Fig 1.
Distribution of dengue, population and climate among the 120 Health Macro Regions, Brazil.
(A) Dengue cases reported from 2010 to 2023; (B) population size (2022); (C) mean annual temperature (1950-1990); (D) mean annual precipitation (1950-1990); (E) mean altitude. Service Layer Credits: Sources: https://www.ibge.gov.br/geociencias/organizacao-do-territorio/malhas-territoriais/15774-malhas.html?=&t=downloads.
Fig 2.
Köppen’s climate system of Brazil at HMR level derived from the original municipality classification by Alvares et al. (2013) [
29]. Service Layer Credits: Sources: https://www.ibge.gov.br/geociencias/organizacao-do-territorio/malhas-territoriais/15774-malhas.html?=&t=downloads. Köppen’s climate system of Brazil: Sources: https://www.ipef.br/publicacoes/acervohistorico/geodatabase/ as open access data and in https://zenodo.org/records/15281683 under a Creative Commons Attribution 4.0 International.
Fig 3.
Diagram illustrating the MEM method applied to the HMR RRAS 12, in São Paulo state; A: The data input contains the number of cases per week per year, needed to describe the disease’s historical pattern (the dengue seasons used were those inside the percentiles 10% and 90%); B: MEM outputs: The year is split into three periods: pre-epidemic, epidemic and post-epidemic period, with their corresponding dates, duration and incidence thresholds. Intensity levels are computed based on historical quantiles. The grey region represents the epidemic period; red dashed line marks the high-intensity period.
Fig 4.
Characteristics of typical dengue seasons in the Brazilian health macroregions
; A: pre-epidemic threshold is defined as the incidence that marks the beginning of the season (cases/week x 10^5); B: post-epidemic threshold is the incidence that marks the end of the season (cases/week x 10^5); C: High-intensity level is the incidence threshold above which a season is considered a severe epidemic (cases/week x 10^5). Service Layer Credits: Sources: https://www.ibge.gov.br/geociencias/organizacao-do-territorio/malhas-territoriais/15774-malhas.html?=&t=downloads.
Fig 5.
Characteristics of typical dengue seasons in the Brazilian health macroregions; A: Season onset date is the epidemiological week at which the pre-epidemic threshold is surpassed.; B: Mean season duration is the number of weeks with incidence above the pre-epidemic threshold; C: variation of onset date, measured as the width of the onset’s 95% confidence interval; D: variation of the season duration, measured as the width of the season duration 95% confidence interval.
Service Layer Credits: Sources: https://www.ibge.gov.br/geociencias/organizacao-do-territorio/malhas-territoriais/15774-malhas.html?=&t=downloads.
Fig 6.
Results of the multivariate cluster analysis for the classification of health macroregions into homogeneous groups based on their climate and dengue transmission seasons
; (A-D) silhouette plot for analysis considering 3 to 6 clusters. (E) Cluster plot using 4 clusters, the final classification.
Fig 7.
Health Macro Region Clustering and Final Dengue Regionalization of Brazil
; (A) Classification of HMR into 4 clusters, following the k-means algorithm; (B) Final classification of Brazil into 5 dengue regions, after the post processing phase. Service Layer Credits: Sources:https://www.ibge.gov.br/geociencias/organizacao-do-territorio/malhas-territoriais/15774-malhas.html?=&t=downloads.
Table 1.
Mean values and relative contributions (%) of environmental and epidemiological variables to cluster differentiation, quantified by inter-cluster variance.