WRAP RMC Phase II Wind Blown Dust Project ENVIRON International Corporation and University of California, Riverside 15 November 2005 Tempe, AZ

Outline Phase I Project Summary Phase II Project Overview Recent Revisions & Enhancements for Base02a Modeling Results Model Performance Evaluation Next Steps

Phase I Project Summary Objectives – Develop general methodology based on ‘MacDougall Method’ – Develop 1996 gridded PM inventory of WB Dust for the Western States Estimation Methodology – Categorize vacant land & soil types (disturbance, vegetative cover, soil texture) – Identify wind tunnel emission factors – Assign threshold wind velocities, wind & precipitation events, reservoir characteristics – Apply emission factors to vacant land as a function of wind speed & soil texture

Data Sources – Land Use/Land Cover (LULC) BELD3, NALCC, NLCD – Soil Characteristics STATSGO; Soil Landscape of Canada; Intl. Soil Reference and Information Centre – Meteorological Data km MM5 – Agricultural Data BELD3; RUSLE2; CTIC

Phase I Project Summary Limitations – Threshold surface friction velocities – Emission factors – Vacant land stability – Dust reservoirs – Rain events – Vegetation density

Phase II Project Overview Objectives – Develop improved general methodology based on Phase I recommendations and recent literature review – Develop 2002 gridded PM inventory of WB Dust for the Inter-RPO regional modeling domain – Development of surface friction velocities and threshold friction velocities – Develop improved emission flux relationships – Improved characterization of disturbed lands – Characterize vacant lands using more up-to-date databases – Conduct model performance evaluation

General Formulation Dust = f(LULC,z 0,u *,u *th,SC) u * = f(u,z 0 ) u *th = f(z 0 ) z 0 = f(LULC)

Recent Literature and Models Draxler, et al., 2001 – Regional model application – Dust emissions a function of u *, u *th, z 0 – z 0 correlated with soil properties of Persian Gulf region – Assumed all soils disturbed Zender, et al., 2003 – Global model application – Dust emissions a function of u *, u *th, z 0 – Global z 0 = 0.01 cm – Uniform soil texture Shao, 2001 – Dust emissions a function of u *, u *th, z 0 – Emphasis on particle size distribution and micro forces

Revised Methodology Threshold velocities Emission factors LULC characteristics & Soil characteristics Reservoir characteristics Agricultural adjustments

Threshold Friction Velocities u *th determined from relations developed by Marticorena, et al, (1997)

Emission Rates Determined from results of Alfaro and Gomes (2001) Dependent on soil type

Dust Category3467 Land useAg.GrassShrubsBarren Surface roughness (cm) Threshold friction Velocity (mile/h) Threshold wind velocity at 38m height (mile/h) Characteristics of Dust Categories

Soil Characteristics Soil Disturbance (Sensitivity Simulations) – Disturbed soils have a greater potential for erosion – Results in reduced threshold surface friction velocities – Soils are assumed to be undisturbed – Assumed percentage of disturbance by land use type – Threshold velocities reduced from ~20% – 90%

Reservoir Characteristics Amount and condition of soils impact dust emissions Climatological effects (rain, snow, etc.) Sandy soils dry fastest, loams medium range, clays dry slowest. Evapotranspiration rates ~ twice as high in summer than winter; ~ 1.4 times higher than spring/fall

Agricultural Adjustments Non-climatic factors significantly decrease soil loss from agricultural lands Similar approach to CARB, 1997 (as in Phase I) Five “adjustment” factors simulate these effects: – Bare soil within fields – Bare borders surrounding fields – Long-term irrigation – Crop canopy cover – Post-harvest vegetative cover (residue)

Data Sources Land Use/Land Cover – National Land Cover Database (NLCD) – Biogenic Emissions Landcover Database (BELD3) Soils – State Soil Geographic Database (STATSGO) – Soil Landscape of Canada (SLC) – International Soil Reference and Information Centre Meteorology – km Gridded MM5

Recent Revisions for Base02a Inventory Updated LULC database Dust Transport Fractions Dust Fine Fraction (PM10/PM2.5)

Updated LULC Database (2000)

LULC Summary/Comparison

Fine Fraction of Dust Original methodology used 0.78/0.22 (PMC/PMF) Revised methodology uses 0.90/0.10

Latest Transport Fractions (Tom Pace, EPA, 6/2005) Effects of near-source dust removal & deposition Varies by LULC; Applied at grid-cell level (12- km) LULC Category Pre02dBase02a Urban Agricultural Grasslands Shrublands Forest Barren/Water

Model Results WRAP 36-km Modeling Domain

Model Results WRAP 12-km Modeling Domain

WRAP 12-km Modeling Results

2002 Monthly WB PM10 Dust Emissions WRAP States

WRAP Dust Comparison 12-km Results

WRAP Dust Comparison 36-km Results

WRAP Dust Comparison

Model Results 2002 Annual WB Dust PM km

36km Model Results

Model Results 2002 Annual WB Dust PM km

12km Model Results

Model Results 2002 Annual – 12-km

Model Results County-level km

Model Performance Evaluation Evaluate model results for reasonableness and accuracy Compare predicted WB dust emissions near IMPROVE monitors with measured IMPROVE dust extinction (B dust ) – Identify occurrences of: 1) Zero WB dust and near-zero B dust 2) Enhanced WB dust and near-zero B dust 3) No WB dust and elevated B dust 4) Enhanced WB dust and elevated B dust Enhancements to CMAQ to track WB and other dust Evaluate model CMAQ model performance with and with out WB dust emissions

Saguaro West, AZ, Coarse Mass

Saguaro West, AZ, Soil

Salt Creek, NM, Coarse Mass

Salt Creek, NM, Soil

Next (Final) Steps Update/Complete Model performance Evaluation – Evaluate CMAQ model performance with & without WBD Emissions Update Project Final Report