Table 3 Overview and discussion of major sources of data uncertainty
Source of uncertainty | Evaluation |
|---|---|
Exposure-response function | Epidemiological evidence is suggestive but far from conclusive. |
Differences in water supply and life style factors for colon cancer incidence between Iowa and Europe. | Meat intake is an important risk factor for cancers. Total meat consumption in Europe and the US are comparable, but beef consumption in Iowa is higher. |
The assumption that exceedance of 25 mg/L NO3 in groundwater samples at 5-20 m depth is equivalent to exposure in all drinking water from small public supply and private wells. | No data were available about extraction depth and water treatment for this type of supply. Local data on nitrate in groundwater and actual use for drinking water were not available. |
Focus on groundwater that is affected by agricultural nitrogen loading. | Relatively unpolluted aquifers overlain by forest or semi-natural vegetation are underrepresented. Therefore exposure probably is overestimated. |
Not considering surface water based drinking. | Considering non-compliance in surface water based public drinking water increases health cost by about 15%. Although about 40% of EU surface waters exceed 25 mg/L NO3[41], we assume that private use of surface water for drinking water is negligible compared to groundwater. |
Not considering consumption of bottled water. | In EU27, the consumption of bottled drinking water, that is very low in nitrate, increased from around 12% of total intake in 2001 to 15% in 2007 and consideration would slightly lower exposure estimates. In fact total beverage consumption is relevant; fruit juices can be high in nitrate and beers high in nitrosamines [42]. |
The assumption that percentage of drinking water samples from large public facilities not complying with standards for nitrate (50 mg/L) or nitrite (0.5 mg/L NO2) is equivalent to exposure, and identical for groundwater and surface water sources | Non-compliance may be incidental and assumption may overestimate exposure. Estimates of exposure to exceedance of 25 mg/L NO3 will be more robust. |