next up previous contents
Next: 6.6 Discussion Up: 6. CHARACTERIZATION AND REDUCTION Previous: 6.4 Formulation of the

6.5 Case Studies

The RPM-IV and RPM-3D are both applied to two significantly different case studies. An attempt is made to understand the extent of the uncertainty associated with the model assumptions and model resolution, by comparing the predictions of these two models.

The first case study uses emission estimates and field data measured near Marathon Oil Refinery at Robinson, Illinois, during June and July, 1977 [183]. This case study involves a primarily VOC dominant source (i.e., the volatile organic compounds emissions are significantly higher than the NO/NO$_2$ emissions). The meteorological data and initial plume concentrations were obtained from the work of Georgopoulos and Roy [79], which was based on the measurements reported by Sexton et al. [183].


  
Figure 6.4: Dependence of plume average O$_3$ concentration on horizontal resolution of the RPM-IV in a VOC dominant regime
\begin{figure} \centerline{\epsfig{figure=marhorres.eps,width=5in}} \par\end{figure}


  
Figure 6.5: Dependence of plume average O$_3$ concentration on vertical resolution of the RPM-3D in a VOC dominant regime (4 horizontal cells)
\begin{figure} \centerline{\epsfig{figure=marverres4.eps,width=5in}} \par\end{figure}


  
Figure 6.6: Horizontal profile of the O$_3$ concentration in the plume in a VOC dominant regime (6 horizontal cells, RPM-IV)
\begin{figure} \centerline{\epsfig{figure=marhorpro6.eps,width=5in}} \par\end{figure}


  
Figure 6.7: Vertical profile of the O$_3$ concentration at the plume centerline (w.r.t. horizontal) in a VOC dominant regime (4 horizontal cells and 6 vertical cells, RPM-3D)
\begin{figure} \centerline{\epsfig{figure=marverpro4.eps,width=5in}} \par\end{figure}

Figure 6.4 shows the average ozone concentrations in the plume, calculated by RPM-IV, as a function of the plume evolution time for different horizontal resolutions, ranging from two cells to ten cells in the horizontal. Similarly, Figure 6.5 shows the average plume concentrations calculated by RPM-3D for different vertical resolutions with four cells in the horizontal. The calculations from RPM-IV, which has no vertical resolution, and RPM-3D with vertical resolution corresponding to two cells, four cells, and six cells, are shown in the figure. The results indicate that the different vertical and horizontal resolutions produce similar estimates of the average plume concentrations. Hence, it is sufficient to use a low-resolution two-dimensional model, for studying the plume average concentrations in such cases.

Figure 6.6 shows the ozone concentrations within each horizontal cell calculated by RPM-IV when the plume cross section is divided into six cells. The plots indicate the average ozone concentrations in the different horizontal cells (indicated by ``A'', ``B'', and ``C'' in Figure 6.3 for RPM-IV). Similarly, Figure 6.7 shows the vertical profile of the ozone concentrations calculated by RPM-3D for the centerline horizontal cell when the plume is divided into four horizontal cells and each cell is in turn divided into six vertical cells. The plots indicate the average of the vertical cells along the horizontal centerline, and individual vertical cells at the horizontal line. The cells are illustrated in Figure 6.3. The horizontal centerline cell is given by ``A'', and the corresponding vertical slices are ``AZ'' (outermost), ``AY'' (middle), and ``AX'' (centerline). Examining the results, it is clear that in a NO$_x$ dominant regime, this would clearly produce an under-estimation of the ground level ozone levels. Examining the results, it is clear that by ignoring the vertical structure of a plume, the RPM-IV approximates the highest possible ground level concentration by the average in the vertical. Hence, in a VOC dominant regime, this would clearly produce an over-estimation of the ground level ozone levels.

The second case study considers a single point source in Mercer County, New Jersey. The simulation is performed for the period of July 7, 1988, 8:00 am to 10:00 am, corresponding to the occurrence of a severe ozone episode in the region. This case study involves a NO/NO$_2$ dominant regime. Ambient pollutant concentrations were obtained from the results of an Urban Airshed Model (UAM-IV) simulation for the Philadelphia/New Jersey modeling domain for the time period of July 6 to July 8, 1988, described by Georgopoulos et al. [81]. The meteorological inputs were obtained from the model inputs of the UAM-IV simulation for the modeling domain [81]. The model results are analyzed in the same manner as described in the earlier case study involving a VOC dominant regime.


  
Figure 6.8: Dependence of plume average O$_3$ concentration on horizontal resolution of the RPM-IV in a NO$_x$ dominant regime
\begin{figure} \centerline{\epsfig{figure=merhorres.eps,width=5in}} \par\end{figure}


  
Figure 6.9: Dependence of plume average O$_3$ concentration on vertical resolution of the RPM-3D in a NO$_x$ dominant regime
\begin{figure} \centerline{\epsfig{figure=merverres4.eps,width=5in}} \par\end{figure}


  
Figure 6.10: Horizontal profile of the O$_3$ concentration in the plume in a NO$_x$ dominant regime (6 horizontal cells, RPM-IV)
\begin{figure} \centerline{\epsfig{figure=merhorpro6.eps,width=5in}} \par\end{figure}


  
Figure 6.11: Vertical profile of the O$_3$ concentration at the plume centerline (w.r.t. horizontal) in a VOC limited (NO$_x$ dominant) regime (4 horizontal cells and 6 vertical cells, RPM-3D)
\begin{figure} \centerline{\epsfig{figure=merverpro4.eps,width=5in}} \par\end{figure}

Figure 6.8 shows the average ozone concentrations in the plume, calculated by RPM-IV, as a function of the plume evolution time for different horizontal resolutions, ranging from two cells to ten cells in the horizontal. Similarly, Figure 6.9 shows the average plume concentrations calculated by RPM-3D for different vertical resolutions with four cells in the horizontal. The calculations from RPM-IV, which has no vertical resolution, and RPM-3D with vertical resolution corresponding to two cells, four cells, and six cells, are shown in the figure. The results indicate that the different vertical and horizontal resolutions produce similar estimates of the average plume concentrations. Hence, it is sufficient to use a low-resolution two-dimensional model, for studying the plume average concentrations in such cases.

Figure 6.10 shows the ozone concentrations within each horizontal cell calculated by RPM-IV when the plume cross section is divided into six cells. Similarly, Figure 6.11 shows the vertical profile of the ozone concentrations calculated by RPM-3D for the centerline horizontal cell when the plume is divided into four horizontal cells and each cell is in turn divided into six vertical cells. Examining the results, it is clear that in a NO$_x$ dominant regime, this would clearly produce an under-estimation of the ground level ozone levels.

next up previous contents
Next: 6.6 Discussion Up: 6. CHARACTERIZATION AND REDUCTION Previous: 6.4 Formulation of the
Sastry S. Isukapalli
1999-01-19