New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and...
Transcript of New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and...
![Page 1: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/1.jpg)
New AERMOD Meteorological Data - Sensitivity Analysis
Brad Ashton
Iowa DNR – Air Quality Bureau
![Page 2: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/2.jpg)
Old Meteorological Data
• Covers period from 2000 – 2004
• 10 meteorological measurement sites:
•Burlington •Cedar Rapids •Des Moines •Dubuque •Mason City
•Moline •Omaha •Sioux City •Sioux Falls •Waterloo
• Iowa City data for the period 1999 – 2003 is used for University of Iowa
![Page 3: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/3.jpg)
New Meteorological Data
• Covers period from 2005 – 2009
• 19 meteorological measurement sites:
•Ames •Burlington •Cedar Rapids •Davenport •Des Moines •Dubuque •Estherville
•Iowa City •La Crosse •Lamoni •Marshalltown •Mason City •Moline •Omaha
•Ottumwa •Sioux City •Sioux Falls •Spencer •Waterloo
![Page 4: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/4.jpg)
Key Differences
• Correction of Automated Surface Observation System (ASOS) wind speed bias
• Inclusion of lower wind speeds
• Use of 1-minute wind data
• Different representative areas
• Each has an impact on model concentration…
![Page 5: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/5.jpg)
Correction of ASOS Wind Speed Bias
![Page 6: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/6.jpg)
Correction of ASOS Wind Speed Bias
• All ASOS wind speeds are truncated instead of being rounded:
(Results in a -0.5-knot wind speed bias on average)
Measured Wind Speed
(knots)
ASOS-Recorded Wind Speed
(knots)
2.1 2
1.8 1
2.6 2
3.1 3
![Page 7: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/7.jpg)
Correction of ASOS Wind Speed Bias
• Newest version of AERMET corrects bias:
“Adjustment of ASOS-based wind speeds (including winds derived from 1-minute ASOS data) by +0.5 knot to account for the bias in reported ASOS wind speeds due to winds being truncated (rather than rounded) to whole knots.”
~ AERMET Model Change Bulletin - 2/28/2011
![Page 8: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/8.jpg)
Correction of ASOS Wind Speed Bias
• Model concentration is inversely proportional to wind speed.
– Correction of bias results in increased wind speed and therefore decreased concentration.
Higher Wind Speed
Lower Concentration
![Page 9: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/9.jpg)
Inclusion of Lower Wind Speeds
![Page 10: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/10.jpg)
Inclusion of Lower Wind Speeds
• More sensitive wind instruments introduced to Iowa ASOS network beginning in 2005.
(> 2 knots)
(~ 0 knots +)
![Page 11: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/11.jpg)
Inclusion of Lower Wind Speeds
Old Data New Data
![Page 12: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/12.jpg)
Inclusion of Lower Wind Speeds
40 CFR Part 51 – Appendix W:
“AERMOD can produce model estimates for conditions when the wind speed may be less than [2 knots], but still greater than the instrument threshold”
“…in steady-state Gaussian plume models…wind speeds of less than [2 knots] but higher than the response threshold of the instrument should be input as [2 knots]…for input to AERMOD, no adjustment should be made…”
![Page 13: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/13.jpg)
Inclusion of Lower Wind Speeds
• Model concentration is inversely proportional to wind speed.
– Improved instrument sensitivity results in an increased occurrence of low wind speeds, and therefore increased concentration.
Lower Wind Speed
Higher Concentration
![Page 14: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/14.jpg)
Use of 1-minute Data
![Page 15: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/15.jpg)
Use of 1-minute Data
• The wind data included in the hourly ASOS record are not ideal for dispersion modeling.
– Based on a single 2-min average observed approximately 10 minutes before each hour.
– Non-zero wind speeds below 3 knots (which are valid in AERMOD) are recorded as calm.
– Variable winds recorded as non-zero wind speed with no direction.
![Page 16: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/16.jpg)
Use of 1-minute Data
• Hours with calm or variable winds are treated as missing by AERMOD.
• Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded hours are generally those that would have otherwise produced high concentrations.
• The exclusion of periods of high concentrations results in an under-prediction bias.
![Page 17: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/17.jpg)
Use of 1-minute Data
• The wind data included in the 1-minute ASOS record is more ideal for dispersion modeling. – Can be used to calculate an average wind direction
and speed (more appropriate for dispersion modeling applications, and addresses the issue of variable winds).
– Wind speeds less than 3 knots are included (helps address the issue of calms).
• Results in fewer calm and variable wind hours, which helps correct the under-prediction bias that occurs when only the hourly ASOS record is used.
![Page 18: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/18.jpg)
Use of 1-minute Data
![Page 19: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/19.jpg)
Use of 1-minute Data
• Model concentration is inversely proportional to wind speed.
– Correction of calm and variable wind bias results in an increased occurrence of low wind speeds, and therefore increased concentration.
Lower Wind Speed
Higher Concentration
![Page 20: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/20.jpg)
Different Representative Areas
![Page 21: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/21.jpg)
Different Representative Areas
![Page 22: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/22.jpg)
Different Representative Areas
– Additional sites
– River valleys
• Representative areas have changed for many locations.
![Page 23: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/23.jpg)
Different Representative Areas
• River valleys that significantly affect atmospheric flow are generally characterized by less disperse wind fields due to channeling effects.
Des Moines, IA (non-valley) Omaha, NE (valley)
• Less disperse wind fields cause increased frequency of impacts at locations downwind of primary wind directions, resulting in higher average long-term concentrations.
![Page 24: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/24.jpg)
Different Representative Areas
• Sites within one of the river valley subdivisions previously represented by meteorological data collected at a site outside the river valley can expect an increase in long-term concentration.
• Mississippi River valley portions of Allamakee, Clayton, Dubuque and Jackson counties
![Page 25: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/25.jpg)
Different Representative Areas
• Sites outside one of the river valley subdivisions previously represented by meteorological data collected at a site within the river valley can expect a decrease in long-term concentration.
• Non-Mississippi River valley portions Clinton and Scott counties
• Non-Missouri River valley portions of Fremont, Harrison, Mills, Monona, Pottawattamie and Woodbury counties
• Montgomery, Page and Shelby counties)
![Page 26: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/26.jpg)
Comparison of Model Results
![Page 27: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/27.jpg)
Comparison of Model Results
• Modeled a variety of sources:
– Point (stacks, vents)
– Volume (haul roads, indoor emissions)
– Area (storage piles)
• Varied release height between 0 and 65 meters
• Included a variety of averaging periods
• Used both old and new meteorological data
![Page 28: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/28.jpg)
Comparison of Model Results
• Calculated the ratio of maximum model result using the new data to the maximum model result using the old data on a like-station basis
(does not address change caused by change in representative measurement site).
Ratio > 1 Increased concentration
Ratio = 1 No Change
Ratio < 1 Decreased concentration
![Page 29: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/29.jpg)
Comparison of Model Results (by averaging period)
![Page 30: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/30.jpg)
Comparison of Model Results (by release height)
![Page 31: New Meteorological Datamissing by AERMOD. •Missing hours are excluded from the model average, and because calms – by definition – occur during periods of light winds, the excluded](https://reader033.fdocuments.us/reader033/viewer/2022042018/5e7628f7de555b109365b37b/html5/thumbnails/31.jpg)
Availability and Usage
• Available now (www.iowacleanair.com): – Meteorological data files
– Technical support document • Methodology
• Representivity analysis
• Detailed sensitivity analysis
• Station summary
• Begin using immediately
• PSD projects with approved protocols may use the 2000 – 2004 data until 6/30/2012.