< back

3SAQS Sensitivity Modeling Plan

This memo presents the sensitivity simulations to be completed under the Three-State Air Quality Study (3SAQS) Pilot Project cooperative agreement. These simulations are covered under Task 2013-7 (Air Quality Modeling and Analysis) of the existing project scope. We are proposing these simulations to gain additional insight into the 3SAQS modeling platforms with respect to boundary conditions, background air quality estimates, and oil and gas emissions impacts on local and regional air quality. These simulations will be conducted through the end of the current period of performance, which ends April 15, 2015.

We first present descriptions of the different scenarios and then a table recommending the order of priority for the simulations.

All these sensitivity simulations will be derived from the 3SAQS Base 2011 version A CAMx configuration.

Sensitivity: GEOS-Chem Boundary Conditions

Objective: Use the 3SAQS 2011 version A CAMx modeling platform to test the impacts of using GEOS-Chem Initial and Boundary Conditions (ICBC).

Approach: Replace the 2011 MOZART ICBC data with GEOS-Chem data obtained from EPA. Prepare the GEOS-Chem global model results for input to the 3SAQS 2011 CAMx modeling platform. In the preparation of the GEOS-Chem BCs, include tracer species for ozone and dust. Run CAMx for the entire year of 2011 on the three two-way nested (36/12/4-km) modeling domains. Conduct a full model performance evaluation of the CAMx results using GEOS-Chem ICBCs and compare these results to the 3SAQS CAMx simulation that used MOZART BCs. Include analysis of the BC ozone and dust tracers on the 3SAQS 4-km modeling domain. Analyses will also include comparisons of the GEOS-Chem and MOZART vertical profiles of ozone at the boundary and key sites (such as Boulder, CO) that are impacted by long-range transport. If the model performance is improved by using GEOS-Chem, we will consider adopting GEOS-Chem boundary conditions for subsequent regional modeling runs.

Simulation ID: 3SAQS_CAMx_Base11a_sensGCBC

Sensitivity: US Background Air Quality with MOZART BCs

Objective: Use the 3SAQS 2011 version A CAMx modeling platform to estimate US background air quality with MOZART boundary conditions.

Approach: Zero out all U.S. anthropogenic emissions in the 3SAQS Base 2011 version A modeling platform. The list of emissions sectors that will remain in the background simulation include:

All Canada and Mexico anthropogenic sources
Biogenic
Windblown Dust
Lightning
Seasalt
Wildfires (exclude agricultural and prescribed burn fires)

Run CAMx for the entire year of 2011 on the outer two-way nested (36/12-km) modeling domains. Calculate the domain-wide and subdomain background ozone concentrations during the ozone season. Calculate differences against the 3SAQS 2011 CAMx base simulation to estimate local vs. transported ozone.

Simulation ID: 3SAQS_CAMx_Base11a_sensUSGB

Sensitivity: Oil & Gas Emissions Sensitivity

Objective: Adjust the 3SAQS basin O&G VOC and/or NOx inventory using factors reported in the literature.

Approach: Conduct a literature review for emissions estimates and scaling factor by 3SAQS O&G basin. Scale the 3SAQS 2011 O&G VOC, NOx, and methane inventories using factors reported in the literature. Use the adjusted inventory in an ozone season CAMx simulation for the three two-way nested (36/12/4-km) 3SAQS modeling domains. The simulation will be based on the 3SAQS_CAMx_Base11a simulation, which used MOZART boundary conditions. Compare the results of the O&G VOC sensitivity to the reference simulation and observations.

Simulation ID: 3SAQS_CAMx_Base11a_sensOGEmis

Sensitivity: Winter Ozone Formation

Objective: Conduct emissions and CAMx configuration sensitivity simulations and evaluate the changes in model performance for estimating winter ozone events

Approach: Combine the result of Sensitivity 3 with changes to parameters affecting dynamics and chemistry in CAMx. Model parameters that we will explore in this sensitivity include VOC and NOx emissions magnitudes, VOC speciation, vertical diffusivity, snow cover, and surface albedo. The sensitivity(s) will target limited time periods and spatial domains that focus on winter ozone formation in the three-state region.

Simulation ID: 3SAQS_CAMx_Base11a_WinOG[n]; where n = sensitivity #

Sensitivity: Ammonia Emissions

Objective: Adjust the temporal distribution and mass emissions rates of key ammonia emission sources and study the impact on model performance

Approach: Work with the NPS Air Resources Division to design emissions sensitivities and evaluation procedures to assess and improve the performance of simulating ammonia (NH3) concentrations and deposition in the Three State study area. Potential evaluation improvements could involve integrating TES satellite retrievals and AMoN surface observations into the suite of 3SAQS model performance evaluation products. We may also consider adding NH3 emissions tracers from different sources, such as livestock, fertilizer, mobile, fires, and industrial sources. While a model inversion study is beyond the scope of this task, constraints on the emissions magnitudes could use published data on the regional underestimation of ammonia emissions. Finally, we will explore adding missing sources of NH3, such as a bidirectional flux from agricultural sources.

Simulation ID: 3SAQS_CAMx_Base11a_sensNH3

Optional Sensitivity: US Background Air Quality with GEOS-Chem BCs

Objective: Use the 3SAQS 2011 version A CAMx modeling platform to estimate US background air quality with GEOS-Chem boundary conditions.

Approach: We will only run this simulation if Sensitivity 1 shows large differences in the CAMx results between the MOZART and GEOS-Chem BCs. We will use the same approach as Sensitivity 2 except replace the MOZART BCs with GEOS-Chem BCs. Compare the background air quality estimates between this simulation and Sensitivity 2.

Simulation ID: 3SAQS_CAMx_Base11a_sensGCUSBG

Sensitivity Recommendations

Sensitivity	Priority	Order	Recommendation
3SAQS_CAMx_Base11a_sensGCBC	High	2	Important simulation to determine the impact of the nested BCs on model performance; a comparison of the MPE results between this run and the base simulation (with MOZART BCs) will dictate the ICBC data that we will use in the final 2011 and possibly 2014 simulations
~~3SAQS_CAMx_Base11a_sensUSGB~~3SAQS_CAMx_Base11a_sensUSBG	High	1	Start with this simulation to get a sense of the influence of background emissions and transport on the air quality in the Three-State Region; will lend insight on how well the model is capturing Stratosphere-Troposphere Exchange
3SAQS_CAMx_Base11a_sensOGEmis	Med	3	This simulation should really be combined with the WRF wintertime configuration. Since the WRF wintertime simulation will not be available until early 2015, we can proceed with this simulation to get a sense of how the model is responding to changes in emissions and whether the chemistry is NOx or VOC limited
Winter Ozone Formation	High	4	This is a higher priority than sensOGEmis because it will combine the meteorology changes with emissions changes. Because it will take time to develop the wintertime meteorology this run slides lower in the order of sensitivities. We will use the emissions sensitivities developed under sensOGEmis to inform the emissions perturbations to apply here.
3SAQS_CAMx_Base11a_sensNH3	Med	5	Since it's not clear what is driving the poor NH3 model performance, this sensitivity will likely be a diagnostic sensitivity to gain insight into what parts of the model need work to improve performance.
3SAQS_CAMx_Base11a_sensGCUSBG	TBD	6	This simulation will only be run if sensGCBC yields significantly different and compelling results relative to the base simulation.