Remote sensing for risk mapping of Aedes aegypti infestations: Is this a practical task?

Our plain-English summary of Remote sensing for risk mapping of Aedes aegypti infestations: Is this a practical task?, by C. Lorenz, F. Chiaravalloti-Neto, M. de Oliveira Lage, et al (Acta Tropica, 2020).

The study

• Authors: C. Lorenz, F. Chiaravalloti-Neto, M. de Oliveira Lage, J.A. Quintanilha, M.C. Parra, M.R. Dibo, E.A. Fávaro, M.M. Guirado, M.L. Nogueira
• Journal: Acta Tropica
• Published: 2020;205:105398
• Cited by: 17 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Mosquito-borne diseases affect millions of individuals worldwide; the area of endemic transmission has been increasing due to several factors linked to globalization, urban sprawl, and climate change. The Aedes aegypti mosquito plays a central role in the dissemination of dengue, Zika, chikungunya, and urban yellow fever. Current preventive measures include mosquito control programs; however, identifying high-risk areas for mosquito infestation over a large geographic region based only on field surveys is labor-intensive and time-consuming. Thus, the objective of this study was to assess the potential of remote satellite images (WorldView) for determining land features associated with Ae. aegypti adult infestations in São José do Rio Preto/SP, Brazil. We used data from 60 adult mosquito traps distributed along four summers; the remote sensing images were classified by land cover types using a supervised classification method. We modeled the number of Ae. aegypti using a Poisson probability distribution with a geostatistical approach. The models were constructed in a Bayesian context using the Integrated nested Laplace Approximations and Stochastic Partial Differential Equation method. We showed that an infestation of Ae. aegypti adult mosquitoes was positively associated with the presence of asbestos roofing and roof slabs. This may be related to several other factors, such as socioeconomic or environmental factors. The usage of asbestos roofing may be more prevalent in socioeconomically poor areas, while roof slabs may retain rainwater and contribute to the generation of temporary mosquito breeding sites. Although preliminary, our results demonstrate the utility of satellite remote sensing in identifying landscape differences in urban environments using a geostatistical approach, and indicated directions for future research. Further analyses including other variables, such as land surface temperature, may reveal more complex relationships between urban mosquito micro-habitats and land cover features.

Copyright © 2020 Elsevier B.V. All rights reserved.

In plain English

This study from Brazil was about mosquitoes, not about the health effects of asbestos. Asbestos roofing appeared only as one of the land features measured. The authors set adult mosquito traps at 60 houses in the Vila Toninho neighborhood of São José do Rio Preto over four summers (2016 to 2019). They caught 705 Aedes aegypti adults (460 females and 245 males). They sorted a WorldView-3 satellite image into 10 land cover types, including asbestos roof and roof slab. They then compared the land cover within 30, 50 and 100 m of each trap with mosquito counts, using a Bayesian spatial model. In the best-fitting models, a combined factor made up mainly of asbestos roof and roof slab, measured within 30 m or 50 m of each trap, was linked to more mosquitoes; within 100 m the link was unclear. The authors wrote that these results did not show statistical significance and called them preliminary. They suggested the link may reflect poorer areas, where asbestos roofing may be more common, or heat and rainwater pooling. They noted that they did not measure surface temperatures and that 60 traps limited the analysis.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Spatial (geostatistical) study relating satellite-derived land cover around mosquito traps to adult mosquito counts, using a Bayesian Poisson model
• Who was studied: Adult Aedes aegypti mosquitoes caught in traps at up to 60 houses in the Vila Toninho neighborhood (about 2000 households, about 6000 people), São José do Rio Preto, Brazil, from December to February each year, 2016 to 2019
• Exposure: Percentage of 10 satellite-classified land cover types (including asbestos roof and roof slab) within 30 m, 50 m and 100 m of each trap. No asbestos fiber type or human asbestos exposure was studied.
• Main results:
  • The traps caught 705 Ae. aegypti adults, 460 females and 245 males, over the four summers
  • The Maximum Likelihood classifier correctly classified 90.6% of the urban land cover, with a Kappa index of 0.89
  • 30 m buffer, females only: the land-cover component RC3 (mainly asbestos roof and roof slab) had a posterior mean of 0.12 (95% credible interval 0.05 to 0.2). This was the best model overall: DIC fell from 979.5 to 977.8 for females and from 1202.66 to 1200.47 for both sexes
  • Both sexes: RC3 posterior mean 0.075 (0.01 to 0.17) at 30 m and 0.08 (0.003 to 0.18) at 50 m. Females only at 50 m: 0.09 (0.02 to 0.18)
  • 100 m buffer: RC3 was also the best-fitting component, but its posterior mean was close to zero: −0.004 (−0.08 to 0.07) for females and 0.005 (−0.07 to 0.08) for both sexes
  • The authors stated these results did not show statistical significance but were consistent with more mosquitoes in areas with a large proportion of asbestos roofing and roof slabs
  • In models using only male mosquitoes, no model fit better than the intercept-only model
• The authors' conclusion: The authors concluded that adult Ae. aegypti were positively associated with asbestos roofing and roof slabs, possibly through socioeconomic or environmental factors. They called the results preliminary and descriptive, and said the findings need testing over longer periods and larger areas.
• Limitations the authors state:
  • Environmental data, such as the surface temperature of each urban material, were not included in the model
  • The sample of 60 traps was too small to allow more independent variables in the analysis
  • The buffers assumed an average flight radius, but mosquitoes might fly farther depending on conditions
  • The authors described the results as preliminary

Topics

Medical terms in this study, explained in our Lexicon: Endemic.

Related studies

• Predicting Aedes aegypti infestation using landscape and thermal features — C. Lorenz, M.C. Castro, et al. (2020)
• Mapping asbestos-cement roofing with hyperspectral remote sensing over a large mountain region of the Italian Western Alps — F. Frassy, G. Candiani, et al. (2014)
• Job tasks, potential exposures, and health risks of laborers employed in the construction industry — G. Burkhart, P.A. Schulte, et al. (1993)
• A task-based assessment of parental occupational exposure to organic solvents and other compounds and the risk of childhood leukemia in California — C. Metayer, G. Scelo, et al. (2016)
• A Bayesian Model and Stochastic Exposure (Dose) Estimation for Relative Exposure Risk Comparison Involving Asbestos-Containing Dropped Ceiling Panel Installation and Maintenance Tasks — F.W. Boelter, Y. Xia, et al. (2017)
• Disease mapping of zero-excessive mesothelioma data in Flanders — T. Neyens, A.B. Lawson, et al. (2017)

Browse return to medical research.