Section 3.0 Air Quality Data and Trends
Ozone monitoring stations in the Poughkeepsie, NY ozone non-attainment area are listed in Table 1 in Appendix A. It should be noted that there are no ambient ozone precursor measurements in this area.
As listed in Table 1 in Appendix A, ozone is measured at three locations -Millbrook in Dutchess County (3602790007), Valley Central in Orange County (360715001), and Mt. Ninham in Putnam County (360790005). The monitored four highest concentrations, both for 1-hr and 8-hr, and the number of exceedance days from 2000 to 2006 are listed in Table 2. Table 3 in Appendix A lists the calculated design values for the period of 2000 to 2004 that are averaged to yield the base case design value (DVC).
3.2 Emissions - Anthropogenic
The 2002 base year emissions inventory has been compiled as part of the regional modeling effort and the details are reported in TSD-1c (2007) in Appendix E and in Pechan (2006). Table 4 in Appendix A lists 2002 emissions by major source category and Figure 3-1 in Appendix A displays emissions in graphical form. The 2009 projected year emissions inventory (2009OTW) was compiled as part of the regional modeling effort and the details are reported in TSD-1d (2007), TSD-1f (2007), TSD-1j (2007) (all in Appendix E) and MACTEC (2007). Emissions were projected based on growth and control and in the case of point sources, are provided as 3 distinct sectors, namely 1) emissions from electric generation units (EGU), 2) emissions from other point sources (Non-EGU), and 3) emissions from non-fossil fuel units (Non. Foss.). Table 5 in Appendix A lists 2009OTW emissions by county and by source sector. In addition to 2009 OTW emissions, there were emissions generated which have been identified as 2009 beyond on the way (2009BOTW) that reflects additional emissions reduction measures being undertaken by the OTC states. In this case, emissions changes were limited to the non-EGU and Area sectors only and the resulting 2009BOTW emissions for these sectors are listed in Table 6 in Appendix A. It should be noted that these emissions data are then processed using SMOKE for use as input to the photochemical model, CMAQ, to simulate ozone over the domain.
3.3 Emissions - Biogenic
Biogenic emissions over the modeling domain were calculated using SMOKE2.1 that incorporated Biogenic Emissions Inventory System (BEIS) v3.1.2. Details of the approach are described in TSD-1b (2007) in Appendix E. Briefly, the method utilized surface temperatures generated by the mesoscale meteorological model (TSD-1a 2007) and gridded land use and emissions factors data provided in SMOKE. These estimated emissions were used in all photochemical model (CMAQ) applications. Table 7 in Appendix A lists the annual emissions for the Poughkeepsie non-attainment area.
The 2002 annual meteorology using MM5 was developed as input data for photochemical model CMAQ. Details of MM5 setup and assessment can be found in TSD-1a (2007) in Appendix E.
3.5 Photochemical Model Application
3.5.1 Base Year
The five month period covering May through September 2002 was examined explicitly for ozone, with a 15-day ramp-up period since this simulation would also be utilized as part of the annual simulation for PM2.5 under the one-atmosphere modeling approach. The model assessment on a regional basis can be found in TSD-1e (2007) in Appendix E, which shows that the simulation can be considered satisfactory in reproducing the observed ozone distribution. Eder et al (2003) suggested that overall normalized mean bias (NMB) should be less than 10% and normalized mean error (NME) of 20% as possible indicators of acceptable model performance for ozone.
The statistical measures applied in this analysis are
Observed average, in parts per billion (ppb):
Predicted average, in ppb (only use Pi when Oi is valid):
Correlation coefficient, R2:
Normalized mean error (NME), in %:
Root mean square error (RMSE), in ppb:
Fractional error (FE), in %:
Mean absolute gross error (MAGE), in ppb:
Mean normalized gross error (MNGE), in %:
Mean bias (MB), in ppb:
Mean normalized bias (MNB), in %:
Mean fractionalized bias (MFB), in %:
Normalized mean bias (NMB), in %:
In particular for this non-attainment area, the assessment is performed with measurements based on the three ozone monitors and the results of the statistical measures are listed in Table 8 in Appendix A for two observed threshold levels of 40 and 60 ppb. Note that these threshold levels restrict the number of days with observed daily maximum 8-hr ozone that is used in this assessment. Results listed in Table 8 in Appendix A indicate that the estimated NME at these monitors is at an acceptable level while the NMB values are slightly higher than that suggested by Eder et al (2003).
3.5.2 Future Year 2009
Photochemical modeling was performed in a manner similar to that of the base year. The intent of this modeling is to use the predicted ozone concentrations relative to the base year and estimate the future design value at the monitored locations as well as other areas of the non-attainment area. The approach used has been documented in EPA Guidance documents (EPA 2005 and EPA 2006) and its application is described in TSD-1g (2007) and in TSD-1h (2007) in Appendix E. Table 9 in Appendix A summarizes the information on the estimated relative reduction factor (RRF) and the projected future design values that are estimated to be below the 8-hr ozone NAAQS of 84 ppb.
3.5.3 Unmonitored Area Analysis
As per EPA guidance (2005, 2006a), the potential occurrence of currently unmonitored locations that are projected to be above the 8-hr ozone NAAQS for this non-attainment area was investigated. The procedure suggested is to examine all grid cells for all counties within and immediately surrounding the non-attainment area using the spatial interpolation and gradient adjustment techniques implemented in the EPA-MATS (Model Attainment Test System) software (Timin, 2006).
In this application, MATS was utilized to spatially interpolate base year observed design values. MATS was also utilized to estimate gradient adjustment factors that were based on the CMAQ predictions of the top-30 daily maximum 8-hr ozone concentrations at each grid cell for the 2002 base case. The relative effect of the emission reduction under the 2009BOTW scenario on daily maximum 8-hr concentrations was then estimated by calculating a gridded field of RRF by treating each grid cell as a monitor location. Two approaches were used for calculating the RRF. MATS is used to provide RRF at each grid cell, and the other approach is based on 9-grid cells as described in TSD-1g and TSD-1h in Appendix E. Finally, DVF for each grid cell is estimated by multiplying the spatially interpolated DVB (from MATS) with the gridded gradient adjustment factors (from MATS) and with the gridded RRF fields estimated by the two methods.
The Poughkeepsie non-attainment area, comprised of Dutchess, Orange and Putnam counties abuts the counties of Sullivan and Ulster in the north and west and Rockland and Westchester counties to the south. It should be noted the Rockland and Westchester counties are part of another ozone non-attainment area and as such are not considered in this analysis.
Table 10 in Appendix A lists all the counties pertaining to the Poughkeepsie non-attainment area and surrounding counties by the corresponding grid cells in the CMAQ modeling domain. The Table also provides information as to whether or not the grid cell is associated with an ozone monitor and the percent of the grid area located over water based upon the land classification used in the meteorological modeling with MM5. All grid cells listed indicate that the projected future design value estimated by either of the methods would be at or below the 8-hr ozone NAAQS of 84 ppb.
3.6 Weight of Evidence
The model-based attainment projects that the 8-hr ozone design values for 2009 for Millbrook (360270007), Valley Central (360715001) and Mt. Ninham (360790005) are 80 ppb, 73 ppb and 81 ppb, respectively. The current design values through 2006 are listed in Table 12 in Appendix A, with 2006 DVs of 74 ppb, 80 ppb and 84 ppb for Millbrook, Valley Central and Mt. Ninham, respectively. These levels are indicative that the area is currently in attainment of the 8-hr ozone NAAQS.
The EPA recommended method of estimating the base year design value (DVc) for the period of 2000 to 2004 is a weighted average approach that weighs 2002 measurements much more than the other years. Another method is to estimate the base year design values is the use of average of the 5-year period. For this approach the 4th highest concentration listed in Table 2 in Appendix A are utilized and average DVc is listed in Table 12 in Appendix A for each of the monitors. The projected design values are estimated using the RRF values from Table 10 in Appendix A and are included in Table 12 in Appendix A. The estimated design values using this method are well below the 8-hr ozone NAAQS, suggesting that this area should be in attainment of the 8-hr ozone NAAQS in 2009.
In addition, trends were examined in the hourly ozone concentrations at these three monitoring stations (TSD-aa 2007) and the results are listed in Table 13 in Appendix A. The estimated trend is found to be strongly dependent upon the length of the data that is being considered in the analysis. The estimated trend (percent per year) at Millbrook and Mt. Ninham is downward for both with and without meteorological adjustment for the overall monitoring period, while the estimated trend for Valley Central is found to vary depending upon the length of the data with both a downward and upward trend. However, it should be noted that the measured design values for Valley Central (see Table 11 in Appendix A) vary over the 5-year period in contrast to the other two monitors in the region and this is reflected in the estimated trend. The downward trend in the 2000 to 2005 period for the two of the three monitors appears to be consistent with that reported by the EPA (2006b) in which different methods were utilized to estimate the percent change in 8-hr ozone over the NOx budget trading program states.
This study shows that based upon the projected emissions inventory and the photochemical modeling, the Poughkeepsie 8-hr ozone non-attainment area would be in attainment.
Eder, B., and S. Yu (2003) An evaluation of the 2003 release of Models-3/CMAQ, presented at the 2003 Annual CMAS Workshop, Research Triangle Park, NC.
EPA (2005) Guidance on the Use of Models and Other Analyses in Attainment Demonstrations for the 8-hour Ozone NAAQS. EPA-454/R-05-002.
EPA (2006a) Guidance on the use of Models and Other Analyses for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.5 and Regional Haze. Draft 3.2-September 2006.
EPA (2006b) http://www.epa.gov/air/airtrends/2006/ozonenbp/
MACTEC (2007) Development of Emission Projection for 2009, 2012, and 2018 for nonEGU point, area, and nonroad sources in the MANE-VU region. www.marama.org/reports
Pechan: (2006) Technical Support document for 2002 MANE-VU SIP Modeling inventories, version 3. Prepared by E. H. Pechan & Associates, Inc. 3622 Lyckan Parkway, Suite 2005, Durham, NC 27707.
Timin, Brian (2006 Communication (e-mail) of release of beta version of MATS
TSD-1a (2007) Meteorological modeling using Penn State/NCAR 5th generation mesoscale model (MM5)
TSD-1b (2007) Processing of Biogenic Emissions for OTC/MANE-VU Modeling
TSD-1c (2007) Emissions processing for the revised 2002 OTC Regional and Urban 12km base case Simulation
TSD-1d (2007) 8hr ozone modeling using the SMOKE/CMAQ system
TSD-1e (2007) CMAQ model performance and assessment 8-hr OTC Ozone Modeling
TSD-1f (2007) Future Year Emissions Inventory for 8-hr OTC Ozone Modeling
TSD-1g (2007) Relative response factor (RRF) and "modeled attainment test"
TSD-1h (2007) Projected 8-hr ozone air quality over the ozone transport region
TSD-1j (2007) Emission Processing for OTC 2009 OTW/OTB 12km CMAQ simulations
TSD-aa (2007) Trends in Measured 1-hr Ozone Concentrations over the OTR modeling domain