Technical Support Document (TSD) for the Response to the September 2010 Section 126 Petition from New Jersey Regarding Sulfur Dioxide (SO2) Emissions from the Portland Generating Station
                                       
                      Docket ID No. EPA-HQ-OAR-2011-0081
                                       
                                       
                                       
                                       
                                       
                                       
  Trajectory Analysis of High SO2 Episodes at the Chester, New Jersey Monitor
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                      U.S Environmental Protection Agency
                                        
                  March 2011ADDITIONAL SUPPORTING INFORMATION

Trajectory Analysis of High SO2 Episodes at the Chester, NJ Monitor
   
A. Summary of trajectory analysis submitted by NJDEP

   The New Jersey Department of Environmental Protection (NJDEP) submitted an analysis using air parcel trajectories on days with high concentrations to support its assertion that the Portland Plant causes violations of the 1-hour SO2 national ambient air quality standard (NAAQS) in New Jersey.  NJDEP examined air trajectories from two episodes with high SO2 concentrations recorded at the nearby Chester ambient air monitor in 2008 and 2009.  The Chester monitor (34-027-3001) is located in Morris County, New Jersey, approximately thirty-four kilometers from the Portland Plant.  The Chester monitor is part of New Jersey's regular State and Local Air Monitoring Station network and was situated to monitor the exposure of the general population of the area to SO2 and was not designed to monitor the specific impact of the Portland Plant or any other impact of any particular emissions source. 
   
   
B.  HYSPLIT Model selection

NJDEP used the Air Resources Laboratory's HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model  as a tool to help determine if high concentrations of SO2 are coming from the Portland Plant.  HYSPLIT is a complete system for computing both simple air parcel trajectories and complex dispersion and deposition simulations.  The model calculation method is a hybrid between the Lagrangian approach, which uses a moving frame of reference as the air parcels move from their initial location, and the Eulerian approach, which uses a fixed three-dimensional grid as a frame of reference.  In the model, advection and diffusion calculations are made in a Lagrangian framework following the transport of the air parcel, while pollutant concentrations are calculated on a fixed grid.  

In this case, NJDEP used the Lagrangian features of the model to determine the direction and speed of the air.  When NJDEP applied the HYSPLIT model, it used a meteorological data set with a grid size of 12 kilometers, which is the finest resolution data set available for use with the model.   The HYSPLIT model's grid size is less than the distance between the Portland Plant and the Chester monitor, so the model is using a data set that is more likely to be an appropriate tool for this analysis.   EPA agrees that using the HYSPLIT model as NJDEP applied it provides useful information on the direction and speed of air parcels in the atmosphere. 

C.  NJDEP's Review of Monitoring Data and Nearby Emission Sources
      NJDEP reviewed air monitoring data for days when high concentrations occurred at its Chester, NJ monitor located 34 kilometers east-southeast of the Portland Power Plant.  The monitoring data from the Chester, NJ monitor have been fully quality controlled, validated and certified by NJDEP, as required by EPA.  The data are available in EPA's Air Quality System (AQS) database.
      During the years 2008 and 2009, the highest 1-hour SO2 concentrations occurred in two episodes, July 17-18, 2008 and December 7, 2009.  During the July 2008 and the December 2009 episodes, the Portland reported that the Plant was operating at 85 and 55 percent of capacity, respectively.  NJDEP's review of emission inventories from its emissions database and Pennsylvania's PADEP eFACTS database show that yearly emissions from Portland are 1000 times greater than any other sources between the Chester monitor and the Portland Plant.  In the area around surrounding the Portland Plant, the Portland Plant's yearly emissions are 1000 times greater than any other source in the area, except for the Martins Creek Power Plant.  Hourly continuous emissions monitoring systems (CEMS) data from EPA's Clean Air Markets database (see http://www.epa.gov/airmarkets/) show that the Martins Creek Power Plant was operating at very low levels during these two episodes, with actual emissions recorded well under the level of emissions measured at the Portland Plant. 
      
D.  Episode I:  July 17-18, 2008

The 1-hour SO2 concentrations measured during Episode I were the highest three hourly concentrations recorded at the Chester monitor during 2008-2009.  Data in EPA's Air Quality System show that during three consecutive hours beginning at 10 pm local time on the 17[th], hourly SO2 concentrations ranged from 85 parts per billion (ppb)  to 53 ppb.  The first two hours of the episode exceeded the 75 ppb 1-hour SO2 NAAQS.  NJDEP included plots of these data showing a rapid rise at the beginning of the episode and drop off at the end of the episode of SO2 values; values were at or near zero ppb except during the air quality episode.   

NJDEP used the HYSPLIT model to calculate forward trajectories from the Portland Plant at a height representative of the plume emitted from its stacks (i.e., 100 meter plume rise added to the 121 meter stack height), and backward trajectories at a height representative of the Chester monitor (i.e., 10 meters above ground) during the time periods associated with high SO2 concentrations at the Chester monitor on July 17, 2008.  New Jersey submitted forward and back trajectories for a 6 hour period (to allow for the transport of the plume from the stack to the Chester monitor) based on wind speeds of 4.5 meters per second (m/s) at 221 m above ground level (near plume height) to 2.7 m/s at 10 m above ground level (near the monitor's height) during the episode.  

E.  Episode II:  December 7, 2009
   
The 1-hour SO2 concentrations measured during Episode II included four of the twelve highest hourly concentrations in the Air Quality System for the Chester monitor during 2008-2009.  During four consecutive hours beginning at 1am local time on the 7[th], hourly SO2 concentrations ranged from 28 to 33 ppb.  Hourly SO2 during this episode rose and fell more gradually; concentrations were less than 10 ppb on either side of the seven-hour episode.   

NJDEP used the HYSPLIT model to calculate forward trajectories from the Portland Plant at a height representative of the plume emitted from its stacks (i.e. 100 meter plume rise added to the 121 meter stack height), and backward trajectories at a height representative of the Chester monitor (i.e. 10 meters above ground) during the time periods associated with high SO2 concentrations at the Chester monitor on December 7, 2009.   NJDEP submitted forward and back trajectories for a 3 hour period (to allow for the transport of the plume from the stack to the Chester monitor) based on wind speeds of 11.2 m/s at 221 m above ground level (near plume height) to 3 m/s at 10 m above ground level (near the monitor's height) during the episode.  

F.  Possible Sources Affecting the Readings Observed at Chester

At the time of these two episodes, emissions from the Portland Plant were vastly greater than other sources between Chester and the Portland Plant or around the Portland Plant.  The only other facility in the vicinity showing potentially significant amounts of SO2 is the Martins Creek Power Plant, however, it was operating at levels well less than the Portland Plant, during the two episodes. 

The following tables show the hourly emissions from Portland Units 1 and 2, and from Martins Creek Units 3 and 4 during the two air quality episodes. 

Table 1 Hourly Sulfur Dioxide Emissions on July 17, 2008

Hour (EDT)
Portland Unit 1
(lbs/hr)
Portland Unit 2
(lbs/hr)
Martins Creek Units 3 & 4 
(lbs/hr)
15
5,024
7,574
35
16
5,041
7,536
42
17
5,051
7,436
33
18
5,013
7,569
21
19
5,003
7,556
16
20
4,969
7,580
16
21
4,913
7,532
9
22
4,986
7,548
0
23
4,475
6,792
0


Table 2 Hourly Sulfur Dioxide Emissions on December 6 and 7, 2009

Hour (EDT)
Portland Unit 1
(lbs/hr)
Portland Unit 2
(lbs/hr)
Martins Creek Units 3 & 4
(lbs/hr)
Dec 6-18
3,503
5,326
0
19
3,554
5,818
0
20
3,493
5,819
0
21
3,352
5,611
0
22
2,830
4,612
0
23
2,856
4,535
0
Dec 7-00
3,192
4,617
0
01
3,305
4,653
0
02
3,423
4,532
0


The forward trajectories provided by NJDEP show that the SO2 emissions from the Portland Plant arrive in the vicinity of the Chester monitor during high concentrations at the monitor.  For example, winds from the Portland Plant at hour 15 (4 pm EDT) on July 17, 2008, end during the three-hour episode of high SO2 concentrations at the Chester monitor.  In addition, trajectories starting at hour 17 (5 pm EDT) on July 17, 2008 at Chester and going backward end up near the Portland Plant.  The trajectories submitted by NJDEP visibly show, that high concentrations at Chester were associated with a source in the vicinity of the Portland Plant.   Since the Portland Plant is the only large emission sources in its vicinity, it is likely that emissions from the Portland Plant caused the high concentrations during these episodes.  

G.   EPA analysis
      NJDEP's method of using HYSPLIT trajectory analyses during the two episodes of high concentrations of SO2 at the Chester NJ air monitoring station uses acceptable methods that support the conclusion that SO2 that was recorded at the Chester monitor was mainly due to emissions from the Portland Plant.  The trajectories, run forward and backwards, show air coming from the vicinity of the Portland Plant to the area near the Chester monitor.  The Martins Creek Power Plant is the only facility in the area that has potential SO2 emissions on the order of the potential emissions of the Portland Plant.  However, during these episodes, emissions from the Martins Creek Power Plant are less than one percent of the emissions of the Portland Plant.  Thus, it is unlikely that any facility other than the Portland Plant caused these high concentrations.    
   
      NJDEP's analysis of monitoring data and wind trajectories provides supporting evidence that the Portland Plant has a noticeable influence over a wide area around the facility.  For reasons explained in the preamble to the proposed rule, EPA continues to rely primarily on modeling to determine whether downwind areas constitute nonattainment or maintenance receptors for purpose of analyzing a section 126 petition with respect to the 1-hour SO2 NAAQS.  Nonetheless, elevated concentrations, of SO2 appear to come from the Portland Plant,  observed at a monitor distant from the area of Portland Plant's peak impact.  This finding underscores the need for EPA to approve the petition and to establish limits and a compliance schedule to address these elevated concentrations of SO2 and  is consistent with EPA's finding that the Portland Plant contributes significantly to and interferes with maintenance of the 1-hour SO2 NAAQS in New Jersey.  




