干旱区生态环境知识资源中心

We modified an existing simulation model to evaluate detection of cattle-fever ticks (Boophilus annulatus and Boophilus microplus) infesting rangeland in the semi-arid thornshrublands of south Texas. This model contained submodels representing (1) tick development in a pasture and (2) tick development on one cow; we added a third submodel that calculated the probability of tick detection on a herd of cattle by human inspectors. This detection submodel randomly distributed ticks to each cow in the herd from a normal distribution with a mean equal to the number of ticks of detectable size predicted by the submodel of tick development on one cow. We controlled biophysical factors such as habitat type, starting season, number of ticks initially infesting the system, number of cows examined, and a human factor, ability of inspectors to detect ticks on individual cows. We simulated all combinations of factors to estimate their combined effect on the probability of detection over a 2-year period under continuous grazing. Results showed the predominant influence of biophysical factors: season of initiation had the largest influence on tick populations (and hence, subsequent detection probabilities), followed by habitat type and level of initial tick infestation. Detectable populations of ticks occurred in fall-initiated simulations in grass only at infestations beginning with greater than or equal to250 larvae per animal, while in mesquite detectable populations occurred with initial infestations greater than or equal to50 larvae per animal. In contrast, detectable populations occurred in spring-initiated simulations initiated with as few as 10 larvae per animal in both grass- and mesquite-dominated pastures. Mean percentage of potential detection days in which simulated inspectors successfully detected ticks ranged from 0 to 99.9% depending on the number of animals inspected and level of detection probability evaluated. The distribution of days in which ticks remained detectable varied with the dynamics of each combination of factors and illustrated that opportunities for detection are not evenly distributed in time. Variation of the human factor, the curve used to estimate ability of inspectors to detect ticks on individual cows, suggest that human detection abilities usually have a relatively minor influence on detection probabilities. Human detection ability, however, becomes more important when on-host tick populations reach low levels or when few cattle are inspected. At these times, increased detection abilities yield windows of detection that otherwise would not exist. These simulations provide the first analysis of tick detection in scenarios under U.S. conditions, and results confirm the concerns regulatory agencies express regarding early detection of cattle-fever ticks. (C) 2003 Elsevier B.V. All rights reserved.