Estimating Air-Water Exchange of Nitric Acid

Valigura, R.A. 1995. An iterative bulk exchange model for estimating air-water transfer of HNO3. J. Geophys. Res. 100:26,045-26,050.

Background

Air-water exchange rates have been estimated for most nitrogen species over open ocean, however, these rates may not apply to coastal areas due to different meteorological conditions. In coastal areas, the primary obstacle to estimating the exchange rate has been the lack of near surface, over-water meteorological data for mesoscale model input. In response to this need, there has been an increased deployment of measurement buoys along the east coast of the United States. One such network of buoys is the Chesapeake Bay Observing System (CBOS) owned and operated by the University of Maryland's Horn Point Laboratory. The two objectives of this project were to, i) develop and evaluate an iterative bulk exchange model to estimate air-water exchange of heat, water and momentum from buoy data, and ii) use the model outputs to estimate air-water transfer rates of nitric acid (HNO3). In a previous study, a similar approach to that used in this study was applied to sulfur dioxide (SO2). Because of a high affinity for water and relatively high ambient concentrations in coastal areas, HNO3 is considered to be the primary nitrogen species of interest for deposition directly to water surfaces. Given its affinity for water, HNO3 transfer can be considered uni-directional (i.e., downwards). Other than the use of bulk-water temperatures, the parameterizations used here are essentially identical to those used by Hicks and Liss [1976] and Wesely et al. [1981]. Three days of mean meteorological data and eddy correlation measurements of heat, moisture and momentum fluxes were used to test the CBOS model. These data were collected on tower, boat and airplane platforms from 16-20 June, 1990 near NOAA's Looe Key National Marine Sanctuary in Florida [Crawford et al., 1993]. The mean data were incorporated into the bulk exchange model and the resulting output was compared, favorably, against the eddy correlation data collected from the airplane [Crawford et al., 1993].

Experimental Set-Up

In late March 1992, the CBOS buoy was anchored in the north Chesapeake Bay, off Howell Point, Maryland (39.36oN 76.1oW) for 3 periods between April 1992 and July 1994 (excluding the winter months) totaling 18 months. The buoy is of the wave-follower variety, and allowed data to be telemetered to a central computer system for remote access. The cross section of the Bay narrows to 4.5 km in the area of the buoy, making the fetch 4.5 km or greater in the geographical window from 210o to 50o, and 1 km or less in the window from 51o to 209o. Water depth is variable across this section of Chesapeake Bay (average depth 5 m), but the buoy is anchored in the deep channel where the depth of water is 12.5 m. The buoy was instrumented to provide input data for the bulk-exchange model. The instruments included: i) air temperature at 4 m above the water, ii) water temperature at 0.75 m below the surface, iii) wind speed at 4 m, and iv) relative humidity at 4 m.

Buoy Results and Discussion

Deposition velocities were calculated for the approximately 25,000 10-minute periods that comprise the CBOS dataset to date. The overall frequency distribution is presented in Figure 1. When viewed on shorter time scales, the distribution begins to change, as is shown for the first week in December 1993 (Figure 1). The potential danger of using average/general deposition velocities in short term analysis becomes apparent when the actual time series of Vd is reviewed for the same week in December 1993, Figure 2. Meyers and Yuen [1987] found that concurrent, high resolution Vd and concentration data improved estimates of O3 deposition, but did not improve estimates of SO2 deposition. The primary reason for the difference between the two chemicals was that the measured variability in O3 concentration was significantly correlated with the corresponding measured variability in Vd (i.e, concentrations were proportional to deposition rates), and concentrations of SO2 were not correlated with corresponding Vd (i.e., concentrations were not proportional to deposition rate). It is unknown if HNO3 concentrations ([HNO3]) over Chesapeake Bay are correlated with Vd ([HNO3] was not measured during this study). To illustrate the potential errors associated with using average or time series Vd, a simple matrix analysis was performed using the December 1993 time series, Table 1. Deposition was estimated using three different Vd regimes (the actual times series Vd, the time series average Vd, and the 1993 average Vd), and three different HNO3 concentrations representative of [HNO3] commonly measured within the Bay region. The concentrations were i) switched from a constant low (1.2 ppb) to a constant high (2.7 ppb) after the front, ii) switched from a constant high to a constant low after the front, and iii) maintained at the average (1.95 ppb). The deposition estimate derived from using the time series average Vd and average [HNO3] was used as the reference value for intercomparisons, Table 1.

Tabulated comparison between deposition (g HNO3 m-2) estimates (Vd * [HNO3]) derived using different averaging schemes and the time series average* estimate for the first two weeks of December 1993 (percent differences are shown in parenthesis).

                            Low-High          High-Low          Average     
                         1.2 to 2.7 ppb    2.7 to 1.2 ppb      1.95 ppb     
 Actual Time Series Vd     .85 (4.5)        .443 (-4.5)       .464 (0.0)    
Time Series Average Vd    .429 (-7.5)        .499 (7.5)          .464*      
      (.0049ms-1)                                                           
1993 Annual Average Vd     .567(22.0)       .661 (42.0)       .614 (32.0)   
      (.0065ms-1)                                                           

The analysis shows that if Vd and [HNO3] are not correlated, the error in deposition estimates is primarily driven by errors in estimating mean values of Vd and [HNO3]. If they are correlated, determining the source the associated errors is more complex. If [HNO3] and Vd are correlated, adequate estimation of deposition can only be obtained by concurrent measurements of [HNO3] and Vd.

The Vd distribution (Figure 1) is likely to be conservative for two reasons. The first and most important reason is that the winter months are not accounted for because the buoy is removed due to ice. Another reason to believe that these estimates are conservative is that assuming equivalent transfer rates for HNO3 and heat does not adequately account for scavenging of HNO3 by aerosol water droplets and particles, which tend to increase deposition rates. There are theoretical limitations with this approach as well: lack of homogeneous conditions in the coastal zone and the inadequacy of similarity theory to describe turbulent conditions measured. Because the northern Chesapeake Bay is narrow, the local landscape tends to affect the meteorology making fetch assumptions unreasonable, thereby making it difficult to assume that the "local" buoy measurements are representative of any sizeable area. The second theoretical consideration concerns the Monin-Obukhov similarity theory upon which the bulk transfer equations are based. This theory has been evaluated, and is considered valid, over a certain range (-1 z/L 1) of meteorological conditions. During the two years of data collection, there were periods where model outputs showed that conditions were too stable/unstable to be considered in the "normal" Monin-Obukhov frame. These periods were closely related to low wind speeds.

The approach used in this study has been shown to be applicable to SO2 and should be applicable to other hygroscopic chemicals such as ammonia. To improve upon this technique, further eddy correlation projects must be performed to evaluate/modify the bulk transfer equation assumptions under low wind conditions. In future investigations, concurrent evaluation of HNO3 concentrations will allow for the quantification of the actual differences between the time-series and the single deposition velocity approaches.

Acknowledgments

This work was funded through the National Research Council Assistantship Program, and was made possible by the Chesapeake Bay Observing System buoy network owned and operated by the University of Maryland Horn Point Environmental Laboratory.

References

Crawford, T.L., R.T. McMillan, T.P. Meyers and B.B. Hicks, The spatial and temporal variability of heat, mass, and momentum air-sea exchange in a coastal environment, J. Geophys. Res., 98, 12869-12880, 1993.

Hicks, B.B. and P.S. Liss, Transfer of SO2 and other reactive gases across the air-sea interface, Tellus, 28, 1976.

Meyers, T.P. and T.S. Yuen, An assessment of averaging strategies associated with day/night sampling of dry-deposition fluxes of SO2 and O3, J. Geophys. Res., 92, 6705-6712, 1987.

Wesely, M.L., Heat transfer through the thermal skin of a cooling pond with waves, J. Geophys. Res., 84, 3696-3700, 1979.