Fig 1.
Flow chart of steps in methodology.
GAM = Generalized additive model; ODE = Ordinary differential equations; ABM = Agent-based model.
Table 1.
Parameter names, definitions, and sources.
Fig 2.
Smoothed map of adult female abundance on 1st January 2000 (the start of the simulation).
Warmer colors (in the Southwest and Northeast of the city) indicate regions of highest abundance. The color scale represents the average number of mosquitoes per household. The shape files for the underlying maps can be found at github.com/scavany/mosquito_dynamics.
Fig 3.
The total predicted city-wide abundance of female adult Ae. aegypti is very similar across the three models.
The periwinkle blue line shows daily values of abundance predicted by the GAM, the purple line shows those predicted by the ODE model, and the pink line those predicted by the ABM. GAM: generalized additive model; ODE: ordinary differential equation model; ABM: agent-based model.
Fig 4.
Normalized female adult abundance over time under the agent-based model: the relative abundance across zones does not substantially vary through time.
Each column represents the daily abundance every 100 days from 2000–2010. Each row is a Ministry of Health zone in Iquitos. Columns are normalized by the total female adult abundance across all zones that day, so each column sums to one and the color represents the proportion of mosquitoes in that zone that day.
Fig 5.
Response of female adult abundance to a city-wide residual insecticide campaign (TIRS) and a city-wide campaign with no residuality (ULV), both initiated on the 1st of January 2002.
The orange line shows the baseline scenario of no-spraying, the purple line the ULV scenario, and blue line the TIRS scenario.
Fig 6.
Normalized female adult abundance over time, following a city-wide ULV campaign initiated on the 1st January 2002, indicated by the red dashed line.
ULV spraying does not have a discernable effect on the spatial distribution of abundance. Each column represents the daily abundance every 100 days from 2000–2010. Each row is a Ministry of Health zone in Iquitos. Columns are normalized by the total female adult abundance across all zones that day, so each column sums to one and the color represents the proportion of all female adult mosquitoes in that zone that day.
Fig 7.
Normalized female adult abundance over time, following a city-wide TIRS campaign initiated on the 1st January 2002, indicated by the red dashed line.
A. TIRS has a substantial effect on the spatial distribution of abundance for most of the year following spraying, as some zones are reduced to zero abundance. Each column represents the daily abundance every 100 days from 2000–2010. Each row is a Ministry of Health zone in Iquitos. Columns are normalized by the total female adult abundance across all zones that day, so each column sums to one and the color represents the proportion of mosquitoes in that zone that day. B. As in A, but with the same color-scale as Fig 4.
Fig 8.
Normalized female adult age distribution over time: mosquito ages were roughly exponentially distributed.
Each column shows the daily proportion of mosquitoes of the age shown on the y-axis. Warmer colors represent higher proportions.
Fig 9.
Normalized difference in female adult ages over time between baseline and a city-wide ULV campaign initiated on 1st January 2002, indicated by the red dashed line.
Following spraying there are proportionally more young Ae. aegypti. Each column shows the difference in the daily proportion of mosquitoes of a given age, shown on the y-axis. Warmer colors represent higher proportions.
Fig 10.
Normalized difference in female adult ages over time between baseline and a city-wide TIRS campaign initiated on 1st January 2002, indicated by the red dashed line.
Following spraying there are proportionally more young Ae. aegypti, and this effect lasts longer than for ULV. Each column shows the difference in the daily proportion of mosquitoes of a given age, shown on the y-axis. Warmer colors represent higher proportions.