300-m zones from 300 m to 6,900 m. These values
longitude. Founded by the Federal Institute of Coloniza- were selected considering knowledge of the feeding and
tion and Land Reform [9] in 1998, the settlement com- reproduction habits of Anopheles species, which can fly
prises 14,400 ha. One half of the area was designated for further than 1.5 km but generally remain near their
settlement by 250 families, whereas the other was set reproductive habitat [13].
aside as a permanent protection area. Each land unit has This study is part of the “Malaria spatial analysis in
an area of about 26 ha, and units are distributed along rural land reform settlements project” and was approved
projected, linear roads numbered from 01 to 14 by the Julio Muller Hospital Ethics Committee.
(Figure 1). The average demographic density is 10.41
inhabitants/km2 . The main economic activities of the Results
settlement are cattle farming and agricultural subsis- In 2005, the settlement was occupied by 718 inhabitants,
tence and, mainly along the streams, some informal gold 394 (54.87%) of them male and 324 (45.13%) female. A
mining. total of 359 malaria cases were reported, 23 of which
were excluded from the sample because no residence
Data acquisition and analysis address was given and 18 because the patient’s residence
Epidemiological data was located outside of the settlement.
A total of 585 notification forms from the Epidemiologic From the remaining total of 336 cases, 133 positive
Surveillance Information System [10] from the year 2005 smears were from residents of Road 08, which corre-
were obtained from the municipal health care adminis- sponds to 37.60% of all reported cases. For Roads 13
tration. All registries were cross-checked and georefer- and 05, 124 (35.10%) and 58 (16.40%) malaria cases
enced in a TerraView 3.1.4 [11] spatial database. were reported, respectively (Figure 2).
Spatial data layers Cases were concentrated in the southern part of the
Spatial data sets were elaborated using TerraView 3.1.4 settlement, where 276 cases were reported, often with
(INPE), ArcGIS 9.2 (ESRI), SPRING 4.3 (INPE) and more than 10 cases per domicile. In this area, 102 cases
ENVI 4.1 (RSI) software. A multispectral TM-Landsat 5 (30.36%) were caused by Plasmodium falciparum and
image from 25-06-2005 (WRS 229/67) was obtained 174 (51.79%) by Plasmodium vivax. The northern part
from the Instituto Nacional de Pesquisas Espaciais of the settlement presented 60 cases, and most domi-
(INPE) and georeferenced. A land-use map was created ciles reported fewer than 10 cases per domicile. Of these
de Oliveira et al. Malaria Journal 2011, 10:177 Page 3 of 9
http://www.malariajournal.com/content/10/1/177

Figure 1 Study area.

cases, 16 (4.76%) were caused by P. falciparum and 44 malaria cases in terms of absolute numbers, the highest
(13.09%) by P. vivax. infection rates occurred in areas with denser vegetation.
The visual interpretation of the Landsat TM imagery Domiciles located within or near the forested permanent
showed that 38.79 km² of the settlement’s land was protection areas with high NDVI values (> 200) reported
deforested, including terrains used for housing, pastures 2.8 cases per domicile; the domiciles with more than
and agriculture. Forest reminiscents occupied 27.34 km² five reported cases are located in areas of high NDVI
and secondary vegetation 7.01 km² (Figure 3). values as well, mainly near the stream network (Figure
The concentration of malaria cases in the southern 4).
part of the settlement coincides with the predominance In absolute terms, the highest numbers of domiciles
of deforested areas there and its proximity to most (41.78%) and malaria cases (44.94%) were located in
mining areas. As expected, 154 (75.86%) of the 203 areas with intermediate values of TC3. Humid surfaces
domiciles (200 in the settlement proper and 3 in the with high values of TC_3 occurred mainly in densely
permanent protection area) were located in open areas, vegetated areas and near the stream network. The aver-
which are generally characterized by low NDVI values age number of malaria cases per domicile in these areas
(from 0 to 100). This result documents the intense land (1.85) is slightly higher than that in areas with inter-
transformation process initiated with the occupation in mediate TC_3 values (1.82). The number of cases per
2000, which has included the transformation of natural domiciles is much lower (1.30 cases per domicile) in
vegetation into pastures, the opening of mines and the areas with low TC_3 values (Figure 5).
installation of mining equipment since 2002. Residents Domiciles reporting malaria cases were observed
in this area presented 239 (71.13%) malaria cases, with across the whole range of distances from the stream
an average of 1.56 per domicile (Table 1). network. Most malaria cases, however, occurred in dom-
Whereas this area of the settlement, where the popu- iciles up to 900 m from streams. Domiciles with more
lation density is highest, reported the majority of the than 15 cases were concentrated even closer to the
de Oliveira et al. Malaria Journal 2011, 10:177 Page 4 of 9
http://www.malariajournal.com/content/10/1/177

Figure 2 Distribution of malaria cases.

hydrographic network, at a maximum distance of 300 nature of malaria occurrence in the Amazon region,
metres (Figure 6). where malaria is often concentrated in settlements and
All domiciles with malaria cases were located less than areas of mining activity. The average prevalence in the
1.6 km from the nearest potential reproductive habitat Juruena municipality is about five times lower than in
of Anopheles species, including watercourses, standing Vale do Amanhecer.
water, springs, wells and small dam reservoirs. Cases The comparison of the spatial pattern of malaria
decreased on a gradient with increasing distance from occurrence with geographic layers representing vegeta-
these locations (Figure 7). tion distribution and distances to watercourses and
Cases were also negatively related to the distance from mining areas showed variable findings in this fine-scale
mining areas. The 70 domiciles (34.48%) located less epidemiological study.
than 1.2 km from a mining area accounted for 237 Numerous studies have shown that malaria infection
malaria cases (70.5%), with an average of 3.38 cases per is influenced by environmental factors, such as tempera-
domicile. All domiciles with more than five cases were ture, precipitation, humidity and altitude. In tropical
located at most 2.1 km from the nearest mining area, regions, such as the Amazon region, year-round high
whereas all domiciles with fewer cases (between one and temperatures and precipitation favor malaria transmis-
five) were at least 1.2 km from the nearest mining area sion [14-16,2]. Most studies have been large-scale sur-
(Figure 8). veys on a regional or country scale [14,17,7]. As
variations at these scales can be governed by other
Discussion external factors, local-scale studies, such as those carried
In the Vale do Amanhecer settlement, malaria was out within a single settlement [18], are crucial for
highly prevalent in 2005, with about 47% of the settle- understanding the driving factors of malaria transmis-
ment’s 718 inhabitants affected by the disease. There- sion and for community-based health planning [19]. At
fore, the study area provides a clear example of the focal this scale, transmission rates are not only governed by
de Oliveira et al. Malaria Journal 2011, 10:177 Page 5 of 9
http://www.malariajournal.com/content/10/1/177

Figure 3 Land use and malaria cases.

demographic or ecological factors but are also

influenced by the social and cultural contexts of local
Table 1 Absolute (AF) and relative frequencies (RF),
populations [20].
average (Av) and standard deviation (SD) of Land-use,
In the studied settlement, relationships between the
NDVI and TC_3
environment and spatial patterns of malaria occurrence
Variables Classes AF RF (%) AF1 RF1 Av SD N.
D2 were observed. For risk evaluation, the combination of
Land-use Agricultural 239 73.31 239 73.31 1.49 2.84 160
land use and vegetation cover must be considered. More
area cases are reported in deforested areas with low NDVI
Secondary 17 5.21 256 78.53 2.13 4.85 8 and TC_3 values, but this is exclusively due to the
vegetation higher local population density and population fluxes.
Forest 70 21.47 326 100.00 1.71 2.74 41 The highest rates of malaria per domicile were reported
Total 326 100.00 - - 1.56 2.90 209 within or near forest reminiscent with high spectral
NDVI Low 209 64.11 209 64.11 1.49 2.78 140 indices. Additionally, at a local scale, it must be consid-
Intermediate 64 19.63 273 83.74 1.28 2.96 50 ered that infection did not necessarily take place in the
High 53 16.26 326 100.00 2.79 3.44 19 domicile.
Total 326 100.00 - - 1.56 2.90 209 Vegetation cover plays an important role in the biolo-
TC_3 Low 78 23.93 78 23.93 1.24 2.11 63 gical cycles of vectors and infectious agents, particularly
Intermediate 151 46.32 229 70.25 1.70 3.44 89 if other environmental conditions, such as precipitation,
High 97 29.75 326 100.00 1.70 2.75 57 temperature or humidity, are altered [21,22]. This altera-
Total 326 100.00 - - 1.56 2.90 209 tion explains the elevated incidence of malaria cases in
1
- Accumulated domiciles inside or near forested areas with high values
2
- Number of domiciles of NDVI, which may be characterized by higher vector
de Oliveira et al. Malaria Journal 2011, 10:177 Page 6 of 9
http://www.malariajournal.com/content/10/1/177

Figure 4 Vegetation index (NDVI) and malaria cases.

Figure 5 Vegetation and soil humidity index (TC_3) and malaria case distribution.
de Oliveira et al. Malaria Journal 2011, 10:177 Page 7 of 9
http://www.malariajournal.com/content/10/1/177

Figure 6 Distance to nearest stream (buffers) and malaria cases.

densities. The low absolute numbers of cases in these Studies such as [7,21,25,26] have pointed out that the
areas are caused by their low housing density (18.2% of distribution and dynamics of malaria cases are related to
all domiciles) and low population flux. the phase of land occupation. Deforestation, aggregated
High TC_3 values were observed inside forested areas with an influx of migrants, frequently produces new
and areas with dense vegetation succession. The highest reproduction and principal feeding habitats, favoring the
average case rate per domicile (1.82) was found in areas occurrence of epidemics in communities with popula-
with high TC_3 values, possibly due to the presence of tions characterized by high mobility, low immunity and
humid areas favoring vector proliferation. These areas low risk perception [24,3].
mainly coincide with buffer zones near streams, explain- Infection rates are also thought to be elevated due to
ing the concentration of cases in domiciles less than 1 the lack of an adequate sanitation infrastructure, which
km from the hydrographic network, which is dense motivates inhabitants to construct their housing near
throughout the settlement. streams to facilitate domestic tasks and personal
High malaria infection rates in the settlement are sup- hygiene. Vasconcelos et al [7] described a similar spatial
posedly related to the changes in land use carried out pattern of domicile distribution in the Jacundá munici-
by a population that originated partially from non-ende- pality, attributing elevated malaria infection rates to this
mic regions. FERREIRA [23] identified a malaria preva- lack of sanitation infrastructure.
lence of 56.0% in individuals who migrated to the The observed malaria infection rates can be explained
settlement from non-endemic regions, who have an by the great number of potential reproduction habitats
infection probability 2.9 times higher than individuals and mining areas in the settlement; cases are highly
from endemic regions, a difference supposedly caused concentrated near these areas, which favor the survival
by their low immunity and lack of knowledge about and circulation of the main vector, Anopheles darlingi
measures to protect against the disease [16,24]. [7]. Close relationships between mining activities and
de Oliveira et al. Malaria Journal 2011, 10:177 Page 8 of 9
http://www.malariajournal.com/content/10/1/177

Figure 7 Distance to nearest potential reproductive habitat (buffers) and malaria cases.

Figure 8 Distance to nearest mining area and malaria cases.

de Oliveira et al. Malaria Journal 2011, 10:177 Page 9 of 9
http://www.malariajournal.com/content/10/1/177

malaria cases in the Amazon region have also been distribuição da malária na Amazônia brasileira. Cadernos de Saúde Pública
2006, 22:517-526.
reported by Santos et al. [27] and Duarte & Fontes [28]. 8. Eisen RJ, Wright NM: Landscape features associated with infection by a
malaria parasite (Plasmodium mexicanum) and the importance of
Conclusions multiple scale studies. Parasitology 2001, 122:507-513.
9. INCRA (Instituto Nacional de Colonização e Reforma Agrária) Dados do
Configuration of land use with a mosaic of reminiscent Assentamento de Vale do Amanhecer - Juruena-MT. Superintendência
forests and occupied deforested areas, a dense stream Regional de Mato Grosso, Cuiabá; 2006.
network, numerous breeding habitats and mining activ- 10. Brasil. Ministério da Saúde. Secretaria de Vigilância em Saúde: Sistema de
Informação de Vigilância Epidemiológica da Malária.[http://www.saude.
ity, paired with a mobile population with low immunity gov.br/sivep_malaria], Disponível em.
and low-risk perception, can explain the high malaria 11. INPE. Instituto Nacional de Pesquisas Espaciais. São Jose dos Campos;
incidence in the studied settlement. Malaria, which is [http://www.dgi.inpe.br/CDSR/], Disponível em.
12. Crist EP, Kauth RJ: The Tasseled Cap demystified. Photogrammetric Eng
known to be a focal disease, presents a heterogeneous Remote Sens 1986, 521:81-86.
spatial distribution of cases, even in the current fine- 13. Borror DJ, Delong DM: Introdução ao Estudo dos Insetos. São Paulo;
scaled local study. Findings of this study demonstrate Edgard Blücher; 1988.
14. Cox JM, Craig DS: Mapping malaria risk in the Highlands of Africa.
increased exposure to malaria in areas near potential Mapping Malaria Risk in Africa/Highland Malaria Project (MARA/HIMAL)
mosquito vector breeding sites, namely near streams Technical Report MARA/Durban, London School of Hygiene and Tropical
and mining areas and near more densely-vegetated Medicine, London; 1999 [http://www.mara.org.za], Acesso em Outubro/
2008. Disponível em.
areas. 15. Sawyer DR, Sawyer DO: Malaria on the Amazon frontier: economic and
social aspects of transmission and control. Belo Horizonte, Cedeplar; 1987.
16. Sawyer DR, Sawyer DO: The malaria transition and the role of social
Author details science research. Westport: Auburn House; 1992:2:105-122.
1
Epidemiological Surveillance, Health Secretary of Mato Grosso. Rua D, 17. Kazembe LN, Kleinschmidt I, Holtz TH, Sharp BL: Spatial analysis and
Political Administrative Center, Cuiabá, Mato Grosso State, 78.050-970, Brazil. mapping of malaria risk in Malawi using point-referenced prevalence of
2
Department of Geography, School of Philosophy, Literature and Human infection data. Int J Health Geographics 2006, 5:41-49.
Sciences, Av. Prof. Lineu Prestes, 338. Cidade Universitária. University of São 18. Jannuzzi PM: Projeções populacionais para pequenas áreas: método e
Paulo, São Paulo, São Paulo State, 05.508-080. Brazil. 3Department of aplicações. Rio de Janeiro: Escola Nacional de Ciências Estatísticas; 2006.
Geography, Federal University of Mato Grosso, Av. F. Corrêa, Cuiabá, Mato 19. Yazoumé Yé, Hoshen M, Kyobutungi C, Louis VR, Sauerborn R: Local scale
Grosso State, 78.060-900, Brazil. 4Department of Endemic Disease, Brazilian prediction of Plasmodium falciparum malaria transmission in an endemic
National School of Public Health, Oswaldo Cruz Foundation. Rua Leopoldo region using temperature and rainfall. Global Health Action 2009, 2:10.
Bulhões, 1480, Rio de Janeiro, Rio de Janeiro State, 21.041-210, Brazil. 20. Mota EGF: Fatores determinantes da situação da malária na Amazônia.
5
Institute of Public Health, Federal University of Mato Grosso, Av. Fernando Revista da Sociedade Brasileira de Medicina Tropical 1992, 25(suppl 2):13-22.
Corrêa, Cuiabá, Mato Grosso State, 78.060-900, Brazil. 21. Correia VRM, Monteiro AMV, Carvalho MS, Werneck GL: Uma aplicação do
sensoriamento remoto para a investigação de endemias urbanas.
Authors’ contributions Caderno de Saúde Pública 2007, 23:1015-1028.
ECO and ESS contributed to the study design, data acquisition, processing 22. Forattini OP: Culicidologia Médica. São Paulo: Edusp; 20022.
and analysis and manuscript preparation. RSS, MAS, PZ, and ESS participated 23. Ferreira IM: Fatores associados à ocorrência da malária em área de
in the study conception, scientific coordination and revision of the assentamento, município de Juruena/MT [dissertação de mestrado].
manuscript. All authors read and approved the final manuscript. Cuiabá: Instituto de Saúde Coletiva da Universidade Federal de Mato
Grosso; 2007.
Competing interests 24. Confalonieri UEC: Saúde na Amazônia: um modelo conceitual para a
The authors declare that they have no competing interests. análise de paisagens e doenças. Estudos Avançados 2005, 19:221-236.
25. Rodrigues AF, Escobar AL, Souza-Santos R: Análise espacial e
Received: 1 December 2010 Accepted: 26 June 2011 determinação de áreas para o controle da malária no Estado de
Published: 26 June 2011 Rondônia. Revista da Sociedade Brasileira de Medicina Tropical 2008,
41:55-64.
References 26. Souza-Santos R: Distribuição Sazonal de vetores da malária em
1. Barbieri AF, Sawyer DO: Heterogeneity of malaria prevalence in alluvial Machadinho d’Oeste, Rondônia, Região Amazônica, Brasil. Cadernos de
gold mining areas in Northern Mato Grosso State, Brazil. Cadernos de Saúde Pública 2002, 18:1813-1818.
Saúde Pública 2007, 23:2878-2886. 27. Santos VR, Yokoo EM, Souza-Santos R, Atanaka-Santos M: Fatores
2. Castro MC, Singer BH: Meio ambiente e saúde: metodologia para análise socioambientais associados à distribuição espacial de malária no
espacial da ocorrência de malária em projetos de assentamento. Revista assentamento Vale do Amanhecer, Município de Juruena, Estado de
Brasileira de Estudos Populacionais 2007, 24:247-262. Mato Grosso, 2005. Revista da Sociedade Brasileira de Medicina Tropical
3. Barbieri AF, Sawyer DO: Uso do solo e prevalência da malária em uma 2009, 42:47-53.
região da Amazônia Brasileira. Caderno de Geografia 2005, 15:9-30. 28. Duarte EC, Fontes CJF: Associação entre a produção anual de ouro em
4. Fotsing JM, Gurgel HC, Zaninetti JM: Estudo de casos de malária em garimpo e incidência de malária em Mato Grosso-Brasil, 1995-1996.
Roraima a partir de técnicas de análise espacial. Anais XIII Simpósio Revista Sociedade Brasileira de Medicina Tropical 2002, 35:665-668.
Brasileiro de Sensoriamento Remoto - INPE, Florianópolis; Brasil; 2007,
2705-2712. doi:10.1186/1475-2875-10-177
5. Camara G, Monteiro AMV: Geocomputation techniques for spatial Cite this article as: de Oliveira et al.: Spatial patterns of malaria in a land
reform colonization project, Juruena municipality, Mato Grosso, Brazil.
analysis: are they relevant to health data? Cadernos de Saúde Pública
Malaria Journal 2011 10:177.
2001, 17:1059-1071.
6. Graham AJ, Atkinson PM, Danson FM: Spatial analysis for epidemiology.
Acta Trop 2004, 91:219-225.
7. Vasconcelos CH, Novo EMLM, Donalisio MR: Uso do sensoriamento
remoto para estudar a influência de alterações ambientais na