Oracle Crystal Ball GBU

Probabalistic Risk Assessment for Pesticides: Estimating Effect Levels for Target and Non-Target Species

- By:

Courtesy of Courtesy of Oracle Crystal Ball GBU

A probabilistic model was developed for estimating avian mortality associated with field application of pesticides. Pesticide exposures for individuals were estimated by random sampling from species specific non-parametric distributions of pesticide bait consumption and pesticide concentrations in the field applied pesticide baits and food sources. Mortality was predicted from species specific exposure versus mortality relationships. Individual variations in pesticide sensitivity (mortality response) were captured by Monte Carlo sampling from species specific distributions of
slopes and median toxicity values (LD50) for each bird (trial). The model was used to predict percent mortality of the
exposed population for a target and non-target species. The mortality estimates provided by this model provide a means for weighing the risks against the benefits of a pesticide control program. These mortality estimates can be subsequently incorporated into population models to estimate the long-term population effects induced by pesticide application scenarios of interest. This approach affords an efficient means to identify optimal pesticide application strategies.

The United States Department of Agriculture/Animal and Plant Health and Inspection Service/Wildlife Services
(USDA/APHIS/WS) has used CPTH (3-chloro-p-toluidine hydrochloride, 3-chloro-4-methylaniline hydrochloride,
DRC-1339) coated rice baits to reduce red-winged blackbird (Agelaius phoeniceus) damage to rice in Louisiana and
to sunflower in North and South Dakota (Glahn and Wilson, 1992; Linz and Bergman, 1996). Typically, CPTH is coated onto brown rice at a target concentration of 2 % (w/w). At the application site, the 2 % CPTH treated brown
rice is diluted with untreated rice at a ratio of 1:25 (w/w) and subsequently broadcast at a rate of 12 to 23 kg/ha.
Concern regarding the use of CPTH has focused on hazards to non-target species which have been observed in or near CPTH baited fields (Linz et al., 2002; Custer et al., 2003; Pipas et al., 2003; Plumart, 2003). While the presence
of non-target bird species suggests a potential hazard, a more detailed analysis is required to quantify the risk of acute toxicity.

In order to maximize target and minimize non-target CPTH induced mortality, sites are pre-baited to attract target species and bait application is timed to minimize exposure of non-target birds. More than a decade of monitoring baits sites during and after CPTH baiting operations has produced evidence of very few non-target causalities. However, this practice has also led to the discovery of very few target species as well. This is likely due to the mobility of target and non-target avian species and the slow acting nature of CPTH, death typically occurs one to several days post ingestion (Johnston et al., 1999). This situation makes it very difficult to quantify target and nontarget impacts of field application of the pesticide CPTH. As USDA has no intention of applying a pesticide with low efficacy to target species and/or significant adverse effects to non-target species, a probabilistic model was developed using toxicity data from confined feeding studies and bait seed consumption data from field observations as a means of estimating both target and non-target mortality associated with CPTH baiting programs.

To model CPTH exposure for multiple species simultaneously feeding on a bait site, a species distribution was prepared
using the custom distribution function in Crystal Ball® (Decisioneering®, Inc., Denver, CO). The distribution contained the proportions of each species observed in the CPTH baited field. For each bird feeding in the pesticide baited field (a single iteration of the model), a species was selected based on the relative probabilities in this distribution. The bird was then randomly assigned a bodyweight from a normally distributed species specific bodyweight distribution (Denning, 1984).

Customer comments

No comments were found for Probabalistic Risk Assessment for Pesticides: Estimating Effect Levels for Target and Non-Target Species. Be the first to comment!