Evaluation of Data Replacement Strategies for CASTNET Dry Deposition Modeling

Abstract: The U.S. EPA (Environmental Protection Agency) established the CASTNET (Clean Air Status and Trends Network) and its predecessor, the NDDN (national dry deposition network), as national air quality and meteorological monitoring networks. Both CASTNET and NDDN were designed to measure concentrations of sulfur and nitrogen gases and particles. Both networks also estimate dry deposition using an inferential model. The design was based on the concept that atmospheric dry deposition flux could be estimated as the product of a measured air pollutant concentration and a modeled deposition velocity (Vd). The MLM (multi-layer model), the computer model used to simulate dry deposition, requires information on meteorological conditions and vegetative cover as model input. The MLM calculates hourly Vd for each pollutant, but any missing meteorological data for an hour renders Vd missing for that hour. Because of percent completeness requirements for aggregating data for long-term estimates, annual deposition rates for some sites are not always available primarily because of missing or invalid meteorological input data. In this work, three methods for replacing missing on-site measurements are investigated. These include (1) using historical values of deposition velocity or (2) historical meteorological measurements from the site being modeled or (3) current meteorological data from nearby sites to substitute for missing inputs and thereby improve data completeness for the network’s dry deposition estimates. Results for a CASTNET site used to test the methods show promise for using historical measurements of weekly average meteorological parameters.

Key words: Dry deposition, deposition velocity, leaf area index, MLM (Multi-Layer Model).

1. Introduction

The U.S. EPA (Environmental Protection Agency) established the CASTNET (Clean Air Status and Trends Network) to provide data for determining relationships between changes in emissions and any subsequent changes in air quality, atmospheric deposition, and ecological effects. The monitoring network emphasized the selection of rural sites. CASTNET has its origins with the National Dry Deposition Network (NDDN), which was established in 1986 and began operation in 1987. Many of the original NDDN sites are still operational after more than 20 years and provide useful information on trends in air quality.

CASTNET was designed primarily to provide long-term measurements for seasonal and annual average concentrations and depositions. Consequently, measurements of weekly average concentrations were selected as the basic sampling strategy. An open-face, three-stage filter pack was employed to measure gaseous and particulate sulfur and nitrogen pollutants as well as concentrations of other pollutant species. The filter pack technology and sampling protocol have been used consistently over the network’s history, providing a comparable data set each year and allowing for the analysis of long-term trends.

values by constructing histograms of percent differences. Comparisons were made using annual mean and 10-year mean values when either three or four quarters were valid (one quarter of the year may not be represented in the annual mean) and a second set of comparisons were made using 10-year mean and annual mean values when all four quarters were valid(the entire year is covered). Figs. 2 and 3 present histograms of percent differences versus number of site-years based on annual mean values when all four quarters were available. Fig. 2 is based on annual mean deposition velocities, and Fig. 3 is based on fluxes estimated from 10-year mean deposition velocities. Each histogram includes a bar for each of five pollutants: SO2, HNO3, SO42-, NO3- and NH4+.

Table 3 summarizes the results from the use of 10-year and annual mean fluxes. The two columns labeled percent of site-years within ±20% of the historical flux estimate show that for the analysis done when all quarters were valid, well over 90% of all site-years have a percent difference within the ±20% window, except for NO3-, which is lower.

The results for SO2, HNO3, NO3- and NH4+ show that the calculated fluxes are somewhat higher than the archived MLM results. The exception is for SO42-which shows lower modeled values. Use of the 10-year mean allowed for the recovery of 131 additional site-years (37%). Many of the missing site-years are the start-up or shut-down years for a specific site. Use of the annual mean allowed for the recovery of only 8 additional site-years (2%) but was

useful in examining how flux estimates calculated with annual average concentrations and deposition velocities are comparable with those aggregated using the current protocol.

Table 3 and Fig. 2 show that the simulations based on individual annual Vd values better match the MLM results. For SO2, approximately 99% of the calculated fluxes are within 20% of the MLM deposition rates, and about 60% of the calculated values are within 4% of the MLM results. Table 3 and Fig. 3 show that SO2 results based on the 10-year mean deposition velocities indicate that approximately 95% of the calculated fluxes are within 20% of the MLM values and about 36% are within 4%.

For HNO3, SO42-, NO3- and NH4+, Figs. 2 and 3 again show the flux estimates based on individual annual Vd values better match the MLM results. The fluxes based on 10-year mean Vd values for HNO3, SO42- and NH4+ are somewhat better than the SO2 results. More than 95% of the flux values based on 10-year mean Vd values are within 20% of the MLM annual values. The results for NO3- are not as good. Particulate nitrate concentrations are lower and more uncertain than the other parameters.

The analysis of the historical deposition velocities demonstrates that the use of 10-year mean Vd values and annual mean pollutant concentrations provides acceptable estimates of annual mean fluxes if a 20% uncertainty is acceptable.

3.3 Analysis of Replacement of Wind Speed and Sigma Theta Data with Historical Values

Model runs were performed by replacing missing vector wind speed and sigma theta data with either historical weekly mean values or “near site” values. The “near site” pair of LYK123 and DCP114 was selected for the analysis with LYK123 as the primary site. Model runs were completed for 2007 and then compared with the archived deposition velocity