Introduction

Millions of United States residents live in areas where Aedes aegypti, the primary vector of dengue, Zika, and chikungunya viruses, and Culex quinquefasciatus, the primary vector of West Nile virus (WNV) and St. Louis Encephalitis virus (SLEV), are established [1–3]. The risk of local transmission from these mosquitoes is not theoretical. WNV is now endemic in the United States [4], having been documented in 96% of counties since its introduction in 1999 and having caused an estimated 7 million infections and more than 50 thousand neuroinvasive cases, although many infections remain asymptomatic [4,5]. Similarly, locally acquired dengue infections are increasingly reported, with recent outbreaks in Florida, California, and Arizona, highlighting a growing public health threat [6,7]. Specifically, in Miami-Dade County, Florida, 314 locally transmitted dengue virus infections were reported between 2010 and 2024 [8]. However, CDC estimates suggest that actual case numbers may be substantially higher due to underreporting, with an estimated reporting rate between 1.0% and 4.8% [9].

While essential for mitigating arboviral disease threats, mosquito control operations are often reactive, triggered by disease cases or public complaints [10]. A shift toward a more proactive strategy, where interventions can be planned in advance, must be underpinned by a foundational understanding of long-term vector dynamics. This involves identifying consistent seasonal trends in key vector species and persistent high-risk areas (“hotspots”) that can inform data-driven operational planning.

A key step toward proactive control is to analyze historical mosquito surveillance data to identify consistent spatial and temporal patterns in vector populations, including the identification of areas with substantially higher mosquito abundance. To provide this evidence for Miami-Dade County, Florida, we analyzed eight years of surveillance data (2017–2024). Our objectives are to: (i) characterize the spatiotemporal distribution of Ae. aegypti and Cx. quinquefasciatus and (ii) identify hotspot areas for their proliferation. The identification of areas conducive to mosquito proliferation serves three main purposes: to identify resource availability at the local scale responsible for supporting mosquito populations; to provide an early warning system for increased mosquito activity; and to inform targeted control strategies to mitigate mosquito proliferation. Therefore, the identification of hotspot areas is relevant for proactively informing mosquito surveillance and control efforts.

Materials and methods

In this study, we used 277 BG-Sentinel 2 traps (Biogents AG, Regensburg, Germany) and 31 CDC light traps that are part of the Miami-Dade County mosquito surveillance system (Fig 1). Each trap is deployed once weekly for 24 hours over approximately 50 weeks per year. Traps are placed in shaded areas protected from direct sunlight, wind, and precipitation to optimize mosquito collection. Collected mosquitoes are morphologically identified to species using standard taxonomic keys [11]. Surveillance protocols remained consistent throughout the study period, with traps deployed at fixed locations using the same trap models and CO₂ as bait. BG-Sentinel and CDC traps baited with CO₂ primarily attract host-seeking female mosquitoes [12] and are not specifically designed to attract males. Although some male mosquitoes were collected, they were considered accidental catches and were excluded from the analyses. Even though the Miami-Dade Mosquito Control surveillance system collected several mosquito species from 2017 to 2024, this study focused on the two primary vector species of epidemiological importance: Ae. aegypti and Cx. quinquefasciatus. No specific permits were required for this study, as the entomological data utilized in this analysis are the property of the Miami-Dade Mosquito Control Division and were collected for entomological surveillance purposes.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Spatial distribution of traps in the Miami-Dade Mosquito Control Surveillance System, 2024.

BG-Sentinel traps are shown as blue circles, and CDC traps as yellow squares. The figure was produced using ArcGIS Online (Esri, Redlands, CA), using freely available layers from Miami-Dade County’s Open Data Hub - https://gis-mdc.opendata.arcgis.com/.

https://doi.org/10.1371/journal.pone.0334586.g001

Monthly species totals were paired with sampling effort, defined as the number of trap-nights per month (calculated by multiplying the number of traps deployed by the number of 24-hour periods they were operated). Mean female mosquitoes per trap-night were calculated for the entire study period.

The identification of areas conducive to mosquito proliferation serves three main objectives: (1) identify resources at the local scale responsible for supporting mosquito populations; (2) provide an early warning system for increased mosquito activity; and (3) inform targeted control strategies to mitigate mosquito proliferation. Therefore, to identify areas conducive to the proliferation of Ae. aegypti and Cx. quinquefasciatus, we created a 4-km buffer around each trap and calculated the number of traps within each buffer, along with the total and mean number of mosquitoes collected by each trap in the buffer. This buffer size ensured inclusion of a sufficient number of neighboring traps (mean: 15 per buffer) to enable compelling statistical analysis while maintaining operational spatial resolution. Traps were classified as hotspots if their average mosquito count was above the 97.5th percentile of their respective buffer. Alternative thresholds for hotspot definitions can be used based on the distance from the buffer mean. We explored the following options: i) above 97.5th percentile of the distribution of distance from buffer mean; ii) 97.5th percentile of the non-outlier traps within respective buffer. To calculate the mean mosquito abundance per buffer, we follow the methods used in Wilke et al. [13]. First, we removed outliers using a standard method based on , where IQR is the interquartile range, n is the number of observations in the buffer, and the 1.58 multiplier is the standard constant for notched boxplots as defined by McGill, Tukey, and Larsen [14].

Figures were produced using ArcGIS Online (Esri, Redlands, CA), using freely available layers from Miami-Dade County’s Open Data Hub - https://gis-mdc.opendata.arcgis.com/. All analyses were conducted in R (version 4.2.2) [15], using “geosphere” package [16] for the distance calculation and “rstudioapi” [17] for dynamic path definition. Additionally, “dplyr” package [17] was used for the calculation of descriptive statistics, but is not required for the data manipulation. The code used to perform the analyses is openly available on GitHub at https://github.com/litvinovamv/Mosquito-Data-Analysis/tree/main/abundance_hotspots.

Results

From 2017 to 2024, a total of 399,418 Ae. aegypti and 1,250,879 Cx. quinquefasciatus mosquitoes were collected in Miami-Dade County, Florida (Table 1). Annual counts and mean abundance per trap-night showed consistent patterns, with Cx. quinquefasciatus remaining more abundant than Ae. aegypti throughout the study period. Yearly relative abundance of Ae. aegypti ranged from 35,612–64,570 mosquitoes, while Cx. quinquefasciatus ranged from 96,154–203,223 mosquitoes. Mean values for Ae. aegypti ranged from 3.39 to 6.06 mosquitoes per trap-night, while Cx. quinquefasciatus ranged from 9.92 to 21.33. The overall mean abundance was 4.25 mosquitoes per trap-night for Ae. aegypti and 13.46 for Cx. quinquefasciatus.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Relative abundance of Aedes aegypti and Culex quinquefasciatus. Mean number of mosquitoes per trap-night shown in parentheses.

https://doi.org/10.1371/journal.pone.0334586.t001

The seasonal population dynamics of Ae. aegypti and Cx. quinquefasciatus showed consistent temporal variations across the study period (Fig 2). Aedes aegypti and Cx. quinquefasciatus showed opposing seasonal trends, with Ae. aegypti abundance concentrated in the summer wet season (June-August) and Cx. quinquefasciatus peaking during the winter dry season (December-February). Culex quinquefasciatus remained more abundant overall compared to Ae. aegypti. Nonetheless, both species were consistently present and collected in relatively high numbers throughout the year, suggesting continuous breeding and survival across seasons. While these patterns represent observational trends rather than results of inferential hypothesis testing, their persistence across the study period underscores their importance for year-round surveillance and situation awareness.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Monthly number of Aedes aegypti (red) and Culex quinquefasciatus (blue) collected in Miami-Dade County, Florida.

The number of trap-nights per month is indicated by the dashed line.

https://doi.org/10.1371/journal.pone.0334586.g002

Among the 306 traps included in this eight-year study, an average of 36 traps were classified as hotspots for Ae. aegypti (Fig 3). Of these, 3 traps were identified as hotspots in all eight years, 3 traps in seven years, 4 traps in six years, and 5 traps in five years. Similarly, an average of 35 traps was identified as hotspots for Cx. quinquefasciatus, with 3 traps classified as hotspots in all eight years, 5 traps in seven years, 5 traps in six years, and 3 traps in five years. Table 2 presents the temporal results of hotspot classification when using a variety of hotspot measures. These results illustrated the flexibility of hotspot classification to accommodate diverse priorities and resources for response by adjusting the sensitivity of situational awareness analysis.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Frequency of hotspot classification for each trap from 2017 to 2024. Hotspot traps were defined as traps with a mean mosquito abundance exceeding the 97.5th percentile within their respective 4-km buffers.

https://doi.org/10.1371/journal.pone.0334586.t002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Spatial distribution of hotspots for female Aedes aegypti and Culex quinquefasciatus in Miami-Dade County, Florida, from 2017 to 2024, based on relative mosquito abundance compared to other traps within a 4 km buffer radius.

Colors represent the number of times a trap has been classified as a hotspot during the study period. The figure was produced using ArcGIS Online (Esri, Redlands, CA), using freely available layers from Miami-Dade County’s Open Data Hub - https://gis-mdc.opendata.arcgis.com/.

https://doi.org/10.1371/journal.pone.0334586.g003

Discussion

The characterization of the spatiotemporal distribution of Ae. aegypti and Cx. quinquefasciatus in Miami-Dade County, Florida, is essential for identifying hotspot areas for mosquito proliferation to guide control operations and improve arbovirus outbreak preparedness and response. Our results show that both Ae. aegypti and Cx. quinquefasciatus have well-established seasonal patterns, with Ae. aegypti peaking in the summer months and Cx. quinquefasciatus in the winter months. The underlying drivers of these seasonal dynamics need further investigation, particularly given the relatively stable temperature in Miami-Dade County throughout the year and the lack of a clear-cut alignment of the relative abundance of the two focus species and rainfall, for example, Ae. aegypti population consistently peaked in July despite temperature and rainfall being virtually the same in July and August [18].

Seasonal trends of mosquito vector species have been reported elsewhere in the United States. In Maricopa County, Arizona, Cx. quinquefasciatus had bimodal peaks aligned with the spring and early fall rainy seasons, while Ae. aegypti populations increased during wetter months of the year and declined sharply at the onset of the dry season [13]. In New Orleans, Louisiana, Ae. aegypti showed consistent peaks from May to August, whereas Cx. quinquefasciatus maintained high abundance year-round, with broad summer peaks [19]. In Los Angeles, California, Cx. quinquefasciatus abundance peaked early in July and declined steadily thereafter, while Ae. aegypti increased gradually from June, reaching a peak in October before decreasing in November [20].

These results indicate that while both species respond to seasonal environmental changes, the timing, intensity, and persistence of peaks vary by location, reflecting differences in local climatic patterns, habitat availability, and species-specific ecological adaptations. The patterns found in Miami-Dade County of Ae. aegypti summer peak abundance and Cx. quinquefasciatus winter peak abundance contrasts with patterns in other regions, underscoring the importance of localized, species-specific surveillance to guide targeted vector control efforts. The operational value of these findings is underscored by the observation that WNV seroconversions in Florida sentinel chickens frequently persist into later in the year [21,22]. This period coincides with peak Cx. quinquefasciatus abundance, providing an entomological link to the sustained risk of spillover. This overlap underscores the need for year-round surveillance and a data-driven framework to allow the early detection of hotspots in urban environments.

Our spatial analysis showed consistent spatial clustering of traps with high mosquito abundance. These spatial trends suggest that local environmental and anthropogenic factors influence species distribution and abundance; however, future studies are needed to empirically validate these observations. The temporal persistence of specific hotspots suggests the presence of stable ecological or structural conditions, as well as resource availability, which support sustained mosquito production. These findings further suggest the presence of priority areas for targeted vector control and support the use of persistent hotspots as sentinel sites for early warning and surveillance.

Miami-Dade County remains particularly susceptible to arbovirus transmission due to its climate, urban landscape, and international connectivity [2,23]. Modeling estimates suggest that, assuming a basic reproduction number of 1.5, the introduction of a single asymptomatic infected individual could lead to a 10% probability of a dengue outbreak resulting in at least 40 symptomatic cases, with a median outbreak size of 73 symptomatic infections [24]. These projections underscore the importance of proactive surveillance and targeted interventions in Miami-Dade County.

Historical spatiotemporal entomological data provide a foundation for the development and implementation of proactive mosquito control measures. However, mosquito populations are highly sensitive to short-term fluctuations in temperature and rainfall, complicating efforts to anticipate and mitigate periods of increased abundance and elevated arbovirus transmission risk [25,26]. This variability presents a challenge for public health agencies responsible for implementing timely and effective mosquito control interventions. Current strategies are primarily reactive, typically triggered by confirmed disease cases, increased mosquito trap counts, or public complaints [27,28].

While this preliminary characterization provides immediate operational utility for public health authorities, further research is needed to elucidate the underlying drivers of these dynamics. Integrating high-resolution historical entomological surveillance data into routine operations could enhance situational awareness and support a more proactive, data-driven approach [29,30]. This integration could facilitate the early identification of high-risk periods and locations, enabling public health authorities to optimize intervention timing and targeting, improve resource allocation, and increase the overall effectiveness of mosquito control programs.

Conclusion

Our findings show that the incorporation of high-resolution entomological surveillance data across spatial and temporal scales enables the identification of key areas and periods of elevated mosquito vector abundance. The consistent detection of hotspot areas across multiple years provides operational value by identifying reliable locations for monitoring mosquito population dynamics, assessing the impact of control interventions, and prioritizing resource allocation. As arboviruses remain a persistent public health threat in urban environments such as Miami-Dade County, the capacity to anticipate and respond to mosquito vector population fluctuations is critical for effective outbreak prevention and control. This study presents data that is instrumental for the improvement of situational awareness, guiding targeted vector control operations, and ultimately reducing the risk of local arbovirus transmission.