Loading...
Publication
Modelo espacialmente explícito de fatores determinantes à transmissão da Dengue e outras doenças arbovirais em Portugal continental
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.89 MB | Adobe PDF |
Authors
Abstract(s)
O vírus da Dengue é uma das doenças arbovirais de maior expansão global, com
potencial de emergência também em regiões temperadas. Em Portugal continental, a
deteção recente do mosquito Aedes albopictus evidencia a necessidade de estudar os
fatores que influenciam a sua distribuição e o risco associado. Esta investigação teve
como objetivo modelar a adequabilidade ambiental do vetor e avaliar a vulnerabilidade
social da população, integrando ambos os fatores na criação de mapas de risco.
Foram testados três algoritmos de modelação: MaxEnt, SVM e ANN. O modelo
ANN foi excluído devido ao fraco desempenho (AUC = 0,5). Os algoritmos MaxEnt e SVM
revelaram padrões distintos de adequabilidade: MaxEnt identificou áreas mais amplas e
generalistas, enquanto o SVM apresentou previsões mais precisas e localizadas, com
AUC externa de 0,83. As regiões costeiras do sul e centro do país destacaram-se como
as mais propensas à presença do vetor.
No que diz respeito à vulnerabilidade social, o método da média permitiu uma
representação mais realista do território, evidenciando fragilidades em zonas rurais e
periferias urbanas. Por sua vez, os métodos fuzzy (produto e Gamma) suavizaram
excessivamente os dados, dificultando a identificação de áreas críticas.
A sobreposição entre adequabilidade ambiental e vulnerabilidade social permitiu
gerar mapas de risco. O método de combinação pela média destacou-se por representar
de forma mais detalhada os padrões espaciais de risco, ao contrário dos métodos fuzzy
que atenuaram as variações.
Este estudo contribui para uma compreensão mais integrada e espacialmente
explícita do risco de transmissão da Dengue em Portugal continental, servindo de
suporte à vigilância epidemiológica e ao planeamento de estratégias preventivas.
Futuras investigações deverão explorar variáveis adicionais e dados de maior resolução
para melhorar a precisão dos modelos.
Dengue virus is one of the most rapidly expanding arboviral diseases worldwide, with potential emergence even in temperate regions. In mainland Portugal, the recent detection of Aedes albopictus highlights the need to investigate the factors influencing its distribution and associated risk. This study aimed to model the environmental suitability of the vector and assess the social vulnerability of the population, integrating both dimensions to create comprehensive risk maps. Three modelling algorithms were tested: MaxEnt, SVM, and ANN. The ANN model was excluded due to poor performance (AUC = 0.5). MaxEnt and SVM produced distinct suitability patterns: MaxEnt identified broader and more generalist areas, whereas SVM offered more precise and localized predictions, with an external AUC of 0.83. Coastal regions in the south and central parts of the country were identified as the most suitable for vector presence. Regarding social vulnerability, the mean-based method provided a more realistic territorial representation, highlighting weaknesses in rural areas and urban peripheries. In contrast, the fuzzy methods (product and Gamma) excessively smoothed the data, limiting the identification of critical areas. The overlay of environmental suitability and social vulnerability enabled the production of risk maps. The mean combination method proved more effective in capturing spatial patterns of risk in detail, unlike the fuzzy methods which attenuated spatial variability. This study contributes to a more integrated and spatially explicit understanding of dengue transmission risk in mainland Portugal, offering valuable insights for epidemiological surveillance and public health planning. Future research should incorporate additional variables and higher-resolution data to enhance model accuracy.
Dengue virus is one of the most rapidly expanding arboviral diseases worldwide, with potential emergence even in temperate regions. In mainland Portugal, the recent detection of Aedes albopictus highlights the need to investigate the factors influencing its distribution and associated risk. This study aimed to model the environmental suitability of the vector and assess the social vulnerability of the population, integrating both dimensions to create comprehensive risk maps. Three modelling algorithms were tested: MaxEnt, SVM, and ANN. The ANN model was excluded due to poor performance (AUC = 0.5). MaxEnt and SVM produced distinct suitability patterns: MaxEnt identified broader and more generalist areas, whereas SVM offered more precise and localized predictions, with an external AUC of 0.83. Coastal regions in the south and central parts of the country were identified as the most suitable for vector presence. Regarding social vulnerability, the mean-based method provided a more realistic territorial representation, highlighting weaknesses in rural areas and urban peripheries. In contrast, the fuzzy methods (product and Gamma) excessively smoothed the data, limiting the identification of critical areas. The overlay of environmental suitability and social vulnerability enabled the production of risk maps. The mean combination method proved more effective in capturing spatial patterns of risk in detail, unlike the fuzzy methods which attenuated spatial variability. This study contributes to a more integrated and spatially explicit understanding of dengue transmission risk in mainland Portugal, offering valuable insights for epidemiological surveillance and public health planning. Future research should incorporate additional variables and higher-resolution data to enhance model accuracy.
Description
Keywords
Dengue Aedes albopictus Vulnerabilidade social Portugal continental