The goal of this project was to develop a variable-grid-resolution atmospheric modeling system that will aid in more-realistic simulation of the interactions between chemical and dynamical processes occurring at differing spatial scales. An additional goal was to develop a modeling tool to accurately assess the impact of emissions related to oil and gas exploration and production (E&P) activities on local and regional air quality. These tools will increase confidence in model results and consequently in their use to develop mitigation strategies and risk-assessment studies.

University of North Carolina, Chapel Hill, NC

Atmospheric processes are inherently complex in that they occur over wide and often disparate spatial and temporal scales. Comprehensive three-dimensional atmospheric models attempt to illustrate the various physical and chemical processes governing the fate of atmospheric pollutants in a consistent manner. They have evolved as useful tools not only to study various atmospheric pollutants but also to aid in developing air pollution abatement strategies.

The accuracy of simulations with such models intricately depends on the size and spatial resolution of the computational mesh on which model calculations are performed. In Eulerian models, emission sources are artificially diluted over the grid volume. Thus, grid resolutions that are too coarse adversely affect the representation of point-source emissions, which account for a significant fraction of total anthropogenic NOx and SO2 emissions in the United States. Onshore, coastal, and offshore oil and gas E&P activities in the Gulf of Mexico region can contribute to increased levels of pollution not only in the coastal regions of the southern states surrounding the Gulf of Mexico but also in interior regions of many of the southeastern states. Since marine boundary layer processes, in general, are less turbulent than their counterparts over land, plumes from these point sources can stay as coherent structures and be transported to land regions without much dilution. The artificial dilution that arises due to coarse grid resolutions in air quality models can lead to significant uncertainties in the model representation of the effects of such emissions. Thus, better representation of these point sources in the Outer Continental Shelf regions is critically important to understanding emission-chemistry-transport processes in the coastal regions neighboring the Gulf of Mexico.

Results 
The project successfully tested a new data-assimilation technique and applied it to a meteorological model (MM5) for improving model predictions/simulations. This is the first technique developed and tested in the community model MM5.

A comprehensive analysis and evaluation of the results was obtained from the final modeling configuration in the MM5. The final meteorological modeling configuration included the researchers’ new surface data assimilation and the ingestion of satellite-derived sea surface temperatures, resulting in improved simulations of meteorological fields. In general, MM5 model simulations for the final modeling configuration compare favorably with observations, indicating overall improvements compared with the “base” simulations.

The emission database has been enhanced using the National Emissions Inventory (NEI) 1999 Version 3. This covers updates to area, non-road, and point sources. For updating mobile sources, researchers used MOBILE6 with updated vehicle miles traveled (VMT) data from the Texas Commission on Environmental Quality (TCEQ). Other databases available from the TCEQ are a 1999 area/non-road inventory and a 1997 Galveston Bay shipping inventory. Further, TCEQ has an offshore point- and area-source inventory that has been used in the past. Plans have been made to use BELD3 land use data to approximate the biogenic emissions (down to 1-km resolution). For the Mexican region, the 1999 BRAVO Mexican inventory will be used. Emissions data available from the MMS for the Outer Continental Shelf in the Gulf of Mexico will also be included in updating the project emissions database. After completing this task, Sparse Matrix Operator Kernel Emissions (SMOKE), the state-of-the-art emissions modeling system, will be utilized to prepare emissions estimates for use in the Multiscale Air Quality Simulation Platform (MAQSIP-VGR).

Testing has been completed in a 1-D model, and implementation of the cloud processor for use in the MAQSIP-VGR is scheduled. The cloud processes will be turned on when performing the MAQSIP-VGR simulations over the Houston-Galveston region.

The latest Emissions inventories from EPA are being used to create input emissions for the Houston and Northeast Gulf area case studies. The project team has completed the preparation of the emissions processor, SMOKE, for variable grid modeling to prepare inputs to the MAQSIP-VGR.

UNC has completed the development and testing of the variable grid emissions processor, the SMOKE-VGR.

Variable grid domain was created for the Eastern US regional domain with a stretch from a 4 Km grid resolution domain over the Houston-Galveston sub-region to 36 Km grid resolution at the boundaries of the regional domain.

UNC completed the preparation of the meteorological data for the August, 2000 episode over the Eastern US which includes the Houston-Galveston domain and the southern Louisiana gulf domains.

The overall objective of this research project was to develop an innovative modeling technique to adequately model the offshore/onshore transport of pollutants. The variable-grid modeling approach that was developed alleviates many of the shortcomings of the traditionally used nested regular-grid modeling approach, in particular related to biases near boundaries and the excessive computational requirements when using nested grids. The Gulf of Mexico region contiguous to the Houston-Galveston area and southern Louisiana was chosen as a test bed for the variable-grid modeling approach. In addition to the onshore high pollution emissions from various sources in those areas, emissions from on-shore and off-shore oil and gas exploration and production are additional sources of air pollution.

Benefits 
The research project will provide two primary benefits: 1) to further develop and refine an advanced variable-grid-resolution air quality modeling tool to provide detailed, accurate representation of the dynamical and chemical processes governing the fate of anthropogenic emissions in coastal environments; and 2) to improve the current understanding of the potential impact of onshore and offshore E&P emissions on ozone and particulate matter non-attainment in the Gulf of Mexico and nearby states.

Summary 
Major portions of this project research were completed successfully. Comprehensive evaluations of various cases have been completed and have generated the meteorological inputs for 36-, 12-, and 4-km grids using a carefully selected final modeling configuration. The variable-grid-resolution meteorology is being generated to drive the MAQSIP-VGR. The emissions model, SMOKE-VGR, has been developed. The development of the MAQSIP-VGR has also been completed.

To perform quality control and analysis of our research results on the VGR domains, the project performers have configured the National Center for Atmospheric Research visualization package, NCL. In addition testing and validation of the SMOKE-VGR emissions processor have been completed.

The following is a summary of all work conducted.

• Identified case studies for which to perform meteorological and air quality model simulations. Our approach included developing and evaluating the meteorological, emissions, and chemistry-transport modeling components for the variable-grid applications, with special focus on the geographic areas where the finest grid resolution was used.
• Evaluated the performance of two atmospheric boundary layer (ABL) schemes, and identified the best-performing scheme for simulating mesoscale circulations for different grid resolutions. Use of a newly developed surface data assimilation scheme resulted in improved meteorological model simulations.
• Successfully ingested satellite-derived sea surface temperatures (SSTs) into the meteorological model simulations, leading to further improvements in simulated wind, temperature, and moisture fields. These improved meteorological fields were important for variable-grid simulations, especially related to capturing the land-sea breeze circulations that are critical for modeling offshore/onshore transport of pollutants in the Gulf region.
• Developed SMOKE-VGR, the variable-grid version of the SMOKE emissions processing model, and tested and evaluated this new system. We completed the development of our variable-grid-resolution air quality model (MAQSIP-VGR) and performed various diagnostic tests related to an enhanced cloud parameterization scheme. We also developed an important tool for variable-grid graphics using Google Earth.
• Completed runs of the MAQSIP-VGR for the Houston-Galveston and southern Louisiana domains for an August 23 to September 2, 2002 time frame.

(February 2009)
Work was conducted to develop a Community Multi-Scale Air Quality Model-Variable Grid Resolution model (CMAQ-VGR) targeting the Houston area for simulation. CMAQ-VGR will have the capability to simulate particulate matter (PM) that is not currently implemented in the MAQSIP-VGR. Development of the model incorporated these three tasks:

• Prepare the VGR meteorology as an input to CMAQ.
• Prepare the emissions processor SMOKE for VGR applications.
• Prepare a CMAQ-VGR.

The University of North Carolina requested and received a no-cost extension due to several tasks needing the extra time to complete their activities. This extension added an additional six months to complete all project work.

Results of the modeling simulations highlighted the usefulness of the variable-grid modeling approach when simulating complex terrain processes related to land and sea close to an urban area. Our results showed that realistic SST patterns based on remote sensing are critical to capturing the land-sea breeze, in particular the inland intrusion of the reversed mesoscale circulation that is critical for simulating air pollution over urban areas near coastal regions. Besides capturing the correct horizontal gradient between land and sea surface temperatures, it is important to use an adequate ABL scheme in order to quantify correctly the vertical profiles of various parameters. The ABL scheme should capture the dynamics of the marine boundary layer, which is not often considered in a typical simulation over land. Our results further showed the effect of using satellite-derived SSTs on the horizontal and vertical extent of the modeled pollution pattern, and the increase in hourly ozone concentrations associated with changes in ABL characteristics resulting from the enhanced mesoscale circulation in the lower troposphere.

UNC has completed all project activities. The final report is available below under "Additional Information".

Funding 
This project was selected in response to DOE’s Broad-Based Announcement, DE-PS26-03NT41613.

$710,025 

$177,506 (20% of total)

NETL - Jesse Garcia (jesse.garcia@netl.doe.gov or 918-699-2036) 
U. North Carolina - Adel Hanna (ahanna@unc.edu or 919-966-1352)

Final Report [PDF-1.67MB]

Publications: 
Preliminary Results on the Development of a Variable-Grid-Resolution Air Quality Model. Presented at the Models-3 Users' Workshop, October 27-29, 2003, Research Triangle Park, NC.

Development of Alternative Methods for Estimating Dry Deposition Velocity in CMAQ. Presented at the Models-3 Users' Workshop, October 27-29, 2003, Research Triangle Park, NC.

Journal articles anticipated:

Development of a Variable-Grid-Resolution Air Quality Model.

Development of the Flux-Adjusting Surface Data Assimilation System (FASDAS) for Applications in Mesoscale Models.

The Effects of Using Satellite-derived Sea Surface Temperatures on Mesoscale Circulations over a Coastal Region and Their Impacts on Air Quality Simulations