UC Berkeley

Background

Few studies of the relationship between residential proximity to agricultural pesticide applications and pesticide levels in the home have incorporated crop location or wind direction. We evaluated the relationship between agricultural pesticide applications using the California Pesticide Use Reporting (CPUR) database and pesticide concentrations in carpet dust accounting for land use and wind direction.

Methods

We measured concentrations (ng/g) of seven herbicides and two fungicides in carpet dust samples from 578 California homes (2001-2007). We created three metrics by computing the density (kg/km²) of use of each pesticide reported in CPUR within 0.5-, 1-, 2-, and 4-km buffers around homes 180- and 365-days before sampling (CPUR metric). We apportioned applications to the crop area within the buffers (CROP-A metric) and weighted CPUR applications by the proportion of days that the home was within ±45° of the downwind direction (W-CPUR metric). We modeled natural-log concentrations (Tobit regression) and dust detections (logistic regression) adjusting for season/year, occupation, and home/garden use.

Results

Detections were >90 % for glyphosate, 2,4-D, and simazine. Detection rates and dust concentrations increased with increasing CPUR densities for all herbicides and one fungicide. Compared to homes without applications within 4 km, the highest tertile of 365-day glyphosate use was associated with ∼100 % higher concentrations (CPUR_T3>9.2kg/km² %_change = 110, 95 %CI = 55, 183; CROP-A_T3>13.4kg/km² %_change = 144, 95 %CI = 81, 229; and W-CPUR_T3>2.1kg/km² %_change = 102, 95 %CI = 50, 171). The highest density tertiles of 2,4-D, simazine, and trifluralin were associated with 2- to 6-times higher concentrations, respectively; that was similar across metrics. Across all metrics, agricultural use of dacthal, dicamba, and iprodione were associated with 5- to 10-times higher odds of dust detections. Associations were unclear for 2-methyl-4-chlorophenoxyacetic acid and null for chlorothalonil.

Conclusions

Agricultural herbicide and fungicide use was an important determinant of indoor contamination within 4 km of homes. Accounting for crops and wind direction did not substantially change these relationships.