Historical and Future Ozone Simulations 
using the MM5/SMOKE/CMAQ System 
in the Portland/Vancouver Area 
                                       
                                       
                                       
                                       
                                       
                                       
                                  Prepared by
                                       
                            Ying Xie and Brian Lamb
                      Laboratory for Atmospheric Research
              Department of Civil & Environmental Engineering
                          Washington State University
                                       
                                       
                                       
                                 Prepared for

                  Oregon Department of Environmental Quality
                          Southwest Clean Air Agency
                    Washington State Department of Ecology
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                               December 31, 2005
                               Table of contents
1 Executive Summary	3
2 Introduction	5
3 Ozone Modeling Study Goal and Design	5
3.1 Previous Ozone Modeling Work	5
3.1.1 Maintenance Plan EKMA modeling	6
3.1.2 Southwest Washington Ozone Study	6
3.1.3 NW Regional Technical Center Demonstration Project	6
4 Episode Selection	7
5 Meteorological modeling	9
5.1 Modeling methods	9
5.2 Meteorological modeling results and evaluation (July 1998 episode)	12
5.2.1 Surface Temperature	12
5.2.2 Surface winds	16
5.2.3 Soil moisture	19
5.2.4 PBL height and vertical profile	21
5.2.5 MM5 statistical evaluation	27
5.2.6 LSM run compared with simple soil model run	28
5.3 Modeling results and evaluation (July 1997 episode)	29
5.4 Back trajectory analysis	35
6 Photochemical air quality modeling	37
6.1 Modeling methods	37
6.2 CMAQ modeling results and evaluation (July 1998 episode)	39
6.2.1 CMAQ statistics evaluation	47
6.2.2 Comparisons with NOx/VOC monitoring data	52
6.3 CMAQ modeling results and evaluation (July 1997 episode)	58
6.3.1 CMAQ statistics evaluation	64
6.3.2 Comparisons with NOx monitoring data	69
6.4 Conclusion for July 1998 and 1997 CMAQ simulation	76
7 Future episode modeling (2015)	76
7.1 Future year base case	76
7.1.1 CMAQ modeling results	76
7.1.2 VOC/NOx ratios and regional increase of ozone	83
7.2 What if scenario - No lower Columbia River point sources	86
Reference	93
Appendix A. Emission Inventory Documentation	95
Appendix B: Area Source Category Emission Calculation Methodologies	108
Appendix C: MOBILE6.2 input files for the 1997 and 1998 episodes.	111
Appendix D:  Acronyms	125
                                       

1 Executive Summary

In order to develop quantitative estimates of ozone levels in future years, a modeling study was completed for the Portland, OR, and Vancouver, WA urban region.  This work was conducted as part of the State Implementation Plan development for this area.  Two historical episodes (July 17-21, 1997 and July 26-28, 1998) were simulated using the Mesoscale Meteorological model version 5/ Sparse Matrix Operator Kernel Emissions Modeling System / Community Multi-scale Air Quality Modeling System (MM5/SMOKE/CMAQ).  These two episodes were selected for this modeling work because the July 1998 episode had the highest ozone levels observed in recent years for the airshed and the July 1997 episode had slightly different meteorological patterns compared to the 1998 case.  In addition to the normal ozone observations, VOC/NOx monitoring data were also available for the July 1997 episode for comparison to the modeled results. 

As part of the base simulation for the July 1998 episode, tests were conducted to investigate the effects of using a more complex land-surface modeling option -- the Noah Land Surface Model, LSM -- within the MM5 meteorological model.  In these tests, using the LSM in MM5 produced a better simulation of surface temperatures and generally lower and more accurate planetary boundary layer (PBL) heights along the west side of the Cascades compared to using the traditional simple soil model.  As a result, air quality modeling results for ozone were also improved using MM5 with the LSM option.  For the CMAQ simulation, model performance statistics for most monitoring sites were within United States Environmental Protection Agency (EPA) suggested 1-hr and 8-hr ozone modeling guidelines.  Time series plots, scatter plots, and Quantile-Quantile plots also showed generally good agreement between model simulations and observations. .While model performance was acceptable for both episodes, the July 1998 episode was selected as the base for evaluation of future year (2015) managed growth maintenance projection simulations because observed ozone concentrations were much higher in 1998.
Two 2015 managed growth maintenance projection simulations were modeled. The first simulation analyzed changes in ozone formation with the emissions inventory changes in the future. The second simulation was a modification of the first designed to examine the influence of the industrial sources located along the lower Columbia River on the ozone formation in the Portland/Vancouver urban area. 

The future year (2015) emission inventory was developed using growth factors depending on the surrogates for each source category.  Federally mandated control strategies and foreseeable changes were also taken into account.   Total emissions for all pollutants were projected to decrease in the 2015 emissions inventory compared to the 1998 emissions inventory.  Furthermore, NOx emissions were projected to have a greater reduction than VOC emissions which resulted in an increase of the VOC/NOx emission ratio.  For 1998, the entire Portland urban area and I-5 corridor appeared to be in the VOC-sensitive region, based upon predicted O3/NOx concentration ratios.  For 2015, the VOC-sensitive areas were much smaller, which is consistent with the increased VOC/NOx ratio.  A large reduction in ozone was predicted to occur downwind of urban Portland in 2015 CMAQ simulation results compared to the 1998 modeling results.  The reduction of ozone levels within the urban core was less significant than at the downwind sites.  The second projected 2015 simulation which  assumed no large industrial activity along the lower Columbia River showed  reductions up to 12 ppb at the southwestern Washington observation sites when compared to the first 2015 managed growth maintenance projection;  sites in Oregon showed small or no differences associated with this simulation.

Further analysis of other potential control strategies, based upon modifications to the 2015 managed growth maintenance projection, will be conducted by ODEQ and SWCAA as part of the overall maintenance plan development.  






2 Introduction 

The current 1-Hour Ozone Maintenance Plans for the Portland/Vancouver airshed were adopted as part of the Oregon and Washington State Implementation Plans in 1997.  The control measures in those plans have helped the region maintain the now-revoked 1-hour ozone National Ambient Air Quality Standard (NAAQS).  The region is also in attainment of the 8-hour ozone NAAQS, the current ozone standard.  

However, due to exceedances of the 1-hour ozone standard in the recent past, the unique meteorological influences from the Columbia River Gorge, and the potential for ozone precursor transport in the region, regional air regulatory agencies proposed performance of photochemical modeling for the area as a basis for improving our understanding of the potential for ozone exceedances in the future.  This report presents the results of this modeling study.  The modeling system used for this work was the MM5/SMOKE/CMAQ system; it was applied to a domain consisting of 4 km x 4 km horizontal grid cells covering an area that extended from north of Puget Sound to south of Salem, OR and from the Pacific coast on the west to beyond the Cascade Mountain range on the east.  The modeling work involved simulation of two historical episodes, July 17-21, 1997 and July 26-28, 1998.  The model results were compared to available observations, and then the 1998 episode was used as a basis for evaluation of future year (2015) managed growth projections.  An important aspect of the modeling study was the evaluation of model sensitivity to use of different land/surface models within the MM5 meteorological model.  Results from these sensitivity studies are also described in this report.  

3 Ozone Modeling Study Goal and Design

The overall goal for this work was to improve understanding of ozone behavior in the Portland/Vancouver airshed and to obtain model results to help evaluate future year conditions and the effectiveness of potential control strategies.  The approach involved the following steps:  1) selection of suitable ozone episodes that represented worst-case or typical ozone conditions, 2) development of the necessary meteorological, emissions, and air quality observations data sets, 3) optimization of the modeling system to yield acceptable model performance for these episodes, and 4) use of the meteorological simulation from one of these episodes as a basis for evaluation of future year conditions and control strategy effectiveness.  This work was guided by the Portland/Vancouver ozone modeling technical committee consisting of representatives from Oregon Department of Environmental Quality (ODEQ), Southwest Clean Air Agency (SWCAA), Washington State Department of Ecology (Ecology), the U.S. Environmental Protection Agency (EPA), and Washington State University (WSU).  


3.1 Previous Ozone Modeling Work 
There have been only limited ozone modeling efforts for the Portland/Vancouver area.  The 1990 Clean Air Act (CAA) Amendments classified the Portland/Vancouver airshed as a marginal ozone nonattainment area.  There is no CAA requirement for a marginal ozone nonattainment area to perform photochemical modeling or any other attainment demonstration.  Furthermore, there have not been other federal requirements for air regulatory agencies to perform ozone air quality modeling in the Portland/Vancouver airshed.  The only ozone modeling work that has been performed has been a result of special project work or large-scale regional modeling efforts that included the Portland/Vancouver airshed in the overall domain.

3.1.1 Maintenance Plan EKMA modeling
The redesignation request/ozone maintenance plans that were submitted to the Oregon and Washington SIPs in 1996 relied on historical ozone monitoring data, future control measures, and emission inventory projections to demonstrate future attainment with the 1-hour ozone standard.  The ozone modeling work for that submittal was limited to a determination of emission levels necessary to prevent future exceedances using the Empirical Kinetics Modeling Approach (EKMA).  This work utilized the Ozone Isopleth Plotting Package with Optional Mechanisms (OSIPM-4) to conduct the EKMA analysis.

3.1.2 Southwest Washington Ozone Study
Following EPA's approval of the Portland and Vancouver ozone SIP submittals, WSU was contracted by Ecology and SWCAA to conduct a photochemical modeling study of the area to determine the potential effects of point sources upwind of the Portland/Vancouver maintenance area on the area's monitored ozone values.  The main purpose of the study was to address the question of the impact of the emissions from the major industrial point sources in Cowlitz County, WA on ozone levels in the Portland/Vancouver maintenance area.  This study also had an objective to help develop the regional photochemical modeling system.  This included running the MM5 prognostic weather model and the CALGRID/CALMET suite of models with a fine resolution grid (5 km).  The study focused on an ozone episode during July 11-14, 1996.  The final report for this study was published in January of 1999 and is available for review on the Ecology website.

3.1.3 NW Regional Technical Center Demonstration Project
This work was the result of a multi-agency collaborative effort to develop regional air quality modeling capabilities.  The contributing agencies for this work were EPA Region 10, Ecology, ODEQ, Idaho Department of Environmental Quality and Environment Canada.  The main goal of this project was to establish a virtual technical center and to demonstrate the use of the EPA CMAQ regional modeling system.  The agencies contracted with WSU to run several CMAQ and CALGRID runs with varying chemical mechanisms and MM5 grid scales.  This has been the most extensive modeling exercise that has encompassed the Portland/Vancouver region to date.  Several modeling runs were completed to look at ozone and visibility at a 12 km grid scale.  The model runs focused on two events,  July 2-14, 1996 and July 23-30, 1996.  The work was conducted from early 2000 to June 2002.  Further details are in O'Neill and Lamb (2004); O'Neill et al. (2005); and Chen (2002).  

These ozone modeling studies provided the basis for the design of the current modeling project.  In particular, the demonstration and evaluation of CMAQ in these previous studies were very valuable for the current study.  Details describing the episode selection and modeling methods are given in the following sections. 

4 Episode Selection 

The Portland/Vancouver Maintenance Area has had relatively mild ozone seasons during the summers of 1999, 2000, 2001, and 2002.  There were monitored exceedances of the 1-hour ozone standard in 1996 and 1998.  There were also times of elevated ozone levels during most years when the standard was not exceeded. 

General EPA guidance for episode selection recommends analyzing observed meteorological patterns to determine which have caused violations for the area in the past.  The Portland/Vancouver ozone modeling technical committee recommended modeling one scenario of each of the major meteorological situations that cause exceedances.  

The following episodes were suggested as potential episodes to model.

 July 11-14, 1996 (Same episode used for the NWRTC Demonstration Project and the Southwest Ozone Study, one 1-hour ozone exceedance in the Portland area)
 July 22-31, 1998 (Series of 3 exceedances in the Portland area during the 25[th] through the 28[th])
 July 17-22, 1997 (Episode with moderate monitored values under the 8-hour standard at the Portland/Vancouver monitors but well monitored)
 July 22-26, 1996 (Episode with the highest ozone measurements in the area for 1996, 1997, and 1998)
 August 7-10, 1996 (Episode with Sauvie Island monitor reaching its highest value for 1996)

The Portland/Vancouver ozone modeling technical committee considered several factors during selection of the episodes to be modeled, including: 

 Ozone episodes selected for modeling should be selected from the most frequent, typical, and representative patterns associated with high ozone in the region.
 Recent episodes are preferred over older episodes because recent episodes represent the current emissions inventory, including mobile and point configurations.  Recent episodes can also be more representative of current reduction strategies.
 Well-monitored episodes (with more meteorology, VOC, and NOx data) are preferred over data poor episodes. Additional data allow for more thorough model evaluation and provide information necessary to understand the processes leading to high ozone.
 Ozone episodes should be selected with high monitored values to represent the meteorology that results in severe ozone.  Furthermore, episodes should have high ozone in the geographical areas where high values typically occur. Such selection will allow modeling potential control strategies with representative ozone concentrations in the areas where the strategies must work.
 Multi-day episodes are the most efficient way to model both long range transport of ozone and the accumulation of local emissions that are associated with ozone.
 Multi-day episodes also allow the possibility of testing model responsiveness. If the model can reproduce both the high and low ozone days within a period, there will be more confidence in the model and its ability to replicate and forecast events.

The Portland/Vancouver ozone modeling technical committee screened the suggested episodes  based on the considerations listed above and recommended two episodes.  The recommended episodes are as follows:

July 22-31, 1998 was selected to be the first episode modeled.  The July 1998 episode accounted for 3 exceedances of both the 1-hour and 8-hour ozone standards at the Carus monitor in the Portland area within a short period of time.  There was also a near exceedance of the 1-hour standard and an exceedance of the 8-hour standard at the Milwaukie monitor.  There were elevated temperatures over the series of days and light NW winds.  The 1998 episode was chosen instead of the July 11-14, 1996 episode because considerable analysis has already been done for the 1996 episode, and there would be more benefit from analyzing a new episode.  Table 1 shows some select data from the 1998 episode.
      
      
Table 1: Selected peak ozone concentrations during the July 22-31, 1998 episode (ppb)

                                   Milwaukie
                                     Carus
Date
                                    1-hour
                                    8-hour
                                    1-hour
                                    8-hour
July 26
                                      124
                                      100
                                      136
                                      117
July 27
                                       
                                       
                                      133
                                      86
July 28
                                       
                                       
                                      137
                                      98
      

July 17-22, 1997 was selected to be the second episode modeled.  This episode did not coincide with any ozone standard exceedances; however there were moderately high ozone levels recorded at the Milwaukie monitor.  This episode also seemed to show a slightly different meteorological pattern than the 1998 episode as indicated by moderate values at both the Vancouver and Carus sites.  The modeling technical committee compared this episode with the July 11-14, 1996 episode and decided that it would be more beneficial to focus the efforts on an event that has not previously been analyzed with a model in this area.  Another reason for the selection of this episode (as opposed to 1996) is that there are additional monitoring data for southwest Washington obtained during an Ecology study conducted in the summer of 1997, 1998, and part of 1999.  In addition to the meteorological and ozone data on the Oregon side, in both 1998 and 1997, Southwest WA had four ozone sites in the domain (Woodland, Mt. View, Castle Rock, and Hockinson), and three NOx sites (Castle Rock, Mt.View and Woodland).  Additionally, in 1997 HC and carbonyl measurements were taken at Castle Rock and Mt.View and there was one additional met site at a Bonneville Power Administration site.  This additional monitoring in 1997 contrasts with the usual situation in southwest Washington where there is a lack of NOx and VOC sites and only a single ozone site.  Table 2 shows some select data from the 1997 episode.


Table 2: Summary of the July 17-22, 1997 episode ozone values (ppb)

                             Mt. View (Vancouver)
                                   Milwaukie
                                     Carus
                                 Sauvie Island
Date
                                    1-hour
                                    8-hour
                                    1-hour
                                    8-hour
                                    1-hour
                                    8-hour
                                    1-hour
                                    8-hour
July 19
                                      59
                                      53
                                      65
                                      55
                                      77
                                      65
                                      59
                                      52
July 20
                                      77
                                      62
                                      101
                                      73
                                      73
                                      68
                                      48
                                      38


5 Meteorological modeling

5.1 Modeling methods
The Mesoscale Meteorological model Version 5 (MM5, Grell et al., 1994) was used to provide the 3-D meteorological field for air quality modeling.  Three one-way nested domains with grid cell horizontal sizes of 36 km, 12 km, and 4 km were applied.  The two outer domains consisted of 98x95 and 133x151 grid cells, respectively.  The innermost domain consisted of 112x112 grid cells which extended from north of Puget Sound in Washington to south of Salem, OR and from the Pacific coast on the west to beyond the Cascade Mountain range on the east.  This domain was larger than the immediate area of Portland/Vancouver and was selected to minimize the influence of boundary conditions.  Vertically, 38 sigma layers were specified. The 12 km and 4 km MM5 domains are shown in Figure 1. 

To obtain meteorological fields that produced acceptable results with CMAQ in terms of ozone performance, several MM5 runs with different physics configurations were conducted as shown in Table 3.  

Table 3. MM5 configuration used for different model runs
                                     Runs
                             5-layer soil model_1 
                              5-layer soil mode_2
                                     LSM_1
                                     LSM_2
                                     Label
                                     "_o" 
                                  "base_sen"
                                 "sen_ini_LSM"
                               "sen_ini_LSM_ns"
                                Initialization
                                  NCEP GDAS 
                                    ETA212 
                                    ETA212
                                    ETA212
                                   Radiation
                                Cloud-radiation
                                CCM2 Radiation
                                Cloud-radiation
                                Cloud-radiation
                                   Moisture
                                  Simple ice
                                   Reisner 2
                                  Simple ice
                                  Simple ice
                                Surface scheme 
                              5-layer soil model
                              5-layer soil model
                                   Noah LSM
                                   Noah LSM
                                Cumulus scheme
                                  KF Cumulus
                                  KF Cumulus
                                  KF Cumulus
                                  KF Cumulus
                                  PBL  scheme
                                      MRF
                                      MRF
                                      MRF
                                      MRF
           Analysis nudging of Tsfc and winds within boundary layer
                                 Not turned on
                                 Not turned on
                                 Not turned on
                                   Turned on

As described later, the "LSM_1" run produced the best overall results for air quality modeling.  In this report, this run is evaluated and compared with the original run ("5-layer soil model_1").  The original 5-layer soil model run was initialized using NCEP Global Data Assimilation System (GDAS) analysis archived on a 2.5x2.5 degree lat/long grid at every 12 hours.  All other runs were initialized using NCEP Eta model analysis archived on a 40 km grid at every 3hrs.  Analysis nudging was applied in the coarse domains (36 km, 12 km) and observational nudging was applied in the inner 4km domain.  For analysis nudging, 3D nudging was employed for winds, temperature, and moisture above the PBL.  Surface nudging was applied for winds and moisture, but not temperature. In some sensitivity runs, boundary layer 3D nudging was turned on for winds and surface nudging was also turned on for temperature.  See Table 3 for configuration details.  Wind speed and direction observations from about 20 surface stations were used for observational nudging of the winds in the inner 4 km domain.  A map of these stations is shown in Figure 2.  Data from four surface stations (Ecology sites as shown in Figure 2) were withheld from the nudging analysis so that an independent evaluation could be conducted at these sites.  For all runs, land use data were derived from 1 km USGS global terrain and land use database. 

The MM5 physics options used in the LSM run included use of the Kain-Fritsch cumulus parameterization scheme (Kain and Fritsch, 1993), the MRF PBL scheme (Hong and Pan, 1996), simple ice moisture, cloud-radiation, and Noah Land Surface Model (Chen and Dudhia, 2001).  In some sensitivity runs, CCM2 radiation, Reisner2 moisture, and the 5-layer soil model were employed as shown in Table 3.  It should be noted that the current configuration used in the University of Washington MM5 forecast system matches the configuration of the soil model run with the CCM2 radiation and Reisner2 moisture treatment.  


                                       
                Figure 1. 12 and 4 km MM5 domain (D03 and D04)















Figure2. Surface stations used in MM5 observational nudging. The stations within the box outlined in blue were used in nudging (expect for four ecology sites which were withheld).  The sites within the inner box outlined in green were used in model evaluation. 

It has been shown that soil moisture has a significant impact on PBL parameters such as surface temperature and PBL height (Zhang and Anthes, 1982).  These processes as well as cloud coverage, profiles of humidity, temperature and wind are all related to surface wetness and vegetation cover (Xiu and Pleim, 2001).  In the traditional MM5 approach (5-layer soil model), soil moisture is set to constant seasonal values depending on land use type.  Therefore, the effects of precipitation and evaporation are not treated explicitly.  In contrast, the Noah LSM has four soil layers with the thickness of 0.1, 0.3, 0.6, and 1.0 m.  It employs one canopy layer and prognostically solves for soil moisture and temperatures in each soil layer.  As a result, the LSM has the ability to account for temporal and spatial changes in soil moisture due to precipitation and evapo-transpiration effects.  




5.2 Meteorological modeling results and evaluation (July 1998 episode)
5.2.1 Surface Temperature
Surface temperature contours from the simple soil-1 and LSM-1 runs are shown in Figure 3a and 3b at 17 PDT on July 28, 1998.  For western Washington and Oregon, the LSM run predicted cooler surface temperatures than the simple soil model, with only a small area around Portland above 35 C.  In the simple soil model, the warm areas were much more extensive around the Puget Sound area and from Portland to Salem.  In eastern Washington, the temperature patterns were opposite, where LSM-1 simulated warmer conditions than the simple soil model.
 
Predicted surface temperature and cloud coverage were compared with observations as shown in Figure 4a-c.  It appears that on the high ozone days (July 26 to July 28), the daily maximum temperatures predicted by both the LSM-1 and simple soil-1 models were fairly close to the observations at Portland and Salem.  On the low ozone days (July 23 to July 25), LSM-1 showed better agreement with the observations which might be related to improved simulation of clouds on these days.  However, in the northwestern part of the domain, such as at Castle Rock(Figure 4c), diurnal variations of surface temperature in LSM-1 were too small compared to the daily variations in the observations.  Similar patterns could also be found at some southwest sites such as UAO and HIO (not shown).  A scatter plot (Figure 5) showed that both model runs tended to overestimate low temperatures and underestimate high temperatures.  However, LSM-1 had slightly better agreement with the observations especially for the peak temperatures.


                                       
                                  Figure 2a. 
Figure 3a. Surface temperature pattern from simple soil model run for 17:00 PDT on July 28, 1998.
                                       
Figure 3b.. Surface temperature pattern from LSM model run for 17:00 PDT on July 28, 1998.


Figure 4a. Surface temperature and cloud cover time series predicted by LSM and simple soil model compared with observations at Portland. 
Figure 4b. Surface temperature and cloud cover time series predicted by LSM and simple soil model compared with observations at Salem. 


Figure 4c. Surface temperature and cloud cover time series predicted by LSM and simple soil model compared with observations at Castle Rock. 

Figure 5. Scatter plot of predicted surface temperature (LSM and simple soil) and observations from all available surface stations

5.2.2 Surface winds
Surface wind patterns are shown in Figure 6a, and 6b for 17 PDT on July 28.  The LSM-1 run predicted northerly winds along the I-5 corridor in western Washington, which veered to northwesterly and moved through the Portland area.  As the flow went further south, there was some apparent convergence with the strong southwesterly onshore flow along the Oregon coast.  In contrast, the winds from the simple soil model run showed very consistent northwesterly winds along the I-5 corridor.  The onshore winds along the Oregon coast flowed northwesterly instead of southwesterly as in the LSM-1 run.  Similar southwesterly onshore flows were also seen in other runs initialized using the Eta analysis, so this might be more related to the difference in initialization than in the surface scheme.

The MM5 first layer (38 m) winds were extrapolated to 10 m using the neutral logarithmic wind speed profile and compared with 10 m surface station observations as shown in Figure 7.  Predicted wind speed and direction showed reasonable agreement with the observations.  Both runs tended to underestimate the daytime surface winds (Figure 7a, 7b, and 7c).  This problem is well known for the MRF PBL scheme (Liu et al. 2004; Zhang et al. 2004.) and was related either to the underestimation of downward momentum flux or to the overestimation of friction velocity according to these researchers. 



                                       
Figure 6a. Surface wind pattern from simple soil model run for 17:00 PDT on July 28, 1998. 
                                       
Figure 6b. Surface wind pattern from LSM run for 17:00 PDT on July 28, 1998. 

Figure 7a. Surface wind speed and wind direction time series predicted by LSM and simple soil model compared with observations at Portland
Figure 7b. Surface wind speed and wind direction time series predicted by LSM and simple soil model compared with observations at Salem.


Figure 7c. Surface wind speed and wind direction time series predicted by LSM and simple soil model compared with observations at Castle Rock.

5.2.3 Soil moisture 
In the simple soil model, soil moisture has constant seasonal values (summer or winter) defined as a function of land use type.  Consequently, in summer most areas on the west side of the domain had volumetric soil moisture values of 0.3, except for urban areas where the soil moisture equaled 0.1.  On the east side of the Cascades, most areas were much drier with soil moisture of 0.1 except for the irrigated land (shown in Figure 8) where soil moistures were equal to 0.5. 

In the LSM-1 run, soil moisture was initialized using analysis fields from the Eta model and solved prognostically.  Figure 8 shows the predicted values from LSM-1 at 17 PDT on July 27. Soil moisture for each layer (0.1, 0.3, 0.6, and 1 m) is shown.  It appears that soil was driest at the surface layer, and had more water in the deeper layers.  On the west side of the Cascades, soil moisture varied from 0.25 at the surface to 0.3 in the deeper layers.  The difference was more substantial on the east side of the Cascades where the water content varied from about 0.1 in the first layer to about 0.3 at 1 m deep.  Some wet areas in the shallow soil layer are obvious in the lower right corner of the domain; these were the result of convective rain storms simulated in this run. 

Modeled soil moisture was also compared with the observations at the Wind River Canopy Crane research site located north of Stevenson, WA approximately 30 km from the Columbia River.  As described on the research site web page (http://cdiac.ornl.gov/ftp/ameriflux/data/us-sites/preliminary-data/Wind-River/1998/WRCCRF_met_meta_data.html), the observed values were measured at 30 cm using Campbell Scientific CS-615 probes.  The station is located in an old growth forest canopy where soil moisture can be highly heterogeneous (personal communication, Timothy Link).  However, the LSM-1 captured the drying trend in the observations fairly well (Figure 9).  Note that this is consistent with the fact that there was little rainfall during the episode for most of the domain, and therefore, soil moisture was mainly controlled by evapotranspiration.  



Figure 8. Volumetric soil moisture defined in the simple soil model and values calculated in the LSM for different layers.

Simple soil
LSM  -  layer 1 (0.1 m)













LSM  -  layer 2 (0.3 m)













LSM  -  layer3 (0.6 m)

LSM  -  layer4 (1 m)






Figure 9. Predicted and observed soil moisture at Wind River Crane site. 




5.2.4 PBL height and vertical profile 
PBL heights on the three peak ozone days are plotted in Figure 10 for the time when daily peak values were reached.  The LSM-1 run generally predicted lower PBL heights along the west side of the Cascades, especially in western Washington.  For this part of the domain, the peak PBL heights from the LSM-1 run were around 800 to 1000 m.  In contrast, the simple soil-1 run predicted average peak heights around 1200 m in this area for all three days.  Different patterns showed up for the east side of the Cascades where the LSM-1 generally predicted higher PBL heights, except for the very low heights in the lower right corner of the domain associated with areas of convective rainfall. 

Mixing heights estimated from upper air soundings at Salem, OR were compared to simulations from two models as shown in Figure 11.  The LSM-1 run predicted lower PBL heights on July 24, 25 and 27 than the simple soil-1 run. On July 26 and 28, the pattern was opposite. Compared to observed values (only available on July 26 to 28), both models tended to overestimate mixing heights by 200 to 300 m on these peak ozone days.


Figure 10. Predicted PBL heights by LSM and simple soil model for July 1998 episode












Figure 11. PBL predicted by simple soil and LSM compared with observed values for July 1998 episode.



Vertical profiles of temperature, wind speed, and wind direction were also compared with the upper air soundings at Salem at 17PDT on July 25 and July 26 as shown in Figure 12.  The LSM-1 run captured the vertical temperature profile fairly well on July 25, whereas the simple soil-1 run appeared to overestimate the temperature and predicted a deeper mixed layer than the soundings.  On July 26, both models matched the temperature profile well up to 700 m but predicted a deeper mixed layer than the observations.  For wind speed, the LSM-1 run underestimated wind speeds up to 600 m and the simple soil-1 run showed some overestimation up to 1500 m on July 25.  On July 26, both models underestimated wind speed.  For wind direction, both models had very good performance on July 25.  On July 26, the LSM-1 run showed better agreement with the observations below the boundary layer.

Figure 12. Observed and predicted vertical profiles of temperature, wind speed and wind direction from the simple soil-1 and LSM-1 runs.  ("old_run" refers to the simple soil-1 run, and "rerun" refers to the LSM-1 run)































































































































5.2.5 MM5 statistical evaluation
The meteorological simulations were also evaluated using the mean bias and mean error for wind speed, direction, and temperature (Table 4).  Four Ecology network sites (Castle Rock, Woodland, BPA, and Mt.Zion) were selected as independent sites and withheld from observational nudging for surface winds.  The mean bias of temperature from the LSM-1 run was 1.4C, about 1C smaller than the bias from the simple soil-1 run.  The mean error of temperature was also smaller in the LSM-1 run compared to the simple soil-1 run (2.9C versus 3.2C). The mean bias and mean error of wind speed and wind direction from the LSM-1 run (ME = 1.5 m/s for WS; ME = 50 for WD) were both larger than those from the simple soil-1 run (ME = 1.2 m/s for WS; ME = 40 for WD).  This could be related to the underestimation of daytime surface winds and resulting error in the wind direction.  The statistics for wind direction and wind speed were calculated separately for stations used for observational nudging and those withheld from observational nudging; these results show that model performance was similar for these two groups of stations.  This might due to the relative low density of the surface stations available for observational nudging in the region.  Nevertheless, for the simple soil-1 run, the results for obs nudging stations had slightly smaller error (38) in wind direction compared with the results for withheld stations (47). 



Table 4. MM5 performance statistics for July 1998 episode
                                     1998
                                      Run
                                    T_bias
                                    T_error
                                    Wd_bias
                                   Wd_error
                                    Ws_bias
                                   Ws_error
                                       
                                       
                                       C
                                       C
                                       
                                       
                                      m/s
                                      m/s
                                      TDO
                                     LSM-1
                                     0.75
                                     2.83
                                      6.5
                                     45.4
                                     -0.90
                                     1.57
                                       
                                 simple_soil_1
                                     2.28
                                     3.05
                                     -1.5
                                     28.9
                                     -0.50
                                     1.19
                                  Castle Rock
                                     LSM-1
                                     1.33
                                     2.88
                                     -2.3
                                     58.3
                                     0.63
                                     1.56
                                       
                                 simple_soil_1
                                     2.85
                                     3.37
                                     -12.1
                                     63.1
                                     1.01
                                     1.27
                                      KLS
                                     LSM-1
                                     1.26
                                     2.89
                                     -9.1
                                     34.3
                                     0.07
                                     1.32
                                       
                                 simple_soil_1
                                     2.85
                                     3.22
                                     -9.4
                                     25.4
                                     0.37
                                     1.16
                                   Woodland
                                     LSM-1
                                     1.89
                                     3.20
                                     -8.5
                                     20.2
                                     -1.60
                                     1.88
                                       
                                 simple_soil_1
                                     2.49
                                     3.23
                                     -0.2
                                     15.6
                                     -1.00
                                     1.46
                                      BPA
                                     LSM-1
                                     2.08
                                     3.10
                                     -9.2
                                     42.8
                                     0.62
                                     1.48
                                       
                                 simple_soil_1
                                     2.95
                                     3.39
                                     -3.3
                                     36.8
                                     1.02
                                     1.29
                                      VUO
                                     LSM-1
                                     1.64
                                     2.86
                                      2.1
                                     52.9
                                     -0.36
                                     1.28
                                       
                                 simple_soil_1
                                     2.50
                                     2.96
                                      3.4
                                     46.9
                                     0.10
                                     1.00
                                      PDX
                                     LSM-1
                                     0.83
                                     2.63
                                     -3.6
                                     46.8
                                     -0.55
                                     1.63
                                       
                                 simple_soil_1
                                     1.92
                                     2.64
                                      5.7
                                     36.0
                                     -0.46
                                     1.25
                                   Mt. Zion
                                     LSM-1
                                     0.97
                                     2.62
                                      6.8
                                     67.7
                                     -0.36
                                     1.18
                                       
                                 simple_soil_1
                                     2.45
                                     3.16
                                     10.7
                                     74.0
                                     0.13
                                     0.97
                                      HIO
                                     LSM-1
                                     1.57
                                     3.25
                                     -18.9
                                     52.3
                                     -0.87
                                     1.40
                                       
                                 simple_soil_1
                                     2.97
                                     3.58
                                     -26.6
                                     39.0
                                     -0.76
                                     1.01
                                      TTD
                                     LSM-1
                                     1.46
                                     2.71
                                      7.3
                                     75.1
                                     -0.74
                                     1.36
                                       
                                 simple_soil_1
                                     2.98
                                     3.27
                                     25.1
                                     54.6
                                     -0.56
                                     1.15
                                      UAO
                                     LSM-1
                                     1.27
                                     2.66
                                      0.8
                                     58.9
                                     -0.44
                                     1.62
                                       
                                 simple_soil_1
                                     2.85
                                     3.22
                                     -7.8
                                     36.9
                                     -0.23
                                     1.09

Table 4. Cont.
                                      SLE
                                     LSM-1
                                     1.65
                                     2.88
                                      1.7
                                     47.6
                                     0.04
                                     1.77
                                       
                                 simple_soil_1
                                     3.00
                                     3.37
                                      5.4
                                     35.2
                                     0.24
                                     1.15
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                    average
                                      Run
                                    T_bias
                                    T_error
                                    Wd_bias
                                   Wd_error
                                    Ws_bias
                                   Ws_error
                                       
                                       
                                       C
                                       C
                                       
                                       
                                      m/s
                                      m/s
                                 all stations
                                     LSM-1
                                     1.39
                                     2.88
                                     -2.19
                                     50.18
                                     -0.37
                                     1.50
                                       
                                 simple_soil_1
                                     2.68
                                     3.21
                                     -0.90
                                     41.03
                                     -0.05
                                     1.17
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                               Withheld stations
                                     LSM-1
                                     1.57
                                     2.95
                                     -3.31
                                     47.23
                                     -0.18
                                     1.53
                                       
                                 simple_soil_1
                                     2.68
                                     3.29
                                     -1.23
                                     47.37
                                     0.29
                                     1.25
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                Nudged stations
                                     LSM-1
                                     1.31
                                     2.84
                                     -1.63
                                     51.66
                                     -0.47
                                     1.49
                                       
                                 simple_soil_1
                                     2.67
                                     3.16
                                     -0.73
                                     37.86
                                     -0.23
                                     1.13




5.2.6 LSM run compared with simple soil model run
The differences from LSM-1 and simple_soil_1 runs are summarized below:

 LSM showed better simulation of surface temperatures. 
 LSM predicted generally lower PBL heights along the west side of the Cascades on the peak ozone days. 
 LSM was able to predict temporally and spatially varied soil moisture field in response to precipitation and evapo-transpiration.  
 The performance for winds were slightly degraded in the LSM run compared to the simple soil model run due to well known problem of the MRF PBL scheme. 
 Ozone simulation was also improved with MM5 including the Noah LSM as shown in the CMAQ modeling section. 

Therefore, LSM-1 run was selected for the MM5 simulation due its overall improved meteorological simulation and better ozone predictions. 


 

5.3 Modeling results and evaluation (July 1997 episode)
Surface temperature contours from MM5/MCIP are shown in Figure 13.  Compared to the July 1998 episode in which extensive high temperatures existed, the surface temperatures in 1997 episode were considerably lower and in the range of 12-35C.  There was generally good agreement between the predicted and observed surface temperatures and cloud coverage at Portland and Salem as shown in Figure 15a and b.  At the KLS site (Figure 15c) located along the I-5 corridor in southwestern Washington, the LSM-1 run under-estimated daytime temperatures on most days.  Nevertheless, there was a fairly large over-estimation of the daytime temperatures on July 20.  The model also missed the overcast condition in that morning, which might explain the overestimate of temperature as well as ozone (shown in the ozone modeling section) around the area.  Overall, however, the LSM-1 performance appears to be better for the 1997 episode compared to 1998, especially for nighttime temperatures. 

The surface wind patterns in the afternoon of July 20 are shown in Figure 14.  There were consistent northwesterly/westerly flows across western Washington and Oregon.  Some convergence close to Milwaukie was apparent between the northwesterly and southwesterly winds in the simulation, which might contribute to the high ozone levels there.  Surface time series of wind speed and wind direction are shown in Figure 16a-c.  The LSM-1 run performed relatively well for wind direction. Similar to the 1998 episode, underestimation of daytime wind speeds were also the case for 1997. 

Modeled PBL heights were compared with observations at Salem as shown in Figure 17.  On three of the four days when data were available (July 15, 17, and 22), the LSM-1 run predicted PBL heights were very close to the observations.  On July 21, the PBL heights were overestimated by about 200 m. 

The LSM-1 performance statistics for the July 1997 episode are summarized in Table 5.  Similar to the 1998 case, Ecology meteorological sites (Castle Rock, BPA, and Mt. Zion) were withheld from observational nudging of surface winds.  Compared with the withheld stations, the stations used for observational nudging had a smaller mean error in wind direction (37 versus 49), but slightly larger mean bias and error in wind speed.  Overall, the statistics for the 1997 episode were better than for 1998, with smaller mean error in temperature (2.1C versus 2.9C) and wind direction (40 versus 50).  It is also worth noting that in 1997 there was a negative temperature bias and positive temperature bias in 1998; this which might be due to the better simulations of nighttime temperature in the 1997 episode. 


                                       

Figure 13. Surface temperature pattern from MM5 for 17:00 PDT on July 20, 1997.

                                       
Figure 14. Surface wind pattern from MM5 for 17:00 PDT on July 20, 1997.
Figure 15a. Surface temperature and cloud cover time series predicted by LSM compared with observations at Portland.

Figure 15b. Surface temperature and cloud cover time series predicted by LSM compared with observations at Salem.

Figure15c. Surface temperature and cloud cover time series predicted by LSM compared with observations at KLS. 


Figure 16a. Surface wind speed and wind direction time series predicted by LSM compared with observations at Portland.

Figure 16b. Surface wind speed and wind direction time series predicted by LSM compared with observations at Salem.


Figure 16c. Surface wind speed and wind direction time series predicted by LSM compared with observations at KLS. 

Figure 17. PBL predicted by LSM compared with observed values for July 1997 episode.



Table 5. MM5/MCIP performance statistics for July 1997 episode
                                     1997
                                      Run
                                    T_bias
                                    T_error
                                    Wd_bias
                                   Wd_error
                                    Ws_bias
                                   Ws_error
                                       
                                       
                                     C
                                     C
                                      
                                      
                                      m/s
                                      m/s
                                      TDO
                                     LSM-1
                                     -0.63
                                     1.74
                                     -9.1
                                     36.1
                                     -1.37
                                     1.74
                                  Castle Rock
                                     LSM-1
                                     -0.16
                                     1.93
                                      2.4
                                     54.4
                                     0.15
                                     1.25
                                      KLS
                                     LSM-1
                                     -1.02
                                     1.82
                                     -20.6
                                     36.4
                                     -0.37
                                     1.42
                                      BPA
                                     LSM-1
                                     0.32
                                     2.04
                                     -3.3
                                     29.7
                                     0.36
                                     1.35
                                      PDX
                                     LSM-1
                                     -1.49
                                     2.14
                                     -4.7
                                     29.8
                                     -1.26
                                     1.71
                                   Mt. Zion
                                     LSM-1
                                      N/A
                                      N/A
                                     -2.5
                                     63.1
                                     -0.11
                                     1.24
                                      DLS
                                     LSM-1
                                     -3.50
                                     3.67
                                     -1.7
                                     35.2
                                     -0.99
                                     2.16
                                      HIO
                                     LSM-1
                                     -1.42
                                     1.96
                                     -22.4
                                     32.2
                                     -0.29
                                     1.50
                                      TTD
                                     LSM-1
                                     -1.44
                                     2.33
                                     29.1
                                     41.3
                                     -0.93
                                     1.48
                                      UAO
                                     LSM-1
                                     -0.39
                                     1.56
                                     -1.4
                                     42.3
                                     -0.82
                                     1.51
                                      SLE
                                     LSM-1
                                     -0.32
                                     1.74
                                      6.3
                                     44.7
                                     -0.87
                                     1.66
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                    average
                                      Run
                                    T_bias
                                    T_error
                                    Wd_bias
                                   Wd_error
                                    Ws_bias
                                   Ws_error
                                       
                                       
                                     C
                                     C
                                      
                                      
                                      m/s
                                      m/s
                                 all stations
                                     LSM-1
                                     -1.01
                                     2.09
                                     -2.54
                                     40.49
                                     -0.59
                                     1.55
                               withheld stations
                                     LSM-1
                                     0.08
                                     1.98
                                     -1.15
                                     49.09
                                     0.13
                                     1.28
                                nudged stations
                                     LSM-1
                                     -1.28
                                     2.12
                                     -3.06
                                     37.26
                                     -0.86
                                     1.65


5.4 Back trajectory analysis 
An analysis of back trajectories for the Portland urban area was conducted by Ecology using the MM5 wind fields (Bowman, personal communication) and the Vis5D software package.  This analysis was done at an early stage of the modeling work, and therefore, the simple soil-1 runs were used for both episodes.  On each day from July 26 to 29 in 1998 and July 18 to 21 in 1997, two 12-hr back trajectories starting 13 PDT and 17 PDT were produced for each monitoring site.  The trajectories terminated at 150 meters above the surface to avoid being kept in the surface layer.  The length of time a parcel spends along the path was accumulated if the parcel was within 500 meters of the surface.  This height was selected to accommodate seasonal and diurnal variations in the mixed layer. 

Two summary plots with all available back trajectories for 1998 and 1997 are shown in Figure 18a and b.  The plots show that for both years the air parcels spent a relatively long time in the immediate area due to the low wind speeds that existed.  In 1998, most back trajectories originated from the north and northwest with a few exceptions coming from the southwest.  It appeared that there was no direct impact from the Puget Sound area in that year.  Compared to 1998, winds in 1997 were more varied.  In addition to trajectories from northwest/north and southwest, there were also trajectories that originated from the northeast in the Puget Sound region.


                                       

18a. Summary plots of back trajectories during July 26 to 29, 1998. 

                                       

18b. Summary plots of back trajectories during July 18 to 21, 1997.




6 Photochemical air quality modeling

6.1 Modeling methods
EPA Models-3 Community Multi-scale Air Quality (CMAQ) Modeling System (Byun and Ching, 1999) version 4.3 was used for photochemical air quality modeling.  Based on state-of-science techniques, CMAQ is a multi-scale, multi-pollutant air quality model that simulates the transportation, transformation, and deposition of atmospheric pollutants including as photochemical precursors and oxidants, particulate matter, and airborne toxics.  It incorporates the output fields from the MM5/MCIP meteorological simulations and emissions derived from the Sparse Matrix Operator Kernel Emissions Modeling System (SMOKE) (http://cf.unc.edu/cep/empd/products/smoke/index.cfm), and then simulates the chemical transport using the CMAQ Chemical Transport Model (CCTM). 

Initial and boundary conditions were primarily derived from GEOS-CHEM global model simulations (personal communication, Rob Elleman).  However, vertical ozone IC/BC concentrations were derived from a combination of data sources.  For the western boundary, ozone levels were based on Trinidad Head ozonesondes data collected during July 98 (personal communication, Dan Jaffe).  For the other three sides (north, east, and south), ozone concentrations were derived from ICTC global model simulations (personal communication, Youhua Tang).  The concentrations of each model species were height dependent.  The concentrations in the first model layer are shown in Table 6.  The CMAQ modeling domain is slightly smaller than the MM5 innermost domain with 99x99 horizontal grid cells at 4 km x 4 km spacing and with 22 layers in the vertical.  The SAPRC99 photochemical mechanism including aqueous chemistry but no aerosol dynamics was employed.  The aerosol module was not employed because of the lack of emission inventory data for key aerosol precursors.  The chemical speciation of the emission inventory was performed by SMOKE according to the SAPRC99 mechanism.  The Modified Euler Backward Interactive (MEBI) solver was used to solve the chemical kinetic equations.  The first two days of each episode were run twice as model spinup to minimize the effects of initial conditions. 


     Table 6. Initial and boundary conditions for CMAQ run (first layer) 
                                    Species
                                  Description
                                     East
                                     North
                                     South
                                     West
                                       
                                       
                                    (ppbV)
                                    (ppbV)
                                    (ppbV)
                                    (ppbV)
                                     HCHO
                                 Formaldehyde
                                   1.98E+00
                                   9.21E-01
                                   1.11E+00
                                   2.01E-01
                                      ALD
                       Acetaldehyde and higher aldehydes
                                   1.39E+00
                                   5.03E-01
                                   7.80E-01
                                   3.30E-02
                                      OP1
                           Methyl hydrogen peroxide
                                   6.03E-01
                                   5.45E-01
                                   7.43E-01
                                   8.22E-01
                                      KET
                                    Ketones
                                   7.51E+00
                                   4.52E+00
                                   4.69E+00
                                   1.92E+00
                                     ONIT
                                Organic nitrate
                                   2.69E-02
                                   1.96E-02
                                   1.23E-02
                                   2.07E-03
                                      ETH
                                    Ethane
                                   1.46E+00
                                   1.25E+00
                                   1.27E+00
                                   1.03E+00
                                      HC3
                 Alkanes w/ 2.7x10-13 > kOH < 3.4x10-12
                                   3.35E-01
                                   2.10E-01
                                   2.06E-01
                                   6.26E-02
                                      HC5
                 Alkanes w/ 3.4x10-12 > kOH < 6.8x10-12
                                   2.70E+00
                                   2.70E+00
                                   2.70E+00
                                   1.39E+00
                                      HC8
                         Alkanes w/ kOH > 6.8x10-12
                                   5.00E-01
                                   5.00E-01
                                   5.00E-01
                                   1.29E-01
                                      OL2
                                    Ethene
                                   5.00E-01
                                   5.00E-01
                                   5.00E-01
                                   1.50E-01
                                      OLT
                               Terminal olefins
                                   2.00E-01
                                   2.00E-01
                                   2.00E-01
                                   8.71E-02
                                      OLI
                               Internal olefins
                                   4.00E-01
                                   4.00E-01
                                   4.00E-01
                                   4.30E-02
                                      TOL
                      Toluene and less reactive aromatics
                                   4.00E-01
                                   4.00E-01
                                   4.00E-01
                                   2.14E-01
                                      XYL
                      Xylene and more reactive aromatics
                                   5.00E-01
                                   5.00E-01
                                   5.00E-01
                                   3.70E-01
                                      PAN
                     Peroxyacetyl nitrate and higher PANs
                                   1.55E-01
                                   1.20E-01
                                   5.89E-02
                                   3.24E-03
                                      ISO
                                   Isoprene
                                   4.00E+00
                                   1.46E+00
                                   3.08E+00
                                   2.06E-02
                                     MACR
                                 Methacrolein
                                   3.30E-01
                                   1.23E-01
                                   2.12E-01
                                   1.03E-03
                                      MVK
                              Methyl Vinyl Ketone
                                   3.82E-01
                                   1.61E-01
                                   2.31E-01
                                   1.38E-03
                                     MPAN
                     PAN analogue formed from Methacrolein
                                   5.83E-02
                                   3.71E-02
                                   2.67E-02
                                   2.67E-04
                                      NO2
                               Nitrogen dioxide
                                   5.67E-01
                                   5.09E-01
                                   1.90E-01
                                   1.17E-02
                                      NO
                                 Nitric oxide
                                   8.69E-02
                                   5.92E-02
                                   2.29E-02
                                   1.17E-03
                                      O3
                                     Ozone
                                   4.00E+01
                                   4.00E+01
                                   4.00E+01
                                   2.34E+01
                                     HNO3
                                  Nitric acid
                                   8.30E-01
                                   3.38E-01
                                   3.42E-01
                                   1.37E-02
                                     HNO4
                               Peroxynitric acid
                                   1.86E-03
                                   1.91E-03
                                   5.50E-04
                                   5.01E-05
                                      NO3
                               Nitrogen trioxide
                                   1.00E-01
                                   1.00E-01
                                   1.00E-01
                                   1.00E-01
                                     H2O2
                               Hydrogen peroxide
                                   2.35E+00
                                   1.26E+00
                                   1.55E+00
                                   9.37E-01
                                      CO
                                Carbon monoxide
                                   1.45E+02
                                   1.21E+02
                                   1.17E+02
                                   8.64E+01
                                     N2O5
                              Nitrogen pentoxide
                                   1.47E-03
                                   1.82E-03
                                   1.87E-04
                                   2.28E-05
                                      SO2
                                Sulfur dioxide
                                   5.00E-02
                                   5.00E-02
                                   5.00E-02
                                   3.00E-02
                                     SULF
                                 Sulfuric acid
                                   3.00E-01
                                   3.00E-01
                                   3.00E-01
                                   2.99E-01




6.2 CMAQ modeling results and evaluation (July 1998 episode)
The local ozone monitoring sites are shown in Figure 19.  Simulated surface ozone contours from the LSM-1 run for July 28 are shown in Figure 20 at hour 6, 10, 14, and 18 PDT.  At 6 PDT, nitric oxide (NO) titration effects were obvious within the urban core and along the I-5 corridor with ozone concentrations below 20 ppb.  At 10 PDT, ozone increases were evident south of Seattle with values above 60 ppb.  However, there was no evident ozone production downwind of Portland at that hour.  At 14 PDT, ozone values above 100 ppb were simulated immediately north and south of the Portland urban core.  Maximum ozone concentrations were simulated at 18 PDT, with the maximum predicted ozone concentration reaching 149 ppb south of Portland.



                                       

Figure 19. Ozone monitoring site map for Southwest Washington and Oregon.


Figure 20. Simulated surface ozone contour on July 28, 1998 at 6, 10, 14, and 18 PDT.

                                       


                                       

                                       



                                       



Observed and simulated 1-hr surface ozone concentration time series are shown in Figure 21. Note that the first two days (July 24 and 25) were used as model spinup, so the results are not included in the statistical evaluation.  All monitoring sites except Hockinson and Wishram are located along the I-5 corridor.  Extending from North to South, Castle Rock, Woodland, and Sauvie Island are rural sites situated north of Portland; Mt View and Milwaukie are two suburban sites, close to Vancouver and slightly southeast of downtown Portland respectively; Carus, another rural site, is located about 30 miles south of Portland at the foot of the mountains; and Turner, the most southern site within the domain, is close to Salem, OR.  Hockinson and Wishram are the two sites that are further away from I-5 and influenced heavily by complex terrain close to the west slope of the Cascade Mountains and along the east edge of the Columbia River Gorge, respectively. 

At Castle Rock, Woodland, Sauvie Island, and Mt View, the maximum observed ozone concentrations were close to 100 ppb during July 26 to 28.  Nighttime ozone titration was obvious at Sauvie Island and Mt View with concentrations decreasing to near zero.  At these four sites, the model simulated the ozone peaks reasonably well, although sometimes the duration of elevated ozone in the simulations was longer than in the observations. 

At Milwaukie, observation data were missing on July 27 and part of July 26 and 28.  The available data show an exceedance of the 1-hr standard on July 26 with maximum value of 124 ppb. (NAAQS 1-hr ozone = 120 ppb, 8-hr ozone = 80 ppb).  The LSM-1 run underestimated the peak ozone on July 26 and slightly overestimated it on July 28.  At Carus, a series of three exceedances of the 1-hr and 8-hr ozone standards occurred on July 26 through 28, with 1-hr peak values of 136, 133, and 137 ppb, respectively.  Nighttime ozone at Carus remained about 20 ppb because it was not affected by local NOx sources.  Compared to the observations, the LSM-1 run missed the maximum ozone on the July 26 but captured the peaks fairly well on the following two days.  The nighttime values were also well simulated, but the timing of the daytime peaks were slightly delayed in the simulations.  At Turner, there was very good agreement between the observed and simulated ozone concentrations although nighttime ozone was overestimated. 

The model performance was different at the complex terrain sites of Wishram and Hockinson.  At Wishram, ozone values were moderate throughout the episode with maximum value of 77 ppb on July 27.  The LSM-1 run performed reasonably well at this site.  At Hockinson, moderately high ozone values (close to 100 ppb) were observed.  However, the LSM-1 run seemed unable to capture the ozone peaks from July 26 to 28. 

Figure 21. Observed and simulated 1-hr surface ozone concentrations for July 1998 episode















6.2.1 CMAQ statistics evaluation 
CMAQ model performance was evaluated by the following set of ozone statistics based on the simulations from July 26 to 28. 
 Normalized bias and normalized gross error between observed and predicted 1-hr/8-hr ozone concentrations.  Data within a selected time window were used in evaluation (10-19 PDT for 1-hr and 14-23 PDT for 8-hr ozone; shown in Table 7).  These time windows were selected to encompass periods of elevated ozone. 
 Normalized bias between observed and predicted 8-hr daily maxima using spatially paired means. (Shown in Table 8)
 Normalized bias between observed and predicted 8-hr daily maxima using all data pairs. (Shown in Table 9)
 Correlation coefficients and scatter plots of observed and predicted 8-hr daily maxima using spatially paired means. (Shown in Figure 22)
 Quantile-quantile plots of observed and predicted 8-hr daily maxima using (a) all data pairs (b) spatially paired means. (Shown in Figure 23a and 23b)

The normalized gross error is defined as:


and the normalized bias is defined as:



At Castle Rock, Woodland, and Sauvie Island, located upwind of Portland, the LSM-1 run had a positive bias especially at Woodland and Sauvie Island (1-hr NB of 24% and 28%, respectively).  This could be related to the fact that simulated ozone peaks are slightly broader than the observed ones.  For spatially paired means, the normalized bias between observed and predicted 8-hr daily maxima are all below 20% at these sites, which meets EPA suggested guidelines (EPA, 1999).  Nevertheless, if evaluated on a daily basis (as shown in Table 9), there was a larger positive bias especially on July 28. 

Mt. View and Milwaukie are two urban sites within the Portland/Vancouver airshed.  At Mt. View, the normalized bias and gross error were reasonable (1-hr NB = 2%, 8-hr NB = 3%, 1-hr NE = 34%, 8-hr NE = 26%).  For 8-hr daily maxima, the bias between the observed and predicted is generally negative when evaluated at the exact grid point of the monitoring site.  However, when evaluated using the surrounding 5x5 matrix of grid cells, the bias becomes positive as a result of steep horizontal ozone gradients around the site.  Due to missing data at Milwaukie, no statistics were calculated for this site. 

At Carus and Turner, downwind of Portland, observed ozone peaks were generally well simulated as shown in the bias between observed and predicted 8-hr daily maxima.  However, the timing of the ozone peaks were slightly shifted at Carus, and consequently, the gross error was moderately high (1-hr NE = 36% and 8-hr NE = 27%).

Hockinson and Wishram are two sites further away from the I-5 corridor and influenced heavily by complex terrain.  The performance statistics were fairly good at Wishram.  At Hockinson, performance was poor with large underprediction.  Nevertheless, there were also steep horizontal ozone gradients around Hockinson so that the predicted 8-hr maxima are considerably higher when evaluated with the surrounding 5x5 grids. 


Table 7. Normalized bias and normalized gross error between observed and predicted 1-hr/8-hr ozone concentrations. (1-hr statistics are calculated within 10-19 PDT, and 8-hr statistics are calculated within 14-23 PDT on July 26, 27, and 28, 1998.)
                                     Site
                                    1-hr NB
                                    1-hr NE
                                    8-hr NB
                                    8-hr NE
                                  Castle Rock
                                      12%
                                      24%
                                      9%
                                      16%
                                   Woodland
                                      24%
                                      31%
                                      23%
                                      26%
                                 Sauvie Island
                                      28%
                                      36%
                                      24%
                                      27%
                                   Mt. View
                                      2%
                                      34%
                                      3%
                                      26%
                                   Milwaukie
                                      N/A
                                      N/A
                                      N/A
                                      N/A
                                     Carus
                                      4%
                                      36%
                                      2%
                                      27%
                                    Turner
                                      10%
                                      29%
                                      5%
                                      16%
                                    Wishram
                                      -6%
                                      13%
                                      -6%
                                      9%
                                   Hockinson
                                     -17%
                                      36%
                                     -15%
                                      28%
Note: Values higher than EPA suggested model performance goal are highlighted in red. 


Table 8. Normalized bias between observed and predicted 8-hr daily maxima using spatially paired means. (based on 5x5 and 1x1 grids)
                                     site
                                   Ave obs 
                                8-hr daily max
                                     (ppm)
                       Ave pre 8-hr daily max 1x1 (ppm)
                                   NB (1x1)
                       Ave pre 8-hr daily max 5x5 (ppm)
                                   NB (5x5)
                                  Castle Rock
                                     0.075
                                     0.079
                                      4%
                                     0.081
                                      8%
                                   Woodland
                                     0.071
                                     0.080
                                      13%
                                     0.082
                                      16%
                                 Sauvie Island
                                     0.071
                                     0.081
                                      14%
                                     0.083
                                      16%
                                   Mt. View
                                     0.075
                                     0.066
                                     -12%
                                     0.083
                                      11%
                                   Milwaukie
                                      N/A
                                     0.085
                                      N/A
                                     0.092
                                      N/A
                                     Carus
                                     0.101
                                     0.096
                                      -5%
                                     0.098
                                      -3%
                                    Turner
                                     0.084
                                     0.085
                                      0%
                                     0.088
                                      5%
                                    Wishram
                                     0.061
                                     0.056
                                      -9%
                                     0.057
                                      -7%
                                   Hockinson
                                     0.078
                                     0.059
                                     -24%
                                     0.069
                                     -12%

Note: 	"1x1"  -  data extracted from the exact grids.
      "5x5"  -  data extracted from the surrounding 5x5 grids.


Table 9. Normalized bias between observed and predicted 8-hr daily maxima using all data pairs. (based on 5x5 and 1x1 grids)
                                     Site
                                     Date
                              Obs 8-hr daily max
                                     (ppm)
                         Pre 8-hr daily max 1x1 (ppm)
                                   NB (1x1)
                         Pre 8-hr daily max 5x5 (ppm)
                                   NB (5x5)
                                  Castle Rock
                                    July 26
                                     0.077
                                     0.073
                                      -6%
                                     0.075
                                      -3%
                                  Castle Rock
                                    July 27
                                     0.077
                                     0.088
                                      14%
                                     0.091
                                      17%
                                  Castle Rock
                                    July 28
                                     0.072
                                     0.075
                                      5%
                                     0.079
                                      10%
                                   Woodland
                                    July 26
                                     0.069
                                     0.071
                                      2%
                                     0.073
                                      5%
                                   Woodland
                                    July 27
                                     0.089
                                     0.091
                                      2%
                                     0.091
                                      2%
                                   Woodland
                                    July 28
                                     0.054
                                     0.079
                                      47%
                                     0.083
                                      55%
                                 Sauvie_Island
                                    July 26
                                     0.074
                                     0.069
                                      -7%
                                     0.070
                                      -4%
                                 Sauvie_Island
                                    July 27
                                     0.078
                                     0.090
                                      15%
                                     0.090
                                      16%
                                 Sauvie_Island
                                    July 28
                                     0.063
                                     0.086
                                      37%
                                     0.087
                                      39%
                                    Mt_View
                                    July 26
                                     0.071
                                     0.060
                                     -15%
                                     0.076
                                      7%
                                    Mt_View
                                    July 27
                                     0.078
                                     0.064
                                     -17%
                                     0.086
                                      10%
                                    Mt_View
                                    July 28
                                     0.076
                                     0.072
                                      -5%
                                     0.088
                                      16%
                                   Milwaukie
                                    July 26
                                     N/A 
                                     0.082
                                     N/A 
                                     0.086
                                     N/A 
                                   Milwaukie
                                    July 27
                                     N/A 
                                     0.085
                                     N/A 
                                     0.091
                                     N/A 
                                   Milwaukie
                                    July 28
                                     N/A 
                                     0.087
                                     N/A 
                                     0.098
                                     N/A 
                                     Carus
                                    July 26
                                     0.117
                                     0.088
                                     -25%
                                     0.090
                                     -23%
                                     Carus
                                    July 27
                                     0.088
                                     0.093
                                      5%
                                     0.097
                                      10%
                                     Carus
                                    July 28
                                     0.098
                                     0.107
                                      9%
                                     0.107
                                      9%
                                    Turner
                                    July 26
                                     0.076
                                     0.070
                                      -8%
                                     0.074
                                      -3%
                                    Turner
                                    July 27
                                     0.098
                                     0.095
                                      -3%
                                     0.101
                                      3%
                                    Turner
                                    July 28
                                     0.078
                                     0.089
                                      13%
                                     0.089
                                      14%
                                    Wishram
                                    July 26
                                     0.060
                                     0.054
                                     -10%
                                     0.055
                                      -8%
                                    Wishram
                                    July 27
                                     0.063
                                     0.054
                                     -14%
                                     0.056
                                     -11%
                                    Wishram
                                    July 28
                                     0.061
                                     0.058
                                      -4%
                                     0.059
                                      -2%
Table 9. cont.
                                   Hockinson
                                    July 26
                                     0.070
                                     0.054
                                     -23%
                                     0.062
                                     -12%
                                   Hockinson
                                    July 27
                                     0.083
                                     0.058
                                     -30%
                                     0.069
                                     -16%
                                   Hockinson
                                    July 28
                                     0.082
                                     0.066
                                     -19%
                                     0.075
                                      -8%

Note: 	"1x1"  -  data extracted from the exact grids.
      "5x5"  -  data extracted from the surrounding 5x5 grids.
 


A scatter diagram of average observed and simulated 8-hr daily maxima (paired in space) is shown in Figure 22.  When plotted with the surrounding 5x5 grid points, the LSM-1 run predicted the peak and the lowest concentrations fairly well with R[2] of 0.63, but there was some overprediction of the moderate values (70-80 ppb).  When looking at the exact grid cell, there was more scatter around the moderate ozone concentrations. 

Quantile-Quantile plots (Figure 23a and 23b) are also plotted for the observed and simulated 8-hr daily maxima which compares the frequency distributions of the observed and predicted ozone concentrations.  The plots show generally good agreement between the observed and predicted concentration distributions, but there was some underperdiction of the peak ozone and overprediction of moderate ozone (80-90 ppb) when evaluated using the range of concentrations within the 5x5 surrounding grid points.  When looking the exact grid point, the bias was similar except for some additional underprediction for the moderate concentrations (around 70 ppb).




Figure 22.  Scatter plot of the average observed and predicted 8-hr daily maxima for 5x5 and 1x1 grids for July 1998 episode

Figure 23a. Quantile-Quantile plot of observed and predicted 8-hr daily maxima using all data pairs for 5x5 and 1x1 grids for July 1998 episode. 



Figure 23b. Quantile-Quantile plot of observed and predicted 8-hr daily maxima using spatially paired means for 5x5 and 1x1 grids for July 1998 episode.

6.2.2 Comparisons with NOx/VOC monitoring data
In 1998, there were four NOx monitoring sites in southwest Washington and Oregon: Lafayette, Castle Rock, Woodland, and Mt. View.  In addition to the three sites mentioned previously, Lafayette is a suburban site very close to the Milwaukie ozone monitoring site.  A major roadway (US-26) is located about 100 m north of the Lafayette site.  Predicted and observed NO/NO2 from 6 to 21 PDT on each day (July 26-28) are shown in Figure 24 for each monitoring site. 

At Lafayette, there were NO peaks in the morning associated with early morning emissions into a shallow mixed layer and lower NO concentrations during the remainder of the daylight hours.  The LSM-1 run appeared to underestimate the early morning peaks and slightly overestimated observed levels from noon into the afternoon.  NO2 concentrations were also generally high in the morning, and lower in the afternoon.  The model captured the trend of NO2 well at most times, especially on July 27.  At Woodland, early morning NO peaks were underestimated and predicted NO2 concentrations were also generally lower than the observed ones.  At Castle Rock, both NO and NO2 were quite low and there was a lack of a distinct diurnal pattern due to its rural location.  The LSM-1 run predicted NO well and also captured the range of NO2 values.  At Mt. View, another urban/suburban site, the LSM-1 run underpredicted morning peaks of NO on the first two days, and overpredicted on the last day. Concentrations of NO2 were also overestimated at this site, especially for the afternoon hours.  One thing worth noticing is that, the grid points of Mt. View and Lafayette both have fairly large NO emissions related to mobile sources, which might have a large impact on the predicted NO/NO2 concentrations.

VOC measurements were also collected at Lafayette on selected days in 1998.  The simulated and observed VOC concentrations are shown at Figure 25.  The VOC data were canister measurements with sampling time from 5-8 and 8-11 PDT on selected days.  The model values were converted from CMAQ output by summing of selected species (alkanes, aromatics, alkenes, terpenes, and isoprene).  The observed NMOC was mostly within 150-200 ppbC except for July 28 when value of 740 ppbC was reported. The LSM-1 run simulated VOC levels perfectly on July 27.  There was a slight underestimation on July 26. On July 28, the LSM-1 run missed the elevated level in the early morning and then slightly overestimated the total VOC concentration at 8-11 PDT.


Figure 24. Observed and simulated NO/NO2 concentrations at Lafayette, Woodland, Castle Rock, and Mt. View from July 26 to 28, 1998.













                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       

                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       


                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       

                                       
                                       
                                       
                                       
                                       
Figure 25. Simulated and observed VOC (ppmC) at Lafayette during the July 1998 episode.



6.3 CMAQ modeling results and evaluation (July 1997 episode)
Surface ozone contours from the LSM-1 run are shown in Figure 26a and 26b for the afternoon of July 19 and 20.  On July 19, elevated ozone concentrations were simulated downwind of Portland with maximum slightly above 100 ppb.  On the next day, high level ozone was simulated right around the Portland urban area with maximum close to 130 ppb.  


                                       

Figure 26a. Simulated surface ozone contour on July 19, 1997 at 17PDT.


                                       
Figure 26b. Simulated surface ozone contour on July 20, 1997 at 17PDT.


Predicted and observed surface ozone time series were plotted in Figure 27 for each monitoring site.  At Castle Rock, Woodland, and Mt. View, ozone concentrations were quite low on most days except July 19 when moderate concentrations (around 60 ppb) occurred.  The LSM-1 run slightly overestimated ozone levels on July 19 and simulated most low ozone days well. Nevertheless, there was significant overestimation on July 20 at all three sites.  As mentioned in the meteorological modeling section, there were also underestimation of cloud cover and overestimation of surface temperatures on that day around these sites, which might explain the large overestimation of ozone. 

At Mt. View and Milwaukie, moderate values occurred on July 19 at both sites and on July 20 at Mt. View.  The peak 1-hr average ozone reached 101 ppb at Milwaukie on July 20, which was the maximum observed in this episode.  The LSM-1 run simulated ozone quite well on low ozone days, and slightly overestimated on moderate and high days.  There also appeared to be some time delay of the ozone peaks at these two sites. 

At Carus and Turner, ozone concentrations were around 80 ppb on July 19 and 20, and quite low on rest of the days.  The LSM-1 run showed some overestimation but captured the overall trend quite well on July 19 and 20.  On the low ozone days, both sites showed very good agreement between observations and simulations. 

At Hockinson and Wishram, ozone levels were low to moderate throughout the episode.  The LSM-1 run simulated ozone quite well at Wishram.  At Hockinson, where simulation was poor in the 1998 episode, model performance was better.  It simulated the right level of ozone on July 20, although the timing of the peak was shifted. 



Figure 27. Observed and simulated 1-hr surface ozone concentrations for July 1997 episode.
















6.3.1 CMAQ statistics evaluation 
Model performance was evaluated using the performance measures based on the simulations from July 17 to 21.  Normalized bias and normalized gross error between observed and predicted 1-hr and 8-hr ozone concentrations are shown in Table 10.  Normalized bias between observed and predicted 8-hr daily maxima using spatially paired means and all data pairs are shown in Table 11 and 12, respectively.

The three upwind sites, Castle Rock, Woodland, and Sauvie Island, all showed positive normalized bias around 20% due to the large overestimation of ozone on July 20.  Normalized bias between observed and predicted 8-hr daily maxima using spatially paired means was around 20% if evaluated at the exact grid, and 30% using the maxima from the 5x5 grids as shown in Table 11. 

Mt. View and Milwaukie, the two urban sites, had a negative (1hr NB = -21%) and positive bias (1hr NB = 15 %), respectively.  The negative bias at Mt. View was mostly due to underprediction on low ozone days.  The two sites both had relatively large normalized errors, partly due to the timing shift of peaks.  On July 19 and 20, the normalized bias between the observed and predicted 8-hr daily maxima was about 20% if evaluated at the exact grid as shown in Table 12. 

The run showed fairly good statistical performance at the Carus and Turner downwind sites (1-hr NB = 0% and -7%; 1-hr NE = 22% and 17%, respectively).  Normalized bias between observed and predicted 8-hr daily maxima using spatially paired means or all data pairs was also mostly below 20%.  Reasonably good performance was found at the Wishram and Hockinson sites. 




Table 10. Normalized bias and normalized gross error between observed and predicted 1-hr/8-hr ozone concentrations. (1-hr statistics are calculated within 10-19 PDT, and 8-hr statistics are calculated within 14-23 PDT from July 17 to 21, 1997.)
                                     Site
                                    1-hr NB
                                    1-hr NE
                                    8-hr NB
                                    8-hr NE
                                  Castle Rock
                                      22%
                                      37%
                                      18%
                                      32%
                                   Woodland
                                      17%
                                      31%
                                      16%
                                      27%
                                 Sauvie Island
                                      22%
                                      31%
                                      20%
                                      27%
                                   Mt. View
                                     -21%
                                      43%
                                     -23%
                                      36%
                                   Milwaukie
                                      15%
                                      46%
                                      5%
                                      27%
                                     Carus
                                      0%
                                      22%
                                      -3%
                                      14%
                                    Turner
                                      -7%
                                      17%
                                      -8%
                                      12%
                                    Wishram
                                      -5%
                                      12%
                                      -5%
                                      9%
                                   Hockinson
                                      -3%
                                      25%
                                      -1%
                                      19%
Note: Values higher than EPA suggested model performance goal are highlighted in red. 



Table 11. Normalized bias between observed and predicted 8-hr daily maxima using spatially paired means. (based on 5x5 and 1x1 grids)
                                     Site
                                   Ave obs 
                                8-hr daily max
                                     (ppm)
                       Ave pre 8-hr daily max 1x1 (ppm)
                                   NB (1x1)
                       Ave pre 8-hr daily max 5x5 (ppm)
                                   NB (5x5)
                                  Castle Rock
                                     0.034
                                     0.040
                                      19%
                                     0.044
                                      32%
                                   Woodland
                                     0.036
                                     0.043
                                      20%
                                     0.045
                                      27%
                                 Sauvie Island
                                     0.036
                                     0.044
                                      24%
                                     0.046
                                      29%
                                   Mt. View
                                     0.040
                                     0.036
                                     -10%
                                     0.050
                                      26%
                                   Milwaukie
                                     0.038
                                     0.044
                                      15%
                                     0.050
                                      29%
                                     Carus
                                     0.045
                                     0.048
                                      6%
                                     0.054
                                      21%
                                    Turner
                                     0.047
                                     0.045
                                      -4%
                                     0.050
                                      7%
                                    Wishram
                                     0.046
                                     0.043
                                      -6%
                                     0.045
                                      -1%
                                   Hockinson
                                     0.042
                                     0.040
                                      -6%
                                     0.045
                                      6%
Note:	 "1x1"  -  data extracted from the exact grids.
       "5x5"  -  data extracted from the surrounding 5x5 grids


Table 12. Normalized bias between observed and predicted 8-hr daily maxima using all data pairs. (based on 5x5 and 1x1 grids)
                                     Site
                                     Date
                              Obs 8-hr daily max
                                     (ppm)
                         Pre 8-hr daily max 1x1 (ppm)
                                   NB (1x1)
                         Pre 8-hr daily max 5x5 (ppm)
                                   NB (5x5)
                                  Castle Rock
                                    July 17
                                     0.028
                                     0.021
                                     -26%
                                     0.023
                                     -16%
                                  Castle Rock
                                    July 18
                                     0.032
                                     0.036
                                      11%
                                     0.044
                                      35%
                                  Castle Rock
                                    July 19
                                     0.058
                                     0.064
                                      12%
                                     0.066
                                      15%
                                  Castle Rock
                                    July 20
                                     0.028
                                     0.056
                                      99%
                                     0.061
                                     114%
                                  Castle Rock
                                    July 21
                                     0.022
                                     0.024
                                      7%
                                     0.028
                                      25%
Table 12. Cont.
                                   Woodland
                                    July 17
                                     0.033
                                     0.025
                                     -24%
                                     0.026
                                     -20%
                                   Woodland
                                    July 18
                                     0.034
                                     0.033
                                      -4%
                                     0.037
                                      6%
                                   Woodland
                                    July 19
                                     0.053
                                     0.069
                                      29%
                                     0.069
                                      29%
                                   Woodland
                                    July 20
                                     0.032
                                     0.060
                                      85%
                                     0.067
                                     107%
                                   Woodland
                                    July 21
                                     0.026
                                     0.027
                                      5%
                                     0.028
                                      10%
                                 Sauvie_Island
                                    July 17
                                     0.029
                                     0.024
                                     -18%
                                     0.026
                                     -12%
                                 Sauvie_Island
                                    July 18
                                     0.033
                                     0.036
                                      6%
                                     0.038
                                      14%
                                 Sauvie_Island
                                    July 19
                                     0.052
                                     0.069
                                      32%
                                     0.069
                                      32%
                                 Sauvie_Island
                                    July 20
                                     0.038
                                     0.065
                                      71%
                                     0.070
                                      83%
                                 Sauvie_Island
                                    July 21
                                     0.025
                                     0.027
                                      8%
                                     0.028
                                      10%
                                    Mt_View
                                    July 17
                                     0.028
                                     0.009
                                     -68%
                                     0.023
                                     -17%
                                    Mt_View
                                    July 18
                                     0.035
                                     0.032
                                      -9%
                                     0.041
                                      18%
                                    Mt_View
                                    July 19
                                     0.053
                                     0.052
                                      -1%
                                     0.070
                                      33%
                                    Mt_View
                                    July 20
                                     0.062
                                     0.075
                                      21%
                                     0.091
                                      47%
                                    Mt_View
                                    July 21
                                     0.023
                                     0.012
                                     -48%
                                     0.026
                                      17%
                                   Milwaukie
                                    July 17
                                     0.022
                                     0.014
                                     -37%
                                     0.019
                                     -10%
                                   Milwaukie
                                    July 18
                                     0.028
                                     0.034
                                      22%
                                     0.038
                                      34%
                                   Milwaukie
                                    July 19
                                     0.055
                                     0.068
                                      25%
                                     0.075
                                      38%
                                   Milwaukie
                                    July 20
                                     0.073
                                     0.089
                                      23%
                                     0.096
                                      33%
                                   Milwaukie
                                    July 21
                                     0.016
                                     0.017
                                      7%
                                     0.020
                                      30%
                                     Carus
                                    July 17
                                     0.030
                                     0.026
                                     -15%
                                     0.027
                                     -10%
                                     Carus
                                    July 18
                                     0.037
                                     0.037
                                      0%
                                     0.043
                                      16%
                                     Carus
                                    July 19
                                     0.065
                                     0.071
                                      10%
                                     0.080
                                      24%
                                     Carus
                                    July 20
                                     0.068
                                     0.080
                                      16%
                                     0.087
                                      27%
                                     Carus
                                    July 21
                                     0.026
                                     0.026
                                      4%
                                     0.035
                                      37%
                                    Turner
                                    July 17
                                     0.028
                                     0.023
                                     -18%
                                     0.028
                                      0%
                                    Turner
                                    July 18
                                     0.047
                                     0.045
                                      -5%
                                     0.049
                                      3%
                                    Turner
                                    July 19
                                     0.067
                                     0.072
                                      7%
                                     0.075
                                      12%
                                    Turner
                                    July 20
                                     0.061
                                     0.058
                                      -4%
                                     0.067
                                      10%
                                    Turner
                                    July 21
                                     0.032
                                     0.027
                                     -15%
                                     0.032
                                      2%
                                    Wishram
                                    July 17
                                     0.040
                                     0.038
                                      -5%
                                     0.039
                                      -1%
                                    Wishram
                                    July 18
                                     0.039
                                     0.040
                                      5%
                                     0.041
                                      7%
                                    Wishram
                                    July 19
                                     0.054
                                     0.048
                                     -11%
                                     0.050
                                      -7%
                                    Wishram
                                    July 20
                                     0.061
                                     0.052
                                     -14%
                                     0.056
                                      -8%
                                    Wishram
                                    July 21
                                     0.037
                                     0.038
                                      5%
                                     0.040
                                      9%
                                   Hockinson
                                    July 17
                                     0.035
                                     0.027
                                     -21%
                                     0.029
                                     -17%
                                   Hockinson
                                    July 18
                                     0.036
                                     0.041
                                      15%
                                     0.047
                                      30%
                                   Hockinson
                                    July 19
                                     0.056
                                     0.044
                                     -21%
                                     0.052
                                      -7%
                                   Hockinson
                                    July 20
                                     0.057
                                     0.054
                                      -5%
                                     0.062
                                      10%
                                   Hockinson
                                    July 21
                                     0.028
                                     0.031
                                      14%
                                     0.033
                                      21%

Note:	 "1x1"  -  data extracted from the exact grids.
       "5x5"  -  data extracted from the surrounding 5x5 grids.



A scatter diagram of average observed and simulated 8-hr daily maxima (paired in space) are shown in Figure 28.  Unlike the July 1998 episode in which only three peak ozone days were used in averaging, both low and high ozone days were included in the 1997 case.  Therefore, all spatially paired means were within the 30-50 ppb range.  When plotted using the exact grid cell, there was some slight overprediction.  When looking at the 5x5 grid cells, ozone at most sites was overestimated. 

Quantile-Quantile plots (Figure 29a and 29b) are plotted for the observed and simulated 8-hr daily maxima using all data pairs and spatially paired means.  When using all data pairs, it appears that the LSM-1 run tended to overpredict for moderate and high ozone levels, and performed better at low concentrations.  When using spatially paired means, the model overpredicted at both the exact grid and the 5x5 grids. 





Figure 28. Scatter plot of the average observed and predicted 8-hr daily maxima for 5x5 and 1x1 grids for July 1997 episode.

Figure 29a. Quantile-Quantile plot of observed and predicted 8-hr daily maxima using all data pairs for 5x5 and 1x1 grids for July 1997 episode. 


Figure 29b. Quantile-Quantile plot of observed and predicted 8-hr daily maxima using spatially paired means for 5x5 and 1x1 grids for July 1997 episode.

6.3.2 Comparisons with NOx monitoring data
NOx data were available at Castle Rock, Mt. View, and Lafayette for the July 1997 ozone episode.  Predicted and observed NO and NO2 concentrations from 6 to 21 PDT are shown for each site in Figure 30 for July 17-21, 1997. 

The observed NOx peaks in the 1997 episode were lower than those in the 1998 episode at Mt. View and Lafayette.  The LSM-1 run showed a large overestimation of NO on the non-episode days (July 17, 18, and 21) at all three sites.  In contrast, the agreement was much better for NO during days with elevated ozone (July 19 and 20).  Note, that these were weekend days with smaller NO emissions from mobile sources.  For NO2, the LSM-1 run showed some overprediction at Mt. View and Lafayette, but matched the general trend at most times.  At Castle Rock, there was a large overestimation of NO2 on the non-episode days.  

The VOC data collected at the southwest Washington sites in 1997 were reported to have abnormally high levels of aromatics due to contamination issues (Westberg et al, 1998), therefore no comparisons were made for this episode.  



Figure 30. Observed and simulated NO/NO2 concentrations at Lafayette, Castle Rock, and Mt. View from July 17 to 21, 1997.



















































































































































































































































































































































































































































































































6.4 Conclusion for July 1998 and 1997 CMAQ simulation
Two historical episodes (July 17-21, 1997 and July 26-28, 1998) were selected as potential base episodes for this modeling work because the July 1998 episode had the highest ozone levels observed in recent years for the airshed and the July 1997 episode had slightly different meteorological patterns compared to the 1998 case.  For the July 1998 episode, CMAQ model performance statistics for most monitoring sites were within EPA suggested 1-hr and 8-hr ozone modeling guidelines as shown in section 6.2.1.  Time series plots, scatter plots, and Quantile-Quantile plots also showed generally good agreement between model simulations and observations.  For the July 1997 episode, the observed ozone levels were much lower than those in the 1998 case.  The CMAQ performance for ozone was also acceptable for the 1997 run, but there was less accurate simulation of ozone levels at different times and locations compared to the 1998 case.  Therefore, the July 1998 episode was selected as the base for evaluation of future year (2015) control strategy simulations due to higher observed ozone concentrations and better simulation of ozone levels. 


7 Future episode modeling (2015) 

7.1 Future year managed growth maintenance projection simulation
7.1.1 CMAQ modeling results
To investigate the impact of changes in emissions due to future growth and various control strategies, a year 2015 emission inventory was developed.  This was then used in a model run with July 1998 meteorology to create a 2015 managed growth maintenance projection for ozone simulation. Emissions comparisons for 2015 and 1998 are shown in Table 13.  For year 2015, the NOx emissions are reduced by 49%. In contrast, the VOC emissions reduction is only 26%, and consequently, the VOC/NOx ratio increases from 0.73 in 1998 to 1.06 in 2015.  There are also 41% and 16% reductions in SOx and CO emissions in 2015 compared to 1998. 

Table 13. Anthropogenic emission comparison between the 2015 managed growth maintenance projection and 1998 run 
                                       
                                     1998
                                  2015 base 
                               Percentage change
                                       
                                   Mole/day
                                   Mole/day
                                       
                                      VOC
                                     1253
                                      924
                                     -26%
                                      NOx
                                     1722
                                      875
                                     -49%
                                      SOx
                                      166
                                      99
                                     -41%
                                      CO
                                     23264
                                     19455
                                     -16%
                                    VOC/NOx
                                     0.73
                                     1.06
                                      45%
Note: Emissions are daily averages (July 24  - 29) summed over the central domain (grid cells 10,20 to 45,55)


Simulated ozone surface contours at 17PDT on July 28, 2015 are shown in Figure 31a.  For comparison, the contours from the 1998 run and the difference plot between the two years are shown in Figure 31b and 31c, respectively.  In the 2015 episode, the maximum 1-hr average ozone reached 128 ppb at 17 PDT on July 28 slightly south of Portland, as compared to 148 ppb in 1998.  The areas with elevated ozone appear to be similar between the two years, although the absolute ozone values in 2015 were lower.  As shown in Figure 31c, large reductions (20-30 ppb) mainly occur downwind of Portland extending from close to Carus to south of Salem.  In contrast to the large decrease of ozone downwind of urban center, the change within Portland itself is smaller (< 15 ppb). 



                                       
                                       
Figure 31a. Simulated ozone contour at 17 PDT on July 28, 2015
                                       
                                       
                                       
                                       
Figure 31b. Simulated ozone contour at 17 PDT on July 28, 1998

                                       

                                       

Figure 31c. Ozone difference contour at 17 PDT on July 28 between 1998 and 2015 (1998  -  2015)


Simulated 1-hr and 8-hr surface ozone concentrations for the 2015 managed growth maintenance projection and July 1998 episode are shown in Figure 32 for each monitoring site.  At the Castle Rock, Woodland, and Sauvie Island sites, there is a daytime decrease and slight night increase of ozone in 2015 compared to 1998. 

At Mt. View and Milwaukie, the daytime reduction is relatively small and sometimes the daily maxima occur slightly earlier in the future year run.  On July 28, for example, the 1-hr ozone peak at Milwaukie is 2-hr earlier and also broader in 2015 compared to 1998.  As a result, the 8-hr maximum ozone on that day increases slightly (89 ppb versus 87 ppb) in the future year run.  

At Carus and Turner, the daytime reductions are large (up to 30 ppb at Turner), but nighttime ozone remains about the same level as 1998.  At Wishram and Hockinson, in addition to the daytime decrease of ozone, nighttime values are also slightly lowered. 

Figure 32. Simulated 1-hr and 8-hr surface ozone concentrations for 2015 managed growth maintenance projection and July 1998 episode.















7.1.2 VOC/NOx ratios and regional increase of ozone
The [O3]/[NOx] concentration ratio has been shown to be a reliable indicator of O3 response to VOC or NOx emissions changes (Tonnesen and Dennis, 2000b).  Values of the ratio < 15 indicate the strongly radical-limited, VOC sensitive condition.  Values > 46 indicate strongly NOx-limited and therefore NOx sensitive region (Arnold et al., 2003).  The predicted [O3]/[NOx] contours in the afternoon of 2015 and 1998 are shown in Figure 33a and b.  For 1998, the entire Portland urban area and I-5 corridor appear to be well into the VOC-sensitive region.  In contrast, for 2015 the VOC-limited areas are much smaller, which is consistent with the increased VOC/NOx ratio.  This indicator helps to explain the ozone increase at Milwaukie in 2015.  The site appears to be strongly VOC limited in the afternoon hours of 1998, and therefore, excess NO titrates O3 and reduces the O3 level.  In 2015, [O3]/[NOx] is slightly above 15 at the site.  Being less VOC-limited and hence having less NO titration, an increase of O3 could be expected. 

Another way to understand the regional increase of ozone is to look at process analysis output from CMAQ.  Process analysis is a feature in CMAQ where the contributions of different chemical/physical processes are tracked by grid cell for each hour of simulations.  For example, ozone chemical production rates are mapped for the two years in Figure 34a and b.  In 1998 there is a net loss of ozone due to NOx titration in the urban core.  In contrast in 2015, ozone is produced at a net rate of 10 to 20 ppb/hr within the urban core.  Almost no NOx titration effects are observed in the 2015 case.  Therefore, higher ozone values might occur in 2015 for the sites within urban center. 



Figure 33a. O3/NOx contour plot at 15 PDT on July 28, 2015


Figure 33b. O3/NOx contour plot at 15 PDT on July 28, 1998



Figure 34a. Ozone chemical production rates from process analysis at 15 PDT on July 28, 2015


Figure 34b. Ozone chemical production rates from process analysis at 15 PDT on July 28, 1998

7.2 What if scenario - No lower Columbia River point sources
Following the 2015 managed growth maintenance projection, a `what-if' control scenario was formulated for analysis.  The scenario assumed no lower Columbia River point sources.  This effectively means no large industrial activity along the lower Columbia River.  Emissions comparisons between the 2015 case and the no lower Columbia River point source case are shown in Table 14.  There is a 6% reduction in VOC emissions and an 11% reduction in NOx emissions due to the removal of the lower Columbia River point sources. 


Table 14. Anthropogenic emissions comparisons between the 2015 managed growth maintenance projection and the no lower Columbia River point source case.
                                       
                               2015 projection 
                              2015 no_columbia_pt
                               Percentage change
                                       
                                   Mole/day
                                   Mole/day
                                       
                                      VOC
                                      924
                                      870
                                      -6%
                                      NOX
                                      875
                                      779
                                     -11%
                                      SOX
                                      99
                                      55
                                     -44%
                                      CO
                                     19455
                                     19046
                                      -2%
                                    VOC/NOx
                                     1.06
                                     1.12
                                      6%
Note: Emissions are daily averages (July 24  - 29) summed over the central domain (grid cells 10,20 to 45,55)



Simulated ozone contours for the 2015 managed growth maintenance projection, the no lower Columbia River point source case, and the difference plots are shown in Figure 35a-c.  Simulated 1-hr and 8-hr surface ozone concentrations for the two cases are shown in Figure 36 for the sites with large ozone changes.  A summary of VOC/NOx emission ratios and predicted 8-hr daily maximum ozone (from the surrounding 5x5 grids) for 1998 episode, 2015 managed growth maintenance projection, and 2015 no lower Columbia River point sources modification are shown in Table 15 and Figure 37.

The amount of O3 reduction is up to 16 ppb for the case with no lower Columbia River point sources.  The maximum reduction does not occur in the same region as the peak ozone is predicted to occur, and therefore the maximum ozone in the Portland/Vancouver airshed is not affected.  The most affected sites are at Woodland and Sauvie Island where up to a 12 ppb decrease was predicted for the afternoon of June 26.  At these two sites, there was also some increase in ozone at night due to less NO titration.  At Mt View and Milwaukie, a small reduction was predicted, but not at the peak ozone hours.  At the other sites, the changes in O3 concentrations were almost negligible.  

Further analysis of other potential control strategies, based upon modifications to the 2015 managed growth maintenance projection, will be conducted by ODEQ and SWCAA as part of the overall maintenance plan development.  


Figure 35a. Simulated ozone contour for 2015 no lower Columbia River point source case at 15 PDT on July 26


Figure 35b. Simulated ozone contour for 2015 projection at 15 PDT on July 26



Figure 35c. Ozone difference plots between the 2015 managed growth maintenance projection and its no Lower Columbia River point sources modification  at 15 PDT on July 26, 2015 
(2015  -  no Lower Columbia point sources)

Figure 36. Simulated 1-hr and 8-hr surface ozone concentrations for the 2015 managed growth maintenance projection and its no lower Columbia River point sources modification.







Table 15. VOC/NOx ratio and predicted 8-hr daily maximum ozone (from the surrounding 5x5 grids) for 1998 episode, 2015 projection, and 2015 no lower Columbia River modification. 
                                       
                                       
                                   1998 base
                                   2015 base
                              2015 no_Columbia_pt
                                    VOC/NOx
                                       
                                     0.73
                                     1.06
                                     1.12
                                     Site
                                     Date
                         Pre 8-hr daily max 5x5 (ppm)
                         Pre 8-hr daily max 5x5 (ppm)
                         Pre 8-hr daily max 5x5 (ppm)
                                  Castle Rock
                                    26-Jul
                                     0.075
                                     0.060
                                     0.060
                                  Castle Rock
                                    27-Jul
                                     0.091
                                     0.077
                                     0.077
                                  Castle Rock
                                    28-Jul
                                     0.079
                                     0.068
                                     0.068
                                   Woodland
                                    26-Jul
                                     0.073
                                     0.062
                                     0.057
                                   Woodland
                                    27-Jul
                                     0.091
                                     0.075
                                     0.073
                                   Woodland
                                    28-Jul
                                     0.083
                                     0.072
                                     0.072
                                 Sauvie_Island
                                    26-Jul
                                     0.070
                                     0.061
                                     0.054
                                 Sauvie_Island
                                    27-Jul
                                     0.090
                                     0.073
                                     0.071
                                 Sauvie_Island
                                    28-Jul
                                     0.087
                                     0.075
                                     0.074
                                    Mt_View
                                    26-Jul
                                     0.076
                                     0.071
                                     0.070
                                    Mt_View
                                    27-Jul
                                     0.086
                                     0.072
                                     0.069
                                    Mt_View
                                    28-Jul
                                     0.088
                                     0.082
                                     0.081
                                   Milwaukie
                                    26-Jul
                                     0.086
                                     0.084
                                     0.084
                                   Milwaukie
                                    27-Jul
                                     0.091
                                     0.085
                                     0.084
                                   Milwaukie
                                    28-Jul
                                     0.098
                                     0.095
                                     0.095
                                     Carus
                                    26-Jul
                                     0.090
                                     0.077
                                     0.077
                                     Carus
                                    27-Jul
                                     0.097
                                     0.085
                                     0.084
                                     Carus
                                    28-Jul
                                     0.107
                                     0.095
                                     0.095
                                    Turner
                                    26-Jul
                                     0.074
                                     0.060
                                     0.059
                                    Turner
                                    27-Jul
                                     0.101
                                     0.078
                                     0.076
                                    Turner
                                    28-Jul
                                     0.089
                                     0.068
                                     0.068
                                    Wishram
                                    26-Jul
                                     0.055
                                     0.050
                                     0.052
                                    Wishram
                                    27-Jul
                                     0.056
                                     0.049
                                     0.050
                                    Wishram
                                    28-Jul
                                     0.059
                                     0.050
                                     0.051
                                   Hockinson
                                    26-Jul
                                     0.062
                                     0.050
                                     0.049
                                   Hockinson
                                    27-Jul
                                     0.069
                                     0.059
                                     0.059
                                   Hockinson
                                    28-Jul
                                     0.075
                                     0.064
                                     0.065

                                     mean
                                     0.081
                                     0.070
                                     0.069

                                      max
                                     0.107
                                     0.095
                                     0.095

                                      min
                                     0.055
                                     0.049
                                     0.049




Figure 37. Predicted 8-hr daily maximum ozone (from surrounding 5x5 grids) for 1998 episode, 2015 managed growth maintanence projection, and 2015 no lower Columbia River point source modification.



 
References

J.R. Arnold, R.L. Dennis and G.S. Tonnesen, 2003. Diagnostic evaluation of numerical air quality models with specialized ambient observations: testing the Community Multiscale Air Quality modeling system (CMAQ) at selected SOS 95 ground sites, Atmos. Environ. 37, pp. 1185 - 1198.

Byun, D.W., Ching, J.K.S. (Eds.), 1999. Science algorithms of the EPA Models-3 Community Multiscale Air Quality (CMAQ) Modeling System. EPA Report No. EPA-600/R-99/030, Office of Research and Development, US Environmental Protection Agency, Washington, DC.

Chen, F., and J. Dudhia, 2001. Coupling an advanced land surface - hydrology model with the Penn State - NCAR MM5 modeling system. Part I: Model implementation and sensitivity. Mon. Wea. Rev., 129, 569 - 604.

Chen, J.C.  M.S. thesis, Washington State University, Department of Civil and Environmental Engineering, Pullman, WA, 2002.

Grell, G.A., Dudhia, J., Stau!er, D.R., 1994. A description of the "fifth generation Penn State/NCAR mesoscale model (MM5). NCAR Technical Note NCAR/TN-398#STR, 138 pp.

H. Westberg, B. Hopkins, E. Allwine and B. Lamb, 1998. Speciated VOC Measurements in Western Washington during the Summer of 1997, Washington State University Laboratory for Atmospheric Research.

Hong, S.-Y., and H.-L. Pan, 1996. Nonlocal boundary layer vertical diffusion in a medium-range forecast model. Mon. Wea. Rev., 124, 2322 - 2339.

Kain, J. S., and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain - Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 165 - 170.

Liu, Y., and co-authors, 2004: Improvements to surface flux computations in the MRF PBL scheme, and refinements to urban processes in the Noah Land-Surface model with the NCAR/ATEC Real-Time FDDA and forecast system. The 5th WRF / 14th MM5 User's Workshop , Boulder, CO. June 22-25, 2004.

O'Neill, S.M.; Lamb, B.K., 2004.  Inter-comparison of the Community Multi-scale Air Quality (CMAQ) model and CALGRID using Process Analysis. Environ. Sci. Technol., 39, 5742-5763.

O'Neill, S., B. Lamb, J. Chen, C. Claiborn, D. Finn, S. Otterson, C. Figueroa, C. Bowman, M. Boyer, R. Wilson, J. Arnold, S. Aalbers, J. Slocum, C. Swab, M. Stoll, M. DuBois, and M. Anderson, 2005.  Modeling aerosol formation and transport in the Pacific Northwest with the Community Multi-scale air quality (CMAQ) modeling system, Environ. Sci. & Technol. , in press.

Tonnesen, G.S. and Dennis, R.L., 2000. Analysis of radical propagation efficiency to assess ozone sensitivity to hydrocarbons and NOX. Part II: long-lived species as indicators or ozone concentration sensitivity. Journal of Geophysical Research 105, pp. 9227 - 9241.

U.S. EPA, (May 1999). Draft Guidance on the Use of Models and Other Analyses in Attainment Demonstrations for the 8-Hour Ozone NAAQS, EPA-454/R-99-004, http://www.epa.gov/scram001/guidance/guide/drafto3.pdf

Vukovich, J. and T. Pierce, 2002. The Implementation of BEIS3 within the SMOKE Modeling Framework, In Proceedings of the 11thInternational Emissions Inventory Conference, Atlanta, Georgia, April 15-18, 2002. 
(Available online:www.epa.gov/ttn/chief/conference/ei11/modeling/vukovich.pdf)

Xiu, A., and J. E. Pleim, 2001. Development of a land surface model. Part I: Application in a mesoscale meteorological model. J. Appl. Meteor., 40, 192 - 209.

Zhang, D.-L., and R. A. Anthes, 1982. A high-resolution model of the planetary boundary layer -- sensitivity tests and comparisons with SESAME-79 data. J. Appl. Meteor, 21, 1594 - 1609.

Zhang, D.-L., and W. Zheng, 2004: Diurnal cycles of surface winds and temperatures as simulated by five boundary-layer parameterizations. J. Appl. Meteor., 43, 157-169.
                                       
                                       
                                       
Appendix A. Emission Inventory Documentation 

1998 and 1997 Methodologies

Biogenic
Biogenic emissions were processed using BEIS v3.09 (Vukovich and Pierce, 2002). The model estimates VOC emissions from vegetation based on land-use categories and nitric oxide emissions from microbial activity based on soil types. The BEIS3 modeling system uses the Biogenic Emissions Landcover Database version 3 (BELD3) which consists of 1 km gridded data 230 land-use types. There are two steps involved in BEIS3 modeling. In the first step, the model inputs the BELD3 land-use data and emissions factors, and outputs gridded, normalized emissions (at 30C and 1000 umol m[-2]·s[-1] PAR). In the second step, the model uses meteorological data, speciation profiles, and normalized emissions generated from the previous step, and outputs gridded, speciated, hourly biogenic emissions. The biogenic emissions for 1998 episode are summarized in Table A1 which shows that 58% of the VOC emissions and 2% of the NOx emissions are from biogenic sources in the central part of the domain that encompasses the Portland/Vancouver area. 

Table A1.  Biogenic emissions summed over the central domain for the July 1998 episode.
                                   Species 
                              Biogenic emissions
                                Total emissions
                               Percent biogenic
                                       
                                   Mole/day
                                   Mole/day
                                       %
                                     ALK2
                                      643
                                      699
                                     92.0
                                   ISOPRENE
                                      225
                                      226
                                     99.6
                                     OLE2
                                      76
                                      131
                                     57.5
                                     TRP1
                                      781
                                      783
                                     99.7
                                      NO
                                      37
                                     1673
                                      2.2
                                      VOC
                                     1724
                                     2977
                                     57.9
                                      NOX
                                      37
                                     1759
                                      2.1
Note: Emissions are daily averages (July 24  - 29) summed over the central domain (grid cells 10,20 to 45,55)



Area
The 1999 inventory, developed for submittal by Oregon to EPA for the 1999 NEI, was the starting point for the Oregon 1997 and 1998 area source emission inventory used for model validation. Using population as a surrogate, all the 1999 emissions were scaled back to estimate 1997 and 1998 emissions for the 15 counties in the modeling domain using MS ACCESS. Table A2 below details the development of the fractions used for each county.  In Appendix B, Table B1 lists the area source categories that were inventoried.

        Table A1: Population and Emission Ratios by County
STATE/COUNTY
July 1
July 1
99-97 County
99-98 County
July 1
1997
1998
reduction fraction
reduction fraction
1999
OREGON
3,217,000
3,267,550
3,300,800
BENTON
76,700
76,600
0.9948
0.9935
77,100
CLACKAMAS
317,700
323,600
0.9720
0.9901
326,850
CLATSOP
34,500
34,700
0.9928
0.9986
34,750
COLUMBIA
41,500
42,300
0.9730
0.9918
42,650
HOOD RIVER
19,200
19,500
0.9746
0.9898
19,700
JEFFERSON
17,100
17,400
0.9688
0.9858
17,650
LINCOLN
42,500
43,200
0.9804
0.9965
43,350
LINN
100,700
102,200
0.9777
0.9922
103,000
MARION
267,700
271,900
0.9726
0.9878
275,250
MULTNOMAH
639,000
641,900
0.9879
0.9923
646,850
POLK
57,400
59,500
0.9551
0.9900
60,100
TILLAMOOK
23,800
24,000
0.9876
0.9959
24,100
WASCO
22,600
22,600
0.9978
0.9978
22,650
WASHINGTON
385,000
397,600
0.9512
0.9823
404,750
YAMHILL
79,200
81,900
0.9531
0.9856
83,100
SWCAA and the Washington Department of Ecology did not have a complete 1999 area source submittal to start with so SWCAA built a 1998 EI for the 8 counties in the core Washington domain following EPA protocols as outlined in EIIP references.  The eight counties in the core Washington domain were Clark, Cowlitz, Klickitat, Lewis, Pacific, Skamania, Wahkiakum, and Yakima counties.  WSU staff then used 1996 area source data for the remainder of the domain counties from the Northwest Regional Technical Center (NWRTC) demonstration project work.  The emission categories, and the methodologies used for calculating emissions for these categories are described in Appendix B.




Nonroad Mobile
To create nonroad mobile estimates, excluding aircraft, rail and marine, for the 1997 and 1998 episodes, DEQ and SWCAA used EPA's 2002 NONROAD model.  Collaborative care was taken to assure the consistency and validity of the model inputs between states and accurately reflect the conditions in each state. The NONROAD model allows users to design various scenarios to generate emissions. The period input parameters for the model runs that generated this data were: seasonal, typical day, summer, weekday to provide a conservative estimate of emissions.  The 2002 NONROAD model was run for 1997 and 1998 episodes, with the same input parameters except for the change in year. Table A2 summarizes the fuel parameter inputs representing both episodes.  
                   
                   Table A2. Oregon NONROAD model fuel parameter inputs
                       
                      * Control efficiency in Multnomah, Washington, and Clackamas counties was set to 81% based on DEQ estimates.  For all other counties within the domain, the control efficiency was set to 0.


Fuel RVP is the July ASTM terminal limit RVP for western Oregon and the Portland/Vancouver area (1997).  The O2 and diesel sulfur weight % are DEQ estimates.  Gasoline sulfur is from a state-by-state fuel properties report prepared for the EPA by ICF consulting.  The CNG sulfur is a conservative estimate for HD5 propane rated for fuel use.

Temperature data for Oregon counties for the 1997 and 1998 episode runs were averaged from the temperatures of the top ten ozone days in the three years leading up to and including the original base year of 1996.  Temperature data were taken from the Oregon Climatological Data Survey.  The temperatures were averaged for the episode; the results are shown in Table A3.

               Table A3.  Oregon NONROAD temperature data inputs
                                       
                                       
Temperature data for WA counties in the 1997 and 1998 episode runs were derived from NOAA data at various sites within those counties.  The temperatures used were average temperatures during the peak days of each respected episode.  Table A4 shows the temperatures used for the 8 core counties in 1998 and for the entire WA domain in 1997.

      Table A4.  WA core domain counties NONROAD temperature data inputs

                                                                   1997 Episode

                                                                   1998 Episode


                                   Avg. High
                                      (F)
                                 Avg. Low (F)
                                   Avg. (F)
                                 Avg. High (F)
                                   Avg. Low
                                      (F)
                                   Avg. (F)
Clark
                                                                           78.9
                                                                           52.6
                                                                           65.8
                                                                           86.8
                                                                           58.6
                                                                           72.7
Cowlitz
                                                                           77.6
                                                                           53.8
                                                                           65.7
                                                                           87.1
                                                                           59.6
                                                                           73.4
Klickitat
                                                                           88.4
                                                                           50.1
                                                                           69.3
                                                                          101.4
                                                                           58.8
                                                                           80.1
Lewis
                                                                           79.9
                                                                           54.0
                                                                           67.0
                                                                           89.1
                                                                           61.4
                                                                           75.3
Pacific
                                                                           72.8
                                                                           50.7
                                                                           61.8
                                                                           74.9
                                                                           57.0
                                                                           66.0
Wahkiakum
                                                                           54.0
                                                                           72.9
                                                                           63.5
                                                                           80.8
                                                                           56.4
                                                                           68.6
Skamania
                                                                           80.4
                                                                           47.2
                                                                           63.8
                                                                           92.3
                                                                           57.2
                                                                           74.8
Yakima
                                                                           89.2
                                                                           53.6
                                                                           71.4
                                                                           99.7
                                                                           66.1
                                                                           82.9
                                       
                                       
Aircraft, rail and commercial marine emissions in Oregon were estimated in the same manner as the area sources.  They were back cast based on the population ratios between the 1999 inventory year and each specific episode year.  ODEQ calculated the Columbia River commercial marine emissions and the emissions were apportioned to the Oregon counties that bordered the Columbia River.  In Washington, aircraft and commercial marine emissions from the 1996 NWRTC demonstration project were used because most of these emissions were outside of the 8 county core WA domain.  Washington locomotive emissions were developed by ODEQ for a 1997 locomotive emission inventory.

Washington Department of Ecology assisted in the Nonroad inventory work by processing the Nonroad model outputs for input into the SMOKE model with a customized script.

On-road Mobile

The 1997 and 1998 mobile emissions were calculated based on episode specific, link-based, daily trip VMT data modeled by RTC and Metro using traffic demand models.  MOBILE6.2 was used to calculate the emission factors with the episode-specific conditions.  The summary of the approach to mobile emissions modeling estimates was to:

 Create link-based daily VMT for the AQMA using traffic demand models, and county-wide VMT for the remainder of the domain.
 Create four total MOBILE6.2 input files representing Clark County, Portland Metro Area Counties (both with I/M features); Washington State counties, and Oregon State counties (both without I/M features) for the remainder of the modeling domain.
 Run MOBILE6.2 at a standard temperature and at individual temperatures. 
 Calculate average daily emissions at the standard temperature. 
 Create a set of temperature adjustment factors from the individual temperature runs. 
 Use SMOKE's gridded temperature file (from MM5) to apply adjustments to the standard emissions.

This method allowed adjustment of hourly emissions based on temperature, but involved post-processing of SMOKE data.  ODEQ completed the MOBILE6.2 work with assistance from ECOLOGY and SWCAA for Washington data, and WSU completed the SMOKE processing. The eight (4 per episode year) MOBILE6.2 input files can be found in Appendix C. 


Industrial Point Sources

Industrial Point Sources from both states were loaded into a common database to ease tracking and model input file generation.  Regional staff for Washington and Oregon compiled annual emission information for permitted facilities.  Industrial Point sources that emit more than 10 tons per year of VOC, 40 tons per year of NOx, or 100 tons per year of CO within the 4 counties of the Portland/Vancouver ozone maintenance area were inventoried for this work.  Point sources that emit more than 40 tpy of NOx or 100 tpy of VOC or CO outside the maintenance area counties were also inventoried.  Additionally, point sources emitting less than those thresholds within 4 kilometers of ozone monitoring sites were included in the episode emission inventories. Stack parameters, activity, and exact location were collected when available to provide the most comprehensive accounting possible.

SOX emissions have an indirect influence on ozone, as such these emissions were included in the model run inventories.  No SOX parameters were used in compiling the point source list for this project.  The emissions were included only for sources that met the NOX, CO, and/or VOC thresholds described above.  Facility SOX data for Washington point sources were included in the 1997 and 1998 runs.  For Oregon, 1997 and 1998 estimated actual SOX emissions for TV point sources were generated from the ODEQ ACSIS point source database.  The emissions were split by SCC, but not by stack or device.  SOX data for the 2002 EI that was used to forecast 2015 emissions estimates was included for all sources in both states.  ODEQ Program Operations staff inventoried the emissions for TV sources; regional staff estimated the emissions for all other sources.

All of these major categories were processed through SMOKE by WSU staff in preparation for the modeling. 

Future Year Documentation

Base Year (2002) methodology for area and point sources

The base year EI chosen to be forecasted to 2015 was the latest emission inventory that was created for submittal to EPA under the Consolidated Emissions Reporting Rule (CERR). This 2002 county by county annual inventory was developed by the respective states following the currently accepted methodologies and submitted as part of the National Emission Inventory (NEI). As required by the submittal process it underwent rigorous QA by ODEQ, Ecology and EPA contractors. These statewide inventories covered many more pollutants than were needed for this ozone modeling. Therefore the ozone precursor pollutants (CO, NOx, VOC) and SOx were selected by source classification code (SCC) from the submitted inventories for the counties in the modeling domain.


Area source forecast method
The selected pollutants comprising the Oregon inventory were input into a MS ACCESS database table to begin the future forecast. Area source forecasts were developed following a method that had been used in the previous Oregon SIPs following EPA guidance in Procedures for Preparing Emission Projections & Guidance for Growth Factors. This method involves creating surrogate growth factors (Table A5) for industrial, commercial, population, and household growth. In addition, there is a no growth option. Each SCC in the inventory is then assigned one of these surrogates. 

Table A5. 2015 Growth Surrogate Factors


The factors themselves were developed from the METRO Economic Report Regional Forecast for the Portland-Vancouver, OR-WA area provided by Chief Economist, Dennis Yee with METRO's Data Resource Center. They represent the average growth rate for the period 2002 to 2015 derived from the industrial and commercial employment projections, population, and resulting household growth. The industrial and commercial decisions were based on SIC grouping from Procedures for the Preparation of Emission Inventories for Carbon Monoxide and Precursors of Ozone, Volume I: General Guidance for Stationary Sources and represent an average of the following sectors.  The same growth factors were used for both the Washington and Oregon source categories (Table A6).  Although, SWCAA did look for WA specific data in other parts of the domain, it was decided that Portland/Vancouver area influences were the greatest and local planning decisions were the most influential in model results.

The rate of growth was calculated with the following formula:

      (Future yr employment - initial yr employment)/initial employment/ # of years of growth.






Table A6.  2002 to 2015 METRO Ozone Summary Industrial Growth Factor Table


Table A7.  2002 to 2015 METRO Ozone Summary Commercial Growth Factor Table


Those factors were then used in the ACCESS database to estimate the 2015 emissions with the following equation:

      2002 base year emissions + (2002 base year emissions * growth factor * 13 growth years)


 
2015 Nonroad Mobile

The NONROAD2004 model was used to estimate the 2015 emissions since it accounts for both growth and federal controls on covered nonroad sources.  According to the May 2004 EPA FAQ on the NONROAD model, the draft NONROAD2004 model includes the following federal emission reduction rules: 

        1) Final Tier 4 nonroad diesel engine standards (also known as the Clean Air Nonroad Diesel Rule), final rule.  However, the model does not yet contain the evaporative emission benefits for the new recreational equipment and large spark-ignition standards. [Adjustments for the evaporative reductions were not done for this modeling effort.]
        
        2) Final nonroad diesel fuel sulfur standards associated with the Tier 4 rule
     
For the 2015 Nonroad Model run inputs, all period parameters except year were retained from the 1997/1998 runs.  

For fuel parameter inputs, a 2015 gasoline sulfur default of 0.0030 weight % from EPA's MOBILE6 was used.  A 2015 diesel sulfur 0.0011 weight % was taken from the inputs for the NONROAD model runs EPA generated to develop the 2004 Nonroad Diesel Engine Final Rule.  Stage II control % was set to 0 since Oregon Stage II phase out is expected to begin sometime in 2008.  The CNG/LPG sulfur weight % and all temperature data were retained from the 1997/1998 runs.  

Fuel O2 weight % and RVP data were improved and revised for the 2015 Nonroad Model episode runs.  Baseline oxygen weight % was revised downward, and it was determined that ethanol is blended year-round at some Oregon terminals.  DEQ assumes that blending will continue through 2015, so the 2015 fuel O2 was estimated to be 1.03% by weight for all counties in the domain.  RVP refinements were made by apportioning terminal RVP to county and subsequently incorporating the revised O2 weight %.  The final RVP estimates are shown by county in Table A8.


  Table A8.  2015 Ozone Modeling Nonroad Model Input RVP  -  Oregon Counties
                                       
                                       

For the Washington portion of the modeling domain, SWCAA used the same temporal parameters that ODEQ used for their portion of the domain.  The fuel parameters were also consistent with the Oregon portion of the domain except for Fuel RVP.  The inputs used for WA counties were consistent with an internal Ecology document summarizing Mobile 6/6.2 input parameters for the different regions of the state entitled Mobile 6/6.1/6.2 Input Parameters and Processing (May 6, 2003 update).  Fuel RVP values used for WA in the 2015 nonroad emissions modeling are shown in Table A9.

Table A9: 2015 Ozone Modeling Nonroad Model Input RVP  - Washington Counties
                                       

Railroad, marine vessels, and airports are estimated independently of the NONROAD model. 2015 locomotive and commercial marine vessel emissions were estimated by the growth factor analysis method similar to area sources. For locomotive engines the national average growth factor (1.3%) from the Federal Clean Air Nonroad Diesel  -  Final Rule Regulatory Impact Analysis Document was applied to grow the emissions. However in the 13 years separating the base and future years that analysis showed that the clean engine technology was working its way into use and lowering the emissions. The reductions from the Tier 0/I/II implementation on emissions work is shown in EPA's Emission Factors for Locomotives (Document EPA420-F-97-051 Dec. 1997) which gives fleet average emission factors for all locomotives from 1999 to 2040. By comparing the 2002 fleet average emission factors to the 2015 fleet average emission factors, the effects on NOx, CO, HC, and PM are given as follows:

      Table A10.  2002 vs. 2015 Locomotive EF Ratios
                                   Pollutant
                                2015/2002 Ratio
                                      HC
                                     0.794
                                      CO
                                       1
                                      NOx
                                     0.607
                                      PM
                                     0.779

These ratios were applied to the grown 2015 emissions to reflect the expected fleet penetration of the Tier 0-II locomotives over time.   In the same manner Nonroad diesel sulfur content is expected to drop to 15 ppm by 2012 for locomotives and marine engines. Chapter 3 of the Clean Air Nonroad Diesel  -  Final Rule Regulatory Impact Analysis gives the anticipated changes to SO2 emissions.  

Table 3.1-6a on page 3-19 gives the Baseline Locomotive fuel sulfur content and Locomotive SO2 emissions in tpy.  This `baseline' sulfur content only fluctuates due to `spillover' from the low-sulfur highway diesel.

Table 3.4-8a on page 3-57 gives the Control Locomotive fuel sulfur content and Locomotive SO2 emissions in tpy.  This table shows substantial reductions for the nonroad low-sulfur diesel impacts as well as highway diesel spillover.

To get the 2002 and 2015 (controlled) numbers on the same basis (lb SO2/gallon diesel), each years annual emissions were divided by the annual fuel usage for that year from table 3.1-6a.  This resulted in the following:

Table A11. Locomotive Baseline and Future Fuel Usage and Emissions
Year
Fuel Usage (gal)
SO2 Emissions (tpy)
Rate (lb SO2/gal)
2002
2.692 x10^9
49,291
0.03662
2015
3.148 x10^9
1,231
0.000782

Therefore the ratio of 2015 to 2002 SO2 emissions due to the low-sulfur fuel is equal to the ratio of 0.000782/0.03662 or:

           Table A12.  Locomotive Sulfur Emissions Reduction Factor
                                   Pollutant
                                2015/2002 Ratio
                                      SO2
                                    0.02135

Commercial marine vessel 2015 emissions received the derived commercial growth factor application. The change in sulfur emissions from large trans-oceanic shipping is not yet clear so it was decided that there would be no reduction factor applied to the commercial marine vessel emissions.

Recreational marine vessels are also calculated "off model" and so the 2002 emissions were grown to 2015 using population as a surrogate following the area source application method.

Aircraft emissions for the Oregon 2002 NEI submitted data had two sources. The Port of Portland went through great detail to prepare emissions for Hillsboro, Troutdale, Mulino, and Portland International airports out to the year 2020. This data was verified and accepted as part of the Oregon CERR submittal. Because the remaining towered airports and the many smaller airports do not fall under the Port's jurisdiction they were estimated by DEQ using the Emissions Dispersion and Modeling System (EDMS) developed by the FAA. Therefore, the 2015 aircraft emissions include those estimated by the Port (where appropriate) and those grown by DEQ using the commercial growth factor as discussed above.

In the Washington part of the modeling domain, the only significant aircraft emissions were in the Puget Sound region, and the most notable of these at the Seattle Tacoma International Airport (Sea-Tac).  Emission data for that facility were compiled by a consultant (CDM) and provided to PSCAA.  PSCAA provided these data to SWCAA for input into the model along with aircraft emissions for Snohomish, Pierce, and Kitsap Counties.  


2015 On-road Mobile

The 2015 mobile emissions were calculated based on link-based, daily trip VMT data modeled by RTC and Metro using traffic demand models.  MOBILE6.2 was used to calculate the emission factors with the 1998 episode-specific conditions.  The summary of the approach to mobile emissions modeling estimates was to:

 Create link-based daily VMT for the AQMA using traffic demand models, 
 ODOT and WDOT provided county-wide VMT for the remainder of the domain. 
 Create four total MOBILE6.2 input files representing Clark County, Portland Metro Area Counties (both with I/M features); Washington State counties, and Oregon State counties (both without I/M features) for the remainder of the modeling domain.
 Run MOBILE6.2 at a standard temperature and at individual temperatures. 
 Calculate average daily emissions at the standard temperature. 
 Create a set of temperature adjustment factors from the individual temperature runs. 
 Use SMOKE's gridded temperature file (from MM5) to apply adjustments to the standard emissions.

This method allowed adjustment of hourly emissions based on temperature, but involved post-processing of SMOKE data.  ODEQ completed the MOBILE6.2 work with assistance from ECOLOGY and SWCAA for Washington data, and WSU completed the SMOKE processing. The eight (4 per episode year) MOBILE6.2 input files can be found in Appendix C.
 
2015 Point  Source Growth Factor Implementation

Point source emissions were grown using 2003 through 2015 annual growth factors supplied by Metro (ref 603).  The growth factors are keyed by Metro number, which was cross-walked to Standard Industrial Classification (SIC) by DEQ staff (Marianne Fitzgerald, AQ-PPD).  The growth factors with matched SICs are shown in Table A.13.

         Table A13.  Point Source SIC growth rates
         

Emissions by facility were grown by matching the facility primary SIC to the SIC in Table 2.7.X.  A linear, non-compounding formula was used:

((Metro growth factor) * (previous year emissions)) + (previous year emissions)

For example, for a selected SIC sub-category for the year 2003, with a 2002 (base year) value of 10 tons per year, and growth factors of 2.5% in 2003 and 3.0 % in 2004:

      2003 tpy = (0.025 growth * 10 tpy) + (10 tpy) = 10.25 tpy
      2004 tpy = (0.03 growth * 10.25 tpy) + (10.25 tpy ) = 10.56 tpy



 Biogenic emissions will be assumed to remain unchanged in the future, although urban development does modify the amount, location, and type of vegetation over time.



Comparisons between 1998 and 2015 future year 

   Figure A1.  Episode Emission Comparison
   




Summary of Assumptions and Strategies built into 2015 EI

The following is a list of the assumptions built in to the 2015 emissions inventory used for the modeling of the base future year scenario.

      Railroad emissions have incorporated low sulfur fuel and the effect of Tier I/II engine penetration into the fleet.
      Final Tier 4 nonroad diesel engine standards and diesel fuel sulfur standards associated with the rule have been included.
      Onroad diesel sulfur was 15 ppm
      Population increased 1.8% 
      Household growth was 2.0%
      Commercial employment growth was 1.0%
      Industrial employment growth was 0.21%


Appendix B: Area Source Category Emission Calculation Methodologies









Appendix C: MOBILE6.2 input files for the 1997 and 1998 episodes.


* Filename: ClrkIM97 -- Clark County I/M run with M6

* This input generates the summer fleet-average VOC, CO, SOx and NOx  emissions from CY1997 for Clark County Washington using Washington 1997 registration data (reg1997) and Vancouver I/M program file description (im1997.van)


*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :
*REPORT FILE        :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 7.8

* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : H:\EPA_PROG\MOBILE62\Mobile6\Run\96PdxSip\97episod\reg1997.txt


I/M PROGRAM        : 1 1993 2050 2 T/O ASM 2525 PHASE-IN
I/M MODEL YEARS    : 1 1968 2050
I/M VEHICLES       : 1 22222 11111111 1    
I/M STRINGENCY     : 1 34.0    
I/M COMPLIANCE     : 1 93.0    
I/M WAIVER RATES   : 1 2.0 4.0   
I/M GRACE PERIOD   : 1 1

I/M PROGRAM        : 2 1993 2050 2 T/O 2500/IDLE
I/M MODEL YEARS    : 2 1968 2050
I/M VEHICLES       : 2 11111 22222222 2    
I/M STRINGENCY     : 2 34.0    
I/M COMPLIANCE     : 2 93.0    
I/M WAIVER RATES   : 2 2.0 4.0   
I/M GRACE PERIOD   : 2 1  

******************** Scenario Section *****************************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1997
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50

ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 260



* Filename: WaNoIM97 -- 1997 Washington No I/M counties 
* (for the remainder of the domain) run with M6.2

* This input generates the summer fleet-average VOC, CO, NOx, SOx and PM10 emissions from CY1997 for all Washington Non I/M counties in themodeling domain using Washington 1997 reg data (reg1997) because it is the registration developed for that year (vs using the 1999 distribution).

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 8.5
* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : H:\EPA_PROG\MOBILE62\Mobile6\Run\96PdxSip\97episod\reg1997.txt


*******************  Scenario Section  **********************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1997
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50 

ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 350.0


* Filename: PDXIM97 -- Tricounty Portland Ozone I/M run with M6

* This input generates the  summer fleet-average VOC, CO, SOx and NOx emissions from CY1997 for Tri-County area using Tri-county average 1996 registration from the reg data we used for the 1996 EI because this data is closer in time to the 97 modeling year and applying the reg data to the LDV and LDT and using the Wa ECOLOGY 1999 vehicle registration for vehicle classes 6-16.

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 7.8

* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : h:\EPA_PROG\MOBILE62\Mobile6\Run\96PdxSip\97episod\RegTRI97.d


* The Metro area has 2 levels of tests (during 1997) so there needs to be 2 I/M program entries
* 1981-1995 Enhanced test, 1975-1980 Basic test, only the LD vehicles get tested.


* 1981-1995 Enhanced (Bar 31) Gary Beyer says this test (2525/5015) models real world better than IM240
I/M PROGRAM        : 1 1975 2020 2 T/O ASM 2525/5015 FINAL
I/M MODEL YEARS    : 1 1981 2020
I/M VEHICLES       : 1 22222 11111111 1
I/M STRINGENCY     : 1 37.4
I/M COMPLIANCE     : 1 90.0
I/M WAIVER RATES   : 1 0.0 0.0 
I/M CUTPOINTS      : 1 ORCUTPT.d
I/M GRACE PERIOD   : 1 2

* 1975-1980 Basic
* According to VIP policy 201 "all 1975-1980 gasoline vehicles are subject to the Basic test."
I/M PROGRAM        : 2 1975 2020 2 T/O 2500/IDLE
I/M MODEL YEARS    : 2 1975 1980
I/M VEHICLES       : 2 22222 22222222 2 
I/M STRINGENCY     : 2 37.4
I/M COMPLIANCE     : 2 90.0
I/M WAIVER RATES   : 2 0.0 0.0



ANTI-TAMP PROG     : 
75 75 80 22222 22222222 1 12 090. 22212221


*******************  Scenario Section  **********************
* A separate scenario must be written for each calendar
* year to be analyzed.
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1997
EVALUATION MONTH   : 7


* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE
* perl script will create a constant temp (min 61 max 61) throughout the range.

MIN/MAX TEMP       : 50 50
ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 260



* Filename: ORNoIM97 -- 1997 Oregon No I/M counties 

For the remainder of the domain run with M6.2
This input generates the summer fleet-average VOC, CO, NOx, SOx and PM10 emissions from CY1997 for all Oregon Non I/M counties in the modeling domain using the 1996 Oregon LDV and LDT vehicle class registration from the 1996 EI  model year end count we received from the DMV and Wa all other registration.

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 9.0
* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : H:\EPA_PROG\MOBILE62\Mobile6\Run\96PdxSip\97episod\regOR97.d

*******************  Scenario Section  ***************************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1997
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50

ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 280

1998 Episode

* Filename: ClrkIM98 -- Clark County I/M run with M6

* This input generates the summer fleet-average VOC, CO, SOx and NOx emissions from CY1998 for Clark County Washington using Washington 1999 registration data.


*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :
*REPORT FILE        :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 7.8

* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : D:\EPA_PROG\MOBILE62\Mobile6N\Run\96PdxSip\reg1999.txt


I/M PROGRAM        : 1 1993 2050 2 T/O ASM 2525 PHASE-IN
I/M MODEL YEARS    : 1 1968 2050
I/M VEHICLES       : 1 22222 11111111 1    
I/M STRINGENCY     : 1 36.0    
I/M COMPLIANCE     : 1 93.0    
I/M WAIVER RATES   : 1 4.0 7.0   
I/M GRACE PERIOD   : 1 1

I/M PROGRAM        : 2 1998 2050 2 T/O GC
I/M MODEL YEARS    : 2 1968 2050
I/M VEHICLES       : 2 22222 22222222 2    
I/M STRINGENCY     : 2 36.0    
I/M COMPLIANCE     : 2 93.0    
I/M WAIVER RATES   : 2 4.0 7.0   
I/M GRACE PERIOD   : 2 1    

I/M PROGRAM        : 3 1993 2050 2 T/O 2500/IDLE
I/M MODEL YEARS    : 3 1968 2050
I/M VEHICLES       : 3 11111 22222222 2    
I/M STRINGENCY     : 3 36.0    
I/M COMPLIANCE     : 3 93.0    
I/M WAIVER RATES   : 3 4.0 7.0   
I/M GRACE PERIOD   : 3 1

******************** Scenario Section *****************************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1998
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50

ABSOLUTE HUMIDITY  : 64.01
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 260


* Filename: WaNoIM98 -- 1998 Washington No I/M counties 

For the remainder of the domain run with M6.2
This input generates the summer fleet-average VOC, CO, NOx, SOx and PM10
emissions from CY1998 for all Washington Non I/M counties in the modeling domain using Washington 1999 reg data (Sally thinks that is better than the 1998).

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 8.5
* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : D:\EPA_PROG\MOBILE62\Mobile6N\Run\96PdxSip\reg1999.txt


*******************  Scenario Section  **********************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1998
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50 

ABSOLUTE HUMIDITY  : 64.01
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 350.0



* Filename: PDXIM98 -- Tricounty Portland Ozone I/M run with M6

* This input generates the summer fleet-average VOC, CO, SOx and NOx emissions from CY1998 for Tri-County area using Tri-county average 2002 registration from Jeff's query of the DMV data (no gas/diesel distinction) and applying the reg data to the LDV and LDT and using the Wa ECOLOGY 1999 vehicle registration for vehicle classes 6-16.

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 7.8

* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : D:\EPA_PROG\MOBILE62\Mobile6N\Run\96PdxSip\IM98OR.d


* The Metro area has three levels of tests so there needs to be 3 I/M program entries
* 1996 to present gets OBD, 1981-1995 Enhanced test, 1975-1980 Basic test
* only the LD vehicles get tested

* 1996+ OBD exhaust benefits
I/M PROGRAM        : 1 1975 2020 2 T/O OBD I/M
I/M MODEL YEARS    : 1 1996 2020
I/M VEHICLES       : 1 22222 11111111 1
I/M STRINGENCY     : 1 37.4
I/M COMPLIANCE     : 1 90.0
I/M WAIVER RATES   : 1 0.0 0.0 
I/M GRACE PERIOD   : 1 2

* 1996+ OBD evaporative benefits
I/M PROGRAM        : 2 1975 2020 2 T/O EVAP OBD
I/M MODEL YEARS    : 2 1996 2020
I/M VEHICLES       : 2 22222 11111111 1
I/M STRINGENCY     : 2 37.4
I/M COMPLIANCE     : 2 90.0
I/M WAIVER RATES   : 2 0.0 0.0 
I/M GRACE PERIOD   : 2 2

* 1981-1995 Enhanced (Bar 31) Gary Beyer says this test (2525/5015) models real world better than IM240
I/M PROGRAM        : 3 1975 2020 2 T/O ASM 2525/5015 FINAL
I/M MODEL YEARS    : 3 1981 1995
I/M VEHICLES       : 3 22222 11111111 1
I/M STRINGENCY     : 3 37.4
I/M COMPLIANCE     : 3 90.0
I/M WAIVER RATES   : 3 0.0 0.0 
I/M CUTPOINTS      : 3 ORCUTPT.d

* 1975-1980 Basic
* According to VIP policy 201 "all 1975-1980 gasoline vehicles are subject to the Basic test."
I/M PROGRAM        : 4 1975 2020 2 T/O 2500/IDLE
I/M MODEL YEARS    : 4 1975 1980
I/M VEHICLES       : 4 22222 22222222 2 
I/M STRINGENCY     : 4 37.4
I/M COMPLIANCE     : 4 90.0
I/M WAIVER RATES   : 4 0.0 0.0

ANTI-TAMP PROG     : 
75 75 80 22222 22222222 1 12 090. 22212221


*******************  Scenario Section  **********************
* A separate scenario must be written for each calendar year to be analyzed.
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1998
EVALUATION MONTH   : 7


* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE
* perl script will create a constant temp (min 61 max 61) throughout the range.

MIN/MAX TEMP       : 50 50
ABSOLUTE HUMIDITY  : 64.01
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 260



* Filename: ORNoIM98 -- 1998 Oregon No I/M counties
 For the remainder of  the domain run with M6.2
This input generates the summer fleet-average VOC, CO, NOx, SOx and PM10
emissions from CY1998 for all Oregon Non I/M counties in the modeling domain using Oregon LDV and LDT vehicle class registration and Wa all other registration.

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA
*******************  Run Section  ***************************
NO REFUELING       :
FUEL RVP           : 9.0
* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : D:\EPA_PROG\MOBILE62\Mobile6N\Run\96PdxSip\NoIM98OR.d

*******************  Scenario Section  ***************************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 1998
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-101 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50

ABSOLUTE HUMIDITY  : 64.01
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 280



* Filename: Q15Clrk -- Clark County I/M run with M6

* This input generates the  summer fleet-average emissions from CY2015
* for Clark County Washington using Washington 2004 registration data.
* This is the file run by DEQ for the CMAQ modeling of the future episode.


*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :


RUN DATA
*******************  Run Section  ***************************
EXPAND EVAPORATIVE :
*STAGE II REFUELING : 92 2 80. 80.

FUEL RVP           : 7.8

* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : H:\EPA_PROG\MOBILE62\Mobile6\Run\Regdist\WaReg\reg2004.txt

* There were not enough programs to define the 1) HDGV2b/3 EVAP OBD & GC program
* anticipated to begin in 2008 and 2) HDGV4+ GC program that began in 2002.
* HDGV gas cap added to program 2 to at least get some of the HD GC test
* Received from W.Strange 10/21/2004 after consultation with his management.

I/M PROGRAM        : 1 1993 2050 2 T/O ASM 2525 PHASE-IN
I/M MODEL YEARS    : 1 1968 1995   
I/M VEHICLES       : 1 22222 11111111 1    
I/M STRINGENCY     : 1 42.4   
I/M COMPLIANCE     : 1 91.0    
I/M WAIVER RATES   : 1 4.9 15.8
I/M EXEMPTION AGE  : 1 25    
I/M GRACE PERIOD   : 1 5    

I/M PROGRAM        : 2 1998 2050 2 T/O GC
I/M MODEL YEARS    : 2 1968 1995   
I/M VEHICLES       : 2 22222 22222222 2    
I/M COMPLIANCE     : 2 91.0    
I/M WAIVER RATES   : 2 4.9 15.8
I/M EXEMPTION AGE  : 2 25    
I/M GRACE PERIOD   : 2 5    

I/M PROGRAM        : 3 1993 2050 2 T/O OBD I/M
I/M MODEL YEARS    : 3 1996 2050   
I/M VEHICLES       : 3 22222 11111111 1    
I/M STRINGENCY     : 3 42.4   
I/M COMPLIANCE     : 3 91.0    
I/M WAIVER RATES   : 3 4.9 5.3
I/M EXEMPTION AGE  : 3 25    
I/M GRACE PERIOD   : 3 5    

I/M PROGRAM        : 4 2002 2050 2 T/O EVAP OBD & GC
I/M MODEL YEARS    : 4 1996 2050   
I/M VEHICLES       : 4 22222 11111111 1    
I/M COMPLIANCE     : 4 91.0    
I/M WAIVER RATES   : 4 4.9 5.3
I/M EXEMPTION AGE  : 4 25    
I/M GRACE PERIOD   : 4 5    

I/M PROGRAM        : 5 1993 2050 2 T/O 2500/IDLE
I/M MODEL YEARS    : 5 1968 2050
I/M VEHICLES       : 5 11111 11222222 2    
I/M STRINGENCY     : 5 42.4   
I/M COMPLIANCE     : 5 91.0    
I/M WAIVER RATES   : 5 4.9 8.3
I/M EXEMPTION AGE  : 5 25    
I/M GRACE PERIOD   : 5 5    

I/M PROGRAM        : 6 1993 2050 2 T/O 2500/IDLE
I/M MODEL YEARS    : 6 1968 2007
I/M VEHICLES       : 6 11111 22111111 1    
I/M STRINGENCY     : 6 42.4   
I/M COMPLIANCE     : 6 91.0    
I/M WAIVER RATES   : 6 4.9 8.3
I/M EXEMPTION AGE  : 6 25    
I/M GRACE PERIOD   : 6 5    

I/M PROGRAM        : 7 1993 2050 2 T/O OBD I/M
I/M MODEL YEARS    : 7 2008 2050   
I/M VEHICLES       : 7 11111 22111111 1    
I/M STRINGENCY     : 7 42.4   
I/M COMPLIANCE     : 7 91.0    
I/M WAIVER RATES   : 7 4.9 5.3
I/M EXEMPTION AGE  : 7 25    
I/M GRACE PERIOD   : 7 5    

******************** Scenario Section *****************************

SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 2015
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-102 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50

ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 15



* Filename: Q15PDXIM-- 2015 Tri County I/M 
This input generates the summer fleet-average VOC, CO and NOx, PM10 emissions from CY2015 for Multnomah, Washington and Clackamas Counties using the tri-county  LDGV and LDDV registration and appropriate extended registration from WA for HDV. This is the file run by DEQ for the CMAQ modeling of the future episode.  This file calculates the refueling emission as being controlled by the canisters on-board the vehicles. There is no longer any stage II vapor recovery systems in place at the gas stations.

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA

*******************  Run Section  ***************************
EXPAND EVAPORATIVE :
*STAGE II REFUELING : 92 2 80. 80.
FUEL RVP           : 7.8

* enter the appropriate registration distribution filename path for the county being modeled.
REG DIST           : R02PDX2.d

* 2 I/M program entries
* 1996 to CY - 2 gets OBD, 1975-1995 Basic test

* 1996+ OBD exhaust benefits
I/M PROGRAM        : 1 1975 2020 2 T/O OBD I/M
I/M MODEL YEARS    : 1 1996 2020
I/M VEHICLES       : 1 22222 11111111 1
I/M STRINGENCY     : 1 37.4
I/M COMPLIANCE     : 1 90.0
I/M WAIVER RATES   : 1 0.0 0.0 
I/M GRACE PERIOD   : 1 4

* 1996+ OBD evaporative benefits
I/M PROGRAM        : 2 1975 2020 2 T/O EVAP OBD
I/M MODEL YEARS    : 2 1996 2020
I/M VEHICLES       : 2 22222 11111111 1
I/M STRINGENCY     : 2 37.4
I/M COMPLIANCE     : 2 90.0
I/M WAIVER RATES   : 2 0.0 0.0 
I/M GRACE PERIOD   : 2 4

* 1975-1980 Basic - light duty gas vehicle basic testing
I/M PROGRAM        : 3 1975 2025 2 T/O 2500/IDLE
I/M MODEL YEARS    : 3 1975 1995
I/M VEHICLES       : 3 22222 11111111 1 
I/M STRINGENCY     : 3 37.4
I/M COMPLIANCE     : 3 90.0
I/M WAIVER RATES   : 3 0.0 0.0

* 1975-1980 Basic  - to correctly model the HDGV basic testing 
I/M PROGRAM        : 4 1975 2025 2 T/O 2500/IDLE
I/M MODEL YEARS    : 4 1975 2025
I/M VEHICLES       : 4 11111 22222222 2 
I/M STRINGENCY     : 4 37.4
I/M COMPLIANCE     : 4 90.0
I/M WAIVER RATES   : 4 0.0 0.0

ANTI-TAMP PROG     : 
75 75 80 22222 22222222 1 12 090. 22212222

*******************  Scenario Section  **********************
* A separate scenario must be written for each calendar
* year to be analyzed.

******************************************************************************************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 2015
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-100 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50

ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 15



* Filename: Q15PXNIM-- 2015 Tri County No I/M 
 This input generates the summer fleet-average VOC, CO and NOx, PM10 and HAPS emissions from CY2015 for Multnomah, Washington and Clackamas Counties using the tri-county  LDGV and LDDV registration and appropriate extended registration from WA for HDV.  This is the file run by DEQ for the CMAQ modeling of the future episode. This file calculates the refueling emission as being controlled by the canisters on-board the vehicles.
There is no longer any stage II vapor recovery systems in place at the gas stations.

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA

*******************  Run Section  ***************************
EXPAND EVAPORATIVE :
*STAGE II REFUELING : 92 2 80. 80.
FUEL RVP           : 7.8

* enter the appropriate registration distribution filename path for the county being modeled.
REG DIST           : R02PDX2.d

*******************  Scenario Section  **********************
* A separate scenario must be written for each calendar
* year to be analyzed.

******************************************************************************************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 2015
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-100 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50

ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 15


* Filename: WaNoIM15 -- 2015 Washington No I/M counties 
(for the remainder of the domain) run with M6.2

This input generates the summer fleet-average VOC, CO, NOx, SOx and PM10  emissions from CY2015 for all Washington Non I/M counties in the modeling domain using latest Washington 2004 reg data. This is the file run by DEQ for the CMAQ modeling of the future episode.

*******************  Header Section  ************************
MOBILE6 INPUT FILE 
PARTICULATES       :
SPREADSHEET        :
DATABASE OUTPUT    :
WITH FIELDNAMES    :
AGGREGATED OUTPUT  :

RUN DATA
*******************  Run Section  ***************************
EXPAND EVAPORATIVE :
*STAGE II REFUELING : 92 2 80. 80.
FUEL RVP           : 8.5
* enter the appropriate registration distribution filename for the county being modeled.
REG DIST           : H:\EPA_PROG\MOBILE62\Mobile6\Run\Regdist\WaReg\reg2004.txt


*******************  Scenario Section  **********************
SCENARIO RECORD    : Summer Fleet-Average Emission 
CALENDAR YEAR      : 2015
EVALUATION MONTH   : 7

* Hourly temp will be varied from 50-100 F. To create temperature profile used by SMOKE

MIN/MAX TEMP       : 50 50 

ABSOLUTE HUMIDITY  : 50.24
PARTICULATE EF     : PMGZML.CSV PMGDR1.CSV PMGDR2.CSV PMDZML.CSV PMDDR1.CSV PMDDR2.CSV
PARTICLE SIZE      : 10.0
DIESEL SULFUR      : 15
                                       
Appendix D:  Acronyms
                                       
BELD3 	Biogenic Emissions Landcover Database - version 3 
BEIS3 		Biogenic Emissions Inventory System  -  version 3
CAA 		Clean Air Act 
CMAQ 	Community Multi-scale Air Quality
CCTM 	CMAQ Chemical Transport Model 
Ecology	Washington State Department of Ecology
EKMA 	Empirical Kinetics Modeling Approach
EPA	 	Environmental Protection Agency 
GDAS 		Global Data Assimilation System 
GEOS		Goddard Earth Observing System 
ICTC	 	Intercontinental Transport and Chemical Transformation
KF 		Kain-Fritsch cumulus scheme
LSM		Land Surface Model
NCEP 	 	National Center for Environmental Prediction
NOx 		Oxides of Nitrogen
MCIP		Meteorology-Chemistry Interface processor
MEBI 		Modified Euler Backward Interactive 
MM5		Mesoscale Meteorological model Version 5
MRF		Medium Range Forecast
NAAQS	National Ambient Air Quality Standards
NWRTC	Northwest Regional Technical Center
PBL		Planetary Boundary Layer 
SAPRC	State Air Pollution Research Center chemical mechanisms
SIP		State Implementation Plan
SMOKE	Sparse Matrix Operator Kernel Emissions Modeling System
SOx 		Oxides of Sulfur
SWCAA	Southwest Clean Air Agency
TV		Title V
VOC		Volatile Organic Compound
                                       
