Figure 1.
This is a schematic of the simulation region where the odor plume is computed. At the edges of the simulation region, the
concentration is carried out by the wind. The patches represent smaller subregions where the hosts are located. The spatial distribution of hosts within each patch is uniform and there can be multiple patches. The hosts are stationary in the following simulations, although this is not a limitation of the model. The mosquitoes are initially placed in a subregion inside the
simulation region but are allowed to leave it (and possibly reenter it) during the simulation. The wind exists everywhere, even outside the
simulation region.
Figure 2.
Examples of mosquito trajectories.
Example mosquito trajectories for upwind (UW), downwind (DW), and crosswind (CW) plume-finding behaviors. The stationary hosts are distributed into two groups and the contour of the odor plume marks a snapshot of the sensing threshold of the mosquito. The UW and CW mosquitoes successfully locate a host; the DW mosquito is unsuccessful. The CW mosquito leaves and re-enters the domain.
Figure 3.
Mosquito flight direction choice.
Schematic of the mosquito direction choice in the host-seeking model. At each discrete flight segment, the angle represents the ideal target direction of the mosquito, which can only be sensed or followed to a precision of
.
Table 1.
Parameter choices.
Figure 4.
The proportion of mosquitoes finding a host within 150 s.
The black bars are the results for a random walk (RW) without wind and , the gray for the gradient method (G) of concentration sensing, and the white for the sampling method (S) of concentration sensing. UW = upwind plume finding, DW = downwind plume finding, and CW = crosswind plume finding. The thin error bars are
standard deviations. If the simulation is allowed to progress beyond 150 s, then no appreciable changes occur in the results for the host-seeking methods, but the random walk continues to accrue contacts.
Table 2.
Most sensitive local variation.
Figure 5.
Examples of straight and meandering plumes with a superposed random velocity field.
The triangles denote host position. The contours show concentration level, with darker bars indicating lower concentration. The lowest contour is and is the same in both plots. The other contours are not the same in both plots because the maximum concentration is roughly three times higher in the meandering plume.
Table 3.
A comparison of plume-finding behaviors in straight and meandering plumes.
Figure 6.
Ratio of contacts between a small and a large host group.
Mean and standard deviation of the ratio of contacts in the small group to that in the large group (S/L) vs the ratio of group sizes. The line denotes an equal per capita contact rate between groups, while the line
denotes an equal number of contacts at each group. The data corresponding to a given ratio of hosts have been separated slightly in the figure for visual purposes only.
Figure 7.
The proportion of mosquitoes finding a host as a function of host density.
The simulation results are given by the solid markers with one standard deviation (UW = upwind, DW = downwind, CW = crosswind). The lines are linear fits to the data using the formula
, where
is a free parameter and
is the length of a side of the square host patch divided by the length of a side of the simulation region. As
increases, the hosts spread out (i.e., decrease in density) and the number of mosquitoes locating a host increases. The
is
labeled as a percentage.
Figure 8.
Growing odor plume in a large domain.
Mosquitoes of all three plume-finding behaviors ( upwind,
downwind,
crosswind) near an evolving plume about 45–70 m long. All distances are in meters and only the outermost contour of the plume is shown. Left: The plume after 250 s. The upwind and downwind mosquitoes are already segregating into the upwind and downwind sides of the domain. Right: The same part of the domain after 500 s. The plume is about 2/3 longer and only crosswind mosquitoes remain in the vicinity of the plume.