Karin Kvale
Precipitation composition trends in the Ohio River Valley
BSES Senior Research 2004

According to the EPA, regulation of sulfur and nitrogen by the Clean Air Act and subsequent amendments resulted in decreases in emissions of 33 and 15% respectively between 1983-2002. However, due to non-linearities of the atmospheric chemistry of these gases, emission reductions need not be linearly related to decreased concentrations of the oxidation products in precipitation. Hence, I analyzed weekly average wet deposition data from six National Atmospheric Deposition Program/National Trends Network (NADP/NTN) locations in and around the state of Indiana (Figure 1) between the dates January 1985 and December 2002 to examine spatial variability of precipitation amount and composition and the magnitude and potential causes of temporal trends. Application of a student t-test for the difference in means and examination of the probability distribution of weekly precipitation indicates the northern three sites (IL11, IN41 and IN20) experience a higher frequency of low precipitation amounts and lower annual total precipitation amounts. Ionic concentrations in precipitation also exhibit statistically significant spatial variability, and in accord with previous research, due to the quasi-exponential dependence of ion concentrations on precipitation amount, the northern sites exhibit a greater frequency of highly concentrated samples. Overall, annual average sulfate concentrations in precipitation decreased and average of 54% from 1985-2002. Over the same time period, nitrate concentrations decreased and average of 14%. Ammonium concentrations exhibited spatial differences, with increases at the southern sites and decreases at the northern sites and an overall increase of 5% (see Figure 2).

Previous research has demonstrated that variations in synoptic-scale meteorology can significantly impact temporal variability of air quality parameters, and this combined with the strong influence of precipitation amount on ionic concentrations led to the hypothesis that some of the temporal trends in precipitation composition may be attributable to meteorological variability rather than emission controls. Analysis of annual precipitation amount at the NADP/NTN sites indicates a high degree of inter-annual variability and some evidence of temporal trends. Hence, weekly ionic concentration data were assigned synoptic types developed by Polderman and Pryor (2004) to identify which synoptic types are associated with highly concentrated precipitation. The annual frequencies of these synoptic types were then examined for evidence that they exhibited trends within the data record. The results indicate that some of the types did exhibit significant changes in annual frequency and hence that meteorological effects may be confounding the trends in ion concentrations.

To further test this hypothesis least squares linear regression equations were developed for annual mean precipitation composition based on only a time index. As an example, the variance was 0.6 for sulfate concentrations at IN22 indicating that 60% of the variance of annual mean sulfate concentration in precipitation at this site may be attributed to a downward trend (and by association emission controls). However, when annual precipitation is added as an addition predictor variable the variance explanation increases to 0.75, indicating that the inter-annual variability of precipitation amount offers additional explanation for variations in annual precipitation chemistry. My presentation will summarize the research and discuss some of the difficulties in ‘declimatizing’ air quality data sets for meteorological variability.