ML082530026

Lr Hearing - Llnl Comparison of Gaussian 2D and 3D Transport Models

ML082530026
Person / Time
Site:	Indian Point
Issue date:	08/04/2008
From:	Office of Nuclear Reactor Regulation
To:	Division of License Renewal
References
Download: ML082530026 (101)

Text

IPRenewal NPEmails From: Robert Palla Sent: Monday, August 04, 2008 1:02 PM To: Beth Mizuno

Subject:

LLNL Comparison of Gaussian 2D and 3D Transport Models Attachments: NUREG CR-6853 LLNL Comparison of Gaussian 2D and 3D Transport Models.pdf The attached document discusses the factor of two agreement between MACCS and the 2D and 3D codes. If you have a minute, read paragraph 1 and 2 on page xii of Executive Summary.

1

Hearing Identifier: IndianPointUnits2and3NonPublic_EX Email Number: 306 Mail Envelope Properties (Robert.Palla@nrc.gov20080804130100)

Subject:

LLNL Comparison of Gaussian 2D and 3D Transport Models Sent Date: 8/4/2008 1:01:49 PM Received Date: 8/4/2008 1:01:00 PM From: Robert Palla Created By: Robert.Palla@nrc.gov Recipients:

"Beth Mizuno" <Beth.Mizuno@nrc.gov>

Tracking Status: None Post Office:

Files Size Date & Time MESSAGE 181 8/4/2008 1:01:00 PM NUREG CR-6853 LLNL Comparison of Gaussian 2D and 3D Transport Models.pdf 5712285 Options Priority: Standard Return Notification: No Reply Requested: No Sensitivity: Normal Expiration Date:

Recipients Received:

NUREG/CR-6853 Comparison of Average Transport and Dispersion Among a Gaussian, a Two-Dimensional, and a Three-Dimensional Model Lawrence Livermore National Laboratory U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research Washington, DC 20555-0001

NUREG/CR-6853 Comparison of Average Transport and Dispersion Among a Gaussian, a Two-Dimensional, and a Three-Dimensional Model Manuscript Completed: October 2004 Date Published: October 2004 Prepared by C.R. Molenkamp (LLNL), N.E. Bixler, C.W. Morrow (SNL),

J.V. Ramsdell, Jr., (PNNL), J.A. Mitchell (NRC)

Atmospheric Science Division Sandia National Laboratories Lawrence Livermore National Laboratory Albuquerque, NM 87185-0748 Livermore, CA 94550 Pacific Northwest National Laboratory Richland, WA 99352 J.A. Mitchell, NRC Project Manager Prepared for Division of Systems Analysis and Regulatory Effectiveness Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 NRC Job Code Y6785

ABSTRACT The sim p lifying atm ospheric transp ort and d isp ersion assu m p tion u sed by MACCS2, the N u clear Regulatory Com m issions cod e for p red icting off-site consequ ences, is tested by com parison to ADAPT/ LODI, a state-of-the-art, three-d im ensional ad vection-d isp ersion cod e. Also includ ed in the com p arison is the N u clear Regu latory Comm issions cod e for rap id em ergency resp onse, RASCAL, and a new er related cod e w ith u p grad ed d ispersion and d ep osition m od u les, RATCH ET. Meteorological d ata for the test w ere provid ed by the Departm ent of Energys Atm osp heric Rad iation Measu rem ent Program Southern Great Plains site in central Oklahom a and Kansas, a site w ith a unique and com prehensive set of m esoscale meteorological d ata. Each m od el w as ru n in its norm al m anner to p rod u ce the annu al average integrated exp osu re and d ep osition for a series of rings at 16.1, 32.2, 80.5, and 160.9 km (10, 20, 50, and 100 m iles) from a hyp othetical release, and the integrated exp osu re and d ep osition for arc-sectors at the sam e set of d istances and the 16 com p ass d irections. N early all the annu al average ring exposures and d epositions and the great m ajority of the arc-sector valu es for MACCS2, RASCAL, and RATCH ET w ere w ithin a factor of tw o of the correspond ing ADAPT/ LODI values.

iii

iv CON TEN TS ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii APPEN DICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix EXECUTIVE

SUMMARY

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi FOREWORD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv ACKN OWLEDGMEN TS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii ACRON YMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix

1. BACKGROUN D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2. SELECTION OF TH E STUDY SITE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3. MODELS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.1 MACCS2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.1.1 Meteorological Rep resentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.1.2 Atm ospheric Transp ort and Disp ersion . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.1.3 Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 3.2 RASCAL and RATCH ET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 3.2.1 Meteorological Rep resentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.2.2 Atm ospheric Transp ort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.2.3 Atm ospheric Disp ersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.2.4 Dry Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.2.5 Wet Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.3 N ARAC Mod els - ADAPT and LODI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.3.1 ADAPT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.3.2 LODI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.3.2.1 Dry Dep osition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.3.2.2 Wet Dep osition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4. METEOROLOGICAL DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4.1 ARM Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4.2 Surface Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 v

4.2.1 ARM SMOS Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.2.2 Oklahom a Mesonet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.3 Upper Air Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.3.1 Sond es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4.3.2 Rad ar Wind Profiler/ Rad io Acou stic Sou nd ing System . . . . . . . . . . 23 4.3.3 N OAA Profiler N etw ork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 4.3.4 Atm ospheric Em itted Rad iance Interferom eter . . . . . . . . . . . . . . . . . 24

5. METEOROLOGICAL DATA PROCESSIN G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 5.1 Pasquill-Gifford (P-G) Stability Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 5.2 Mixing H eight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 5.3 Low -Level N octurnal Jet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 5.4 MACCS2 Input Meteorology File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 5.5 RASCAL/ RATCH ET Inp u t Meteorology Files . . . . . . . . . . . . . . . . . . . . . . . . . 28 5.6 ADAPT/ LODI Input Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

6. DATA QUALITY ISSUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 6.1 SMOS Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 6.2 N PN and 915 MH z RWP Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

7. SOURCE TERM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

8. SIMULATION PROCEDURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 8.1 MACCS2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 8.2 RASCAL/ RATCH ET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 8.3 ADAPT/ LODI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 8.3.1 Grid s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 8.3.2 Solution Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 8.4 Com puter Tim e Requ irem ents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 8.4.1 MACCS2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 8.4.2 RASCAL/ RATCH ET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 8.4.3 ADAPT/ LODI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

9. WIN D CH ARACTERISTICS OF TH E ARM SGP SITE . . . . . . . . . . . . . . . . . . . . . . . . . . 47

10. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 10.1 Arc Averages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 10.2 Arc-Sector Averages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 10.3 Tw o-Dim ensional Exposu re and Dep osition . . . . . . . . . . . . . . . . . . . . . . . . . . 62 10.4 Sum m ary of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

11. REFEREN CES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 vi

FIGURES Figure 1. Location of Surface Meteorological Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 2. Location of ARM Upper Air Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 3. Cu m u lative N u m ber of H ou rs w ith Mixing H eights less than a Given H eight

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Figure 4. Exam ple of a N octurnal Jet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Figure 5. Central Facility Sond e Lau nched at 5:29 am CST on Janu ary 31, 2000. . . . . . 34 Figure 6. 915 MH z RWP Wind Profile for the Central Facility at 10:00 p m CST on March 18, 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 7. Central Facility Sond e Lau nched at 5:33 am CST on N ovem ber 15, 2000 . . . . 36 Figure 8. 915 MH z RWP Wind Profile for Beau m ont at 11:00 am CST on May 14, 2000

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Figure 9. Topography of the ARM SGP Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Figure 10. Wind Rose for the ARM Central Facility at Lam ont, OK . . . . . . . . . . . . . . . . . 47 Figure 11. Surface Wind Rose for the ARM Central Facility for Winter, 2000 . . . . . . . . 48 Figure 12. Surface Wind Rose for the ARM Central Facility for Sp ring, 2000 . . . . . . . . . 48 Figure 13. Surface Wind Rose for the ARM Central Facility for Su m m er, 2000 . . . . . . . 48 Figure 14. Surface Wind Rose for the ARM Central Facility for Au tu m n, 2000 . . . . . . . 48 Figure 15. Surface Wind Rose for the Elk Falls, KA Site . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 16. Surface Wind Rose for the Ashton, KA Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 17. Surface Wind Rose for the Byron, OK Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 18. Surface Wind Rose for the Ringw ood , OK Site . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 19. Surface Wind Rose for the Meeker, OK Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 20. Wind Rose from 915 MH z Profiler at 87 m H eight at Lam ont, OK . . . . . . . . 50 Figure 21. Wind Rose from 915 MH z Profiler at 87 m H eight at Beau m ont, KA . . . . . . 50 Figure 22. Wind Rose from 915 MH z Profiler at 87 m H eight at Med icine Lod ge, KA

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 23. Wind Rose from 915 MH z Profiler at 87 m H eight at Meeker, OK . . . . . . . . 50 Figure 24. Wind Rose from 915 MH z Rem ote Wind Profiler (RWP) at 495 m height at Lam ont, OK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 25. Wind Rose from N OAA Wind Profiler (N WP) at 500 m H eight at Lam ont, OK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 26. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 16.9-km (10-mile) Arc

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Figure 27. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 32.2-km (20-mile) Arc

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Figure 28. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 80.5-km (50-mile) Arc

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Figure 29. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 160.9-km (100-mile) Arc

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Figure 30. Arc-Sector Exposure for Dep ositing Sp ecies on 16.1-km (10-mile) Arc . . . . . 58 vii

Figure 31. Arc-Sector Exposure for Dep ositing Sp ecies on 32.2-km (20-mile) Arc . . . . . 58 Figure 32. Arc-Sector Exposure for Dep ositing Sp ecies on 80.5-km (50-mile) Arc . . . . . 59 Figure 33. Arc-Sector Exposure for Dep ositing Sp ecies on 160.9-km (100-mile) Arc . . . 59 Figure 34. Arc-Sector Deposition on 16.1-km (10-mile) Arc . . . . . . . . . . . . . . . . . . . . . . . . 60 Figure 35. Arc-Sector Deposition on 32.2-km (20-mile) Arc . . . . . . . . . . . . . . . . . . . . . . . . 60 Figure 36. Arc-Sector Deposition on 80.5-km (50-mile) Arc . . . . . . . . . . . . . . . . . . . . . . . . 61 Figure 37. Arc-Sector Deposition on 160.9-km (100-mile) Arc . . . . . . . . . . . . . . . . . . . . . . 61 Figure 38. N orm alized Exposure for N on-Dep ositing Material on the 16.1-km (10-m ile)

Arc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Figure 39. N orm alized Exposure for N on-Dep ositing Material on the 32.2-km (20-m ile)

Arc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Figure 40. N orm alized Exposure for N on-Dep ositing Material on the 80.5-km (50-m ile)

Arc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Figure 41. N orm alized Exposure for N on-Dep ositing Material on the 160.9-km (100-m ile) Arc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Figure 42. N orm alized Exposure for Dep ositing Material on the 16.1-km (10-mile) Arc.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Figure 43. N orm alized Exposure for Dep ositing Material on the 32.2-km (20-mile) Arc.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Figure 44. N orm alized Exposure for Dep ositing Material on the 80.5-km (50-mile) Arc.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Figure 45. N orm alized Exposure for Dep ositing Material on the 160.9-km (100-m ile)

Arc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Figure 46. N orm alized Deposition on the 16.9-km (10-mile) Arc. . . . . . . . . . . . . . . . . . . . 67 Figure 47. N orm alized Deposition on the 32.2-km (20-mile) Arc. . . . . . . . . . . . . . . . . . . . 67 Figure 48. N orm alized Deposition on the 80.5-km (50-mile) Arc. . . . . . . . . . . . . . . . . . . . 68 Figure 49. N orm alized Deposition on the 160.9-km (100-mile) Arc. . . . . . . . . . . . . . . . . . 68 Figure 50. Annual Average Exposu re for N on-Dep ositing Sp ecies . . . . . . . . . . . . . . . . . 69 Figure 51. Annual Average Exposu re for Dep ositing Sp ecies . . . . . . . . . . . . . . . . . . . . . . 70 Figure 52. Annual Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 TABLES Table 1. Description of Weather Bins Used in MACCS2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Table 2. RASCAL and RATCH ET Precip itation Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Table 3. RASCAL Washout Coefficients (s -1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Table 4. RATCH ET Wet Deposition Mod el Param eters . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Table 5. N um ber of Sond es at Each Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Table 6. N um ber of H ours in Each Stability Category for MACCS2 Site . . . . . . . . . . . . 28 Table 7. Seasonal Average Mixing H eights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Table 8. N um ber of H ours in Each Stability Category for RASCAL Sites . . . . . . . . . . . . 29 Table 9. N um ber of H ours vs N um ber of Valid Su rface Observations . . . . . . . . . . . . . . 29 viii

Table 10. N um ber of H ours vs. N u m ber of Up p er Air Profiles . . . . . . . . . . . . . . . . . . . . 31 Table 11. H ours w ith Valid Data for Profiler and Sond e Sites . . . . . . . . . . . . . . . . . . . . . 31 Table 12. Source Term Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Table 13. Meteorology Grid Vertical Levels and Corresp ond ing Central Facility Altitud es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Table 14. N on-Dep ositing Species Arc Average Exp osure . . . . . . . . . . . . . . . . . . . . . . . . 53 Table 15. Depositing Species Arc Average Exp osu re . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Table 16. Arc Average Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 APPEN D ICES APPEN DIX A - CON TEN TS OF DATA CD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 ix

x EXECUTIVE

SUMMARY

The N u clear Regulatory Com m issions cod e for p red icting off-site consequ ences and p robabilistic risk assessm ents, MACCS2, u ses a straight-line Gau ssian m od el for atm osp heric transport and d ispersion that has been criticized as overly sim p listic.

Becau se of an increased interest in level-3 probabilistic risk analyses, testing of the sim p lifying transport and d ispersion assu m p tion is p erform ed here u sing com p arison of MACCS2 to ADAPT/ LODI, a state-of-the-art, three-d im ensional ad vection-d isp ersion cod e. Also includ ed in the com p arison is the N u clear Regu latory Com m issions cod e for rapid em ergency resp onse, RASCAL3.0, and RATCH ET, a new er related cod e w ith upgrad ed d ispersion and d ep osition m od u les.

The site chosen for the test w as the Dep artm ent of Energys Atm osp heric Rad iation Measurem ent Program Southern Great Plains site in central Oklahom a and Kansas, w hich w as selected prim arily on the basis of the available atm osp heric d ata. The d ata set consisted of hourly average m easu rem ents of w ind , tem p eratu re, and tu rbu lence both at the surface and aloft, and hou rly p recipitation d u ring the entire year 2000. This is the only site w e know of that has m u ltip le u p p er air w ind and tem p erature m easu rem ents w ithin a 322-km (200-m ile) squ are area over a p eriod longer than one year, and it allow ed us to perform the com p arison w ithou t u sing any p seud o observations from forecast or pred iction m od els. We w ou ld have p referred a site w ith greater topographical and d iurnal heterogeneity, bu t w ith frequ ent low -level noctu rnal jets and occasional severe storm s there w as su fficient variability.

Each m od el w as run in its norm al mod e by p ersonnel fam iliar w ith it, and each u sed its typ ical set of m eteorological observations. MACCS2 u sed only the su rface observations (w ind s at 10 m , and surface precip itation) at one site, the Central Facility. RASCAL and RATCH ET used surface d ata at the Central Facility and at five ad d itional sites.

ADAPT/ LODI used the entire three-d im ensional d ata set consisting of 100 su rface sites and fifteen u p p er air sites (ten rem ote w ind p rofiler and five sond e sites). ADAPT is an ad vanced d ata assim ilation m od el d esigned to convert su ch large and d iverse sets of observations into grid d ed three-d im ensional w ind and turbulence field s that agree w ith the observations and are m ass-consistent; these grid d ed m eteorological d ata allow LODI to accu rately calcu late atm osp heric transp ort and d isp ersion for ind ivid u al releases.

MACCS2 used its binning proced ure to select a rep resentative set of 610 release start times and associated w eights from the 8760 hou rly observations. A hyp othetical point sou rce w as d efined that released 1016 Bq each of a d ep ositing and non-d ep ositing m aterial u niform ly over a period of 30 m inu tes at a height of 50 m w ith a bu oyant heat flux of 106 W. RASCAL, RATCH ET, and LODI calcu lated ind ivid u al exp osu re (air concentration at the surface integrated over the p assage of the p lu m e) and d ep osition xi

(total material d eposited on the grou nd by w et and d ry d ep osition d u ring p lum e passage) spatial d istributions for each of these releases and averaged them u sing the associated w eights to prod uce an annu al average for com p arison w ith MACCS2.

Tw o com parison m etrics w ere u sed . The first metric consisted of the average exp osure and d ep osition in four circular rings arou nd the sou rce at d istances betw een 14.4 and 16.1, 30.6 and 32.2, 78.7 and 80.5, and 159.3 and 160.9 km (9 and 10, 19 and 20, 49 and 50, and 99 and 100 m iles). The second m etric consid ered the average exp osu re and d ep osition in arc-sectors using the sam e fou r d istances for the arcs and the 16 com pass sectors from N clockw ise around to N N W, a total of 64 valu es for each exp osure (d ep ositing and non-d epositing m aterial) and d ep osition (for d ep ositing m aterial).

Sim ilar com parisons w ere perform ed w ith RASCAL and RATCH ET bu t only for the inner three rings because RASCAL and RATCH ET only follow ed the p lu m e for 80.5 km (50 miles).

MACCS2's ring average values ranged from a m inim u m of 0.64 to a m axim u m of 1.58 tim es the correspond ing LODI ring average w ith higher ratios occurring for the 16.1-km (10-m ile) ring and low er for the 80.5- and 160.9-km (50- and 100-m ile) rings. All these ratios are w ell w ithin a factor of tw o. The arc-sector exp osu res and d ep ositions for MACCS2 w ere also usually w ithin a factor of tw o of the corresp ond ing valu e for LODI.

Of the 192 exposures and d epositions (4 arcs, 16 sectors, 2 exp osu res and 1 d ep osition),

only nine w ere m ore than tw ice as large (all in the 16.1-km arc) and tw elve w ere less than half as large (four in the 80.5-km and eight in the 160.9-km arc), and these w ere u su ally in sectors w here the exposu re or d eposition w as sm aller. Differences greater than a factor of three occurred only tw ice. Overall, the arc average and the great m ajority of the arc-sector average exp osu res and d ep ositions w ere w ithin a factor of tw o w hen com paring MACCS2 to LODI.

RASCAL calculated exp osures and d ep ositions consistently larger than LODI (ratio range for rings from 1.12 to 1.65) w hile RATCH ET calcu lated values sm aller than LODI (ratio range for rings from 0.48 to 0.88). Still, nearly all these valu es are w ithin a factor of tw o of LODI. The larger ratios for RASCAL and the sm aller ratios for RATCH ET m ust be d ue to the d ispersion and d ep osition m od u les since those are the only d ifferences betw een the m od els. The arc-sector exp osu res and d ep ositions for RASCAL and RATCH ET often d iffer from LODI by m ore than a factor of tw o, bu t this is p artly related to the fact that RASCAL tend s to consistently p rod u ce higher and RATCH ET low er values than LODI. RASCAL has 33 of 144 exp osu res and d ep ositions (3 arcs, 16 sectors, 2 exposures and 1 d ep osition) m ore than tw ice as large as LODI, and none less than half as large. Ten of these are m ore than three tim es as large as LODI. RATCH ET has three of 144 exposures and d ep ositions more than tw ice as large as LODI and 33 less than half as large. Of these 33, ten are less than one-third as large as LODI. Differences xii

betw een RASCAL/ RATCH ET and LODI m ay be d u e to d ifferent p aram eterizations of d isp ersion and d eposition, to d ifferent rep resentations of transp ort, or to both.

xiii

xiv FOREWORD The com p arisons presented in this rep ort are am ong three d ifferent m od els that can be used for d eterm ining how rad ioactive m aterial that cou ld be released follow ing a postu lated incid ent at an N RC-licensed facility cou ld m ove throu gh and sp read w ithin the atm osp here. This process of m oving and sp read ing is called atm osp heric transp ort and d isp ersion (ATD). Some m aterials are gaseous in nature, w hile others are in the form of sm all particles, w hich can fall ou t onto the grou nd , a p rocess called d ep osition.

This com p arison w as und ertaken becau se the simp le mod els have been criticized as being too-sim ple, w hich could lead to inaccu rate estim ates of risk. Since w eather is a key factor in the transport and d ispersion of rad ioactive material, the resu lts that w ere com p ared w ere averages over hund red s of w eather trials. A w eather trial is a set of sequ ential (tim ew ise) w eather observations used to m ove and sp read a p lu m e of m aterial from the place of release to the p lace w here it exits the region of interest.

ATD m od els are referred to as one-d im ensional, tw o-d im ensional, or three-d im ensional. The d ifficulty in d efining these terms for ATD m od els is that, for all three, d isp ersion actu ally takes place in all three d im ensions. The d ifferences am ong the m od els occu r in the estim ation of the transp ort. A one-d im ensional m od el assu m es that a p lum e m oves d ow nw ind along a straight line at the sp eed of the w ind . As the p lum e m oves d ow nw ind , it broad ens (in the crossw ind d irection) and grow s taller (in the vertical d irection). A tw o-d im ensional m od el allow s the p lu m e to bend and change d irection. Again, the plum e broad ens and grow s taller as it m oves d ow nw ind . A three-d im ensional m od el is the m ost comp lex. It allow s ind ivid u al p articles (m aking u p the plum e) to m ove in any d irection. With the three-d imensional mod el, the p lum e can sp lit into tw o p lu m es as it encou nters a hill, a canyon, or another com p lex w ind p attern.

The mod els are:

(1) the one-d im ensional, straight-line Gau ssian m od el that the N RC uses for cost/ benefit calculations, for em ergency p lanning, and for estimating off-site consequ ences for a Probabilistic Risk Assessm ent (PRA). The cod e is MACCS2, the MELCOR Accid ent Consequence Cod e System, Version 2. Sand ia N ational Laboratories ran this cod e.

(2) tw o tw o-d im ensional m od els w ith slightly d ifferent rep resentations of d isp ersion and of d eposition. The first cod e is RASCAL, Rad iological Assessm ent System for Consequ ence Analysis, w hich is used tod ay in N RCs Incid ent Resp onse Center for resp onse to rad iological em ergencies. The second cod e is RATCH ET, Regional Atm osp heric Transport Cod e for H anford Em ission Tracking, w hich w as d evelop ed for u se in the H anford Environm ental Dose Reconstru ction Project. Develop m ent of RATCH ET em phasized upgrad ing the m ethod s in RASCAL for calcu lating d isp ersion and d ep osition. Pacific N orthw est N ational Laboratory ran these cod es.

xv

(3) a three-d im ensional m od el that emp loys tw o cod es, one that w as u sed in this case to estim ate the w ind field in three d im ensions based on thou sand s of d ata points and another that w as used to estim ate the gaseou s and p articulate m aterial transp ort. The w ind -field cod e w as ADAPT (Atm osp heric Data Assim ilation and Param erterization Techniqu es) and the d ispersion and d ep osition cod e w as LODI (Lagrangian Op erational Dispersion Integrator). Law rence Liverm ore N ational Laboratory ran these cod es.

The resu lts from ADAPT/ LODI calcu lations w ere used as a benchm ark for the sim pler (1-D and 2-D) cod es. It w ould have been p referable to com p are the simp ler cod es w ith m easu red values, but such m easurem ents d o not exist over the d istance of interest to the N RC.

The location selected for this com parison, the Dep artm ent of Energys Atm osp heric Rad iation Measurem ent Southern Great Plains site (ARM SGP), w as chosen to p rovid e the most realistic test of the mod els cap abilities. N o other site in the United States p rovid es the regularly-collected m easu rements at the su rface and above the su rface at m ore than one location w ithin 160.9 km (100 m i) necessary to allow this com p arison to be based solely on m easured w eather d ata.

Each m od el w as run by users fam iliar w ith its op eration, u sing each cod e in its norm al m anner and relying on its typical set of observations of the w eather. Com p arisons w ere m ad e for 4 d ifferent d istances from the site of the assu m ed released (3 for RASCAL and RATCH ET) and for 16 d ifferent d irections, rep resenting com p ass d irections. Good agreem ent w as found am ong all the m od els tested , consid ering the p u rp oses of the variou s cod es. The averages over one-m ile w id e rings at the fou r d istances w ere w ithin a factor of tw o for both the MACCS2 and the RASCAL/ RATCH ET resu lts com p ared w ith the LODI/ ADAPT results.

Cheryl Trottier, Chief Rad iation Protection, Environm ental Risk, and Waste Managem ent Branch Division of System s Analysis and Regu latory Effectiveness Office of N u clear Regu latory Research xvi

ACKN OWLED GMEN TS We gratefully acknow led ge that the m eteorological d ata w ere obtained from the Atm osp heric Rad iation Measurem ent (ARM) Program sp onsored by the U.S.

Dep artm ent of Energy, Office of Science, Office of Biological and Environm ental Research, Environm ental Sciences Division. A p ortion of the su rface meteorological d ata w as from the Oklahom a Mesonet Program , a netw ork of environm ental m onitoring stations d esigned and im p lem ented by scientists at the University of Oklahom a and at Oklahom a State University and op erated by the Oklahom a Clim atological Su rvey.

A p ortion of this w ork w as perform ed u nd er the au sp ices of U. S. Dep artm ent of Energy by the University of California, Law rence Liverm ore N ational Laboratory u nd er contract W-7405-EN G-48.

The Pacific N orthw est N ational Laboratory is op erated by Battelle for the U. S.

Dep artm ent of Energy und er Contract DE-AC06-76RL01830. The p ortion of the stu d y involving RASCAL and RATCH ET w as fu nd ed by the Incid ent Resp onse Directorate of the U.S. N uclear Regulatory Com m ission Office of N u clear Secu rity and Incid ent Resp onse.

xvii

xviii ACRON YMS ADAPT Atm ospheric Data Assim ilation and Param eterization Techniqu es AERI Atm ospheric Em itted Rad iance Interferom eter ARM Atm ospheric Rad iation Measu rem ent ATD Atm ospheric Transp ort and Disp ersion DOE Dep artment of Energy GMT Greenw ich Merid ian Time LLN L Law rence Liverm ore N ational Laboratory LODI Lagrangian Op erational Disp ersion Integrator LST Local Stand ard Tim e MACCS2 MELCOR Accid ent Consequ ence Cod e System , Version 2 MSL Mean Sea Level N ARAC N ational Atm osp heric Release Ad visory Center netCDF N etw ork Com m on Data Form at N OAA N ational Oceanic and Atm osp heric Ad m inistration N PN N OAA Profiler N etw ork N RC N uclear Regulatory Comm ission PRA Probabilistic Risk Analyses RASCAL Rad iological Assessm ent System for Consequ ence Analysis RASS Rad io Acoustic Sou nd ing System RATCH ET Regional Atm osp heric Transp ort Cod e for H anford Em ission Tracking xix

RWP Rad ar Wind Profiler SGP U.S. Southern Great Plains Site UTC Coord inated Universal Time xx

1. BACKGROUN D The N u clear Regulatory Com m issions (N RCs) cod e for p red icting off-site consequ ences, MACCS2 (Chanin, et al. 1998) (MELCOR Accid ent Consequ ence Cod e System , Version 2), uses a sim plified m od el for atm osp heric transp ort and d isp ersion (ATD), that is, a straight-line Gau ssian m od el. The MACCS2 calculations are u sed by the N RC for planning purposes, for cost-benefit analyses, and in level-3 p robabilistic risk analyses (PRAs). The MACCS2 ATD m od el has been criticized as being overly sim p listic, even for its purposes. The ju stification for its u se has been that only average or exp ected values of m etrics of interest are need ed for the N RCs pu rp oses and that a sim p lified m od el, by averaging m etrics of interest obtained u sing nu m erou s w eather sequ ences one-by-one, com pensates for the loss of stru ctu re in the m eteorology that occu rs aw ay from the point of release. The sim p le m od el has been retained becau se of the d esire to have short running tim es on p ersonal comp u ters covering the entire p ath throu gh the environm ent, includ ing the food and w ater p athw ay, and covering essentially a lifetim e of exposure to a contam inated environm ent.

The assu m ption abou t the ad equacy of averaging m etrics of interest over nu m erou s w eather sequences has never been tested for the N RCs p u rp oses. Becau se of increased interest in level-3 PRA, testing of this assu m p tion is p erform ed here u sing com p arison of MACCS2 , the simplified m od el, to LODI (N asstrom , et al. 2000) (Lagrangian Op erational Dispersion Integrator), a state-of-the-art, three-d im ensional ad vection-d isp ersion cod e that uses a Lagrangian stochastic, Monte Carlo m ethod . LODI is cou p led to ADAPT (Sugiyam a and Chan 1998) (Atm osp heric Data Assim ilation and Param eterization Technique), w hich p rovid es tim e-varying, three d im ensional field s of m ean w ind s, turbulence, pressure, temp eratu re, and p recip itation based on, in this case, observed m eteorology.

RASCAL3.0 (Sjoreen, et al. 2001) (Rad iological Assessm ent System for Consequ ence Analysis, Version 3.0) is used by the N RC for em ergency resp onse ap p lications w hich requ ire rapid response. RASCAL3.0 contains ATD com p onents that are interm ed iate in com p lexity betw een MACCS2 and ADAPT/ LODI. It em p loys tim e-varying, tw o-d im ensional m eteorological field s of w ind , stability, and p recip itation based on su rface-level m eteorological observations as inp u t to a Lagrangian trajectory transp ort m od el and a Gau ssian puff d ispersion m od el. The d isp ersion p ortions of RASCAL3.0 are sim ilar to those of MACCS2, w hile the transp ort p ortions are significantly d ifferent.

N RC is consid ering u pgrad ing RASCAL3.0 by rep lacing the d isp ersion p ortions of the cod e w ith d ispersion m od ules from the RATCH ET cod e (Ram sd ell, et al. 1994).

RASCAL3.0 and RATCH ET w ere d evelop ed from the sam e p recu rsor cod e. In this stu d y, RATCH ET refers to a d evelop m ental version of RASCAL that incorp orates d isp ersion and d eposition m od ules from the original RATCH ET cod e. The RATCH ET 1

d isp ersion and d eposition m od ules are m ore m od ern than those in RASCAL. A com p arison of RASCAL and RATCH ET to ADAPT/ LODI has been inclu d ed .

The objective of this stud y is to d eterm ine if the average ATD resu lts from these cod es are su fficiently close that m ore com p lex m od els are not requ ired for the N RC p u rp oses of planning, cost-benefit, and PRA or d ifferent enou gh that one or both of the N RC cod es should be m od ified to p rovid e m ore rigorou s ATD. The d ecision w ill be m ad e by the N RC using results of this stud y and other factors, m ost notably ru n time and inp u t requ irem ents.

It w ou ld be best if MACCS2 and RASCAL/ RATCH ET resu lts cou ld be com p ared w ith m easurem ents over the long d istances and typ es of terrain of interest to the N RC.

H ow ever, such m easurem ents d o not exist, so the less d esirable comp arison w ith a state-of-the-art cod e w as chosen to p rovid e inp u t into the d ecision on the ad equ acy of MACCS2 ATD. The com parison w as also an op p ortu nity to gain ad d itional baseline inform ation on the perform ance of the RASCAL/ RATCH ET cod e. Comp arisons of LODI/ ADAPT results w ith intentional and u nintentional releases can be fou nd in Foster, et al. (2000). These com parisons, althou gh over shorter ranges than those of interest to the N RC, d em onstrate that LODI/ ADAPT is su fficiently accu rate for the p u rp oses of this stud y.

2

2. SELECTION OF THE STUD Y SITE Qu ite a few locations w ere consid ered as possible sites for this stu d y. These inclu d ed cu rrently operating nuclear pow er p lants, several DOE laboratory sites, and a few other locations. The follow ing criteria w ere consid ered in m aking a final selection:

C a d ata set w ith sufficient observations to characterize the horizontal w ind field as a three-d im ensional function of height and p osition from the sou rce ou t at least 160.9 km (100 miles),

C top ography that w ould interact w ith the large-scale flow p rod u cing local m od ification of w ind speed and d irection, and C a site w ith changes in surface p rop erties that cou ld affect the local flow , such as a coastal site w ith a land -sea breeze.

As w e consid ered the p ossible sites, w e cou ld id entify only one that satisfied the first criterion, the Departm ent of Energy (DOE) Atm osp heric Rad iation Measu rement (ARM) Sou thern Great Plains (SGP) site in Oklahom a and Kansas. N o other site p rovid ed regular u p p er air d ata at m ore than one location w ithin 160.9 km (100 m iles) of the sou rce. To use a d ifferent site w ou ld have requ ired u se of a regional mod el to d eterm ine the flow field s, and w e w anted to base this stu d y solely on observations. The top ograp hy of Oklahom a and Kansas is relatively sm ooth and has m inim al effect on the w ind field , and the su rface is fairly u niform and therefore p rod u ces relatively little local therm al forcing. H ow ever, w ind field s in Oklahom a and Kansas are frequ ently affected by low -level nocturnal jets and occasional severe storm s. Therefore, the last tw o criteria w ere only partially satisfied , but there w as su fficient variability for the p u rp ose of this stu d y. At the outset w e realized that if the d ifferences betw een MACCS2 and ADAPT/ LODI w ere large at the ARM site, they w ould be large everyw here, and the transp ort and d ispersion m od ule in MACCS2 w ou ld likely requ ire rep lacem ent. Bu t if the d ifferences w ere sm all, the ad equ acy of MACCS2s atm osp heric transp ort and d ispersion m od ule m ight still be u nresolved for som e special locations.

3

4

3. MOD ELS 3.1 MACCS2 The MELCOR Accid ent Consequence Cod e System Version 2 (MACCS2) (Chanin et al.

1998) w as d eveloped at Sand ia N ational Laboratories for the N RC. Its p rimary u se is in p erform ing consequ ence analyses in su p p ort of level-3 p robabilistic risk assessm ents (PRAs). It is also used by the N RC for p lanning p u rp oses and cost-benefit analyses.

MACCS2 is the latest in a series of N RC-sp onsored cod es for estimating off-site consequ ences follow ing a release of rad ioactive m aterial into the environm ent. The first cod e in the series w as CRAC (Calculation of Reactor Accid ent Consequ ences), w hich w as d eveloped for the Reactor Safety Stu d y (WASH -1400, 1975). The first version of MACCS w as released to the public in 1987. A su bsequ ent version w as u sed in the benchm ark PRA stud y reported in N UREG-1150.

MACCS2 is a versatile cod e, w ith m ost of its p aram eters being u nd er u ser control to facilitate the perform ance of sensitivity and u ncertainty analyses. The p rincipal p henom ena consid ered by MACCS2 are atm osp heric transp ort and d isp ersion (ATD),

short- and long-term m itigative actions, exp osu re p athw ays and d oses, d eterm inistic and stochastic health effects, and econom ic costs. Of these cap abilities, only the ATD processes are consid ered in the p resent stud y.

The atm ospheric m od els in MACCS2 are relatively simp le. Released m aterial is assum ed to travel d ow nw ind in a straight line. The concentration p rofiles in the cross-w ind and vertical d imensions are ap p roxim ated as being Gau ssian. The Gau ssian p lu m e m od el w as chosen for MACCS2 becau se it requ ires m inim al com p u tational effort and allow s large num bers of realizations to be calcu lated . These realizations represent uncertainty in w eather d ata at the time of a hyp othetical accid ent and u ncertainty in other inp u t p aram eters to represent d egree of belief. Large nu m bers of realizations (hu nd red s) are generally need ed to p erform PRA and sensitivity stu d ies.

3.1.1 Meteorological Representation The norm al calculation m od e for MACCS2 is to sam p le from hou rly w eather d ata for one year and to calculate ATD using a Gau ssian m od el in each of 16 d irections. Each d irection correspond s to a 22.5 d egree-w id e sector that is centered on a stand ard com p ass p oint. Each w eather sequence is w eighted by its p robability of occurrence. The w eather sequences are norm ally chosen, and have been chosen for this stu d y, to em phasize sam pling of sequences believed to be imp ortant to the p red iction of early health effects in an exposed population. This emp hasizes selection of w eather sequ ences 5

in w hich it rains w hile the plum e rem ains w ithin abou t 32.2 km (20 m iles) from the p oint of release.

MACCS2 w as u sed to select the w eather sequ ences that w ere u sed in this stud y. A total of 610 sequ ences w as chosen using the stand ard w eather binning ap p roach. This ap p roach bins each of the 8760 hou rs of d ata in an annu al w eather file into 36 bins, as show n in Table 1. The last tw o colu m ns of the table rep resent the valu es for the ARM SGP site.

The colum ns in Table 1 show for each w eather bin the inclu d ed stability class or classes, the w ind speed range, and the range of d istances traveled by the p lu m e w hen rain of a prescribed intensity occurs. It also show s the overall nu m ber of w eather sequ ences in the bin and the num ber of w eather sequ ences selected from the bin in this stu d y. The algorithm used to d eterm ine the nu m ber of sequ ences selected from each bin is the larger of tw o quantities: 12 or 5% of the nu m ber of sequ ences in the bin. In 13 cases, the nu m ber of w eather sequences in the bin is few er than 12. In these cases, all of the sequ ences in the bin are selected . Selection of sequ ences from a bin w here not all sequ ences are chosen is perform ed by a sequ ential Monte-Carlo p rocess.

The p robability associated w ith a w eather trial is calcu lated w ithin MACCS2 u sing the follow ing algorithm . First, the probability that a w eather trial falls into a p articu lar bin, PB, is p rop ortional to the num ber of trials that are assigned to that bin, w here N B is the num ber of w eather trials in bin B and N is the total nu m ber of w eather trials for the year (8760 in a 365-d ay year). The p robability for a w eather trial from bin B is then expressed as w here PT is the probability associated w ith w eather trial T and N SB is the nu m ber of w eather trials sam pled from bin B (given in the last colu m n of Table 1). Thu s, the sum of the p robabilities of the w eather trials selected from bin B is PB. Valu es for the p robabilities for each w eather trial w ere d eterm ined from the MACCS2 ou tp u t and w ere u sed in the averaging p rocess for all the resu lts p resented in this rep ort.

The stand ard practice of allow ing w ind rotation w as used for the MACCS2 results, w hich essentially expand s the num ber of w eather trials by a factor of 16. This p ractice w as not ad opted by the other cod es. For each w eather trial, a set of calcu lations is p erform ed to account for the fact that the w ind could have been blow ing in any of the 16 com p ass d irections. Each of the 16 resu lts for w ind rotation is w eighted by the 6

Table 1. Description of Weather Bins Used in MACCS2 Bin Stability Wind Rain N umber of N umber of N o. Class Speed Weather Weather Range (m/s) D istance Intensity Sequences Sequences (km) (mm/hr) in Bin Selected 1 A/ B 0-3 < 32 0 312 16 2 A/ B >3 < 32 0 194 12 3 C/ D 0-1 < 32 0 13 12 4 C/ D 1-2 < 32 0 100 12 5 C/ D 2-3 < 32 0 361 18 6 C/ D 3-5 < 32 0 1077 54 7 C/ D 5-7 < 32 0 2202 110 8 C/ D >7 < 32 0 2370 119 9 E 0-1 < 32 0 6 6 10 E 1-2 < 32 0 69 12 11 E 2-3 < 32 0 177 12 12 E >3 < 32 0 998 50 13 F 0-1 < 32 0 29 12 14 F 1-2 < 32 0 67 12 15 F 2-3 < 32 0 52 12 16 F >3 < 32 0 3 3 17 all all 0-3 0-2 331 17 18 all all 3-6 0-2 8 8 19 all all 6 - 11 0-2 31 12 20 all all 11 - 21 0-2 108 12 21 all all 21 - 32 0-2 118 12 22 all all 0-3 2-4 39 12 23 all all 3-6 2-4 1 1 24 all all 6 - 11 2-4 1 1 25 all all 11 - 21 2-4 5 5 26 all all 21 - 32 2-4 9 9 27 all all 0-3 4-6 27 12 28 all all 3-6 4-6 0 0 29 all all 6 - 11 4-6 3 3 30 all all 11 - 21 4-6 5 5 31 all all 21 - 32 4-6 7 7 32 all all 0-3 >6 27 12 33 all all 3-6 >6 0 0 34 all all 6 - 11 >6 1 1 35 all all 11 - 21 >6 4 4 36 all all 21 - 32 >6 5 5 Total 8760 610 7

p robability of the w ind blow ing in the sp ecified d irection. The probabilities associated w ith the p ossible w ind d irections are constru cted for each w eather bin and are p rop ortional to the num ber of trials in the bin in w hich the w ind blow s in the sp ecified d irection. This p robability is given by w here PBR is the probability of a sam p le in bin B having w ind d irection R and N BR is the num ber of w eather trials in bin B w ith w ind d irection R. The final p robability for w eather trial T w ith w ind rotation R u sed in the MACCS2 cod e is sim p ly the p rod u ct of the tw o p robabilities, as follow s:

w here PTR is the probability of w eather trial T w ith w ind d irection R.

MACCS2 uses single-point w eather d ata. Thu s, it ap p roxim ates w eather d ata as spatially u niform . The w eather d ata file contains the follow ing inform ation: Ju lian d ay of the year, hour of the d ay, w ind d irection, stability class, and p recip itation rate. It also contains seasonal m ixing heights (d iscu ssed in su bsection 5.2). While MACCS2 d oes not m od el sp atial variation in w ind cond itions, it d oes m od el tim e d epend ence. Once a p lu m e is form ed , its d irection is not allow ed to change; how ever, the w ind sp eed ,

stability class, and precipitation rate can change hou r-by-hou r.

3.1.2 Atmospheric Transport and D ispersion The p lu m e is assum ed to m ove d ow nw ind at the p rescribed w ind sp eed ad ju sted for p lum e centerline elevation. The plum e broad ens by d isp ersion d u e to atm osp heric tu rbu lence as it is transported d ow nw ind . MACCS2 allow s d isp ersion to be treated either by m eans of a lookup table or as a p ow er-law fu nction of d istance. For this w ork, the stand ard Tad m or and Gur looku p tables (Tad m or and Gu r 1969, Dobbins 1979) w ere u sed to d eterm ine cross-w ind and vertical d isp ersion as a fu nction of d ow nw ind d istance and stability class.

Vertical d ispersion is assum ed to occu r only w ithin the mixing layer. MACCS2 u ses four m ixing heights to represent the fou r seasons of the year. These m ixing heights rep resent seasonal averages of the d aily m axim u m valu es of the m ixing heights.

Calculation of the m ixing heights u sed in this stu d y is d iscussed in section 5. The MACCS2 Gaussian plum e m od el treats the grou nd su rface and a su rface at the m ixing height as p lanes of reflective sym m etry.

8

3.1.3 D eposition Dry d ep osition is treated in MACCS2 by m eans of a d ep osition velocity. Aerosols can be d istributed am ong 10 aerosol bins, each w ith its ow n d ep osition velocity. For this stu d y, a single aerosol bin w as used and the d ep osition velocity w as chosen to be 0.01 m / s. The MACCS2 m od el calculates d ep osition rate as the p rod u ct of d ep osition velocity and aerosol concentration in the air at grou nd level.

The m od el for w et d eposition (w ashou t) in MACCS2 accounts for the effect of rain intensity. The m od el has the follow ing form (Brenk and Vogt 1981):

w here 7 is the fraction of aerosol that rem ains in the atm osp here (d im ensionless), C1 is the linear w ashout coefficient (s-1), )t is the d u ration of rainfall (s), I is the rain intensity (m m/ hr), and C2 is the nonlinear w ashou t coefficient (d im ensionless). The valu es of C1 and C2 u sed in this stud y are 7A10-5 and 0.75, resp ectively.

3.2 RASCAL and RATCHET RASCAL (Sjoreen, et al. 2001) is an N RC rad iological assessm ent tool for u se in em ergency response applications. It consists of mod u les that estim ate accid ent sou rce term s for nu clear p ow er p lants and other nu clear fu el cycle facilities; transp ort, d isp ersion, and d eposition of rad ionu clid es; and d oses. It also inclu d es a m eteorological p rep rocessor that p repares m eteorological d ata for u se by the atm osp heric transp ort m od u les. For this stud y, the m eteorological p rep rocessor m od u le and one of the atm osp heric transport m od ules w ere u sed to estimate time-integrated air concentrations of d epositing and non-d ep ositing sp ecies and su rface d ep osition of the d ep ositing sp ecies for com p arison w ith estim ates m ad e by MACCS2 and ADAPT/ LODI. These tw o m od ules are referred to as RASCAL, althou gh they w ere ru n ou tsid e the usual RASCAL fram ew ork to efficiently accom p lish the requ ired calcu lations. Minor m od ifications w ere mad e in the cod es to accu m u late tim e-integrated concentrations for the full p eriod of p lu m e passage for each release rather than for 15-m in intervals and to ru n from the tim e of release u ntil the p lu m e left the m od el d om ain rather than for a specified p eriod of time. These cod ing changes d id not alter the atm ospheric transport, d ispersion, or d ep osition calcu lations.

The atm ospheric m od els in RASCAL are betw een MACCS2 and ADAPT/ LODI in com p lexity. They includ e a m ore com p lete rep resentation of tem p oral and sp atial changes in m eteorological cond itions than MACCS2, bu t they d o not includ e the fu ll three-d im ensional representation of the atm osp here inclu d ed in ADAPT/ LODI.

9

RASCAL is a Lagrangian trajectory Gau ssian-p u ff d isp ersion m od el d erived from the MESORAD m od el (Sherpelz, et al. 1986, Ram sd ell, et al. 1988). Develop m ent of RASCAL emphasized ad d ition of rad ioactive d ecay and d ose calcu lations. Another cod e, RATCH ET, (Ramsd ell, et al. 1994) w as d evelop ed from MESORAD for u se in the H anford Environm ental Dose Reconstru ction Project (Ship ler, et al. 1996). The d evelop m ent of RATCH ET em phasized u p grad ing m ethod s for calcu lating d isp ersion and d ep osition. The N RC is consid ering u sing RATCH ETs d isp ersion and d ep osition m ethod ology in RASCALs atm osp heric m od els. In this stu d y, RATCH ET refers to a d evelop m ental version of the RASCAL atm osp heric m od el that incorp orates d isp ersion and d eposition algorithm s from the original version of RATCH ET. The d evelop m ental m od el includ es the m ethod s of treating atm osp heric transp ort that are inclu d ed in the existing version of RASCAL.

3.2.1 Meteorological Representation ARM SGP m eteorological d ata for the m od el com p arison w ere obtained from Law rence Liverm ore N ational Laboratory, LLN L. The follow ing p aragrap hs briefly d escribe the p rocessing of the m eteorological d ata for u se by RASCAL and RATCH ET, w hich both use the sam e m eteorological input. A sep arate m eteorological file containing hou rly d ata w as received for each of six m eteorological stations. These files w ere com bined into a single file containing record s w ith d ata for all stations for each hou r. In creating this file, the original d ata w ere cop ied as received , excep t for p recip itation d ata. The original files contained temperature and p recip itation rates, this inform ation w as u sed to estim ate precipitation form and intensity, for exam p le light rain or m od erate snow ,

w hich is the input need ed for RASCAL and RATCH ET.

The m eteorological preprocessor program converted the hou rly d ata from the m eteorological stations into the spatially and tem p orally varying m eteorological field s u sed by RASCAL and RATCH ET (Sjoreen, et al. 2001). The w ind field is d erived by interp olation of available surface w ind d ata. Field s for atm osp heric stability and p recip itation type and rate are generated u sing rep orted valu es for the closest observation to each nod e. Finally, the m ixing height field is generated in tw o step s. In the first step, an initial field is prep ared u sing valu es calculated from su rface observations at the available stations. Then a low p ass, sp atial filter is u sed to sm ooth the initial field .

Su rface w ind d ata, initially record ed as w ind d irection and sp eed are converted to Cartesian com ponents of the transp ort vector. A w ind field is generated for each of these com p onents using 1/ r 2 w eighted interp olation, w here r is the d istance betw een the nod e on the Cartesian grid and the m eteorological observation location. Wind s at the release height, w hich are used for transp ort calcu lations, are estim ated from the su rface layer w ind s using bound ary layer w ind sp eed p rofiles that are a fu nction of 10

atm osp heric stability and surface rou ghness. A u niform su rface rou ghness of 0.2 m w as used for this stud y. The RASCAL m eteorological processor has a relatively simp le routine to ad just w ind field for the effects of top ograp hy in stable atm osp heric cond itions; how ever, that routine w as not u sed in this stu d y.

There are tw o atm ospheric stability field s. One consists of Pasqu ill-Gifford stability classes (Pasquill 1961, Gifford 1961, Tu rner 1964), and the other consists of the inverse Monin-Obukhov length (Monin and Obu khov 1954). The Monin-Obu khov length is estimated from the Pasquill-Gifford stability class u sing a grap hical relationship betw een Monin-Obukhov length, stability class, and su rface rou ghness d erived by Gold er (1972).

Mixing height is calculated for each m eteorological observation u sing relationships d erived by Zilitinkevich (1972). Calcu lated m ixing heights, w hich are a fu nction of w ind sp eed , stability, surface roughness, and latitu d e, are constrained to a m inim u m of 50 m and a m axim um of 2000 m . When m ixing heights have been calcu lated for all of the observations, a m ixing height field is generated as ind icated above.

RASCAL and RATCH ET accept three p recipitation cond itions - no p recipitation, rain, and snow . Every hour, the p recipitation grid is up d ated u sing the hou rly observations.

Each nod e on the grid is assigned the p recip itation, or lack thereof, from the closest m eteorological station. If the closest station d oesn't have a current observation, the p recip itation from the closest station w ith a cu rrent observation is u sed . If there are no stations w ith current observations, p ersistence is assu m ed and the last p recip itation grid is u sed . N o interpolation or sm oothing is p erform ed on the p recip itation d ata.

Rain, w hich includ es any form of liqu id p recip itation, m ay have an intensity of light, Table 2. RASCAL and RATCH ET m od erate, or heavy. Sim ilarly the intensity Precip itation Rates (mm hr -1 )

of snow , w hich includ es all form s of frozen p recip itation, m ay be light, m od erate, or Intensity Rain Snow heavy. Precipitation intensity d eterm ines the p recip itation rate as listed in Table 2. The Light 0.6 0.3 rates for m od erate and heavy precip itation Mod erate 3.8 1.7 are set to the same valu e, bu t cou ld be m ad e d ifferent. In m ost parts of the country, heavy H eavy 3.8 1.7 p recip itation is sufficiently infrequent that setting precipitation rates equal has little effect on the clim atological d ispersion estim ates.

11

3.2.2 Atmospheric Transport RASCAL and RATCH ET use the sam e m ethod for calcu lation of atmosp heric transp ort.

In both cod es, the plum e is represented by a series of p u ffs released at 5-m inu te intervals. Each puff contains the activity released d u ring a 5-m inu te p eriod . The height of release is the sum of the actu al release height and final p lum e rise. A m od ification, m ad e to RASCALs plum e rise calcu lation for u se in this m od el com p arison, w as to ad d the op tion of calculating plum e rise from the heat of release based on equ ations of Briggs (1984). Puff m ovem ent is controlled by the w ind at the release height w ith the m ovem ent vector upd ated every 5 m inu tes u sing w ind s interp olated to the p u ffs current p osition from the w ind field s (Sjoreen, et al. 2001).

3.2.3 Atmospheric D ispersion RASCALs atm ospheric d ispersion calculations u se m ethod s d evelop ed in the 1950s and 1960s (Sjoreen, et al. 2001). The d isp ersion p aram eters are a function of d istance traveled and atm ospheric stability u sing nu m erical ap p roxim ations to the Pasqu ill-Gifford d ispersion curves sim ilar to those in nu m erou s N RC com p u ter cod es, e.g.,

PAVAN (Band er 1982) and XOQDOQ (Sagend orf 1982).

Disp ersion in RATCH ET represents ad vances in the science d u ring the 1970s and 1980s (Ram sd ell, et al. 1994), and the techniqu e u sed is ou tlined here. RATCH ETs d isp ersion p aram eters are calculated from travel tim e and m easu res of the atm osp heric tu rbu lence.

Du ring the first hour follow ing release, the horizontal d isp ersion p aram eter is p rop ortional to the prod uct of a m easu re of the horizontal com p onent of tu rbu lence in the w ind and time since release. After the first hou r, the rate of increase in the horizontal d ispersion p aram eter is a fu nction only of travel tim e u ntil an u p p er lim it of 105 m is reached .

The vertical d ispersion param eter is calculated as the p rod u ct of a m easu re of the vertical com ponent of turbulence and a fu nction that accou nts for d ecreasing effectiveness of turbulence in d ispersing the p u ffs at long travel time. For neu tral and u nstable atm ospheric cond itions (Pasqu ill-Gifford stability classes A throu gh D), the fu nction is equal to 1.0. For stable cond itions (Pasqu ill-Gifford stability classes E throu gh G), the function d ecreases the rate of grow th of the vertical d isp ersion p aram eter from being p roportional to tim e to the first pow er near the release p oint to being p roportional to the square root of travel tim e after the first few m inu tes.

If m aterial is released w ithin the m ixing layer, the vertical d isp ersion coefficient is limited by the top of the m ixing layer. The coefficient w ill resu m e grow th if the m ixing height increases. H ow ever, the vertical d ispersion p aram eter is not allow ed to d ecrease if the m ixing height d ecreases.

12

Calcu lation of the d ispersion param eters requ ires estim ates of tu rbu lence p aram eters.

These param eters are estim ated as they are requ ired u sing the available m eteorological d ata and atm ospheric bound ary layer relationships (H anna, et al. 1982; Panofsky, et al.

1977). In no case is either of the tu rbu lence p aram eters p erm itted to d ecrease below 0.01 m / s.

3.2.4 D ry D eposition In RASCAL, d eposition is calculated Table 3. RASCAL Washou t Coefficients (s-1).

u sing a sou rce d ep letion m od el w ith a constant d ry d eposition velocity of Intensity Rain Snow 0.01 m / s and w et d eposition is calcu lated using a sim ple w ashout Light 2.2 x 10-4 1.0 x 10-4 m od el w ith constant washout Mod erate 6.1 x 10-4 3.3 x 10-4 coefficients. These method s are d escribed in d etail in Sjoreen, et al. H eavy 1.1 x 10-3 6.4 x 10-4 (2001). The w ashout coefficients used in RASCAL are listed in Table 3.

The d ep osition calculations in RATCH ET are d escribed in Ram sd ell, et al. (1994). Dry d ep osition is still calculated using a d ep osition velocity; how ever, the d ry d ep osition velocity is no longer constant. It is a function of characteristics of the m aterial, su rface roughness, and atm ospheric cond itions. Consequ ently, the d ry d ep osition velocity is a fu nction of both position and tim e. The m od el u sed to calcu lated d ry d ep osition velocity in RATCH ET is based on an analogy w ith electrical system s (Seinfeld 1986).

The analogy assum es that d eposition velocity is inversely p rop ortional to the su m of three resistance com ponents - an aerod ynam ic resistance, a su rface-layer resistance, and a transfer resistance. The aerod ynam ic and su rface-layer resistances are fu nctions of w ind sp eed , stability and surface rou ghness. The transfer resistance is a function of the characteristics of the d epositing m aterial and the su rface typ e. In RATCH ET, transfer resistance is used as a m eans of placing a low er lim it on the total resistance, or, in other term s, placing an upper lim it on d ry d eposition velocities. Assu ming transfer resistances of 10 s/ m for reactive gases and 100 s/ m for fine p articles (. 10 m icrons) yield d ry d eposition velocities that are consistent w ith exp erim entally d eterm ined d ep osition velocities.

3.2.5 Wet D eposition There are tw o w et d ep osition param eterizations in RATCH ET, one for p articles and another for gases. The gas scavenging p aram eterization u ses a sou rce-d epletion m od el sim ilar to the m od el used for d ry d ep osition. The w et d ep osition velocity for scavenging gases by rain is a function of a solu bility coefficient, w hich is related to the 13

H enrys Law constant for the gas, and p recip itation rate to the 3/4 p ow er (Slinn 1984).

The w et d eposition velocity for scavenging of non-reactive gases (e.g., CH 3I) by rain is abou t three ord ers of m agnitu d e low er than for reactive gases (e.g., I2). Snow is not a p articu larly good scavenger of gases. For tem p eratu res above -3°C, w et d ep osition velocities for snow are estim ated using the w ater equ ivalent p recip itation rate. For tem peratures below -3°C, the snow su rface is frozen and the w et d ep osition of gases is low ; therefore, RATCH ET sets the w et d eposition velocity to zero.

Particles are collected by Table 4. RATCH ET Wet Dep osition Mod el Param eters.

precipitation as it falls through the p u ffs. We Reactive Gas Particle assu m e that rain or snow Wet Dep osition Washou t falls through the entire Velocity (m / s) Coefficient (1/ s) vertical extent of the puff, Light Rain 1.7x10-4 2.7x10-4 and w e calculate w et d ep osition of particles w ith Mod erate Rain 1.1x10-3 1.1x10-3 a w ashou t m od el w here the w ashou t coefficient is a Light Snow 8.4x10-5 1.5x10-5 fu nction of precipitation Mod erate Snow 4.8x10-4 8.5x10-5 type and rate as given in Table 4.

3.3 N ARAC Models - AD APT and LOD I A key com ponent of this stu d y is u tilization of a state-of-the-art atm osp heric transp ort and d ispersion m od el, d riven by observed m eteorological d ata, to p rovid e the exp osure (concentration of a m aterial in near-su rface air integrated over p lu m e p assage) and d eposition (total am ount of a m aterial d eposited on the grou nd d u ring p lu m e passage) for each of the sam ple cases. The w eighted average of these accu rate ind ivid u al case results then rep resents a stand ard for ju d ging the app rop riateness of sim p ler m od els.

The N ational Atmospheric Release Ad visory Center (N ARAC) at Law rence Liverm ore N ational Laboratory (LLN L) is a national su p p ort and resou rce center for p lanning, real-time assessm ent, em ergency resp onse, and d etailed stu d ies of incid ents involving the sp read of hazard ous m aterial accid entally or intentionally released into the atm osp here. Within its em ergency resp onse system , N ARAC p rovid es a su ite of m ulti-scale (local-, regional-, continental- and global-scale) atm osp heric flow and d isp ersion m od els. The num erical m ethod s used by these m od els have been verified u sing exact m athem atical solutions, and m od el resu lts have been evalu ated u sing field exp erim ents (Foster, et al. 2000). These evaluations p rovid e confid ence in the accuracy of the N ARAC m od els.

14

The N ARAC em ergency response m od eling system consists of a cou p led su ite of m eteorological and d ispersion m od els. The d ata assim ilation m od el, ADAPT, constru cts field s of m ean w ind s, pressure, precipitation, tem p eratu re, and tu rbu lence, u sing a variety of interpolation m ethod s and atm osp heric p aram eterizations. N on-d ivergent w ind field s are prod uced by an ad ju stm ent p roced u re based on the variational principle and a finite-elem ent d iscretization. The d isp ersion m od el, LODI, solves the 3-D ad vection-d ispersion equation using a Lagrangian stochastic, Monte Carlo m ethod .

LODI includ es m ethod s for sim ulating the processes of m ean w ind ad vection, tu rbu lent d iffusion, rad ioactive d ecay and prod u ction, bio-agent d egrad ation, first-ord er chem ical reactions, w et d eposition, gravitational settling, d ry d ep osition, and bu oyant/ m om entum plum e rise. The mod els are cou p led to N ARAC d atabases provid ing top ograp hy, geograp hical d ata, real-tim e m eteorological observational d ata, and global and m esoscale forecast m od el p red ictions. In this stu d y w e u se ADAPT to convert the observed m eteorological d ata into 3-D grid d ed field s of w ind and tu rbu lence param eters and LODI to calcu late the release, transp ort, d isp ersion, and w et and d ry d eposition of representative p ollu tants and the corresp ond ing exp osu res at the su rface.

The N ARAC m od els ADAPT and LODI have great flexibility and nu m erou s op tions that w ere not necessary for this stu d y. The follow ing d iscu ssion focu ses p rim arily on the mod els as used in this stud y. For a fu ll d iscu ssion of ADAPT/ LODI cap abilities see the references.

3.3.1 AD APT ADAPT is an atm osp heric d ata assim ilation m od el that bu ild s three-d im ensional grid d ed m eteorological field s (Sugiyam a and Chan 1998). It p rovid es a selection of ap p roaches to p rocess input m eteorological d ata p rovid ed in this stu d y by observations at a set of stations. The m od el incorp orates a nu m ber of interp olation and extrap olation techniques, includ ing both d irect and iterative solvers, and atm osp heric p aram eterizations. We used ADAPTs capabilities to coherently blend the su rface d ata w ith the upper air sound ings on a uniform ly sp aced , four km resolu tion, horizontal grid , and to m ass-consistently calcu late the three d im ensional w ind field s and tu rbu lence param eters w hich then w ere u sed to d rive N ARAC's d isp ersion m od el, LODI .

ADAPT p rod uces non-d ivergent (m ass-consistent) w ind s by m inim al ad ju stm ent of inp u t field s d erived from observational d ata. This ad ju stm ent p rovid es not only the horizontal w ind s but the vertical com p onent as w ell. The algorithm is based on a variational form ulation and a finite-elem ent sp atial d iscretization, w hich u ses a grid -

p oint rep resentation of the w ind field s (in contrast to the flux-based staggered grid 15

rep resentation often used in finite-d ifference ap p roaches) and p rovid es a rigorou s, flexible treatm ent of bound ary cond itions. Tw o iterative solvers, the incom p lete Cholesky conjugate grad ient and the d iagonally scaled conju gate grad ient, p rovid e an efficient num erical solution of the Poisson equ ation d erived from the variational p rincip le.

The ou tp u t m eteorological field s are highly d ep end ent on the d ensity and d istribu tion of m easu rem ents, the com plexity of terrain, and the p rop er p aram eterization of atm osp heric structu re to represent p hysical p rocesses not d irectly m od eled . ADAPT d oes not ad d unverified structure to the outpu t. The m ass-consistent w ind algorithm m inim ally ad ju sts the w ind s to ad d stability d ep end ent steerage around top ograp hical featu res; structures su ch as re-circulations not su p p lied by the initial observations are u su ally not generated by such a proced u re.

ADAPT's d ata assim ilation proced u res ingest and blend d ata from a variety of sou rces.

There are tw o broad classes of m eteorological d ata - observational d ata and grid d ed field s. The form er are m easurem ents, forecast sou nd ings, or u ser generated p seu d o-observations for one to m any vertical levels at a single station location (latitu d e and longitu d e location) and tim e. Grid d ed field s are analyses or forecasts either acqu ired from external sources or generated by other m od els. In this stu d y w e u sed only observational d ata.

ADAPT d ivid es observational d ata into three categories - su rface, tow er, and u p p er air.

Su rface d ata consist of m easurements at a single near-grou nd height. Tow er d ata contain m easurem ents at a single elevated height or at m u ltip le levels, the low est of w hich is at or near the surface. Up per air sound ings p rovid e m u lti-level d ata w ith the low est levels in the planetary bound ary layer (that p ortion of the atm osp here from the su rface to the geostrophic w ind level). H ere w e categorized all d ata as either su rface or u p p er air.

ADAPT output files provid e the three-d im ensional w ind com p onents, u (zonal or east-w est w ind ), v (merid ional or north-sou th w ind ), and w (vertical w ind ), and tu rbu lence p aram eters, all of w hich vary in the three spatial d im ensions ! east-w est, x; north-sou th, y; and height, z. The files also p rovid e planetary bou nd ary layer height, Monin-Obukov length, and surface friction velocity, all of w hich w ere assu m ed sp atially u niform in this stud y.

3.3.2 LOD I LODI is an atm ospheric d ispersion m od el that u ses a Lagrangian stochastic, Monte Carlo method w hich calculates possible trajectories of flu id p articles in a tu rbu lent flow to solve the 3-D ad vection d isp ersion equ ation (N asstrom et al. 2000, Erm ak and 16

N asstrom 2000). Particles are m arked at the sou rce w ith an ap p rop riate am ou nt of contam inant m ass based upon prescribed emission rates and geom etry. A large nu m ber of ind ep end ent trajectories are calcu lated by m oving p articles in resp onse to the variou s p rocesses represented w ithin the simu lation, and the m ean contam inant air concentration is estimated from the sp atial d istribu tion of the p articles at a p articu lar tim e. LODI uses a coord inate system w ith a continu ous terrain rep resentation at the low er bound ary.

In general, the tw o m ost im p ortant processes are ad vection by the m ean w ind and d isp ersion by turbulent m otion. To calcu late mean w ind ad vection, 3-D grid d ed m ean w ind field s from ADAPT w ere inpu t into LODI. Tu rbu lent d isp ersion w as m od eled via rand om d iffusive m ovem ents using atm osp heric ed d y d iffu sivity (K) p aram eterizations. Wet and d ry d eposition w ere also sim u lated .

In this stud y the sou rce w as m od eled as a continu ou s p oint release w ith bu oyancy and m om entu m , and the integrated exposu re and com bined w et and d ry d eposition w ere outpu t.

3.3.2.1 D ry D eposition Dry d ep osition is param eterized in term s of a d ep osition velocity at a reference height (1-1.5 m in this stud y over land ). The d ry d eposition flu x onto the su rface is then given by the p rod uct of this d eposition velocity and the concentration at the reference height.

The d ep osition velocity is com posed of tw o ind ep end ent velocities: the non-settling d ep osition velocity, and the gravitational settling velocity. The non-settling d ep osition velocity is usually calculated in term s of a resistance m od el. H ow ever, for this stu d y it w as set to 0.01 m / s to m atch the assu m p tion in MACCS2; the settling velocity w as set to 0.

3.3.2.2 Wet D eposition To m inim ize d ifferences in the m od els by parameterizations other than transp ort and d isp ersion, w et d ep osition w as calcu lated u sing a scavenging rate w ith the same coefficients as MACCS2. The rate of d ep letion of mass, m, from a LODI p article occurs at a rate w here t is tim e and 7 is the scavenging coefficient w hich is given by 17

w here p is the rain rate in m m / hr and the coefficients a and b are set to 7@10-5 s -1 and 0.75, respectively, the sam e values used by MACCS2 in this stu d y.

18

4. METEOROLOGICAL D ATA The u sefu lness of the com parison of MACCS2 or RASCAL/ RATCH ET w ith ADAPT/ LODI is only as good as the ADAPT/ LODI m eteorology for the chosen site; therefore, the preparation of the m eteorology files w ill be d iscussed in d etail. The location of this stu d y is the DOE ARM Sou thern Great Plains site in Oklahom a and Kansas. This site provid es an extensive regional d ata set, w ith d ata from the ARM Program as w ell as the Oklahom a Mesonet and N ational Oceanic and Atm osp heric Ad m inistration (N OAA) Profiler N etw ork (N PN ), covering a p eriod of over 10 years.

We selected the year 2000 for this p roject becau se w e knew from p reviou s exp erience that this year had a com p lete d ata set w ith no extend ed p eriod s of m issing d ata.

4.1 ARM Site The ARM Program is a m ulti-laboratory, interagency p rogram created in 1989 w ith fu nd ing from the U.S. Departm ent of Energy as p art of its effort to resolve scientific uncertainties about global climate change w ith a sp ecific focu s on imp roving the p erform ance of general circulation m od els u sed for clim ate research and p red iction. The ARM Program established and operates field research sites in several climatically significant locations w here scientists collect and analyze d ata obtained over extend ed p eriod s of tim e from large arrays of instru m ents to stu d y the effects and interactions of su nlight, rad iant energy, and cloud s on temp eratu res, w eather, and clim ate.

The U.S. Southern Great Plains (SGP) site w as the first field m easu rem ent site established by the ARM Program . It w as chosen for several reasons inclu d ing its relatively hom ogenous geography and easy accessibility, w id e variability of climate clou d type and su rface flu x p roperties, and large seasonal variation in tem p erature and sp ecific hum id ity. The SGP site consists of in situ and rem ote-sensing instru m ent clu sters arrayed across approxim ately 143,000 squ are kilom eters (55,000 squ are m iles) in north-central Oklahom a and sou th-central Kansas. The heart of the SGP site is the heavily instrum ented Central Facility located on 0.647 km 2 (160 acres) of cattle p astu re and w heat field s southeast of Lam ont, Oklahoma. The instru m ents at the Central Facility and throughout the site autom atically collect d ata on su rface and atm osp heric p rop erties, routinely provid ing d ata to the ARM Archive and Data Center. The Data Center acquires ad d itional d ata from other sou rces, su ch as N ational Weather Service satellites and surface d ata.

With ou r focus on atm ospheric transp ort and d isp ersion, w e selected d ata from the ARM archive that provid e vertical profiles of w ind and tem p eratu re, and su rface w ind ,

temp eratu re, precipitation, and stability. The selected d ata w ere then tailored to m eet the specific need s of each m od el. MACCS2 requ ired only hou rly su rface w ind , stability, and p recip itation d ata at the Central Facility. RASCAL/ RATCH ET used hou rly su rface 19

w ind , stability, precipitation, and tem p erature d ata from the Central Facility and the five other ARM surface sites closest to the Central Facility. ADAPT/ LODI used all the hou rly su rface w ind , tem peratu re and p recip itation d ata and all the vertical p rofiles of w ind and tem perature d escribed later to constru ct three-d im ensional flow field s and tu rbu lent d iffusion coefficients.

4.2 Surface D ata 4.2.1 ARM SMOS Sites The Surface Meteorological Observation System (SMOS) m ostly u ses conventional in situ sensors to obtain 1-m inute and 30-m inu te averages of su rface w ind sp eed , w ind d irection, air tem peratu re, relative hu m id ity, barom etric p ressu re, and p recip itation at the central facility and m any of the extend ed facilities. Detailed inform ation on these observations is available on the ARM Web site (http :/ / w w w .arm .gov). The locations of these sites are show n by the xs in Figu re 1. The w ind s are m easu red at 10 m and the rest of the param eters at 2 m . From the archived ARM N etw ork Comm on Data Form at (netCDF) d ata files w e extracted the 30-m inu te d ata and p rod u ced hou rly average valu es for w ind speed and d irection, temp eratu re, p ressu re, vapor p ressu re, p recip itation, and stand ard d eviation of the w ind d irection at each of the 15 ARM sites.

This d ata set provid ed all the d ata need ed by MACCS2 and RASCAL/ RATCH ET.

4.2.2 Oklahoma Mesonet The Oklahom a Mesonet is network of environm ental m onitoring stations d esigned and im p lem ented by scientists at the University of Oklahoma and at Oklahoma State University. In 2000 it consisted of 114 au tom ated stations covering the state of Oklahom a. At each site, the environm ent is m easu red by a set of instru m ents located on or near a 10 m tower. Detailed inform ation on these observations is available on the Oklahom a Mesonet w eb site (http:/ / w w w .m esonet.org). The ARM d ata archive netCDF files contain five-minute averages of the observations for all these sites. We extracted the w ind speed and d irection, tem p eratu re, p ressu re, p recipitation, solar rad iation, and stand ard d eviation of the w ind d irection at each of the 83 Oklahom a Mesonet sites w ithin the SGP bound aries as show n by the +s in Figu re 1. From the five-m inu te average d ata w e prod uced hourly average observations, and these d ata plu s the 15 ARM SMOS sites p rovid ed the su rface d ata inp u t for ADAPT.

4.3 Upper Air D ata At the ARM SGP site there are several d ifferent m easu rem ent p latform s that p rovid e p rofiles of above-surface w ind s. These d ata sets inclu d e m easu rem ents from balloon sond es at five sites, 915 MH z rad ar w ind p rofiler (RWP) and Rad io Acou stic Sou nd ing 20

Figure 1. Location of Surface Meteorological Sites. + ind icates Oklahom a Mesonet sites,

the Central Facility, x+ the ARM SMOS sites u sed by RASCAL/ RATCH ET, and x the other ARM SMOS sites.

System (RASS) at four sites, and the N OAA Profiler N etw ork (a 404 MH z RWP/ RASS system ) at seven sites in the ARM Sou thern Great Plains area. In ad d ition, Atm osp heric Em itted Rad iance Interferom eter (AERI) vertical tem p eratu re p rofiles at the sam e five sites as the sond es w ere used to help d eterm ine bou nd ary layer heights. The locations of these sites are given in Figure 2. Detailed inform ation on these observations is available on the ARM w eb site (http :/ / w w w .arm.gov).

4.3.1 Sondes Balloon sond es are launched throu ghou t the year at the Central Facility and d u ring Intensive Operational Period s (IOP) at the other fou r sites. The rou tine sond es at the Central Facility are launched every 6 hou rs Mond ay throu gh Frid ay. Du ring IOPs 21

Figure 2. Location of ARM Upper Air Observations. + ind icates sond e and AERI sites, x ind icates N OAA Profiler sites, and Q 915 MH z RWP/ RASS sites. N ote that som e sites have m u ltiple m easurem ents, and the Central Facility has all three typ es.

sond es are lau nched every 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> at all 5 sites, the Central Facility and 4 bou nd ary sites. In 2000 there w ere three IOPs that p rovid ed ad d itional sond e d ata, March 1 to March 22, Septem ber 25 to October 8, and N ovem ber 27 to Decem ber 22. The total nu m ber of sond es at each site is given in Table 5. The sond e netCDF files in the ARM archive p rovid e latitud e, longitud e, altitu d e, p ressu re, tem p eratu re, and d ew p oint temp eratu re every tw o second s d uring ascent and w ind sp eed and d irection every ten second s. The ascent rate is typically abou t 5 m / s, so the vertical resolution of the w ind d ata is about 50 m . On N ovem ber 24 a change in p roced u re started returning w ind d ata every tw o second s w ith the resolu tion changing to abou t 10 m .

22

Table 5. N u m ber of Sond es at Each Site Site ID N u m ber Central Facility SDC1 1333 H illsboro, KS SDB1 439 Vici, OK SDB4 443 Morris, OK SDB5 458 Purcell, OK SDB6 460 4.3.2 Radar Wind Profiler/Radio Acoustic Sounding System The 915 MH z rad ar w ind profiler (RWP)/ rad io acou stic sou nd ing system (RASS) operates by transm itting electrom agnetic energy into the atm osp here and m easu ring the strength and frequency of back-scattered energy. It transm its in tw o d ifferent vertical p lanes and receives back-scattered energy from refractive ind ex flu ctu ations that are m oving w ith the m ean w ind . From the ARM netCDF files w e extracted hourly vertical p rofiles of horizontal w ind , and virtu al tem p eratu re. There are tw o vertical resolutions in the d ata (d epend ent on p ow er level); for low er heights the first level is typ ically 87 m above ground w ith a vertical sp acing of abou t 60 m u p to 2-2.5 km . For higher levels the low est level is abou t 225 m w ith a sp acing of 225 m u p to 6 km .

4.3.3 N OAA Profiler N etw ork The N OAA Profiler N etw ork (N PN ) consists of 35 u nm anned Dop p ler Rad ar sites located in 18 central states and Alaska; they p rovid e hou rly vertical w ind p rofile d ata.

Eleven of these sites also are equipp ed w ith RASS System s that p rovid e virtu al tem p eratu re profiles. Seven of the N PN sites are located in the ARM SGP region, and six of these are equipped w ith RASS systems. Wind N PN p rofilers are d esigned to op erate reliably and unattend ed in nearly all w eather cond itions. To reach the trop op au se, they use a relatively long w avelength. The rad ars d etect fluctu ations in atm osp heric d ensity caused by turbu lent m ixing of volum es of air w ith slightly d ifferent tem peratu re and m oistu re content. The resu lting flu ctu ations of the ind ex of refraction are used as a tracer of the m ean w ind in clear air. From the ARM d ata archive netCDF files w e extracted vertical p rofiles of horizontal w ind and virtu al temp eratu re for the seven sites. Because of the large w avelength u sed by the N PN rad ars, vertical resolution of these profiles is rather cou rse; the low est level is at 500 m w ith a sp acing above that level of 250 m . N evertheless, N PN d ata w ere consistently available and are a m ajor com ponent of the input d ata set u sed by ADAPT.

23

4.3.4 Atmospheric Emitted Radiance Interferometer The atm osp heric emitted rad iance interferometer (AERI) m easu res the absolu te infrared spectral rad iance (w atts per square m eter p er sterad ian p er w avenu m ber) of the sky d irectly above the instrum ent. A calibrated sky rad iance sp ectru m is p rod u ced every ten m inu tes. Am ong other things the AERI d ata can be u sed for calculating vertical atm osp heric p rofiles of tem p erature and w ater vapor. In this stud y w e u sed the vertical tem p eratu re p rofiles to calculate bou nd ary layer heights. The AERI instru m ents are located at the sam e sites w here sond es are lau nched , and these d ata w ere p articu larly u sefu l becau se they p rovid ed high resolu tion tem p erature p rofiles w hen no sond e d ata w ere available.

24

5. METEOROLOGICAL D ATA PROCESSIN G 5.1 Pasquill-Gifford (P-G) Stability Category Both MACCS2 and RASCAL/ RATCH ET requ ire the Pasqu ill-Gifford (P-G) stability category in their input files, but the ARM d ata archive d oes not inclu d e this p aram eter.

The ARM d ata archive d oes includ e the stand ard d eviation of su rface w ind d irection, F2 , for both the ARM and Oklahom a Mesonet surface sites. This p aram eter w as converted into P-G stability category u sing the F2 m ethod given in N RCs Regu latory Gu id e 1.23, "Onsite Meteorological Program s," and in chap ter 6.4.4 of EPA (2000),

available on the w eb at http:/ / w w w .w ebm et.com / m et_m onitoring/ 644.htm l. Briefly, an initial P-G stability category is set based on F2 (F2 $ 22.5, A; 17.5 # F2 < 22.5, B; 12.5 #

F2 < 17.5, C; 7.5 # F2 < 12.5, D; 3.8 # F2 < 7.5, E; F2 < 3.8, F). This initial estim ate is then m od ified based on w ind speed (higher sp eed s giving P-G stability closer to neu tral, category D) and d aytim e/ nighttim e w ith the P-G category in the u nstable to neu tral range (A-D) d uring the d ay and the neu tral to stable range (D-F) at night.

5.2 Mixing Height Mixing height is another d erived atm osp heric parameter that is not d irectly archived in the ARM d ata set. MACCS2 selects its m ixing height from an inp u t list of eight m ixing heights d ep end ing on season and tim e of d ay. The eight mixing heights are the average m orning and afternoon m ixing height for each season. In the cu rrent im p lem entation, the larger of these tw o values (the afternoon height) is u sed by the cod e.

RASCAL/ RATCH ET calculates m ixing height as a fu nction of w ind sp eed , stability, su rface roughness, and latitu d e (Zilitinkevich 1972) for each m eteorological station and for each hour. ADAPT requires m ixing height for each hou rly m eteorological d ata set.

Della Monche (2002) exam ined several techniqu es for d eterm ining mixing height u sing available ARM d ata. The technique that p erform ed best u nd er the w id est p ossible cond itions w as a technique by H effter (1980) based on p otential tem p eratu re p rofiles.

The technique w orks in every stability regim e, and in alm ost every case gave estim ates of the m ixing height that w ere at or above the m ixing height as observed from an aircraft. Based on these results w e ad op ted the Della Monche/ H effter m ethod to estim ate m ixing heights.

The H effter m ethod calculates potential tem p eratu re from the vertical p rofile of tem p eratu re and pressure available from the sond e and AERI d ata. Each p rofile is exam ined for the existence of a critical inversion w hich is assu m ed to m ark the top of the m ixed layer. A critical inversion is d efined as the low est inversion that meets the follow ing two criteria:

25

)2/ )z > 0.001 K/ m and 2t - 2b > 2 K w here )2/ )z is the p otential tem p eratu re lapse rate and 2t and 2b are the p otential tem p eratu re at the top and bottom of the critical inversion layer resp ectively; these criteria give the m ixing height as that p oint in the inversion w here the tem p erature is 2 K greater than the temp eratu re at the inversion base. Low er and u p p er lim its w ere p laced on m ixing height at 40 m and 3000 m respectively. In cases w here the proced ure failed to find a m ixing height, the m ixing height w as set to the clou d base height if there w as a clou d base, otherw ise the m ixing Figure 3. Cu m u lative N u m ber of H ours w ith height w as left und eterm ined Mixing H eights less than a Given H eight.

and that site w as not used . A single m ixing height for each hour w as com p u ted as the average of the calcu lated valid ind ivid u al m ixing heights. These hourly average m ixing heights w ere u sed in the ADAPT inp u t files and w ere also seasonally averaged to p rovid e d aytim e and nighttim e seasonal m ixing heights for MACCS2. The mixing height in ADAPT w as assum ed constant over the entire sp atial d omain. The cu m u lative nu m ber of hou rs w ith m ixing heights less than a given height is show n in Figu re 3.

5.3 Low -Level N octurnal Jet The ARM SGP site lies alm ost in the center of the U. S. region w ith m ost frequ ent low -level noctu rnal jets (Bonner 1968), and these cou ld be an im p ortant transp ort m echanism for released m aterial. Since only ADAPT/ LODI uses u p p er level w ind d ata, these jets could lead to transport d ifferences am ong the m od els. An exam p le of a low level nocturnal jet is given in Figure 4 w hich is a p lot of the low est 1500 m of the RWP w ind p rofile at the ARM Central Facility on the night of Ju ne 7, 2000. N ote the increase 26

Figure 4. Exam ple of a N octurnal Jet. The left p lot show s the vertical profile of w ind d irection and the right w ind speed from the RWP at the ARM Central Facility on the night of Ju ne 7, 2000. Ind ivid ual profiles at 1800, 2100, 0000, 0300, and 0600 local stand ard tim e show the increase in w ind sp eed in the layer betw een 200 and 800 m .

in w ind sp eed from a m axim um of abou t 10 m / s at 1800 local stand ard tim e to over 20 m / s betw een m id night and 0300 in the layer from 200 to 800 m .

5.4 MACCS2 Input Meteorology File The MACCS2 m eteorology input file requ ires 8760 (365 d ays x 24 hou rs) hou rly observations of w ind speed and d irection, stability category (A-F), and p recip itation at the release site, plus eight (four seasons x d ay/ night) m ixing heights. The hou rly observations w ere extracted from the ARM SMOS d ata at the Central Facility. Wind d irection in d egrees w as converted to the sector the w ind is blow ing tow ard , and stability category w as d eterm ined from the stand ard d eviation of the horizontal w ind d irection as d escribed above. The d istribu tion of hou rs in each stability category is 27

given in Table 6. When the Central Facility had Table 6. N u m ber of H ou rs m issing d ata (67 of 13560 half hours), observations in Each Stability Category from the closest neighboring ARM SMOS station for MACCS2 Site (E9 - Ashton, Kansas or E6 - Tow and a, Kansas) w ere used . The eight mixing heights w ere d eterm ined as Category Cou nt d escribed above and are given in Table 7. N ote, A 259 how ever, that MACCS2 round s the d ata to the nearest 100 m . B 290 C 549 The ARM archive stores d ata using Coord inated Universal Tim e (UTC), w hat used to be referred to D 6122 as Greenw ich Merid ian Tim e (GMT). Local Stand ard Tim e (LST) at the SGP site is UTC - 6 E 1358 hou rs. So the first d ata in the ARM archive files for F 182 the year 2000 are for 1800 LST on Decem ber 31, 1999. H ow ever, since 2000 is a leap year, the last six hou rs of the 8760 com e from Decem ber 31, 2000 d ata. The Table 7. Seasonal Average Mixing H eights (m) rem aining 18 hours2.083333e-4 days <br />0.005 hours <br />2.97619e-5 weeks <br />6.849e-6 months <br /> of Decem ber 31, Season \ Tim e Morning Afternoon 2000 d ata allow us to follow the transport and d ispersion of releases Winter 338. 788.

that occu r in the last few hours of the Sp ring 398. 1011.

year in the ADAPT/ LODI calcu lation. Su m m er 276. 1311.

Au tu m n 366. 874.

5.5 RASCAL/RATCHET Input Meteorology Files RASCAL/ RATCH ET m eteorology inp u t u ses observations from several su rface stations in the vicinity of the release. In this stu d y observations from the ARM SMOS sites at the Central Facility and the five other sites closest to the Central Facility (see Figu re 1) w ere u sed . H ou rly average w ind speed , w ind d irection, temp eratu re, and p recip itation w ere extracted from the ARM d ata archive. Atm osp heric stability w as calcu lated from the stand ard d eviation of the w ind d irection as d escribed above. The d istribu tion of stability categories for the six sites is given in Table 8, along w ith the total nu m ber of hou rs of m eteorological d ata. The m issing d ata p eriod s are those tim es w hen neither of the tw o half hours that contribute to an hou rly average w as available. This d ata set also inclu d ed d ata for the last 18 hours2.083333e-4 days <br />0.005 hours <br />2.97619e-5 weeks <br />6.849e-6 months <br /> of Decem ber 31, 2000; d ata that w as ignored for the MACCS2 d ata set.

Meteorological d ata for the Central Facility and the five other sites w ere com bined into a single file for use by the RASCAL m eteorological d ata p rocessor. All available d ata for 28

Table 8. N um ber of H ours in Each Stability Category for RASCAL Sites Station 6 Central EF-7 EF-9 EF-11 EF-15 EF-20 Stability 9 Facility Elk Falls Ashton Byron Ringw ood Meeker A 258 404 174 323 257 356 B 292 529 361 396 390 601 C 551 874 526 625 785 1023 D 6123 4145 5649 4855 5389 4399 E 1367 854 1365 1629 1060 1121 F 183 968 126 340 223 292 Total 8774 7774 8201 8168 8104 7792 Table 9. N u m ber of H ou rs vs each hou r w ere used to generate the N u m ber of Valid Su rface m eteorological d ata field s used by RASCAL and Observations.

RATCH ET. N o effort w as m ad e to rep lace m issing N u m ber of N u m ber of valu es. If all valu es for a given p aram eter are Observations H ou rs m issing, p ersistence is assum ed , and the p reviou s field is u sed . 89 3 90 0 5.6 AD APT/LOD I Input Files 91 0 ADAPT is d esigned to construct a three-92 1 d im ensional w ind field that agrees as closely as possible w ith all the available observations and is 93 2 m ass consistent. The choice of the ARM SGP site w as d riven prim arily by our d esire to have a 94 12 m eteorological d ata set that w ould d efine this 95 27 w ind field based com pletely on observations.

ADAPT exp ects observations to be p rovid ed in 96 217 tw o coord inated text files, an observ.m et file that gives the height, w ind speed , w ind d irection, 97 552 tem p eratu re, d ew point, and pressu re observations 98 1359 for each station, and a stnloc.m et file that gives the x, y, z g , latitud e and longitud e of the station. The x 99 2659 and y p osition is given in the Universal Transverse 100 3694 Mercator (UTM) coord inates used by the m eteorological grid , and z g is the station altitu d e. 101 258 29

In this stu d y the observ.m et file inclu d ed su rface observations from the ARM SMOS and Oklahom a Mesonet sites and u p p er air observations from RWP and N PN p rofilers and sond es.

Since w e have such a large set of m eteorological observations, w e p u t each hou rly average set of observations in a sep arate inp u t file. Most of the tim e nearly all the su rface stations provid e d ata; this is show n in Table 9. N ote that there w ere 101 stations (15 ARM and 86 Oklahom a Mesonet) only throu gh Janu ary 20 w hen one of the Oklahom a Mesonet stations w as closed .

ADAPT requires at least one u pper air sou nd ing; it w as the availability of a d ense netw ork of these profiles that led u s to u se the ARM SGP site for this stu d y. Both the N PN and ARM 915 MH z RWP profiles are fairly reliable and p rovid e valid d ata m u ch of the time. The num ber of hours w ith valid p rofiles (ou t of 8784 hou rs) for each site is given in Table 10. The N PN profiles have fairly cou rse vertical resolu tion, 250 m , and start w ell above the surface, 500 m . The RWP p rofiles have good vertical resolu tion and start near the surface, but they are prone to errors and m any of them have lim ited vertical extent. The sond e p rofiles start near the ground and u su ally extend to great heights, bu t they are only available p art of the time. H ou rly ADAPT inp u t m eteorological d ata files includ e vertical profiles at each N PN and 915 MH z RWP site that had valid d ata and sond e d ata at each site w here a sond e w as lau nched w ithin three hou rs of the ADAPT m et time; i.e., the sam e sond e d ata cou ld be includ ed in u p to six hou rly ADAPT observ.m et inp u t files. The u p p er air p art of these files cou ld and occasionally d id inclu d e profiles from all 16 sites (7 N PN , 4 RWP, and 5 sond e p rofiles);

in the w orst cases only tw o profiles w ere available. The nu m ber of hou rs (d ata sets) vs.

the nu m ber of profiles is given in Table 11.

30

Table 10. N um ber of H ours vs. Table 11. H ou rs w ith Valid Data for Profiler N u m ber of Upper Air Profiles and Sond e Sites N um ber of H ours w ith Site ID Valid Data Profiles H ours H illsboro, KS WX04 8274 2 7 H aviland , KS WX06 8500 3 23 N eod oska, KS WX07 6482 4 25 Lam ont, OK WX08 8264 5 4 Vici, OK WX09 8266 6 0 H askell, OK WX10 8618 7 19 Pu rcell, OK WX11 8156 8 217 Central Facility PRC1 8434 9 670 Beau m ont, KS PRI1 8442 10 1747 Med icine Lod ge, KS PRI2 8339 11 2529 Meeker, OK PRI3 6394 12 2120 Central Facility, OK SDC1 5847 13 134 H illsboro, KS SDB1 1387 14 495 Vici, OK SDB4 1392 15 554 Morris, OK SDB5 1435 16 240 Pu rcell, OK SDB6 1437 31

32

6. D ATA QUALITY ISSUES Although the ARM SGP site has the largest amount and highest qu ality regional d ata available in the w orld , as w e used the d ata w e becam e aw are of several d ata qu ality issues. These includ ed ARM SMOS d ata archive files that w ere corru p ted , N PN and RWP p rofiles that includ ed very large w ind sp eed s, and w ind errors in sond es w hen the height levels w ere uneven.

6.1 SMOS D ata Each ARM SMOS netCDF file contains half-hour average d ata for one d ay at a p articular site. The d ata for all variables for each half-hou r are w ritten sequ entially. As w e p rocessed the d ata, occasionally w e encou ntered a file that at som e p oint in time contained incorrect values, alm ost as if a bit had been d ropp ed or ad d ed . Reread ing the file from the archive d id not remed y the error. So, for the corru p t files, w e extracted the good d ata and left the d ata for the rest of the d ay as m issing. This p roblem occu rred for abou t 75 of 5500 files. In som e cases m ost of the d ata for a d ay w as lost, in others only the last half hour. Even in the archive, not all files contain all 48 half-hou rs, and in som e cases the d ata for a single d ay are sp lit betw een tw o files. The p rocessing algorithm w as w ritten to hand le these anom alies.

6.2 N PN and 915 MHz RWP D ata The first tim e w e ran the sim ulation, ADAPT gave frequ ent w arning m essages that observed w ind speed s w ere greater than 150 m / s, an u nrealistically high valu e, especially since the greatest altitud e inclu d ed in the sim u lation is 6000 m . As w e investigated theses w ind s, w e d iscovered many ad d itional instances of incorrect w ind s, w ith som e of the bad values occurring near the surface. Since low levels w ind s are cru cial for calculating the transport and d isp ersion of low altitu d e p ollu tant releases, w e w ere com pelled to scan the observed p rofiles for errors and attem p t to correct them or rem ove the bad points.

The nu m ber of profiles in the d ata set is qu ite large, over 90,000. Since w e are concerned m ostly w ith w ind , w e w rote a cod e that d isp lays w ind sp eed and d irection in ad jacent p anels and allow s us to either accept the profile or m ake changes by sp ecifying a new valu e for a layer, interpolating betw een layers, extrap olating from above or below , or interp olating betw een a specified low er and u p p er layer. We also had the op tion of com p letely d eleting a profile, and w e au tom atically d eleted all p rofiles w ith few er than fou r layers. The proced ure used actu ally w rote a m od ification file for the observ.m et file. Then, in another pass through the d ata, a new observ.m et file w ith the m od ifications w as prod uced for the simu lation. It took abou t 10 w eeks to look at all 33

90,000+ p rofiles and m od ify the incorrect ones. Sam p les of the m ore frequ ent errors and corresp ond ing corrections are given in Figu res 5-8.

Figure 5. Central Facility Sond e Lau nched at 5:29 am CST on Janu ary 31, 2000. This is an examp le of a sond e that had to be d eleted from the d ata set because it has so m any errors.

34

Figure 6. 915 MH z RWP Wind Profile for the Central Facility at 10:00 p m CST on March 18, 2000. This is an exam ple case w here the rep orted low est level w ind (77.5 m / s) is w rong. The corrected profile is d eterm ined by extrap olating d ow nw ard from the 2 levels above. Where the tw o profiles overlap only the heavy line show s.

35

Figure 7. Central Facility Sond e Lau nched at 5:33 am CST on N ovem ber 15, 2000. Qu ite a few of the sond es exhibited speed anom alies in an otherw ise fairly sm ooth p rofile.

They w ere often associated w ith a m issing layer in the d ata. They w ere corrected by interp olating betw een the bottom and top of the anom aly. N ote that d irection w as hard ly affected .

36

Figure 8. 915 MH z RWP Wind Profile for Beau m ont at 11:00 am CST on May 14, 2000.

This sou nd ing exhibits an error seen occasionally w ith the 915 MH z RWP, having one or more layers w here w ind speed and d irection are offset from the rest of the sou nd ing.

The correction resets values in the offset layer by interp olation. The second anom aly at 2400 m is seen frequ ently at this height w here the high and low resolu tion rad ar m od es m erge.

37

38

7. SOURCE TERM Each cod e requires that the user inp u t a sou rce term , that is, p aram eters giving the time and d u ration of the release, the height of the release, bu oyancy of the released m aterial, and release m agnitud es of d ifferent rad ionu clid es. This last inp u t is d escribed in all the cod es as an inventory of each rad ionu clid e at the start of the p roblem and a release fraction of several rad ionuclid e chem ical elem ent grou p s. This form u lation, w hich allow s each cod e to account for rad ioactive d ecay of the variou s rad ionu clid es from the start of the problem to the release of the m aterial, is not need ed for this stu d y. Ou r sou rce term w as form ulated to be as sim p le as p ossible w hile still allow ing the ATD p rocesses to be com pared : w e chose a single, long-lived rad ionu clid e that d oes not d ep osit and a single, long-lived rad ionu clid e that d oes d ep osit. Fu rther, the inventory of each of these tw o rad ionuclid es w as arbitrarily chosen as 1016 Bq. This d oes not rep resent a realistic release from any N RC-licensed facility.

We chose only tw o rad ionuclid es instead of tens of d ifferent rad ionu clid es (as cou ld be in a release from a severe accid ent at a nu clear p ow er p lant) becau se w e w anted to avoid confound ing the d epositions and exp osu res w ith short-, m ed iu m -, and long-lived m aterial, in case the com p arison w ere to show u nfavorable resu lts. We believed that trou ble-shooting the d ifferences w ou ld be easier w ith only tw o rad ionu clid es. As w ill be seen in the results section, this sim p lification w as unnecessary.

The characteristics of the source term for this stu d y are given in Table 12. The valu es of Fy and Fz , the initial size of the plu m e, are not u su ally consid ered p art of the sou rce term , bu t since they influence the initial p lu m e they are inclu d ed here.

Table 12. Sou rce Term Sp ecification Characteristic Value Location ARM Central Facility Tim e of release 0.0 s Duration of release 1800 s, u niform Am ount of release, each nu clid e 1016 Bq H eight of release 50 m Buoyant energy 106 W 39

40

8. SIMULATION PROCED URE This stud y is aim ed at evaluating the u se of a sim p le atmosp heric transp ort and d isp ersion m od el in pred icting off-site consequ ences of large accid ental releases from nu clear p ow er plants and other N RC-licensed facilities. The p rim ary m etric chosen for com p arison w as the annual average integrated concentration in fou r arcs and sixteen com p ass d irections around the assu m ed release point, the DOE ARM SGP Central Facility near Lam ont, Oklahom a. The arcs w ere at d istances of 14.5-16.1, 30.6-32.2, 78.8-80.5, and 159.3-160.9 km (9-10, 19-20, 49-50, and 99-100 m iles); the com p ass d irections w ere the sixteen 22.5 d egree sectors from N clockw ise arou nd to N N W. Each m od el u sed a proced ure to generate p red icted annu al average exp osu res (near su rface air concentrations integrated d uring p assage of the p lu m e in Bq-s/ m 3) and d ep ositions (total m aterial d eposited on the grou nd in Bq/ m 2) for these 64 areas u sing norm al techniqu es.

8.1 MACCS2 MACCS2 w as run in a single step that involved tw o of the three m ajor m od u les in the cod e. The first m od ule in the sequence is ATMOS, w hich calcu lates atm osp heric transp ort and d ispersion (ATD). ATMOS first bins all of the hou rs of the annu al w eather d ata into 36 bins, as d iscussed in section 3. It then selects w eather trials rand om ly from these bins. In this case the nu m ber of w eather trials w as lim ited to 610 to keep the CPU requirem ents for ADAPT/ LODI from being excessive. H ow ever, 610 w eather trials are more than enough to attain valid statistics, esp ecially for the mean valu es that are presented in this stu d y. After selecting the w eather trials, ATMOS calculates the atm ospheric transport and d isp ersion for each trial in the set. The EARLY m od u le, w hich calculates em ergency resp onse and acu te health effects, w ould not have been need ed except that it also contains the logic to perform w ind rotations and the ou tp u t capability that w as required for com p arison w ith the other cod es.

8.2 RASCAL/RATCHET Meteorological d ata processing involved three step s. In the first step the fu ll year of m eteorological d ata for the six sites w as combined into a single file w ith d ata in the form at required by the RASCAL m eteorological d ata p rocessor; in the second step the m eteorological d ata file w as d ivid ed into tw elve m onthly files. Each m onthly file contained d ata for the m onth plus d ata for the first tw o d ays of the follow ing m onth.

These tw o steps required only a few second s of com p u ter tim e. The third step w as ru nning the RASCAL m eteorological p rocessor. The p rocessor created sets of m eteorological d ata field s for three m od el d om ains - 32.2 km (20 miles) on a sid e, 80.5 km (50 m iles) on a sid e, 160.9 km (100 m iles) on a sid e - for each m onth.

41

8.3 AD APT/LOD I Table 13. Meteorology Grid Vertical Levels and Corresp ond ing The ADAPT/ LODI m od el system is d esigned to Central Facility Altitu d es (MSL).

p rod u ce rap id and accurate estim ates of Layer Sigm a Altitu d e (m) d ow nw ind concentrations from accid ental 1 0. 312 releases of hazard ous and toxic pollu tants. The general app roach u sed in this stud y w as to treat 2 0.003730 332 each postulated release as a sep arate event, and , 3 0.007927 355 after running all cases, calcu late the average 4 0.012651 380 exp osu re and d ep osition from the ind ivid u al 5 0.017967 409 cases w eighted by their ind ivid ual frequ encies. 6 0.023950 442 The ind ivid ual cases w ere 610 releases at tim es 7 0.030683 479 and w ith w eights (frequencies) id entified by 8 0.038261 520 MACCS2 as representative of the entire years 9 0.046789 566 m eteorology. Since the proced ure w as highly 10 0.056387 619 rep etitive, the com puter p rocessing w as 11 0.067189 678 accom p lished using a series of scripts.

12 0.079345 744 8.3.1 Grids 13 0.093026 818 14 0.108423 902 Both ADAPT and LODI perform their 15 0.125751 997 calcu lations using grid s that d efine a fram e of 16 0.145252 1103 reference in the vicinity of the release site. Tw o 17 0.167199 1222 grid s w ere used in this stud y, a three- 18 0.191898 1357 d im ensional m eteorology grid and a tw o- 19 0.219696 1508 d im ensional concentration (exposure and 20 0.250979 1679 d ep osition) grid .

21 0.286187 1871 The m eteorology grid specifies the locations 22 0.325810 2087 w here grid d ed m eteorology is d efined by 23 0.370402 2329 ADAPT and provid es a fram e of reference for 24 0.420587 2603 transport and d ispersion of LODI parcels. It has a 25 0.477067 2911 u niform 4 km spacing in the horizontal covering 26 0.540629 3257 a squ are area 400 km on a sid e centered at the 27 0.612165 3647 ARM SGP Central Facility. In the vertical it u ses 28 0.692672 4085 a non-uniform terrain- follow ing sigm a 29 0.783276 4579 coord inate (A sigm a coord inate gives height as a 30 0.885244 5134 fractional d istance betw een the surface and the 31 1.000000 5760 top of the d om ain, 5760 m in this case.) w ith greater resolution near the ground , and cou rser resolu tion aloft. The bottom of the grid follow s the topograp hy w hile the top is at a constant elevation of 5760 m above m ean sea level (MSL). The list of sigm a levels is 42

Figure 9. Topography of the ARM SGP Site. The view is from the sou theast looking northw est. The black line at 37° N ind icates the Oklahom a/ Kansas bord er. Black half circles ind icate the locations of Oklahom a City and Wichita, and a light one Tulsa. The range of elevations is from 153 to 760 m above sea level.

given in Table 13, and the top ograp hy of the region is illustrated in Figu re 9. Becau se of the vertical stretching in the plot, the topograp hy ap p ears m ore ru gged that it is; the total change in elevation across the d om ain is abou t 600 m over a d istance of 400 km , for a m ean slope 1.5 m / km . The general slop e is from the northw est to the sou theast, bu t there are several river valleys and other irregu larities in the terrain.

A concentration grid is used by LODI to calcu late and accu m u late the exp osu re and d ep osition d uring a run. For this stu d y it has a u niform spacing of one km and covers a squ are region 350 km on a sid e centered on the ARM SGP Central Facility. The final resu lt of a LODI sim ulation is the exp osu re and d ep osition of em itted m aterials at the 350 by 350 points of the concentration grid .

8.3.2 Solution Steps The scrip t that runs the cases allow s one to p rocess all 610 cases in sequ ence or start w ith any given case and run a specified nu m ber. It also assu m es that the ADAPT 43

m eteorology input files based on the ARM SGP d ata are alread y p rep ared . The step s are the follow ing:

1. Read the case start times (Ju lian d ay and hou r) from the w eather d ata file p rovid ed by MACCS2. The case nu m ber is used as an ind ex for saved results.

2. Loop over the requested cases (Step s 3-6).

3. Set the current m onth and d ay. ADAPT and LODI use d ates/ times in the form at yyyyMMMd d _hhm m ss w here yyyy is the year, MMM is the three-character m onth, d d is the d ay, hh is the hou r and the m inu tes and second s (mm and ss) are both 00 for this p roblem because w e u se only hou rly m eteorological d ata and hourly release tim es.

4. Delete ADAPT output/ LODI inp u t m eteorology files from hou rs p rior to this release. These three-d im ensional d ata files are qu ite large (23 Mb each), and since w e process d ates sequentially, those files for tim es p rior to the release tim e are no longer need ed . We keep those files that w ere created for the p reviou s ru n and are still need ed for this run. Of cou rse, the first case has no u nneed ed files to d elete.

5. Create ADAPT output/ LODI inp u t m eteorology files for the next 48 hou rs. To be sure w e have accurate concentrations, all the LODI parcels need to exit the 400-km square d om ain by the end of the ru n. In som e cases 24 hou rs w as too short, so w e u sed 48 hou rs. Since LODIs ru ntim e d ep end s d irectly on the num ber of rem aining parcels, there is no time p enalty for ru nning too long. This step runs ADAPT repeated ly. The m eteorology files are p re-p rep ared , and the ADAPT nam elist file is created by the script. The key m ixing height p aram eter is read from the hourly m ixing height file p rep ared w hile p rocessing the ARM m eteorology d ata by the proced u re d escribed in section 5.2.

6. Run LODI to calculate the exp osu res and d ep osition for this case. The LODI w ind and tu rbulence d ata com e from the files ju st p rod u ced by the p reviou s ADAPT runs and the nam elist files are created in the script. The scavenging rate is read from a file that translated hourly p recip itation rates for the ARM Central Facility into the appropriate scavenging rates as given in section 3.3.2.2. In these sim ulations 25,000 LODI parcels w ere tracked for each species; this is at the low er range of the typical nu m ber of parcels exp erience has show n gives correct concentrations. The m ain ou tp u ts from LODI for each case are tw o files giving exp osures for the d epositing and non-d ep ositing sp ecies and one file of d ep osition for the d ep ositing species.

44

7. The exposures and d epositions calcu lated by LODI are p rocessed by a sm oothing function. Concentrations in LODI are d erived by p eriod ically sam pling the position and p rop erties of p arcels. With a lim ited nu m ber of p arcels, concentrations can be u neven, p articu larly at great d istances from the sou rce, and plots based on these concentrations are often ragged . Using a 1-2-1 sm oothing function prod uces the sm oother p lots w e exp ect w ithout changing the accuracy of the solution.

8. Red uce the 350 by 350 one-kilom eter resolu tion exp osures and d ep ositions to the arc/ sector values. This w as d one by simp ly constru cting a list of p oints in the concentration grid that w ere w ithin each arc/ sector and find ing the average exp osure or d eposition for these p oints. For ind ivid u al arc/ sectors at 16.1, 32.2, 80.5, and 160.9 km (10, 20, 50, and 100 m iles), the nu m ber of p oints w as betw een 9 and 12, betw een 19 and 21, betw een 40 and 52, and betw een 101 and 106, resp ectively.

9. Calculate the annual average exp osu res and d eposition as the w eighted average of the 610 ind ivid ual cases.

We also evaluated perform ing the last tw o step s in reverse ord er, calcu lating the w eighted average annual exp osure on the 350 by 350 grid , and then d eterm ining the arc/ sector values. Both m ethod s give the sam e resu lt.

8.4 Computer Time Requirements 8.4.1 MACCS2 Perform ing all 610 w eather trials w ith the MACCS2 cod e requ ired less than 1 m inu te of CPU tim e. For this problem , only the ATMOS and EARLY m od u les w ere u sed ; long term consequences calculated in CH RON C w ere not need ed and so w ere not p erform ed . Otherw ise, MACCS2 w as ru n in the stand ard w ay, w hich involves w eather binning and w ind rotation. Weather binning is d escribed in section 3. Wind rotation involves accounting for the possibility that the w ind m ight have been blow ing in a d ifferent d irection than the one corresp ond ing to the beginning of the w eather trial. In the w ind rotation m od e, MACCS2 p erform s a calcu lation for each of the 16 com p ass d irections for each w eather trial. Each of these calcu lations is w eighted by the p robability that the w ind m ight have been blow ing in that d irection, as d eterm ined by p erform ing statistics on the w eather file. (Wind rotation is d escribed fu rther in section 4.) Thu s, the 610 w eather trials perform ed by MACCS2 accou nted for 8760 p ossibilities.

45

8.4.2 RASCAL/RATCHET Processing the m eteorological d ata to create the inp u t files for RASCAL requ ired abou t 16 m in of com puter tim e and resulted in 36 files w ith an average size of abou t 12.5 Mb.

RASCAL and RATCH ET w ere run in batch m od e on a netw orked 3 GH z PC w ith 1 Gb RAM. Each case w as run three times, once for each m od el d om ain. The total time requ ired to run RASCAL for all 610 releases for the three m od el d om ains w as about 46 m in (~1.5 s per release). About 65 m in w ere requ ired for the RATCH ET ru ns (~2.2 s p er release).

8.4.3 AD APT/LOD I ADAPT and LODI w ere run on a DEC com p u ter w ith 1 GH z alp ha p rocessors. ADAPT ru ns that prod uced one hourly m eteorology d ata set took an average of 40 s w ith a range of 37-52 s. Each LODI run that prod u ced exp osu res and d ep ositions for one of the 610 cases used an average of 138 s w ith a range of 51 to 360 s. The entire sim u lation, m ad e u p of 8778 ADAPT runs and 610 LODI ru ns, took 435300 s (121 hou rs) of CPU tim e.

46

9. WIN D CHARACTERISTICS OF THE ARM SGP SITE The w ind and its variability is the m ost im p ortant p aram eter in this stu d y, and one of the best w ays to sum m arize w ind s at a location is w ith a w ind rose that show s the relative frequency of w ind s w ith particular d irections and sp eed s at a given site. The w ind rose for the surface d ata (10 m m easurement height) at the ARM SGP Central Facility near Lam ont, OK for the year 2000 is show n in Figu re 10. This is the only w ind d ata used by MACCS2 to calculate exp osu re and d ep osition. It is also im p ortant to p oint ou t, especially to those fam iliar w ith MACCS2, that the w ind roses p lotted here are in the stand ard form at w here the arm s p oint in the d irection the w ind is com ing from rather than using MACCS2's convention show ing the d irection the w ind is blow ing tow ard s. Figure 10 show s a very large pred om inance of southerly and south-southeasterly w ind s; nearly half the time the w ind is from the southeast through south-sou thw est. When the w ind is not from the south it is m ost often from the north. Wind s pred ominately from the east (12.0%) and w est (6.7%) occur relatively infrequently.

The seasonal variability of the w ind s at the Central Facility is show n by the su rface w ind rose plots for each season in Figu res 11-14. In the sum m er the w ind blow s w ith a southerly com p onent (d irection from southw est through southeast) over 70% of the tim e, w hile in w inter the frequency of w ind s w ith a northerly com ponent (36%) is nearly equal to the frequency of w ind s w ith a southerly com ponent Figure 10. Wind Rose for the ARM Central (39%). Facility at Lam ont, OK. The sp eed ranges in m / s are 1-2.5, 2.5-4.5, 4.5-7, 7-10, >10. Wind While MACCS2 d oes not take the sp eed is less than 1 m / s 1.25% of the tim e.

sp atial variability of w ind into account, Wind d irection is the d irection the w ind is RASCAL/ RATCH ET and blow ing from . The circles are at 4, 8, 12, 16, ADAPT/ LODI d o. and 20%. There are 8774 valid hou rly RASCAL/ RATCH ET u ses the five su rface (10 m ) w ind observations at this ad d itional ARM surface sites closest to site. The average w ind speed is 5.73 m / s.

47

Figure 11. Surface Wind Rose for the Figure 12. Su rface Wind Rose for the ARM Central Facility for Winter, 2000. ARM Central Facility for Sp ring, 2000.

Figure 13. Surface Wind Rose for the Figure 14. Su rface Wind Rose for the ARM Central Facility for Sum m er, 2000. ARM Central Facility for Au tu m n, 2000.

48

the Central Facility. The w ind roses for these sites for the year 2000 are show n in Figu res 15-19. While all these sites have w ind roses sim ilar to each other and to the Central Facility, there are d ifferences that are p robably d u e to natu ral variability and perhaps local effects su ch as su rface cond itions and terrain. All the sites have a strong Figure 15. Su rface Wind Figure 16. Su rface Wind Figure 17. Su rface Wind Rose for the Elk Falls, KA Rose for the Ashton, KA Rose for the Byron, OK Site.

Site. Site.

The sp eed ranges are 1-2.5, 2.5-4.5, 4.5-7, 7-10, and >10 m / s. Wind sp eed s less than 1 m / s occur 13.5% (Elk Falls),

0.7% (Ashton), 3.2% (Byron),

1.9% (Ringw ood ) and 5.5%

(Meeker) of the time. Wind d irection is the d irection the w ind is blow ing from . The circles are at m u ltip les of 3%

(Elk Falls), 5% (Ashton), and 4% (Byron, Ringw ood , and Meeker). There are 7774 (Elk Figure 18. Su rface Wind Figure 19. Su rface Wind Falls), 8201 (Ashton), 8166 Rose for the Ringw ood , OK Rose for the Meeker, OK (Byron), 8104 (Ringw ood ),

Site. Site. and 7792 (Meeker) valid hou rly su rface (10 m ) w ind observations. The average w ind speed s are 4.14 m / s (Elk Falls), 5.47 (Ashton), 4.60 (Byron), 4.67 (Ringw ood ), and 4.15 (Meeker).

49

p eak in w ind frequency associated w ith sou therly flow . At all the ad d itional sites w ind sp eed s are low er than at the Central Facility.

In ad d ition to m any m ore surface sites, ADAPT/ LODI u ses u p p er air w ind d ata from p rofilers and sond es. Wind roses from the 915 MH z p rofiler at its low est height of 87 m are show n in Figures 20-23. Valid d ata from the Central Facility p rofiler at the low est The sp eed ranges are 2-4, 4-7, 7,10, 10-14, 14-18, and

>18 m / s. Wind sp eed s less than 2 m / s occur 19.1% (Lam ont), 2.2%

(Beau m ont), 3.6%

(Med icine Lod ge), and 5.9% (Meeker) of the tim e.

Wind d irection is the d irection the w ind is blow ing from . The circles are at m u ltip les of 2%

(Lam ont), 4%

Figure 20. Wind Rose from Figure 21. Wind Rose from (Beau m ont), and 5%

915 MH z Profiler at 87 m 915 MH z Profiler at 87 m (Med icine Lod ge and H eight at Lam ont, OK H eight at Beaum ont, KA. Meeker). There are 1010 (Central Facility). (Lam ont), 4633 (Beau m ont), 7701 (Med icine Lod ge), and 1979 (Meeker) valid hou rly w ind observations at 87 m . The average w ind sp eed s are 5.72 m / s (Lam ont), 9.63 (Beau m ont), 8.65 (Med icine Lod ge), and 7.50 (Meeker). The Central Facility, w ith a rather sm all nu m ber of observations, d oes not Figure 22. Wind Rose from Figure 23. Wind Rose from seem to rep resent the 915 MH z Profiler at 87 m 915 MH z Profiler at 87 m year very w ell.

H eight at Med icine Lod ge, H eight at Meeker, OK.

KA.

50

height are often m issing (valid w ind d ata are available for only 1010 ou t of 8778 hou rs),

and the d ata d o not seem to rep resent the entire year very w ell. Fortu nately, sond es are frequ ently available at this site to m ake up for the lack of p rofiler d ata. The other p rofilers p rovid e 87 m w ind d ata m ost of the tim e and the w ind s are representative.

Sou therly w ind s d om inate, and w ind sp eed s are higher than at the su rface. A final set of w ind roses, for a height of 500 m at the Central Facility, is show n in Figu re 24, w hich p rovid es the observations from the 915 MH z p rofiler, and Figu re 25, w hich p rovid es the d ata from the N OAA w ind p rofiler netw ork. These are m easu rem ents of the same qu antity by tw o d ifferent instrum ents. The p lots are qu ite similar, bu t they also exhibit d ifferences reflecting the variation in the w ind s and the inaccu racy of w ind p rofiling Figure 24. Wind Rose from 915 MH z Figure 25. Wind Rose from N OAA Rem ote Wind Profiler (RWP) at 495 m Wind Profiler (N WP) at 500 m H eight at height at Lam ont, OK (Central Facility). Lam ont, OK (Central Facility).

The sp eed ranges in these figures are 2-4, 4-7, 7,10, 10-14, 14-18, and >18 m / s. Wind sp eed s less than 2 m / s occur 1.7% (RWP) and 4.6% (N WP) of the time. Wind d irection is the d irection the w ind is blow ing from . The circles are at 4, 8, 12, 16, and 20%. There are 7490 (RWP) and 8102 (N WP) valid hou rly w ind observations at 500 m .

51

instru m ents. At this m id -level of the atmosp here the w ind s are stronger than at the su rface and on average have veered (rotated clockw ise) from the p red om inant su rface d irection a bit east of south to a d irection a bit w est of sou th. If w ind sp eed and d irection are im portant param eters in controlling the arc and arc/ sector annu al average exp osure and d ep osition, this change in w ind d irection cou ld be im p ortant becau se som e of the released m aterial m ay be transp orted at heights several hu nd red m eters above the surface.

52

10. RESULTS Ou r p rim ary goal is to p rovid e an evalu ation of the atm osp heric transp ort and d isp ersion (ATD) m od u les in MACCS2, averaged over a rep resentative m eteorological d ata set, and to enable a d iscussion abou t their continu ed u se for N RC p u rp oses. The evalu ation is d one by com paring resu lts from MACCS2 to those from ADAPT/ LODI, a com p lex state-of-the-art ATD cod e. In ad d ition resu lts from RASCAL and RATCH ET w ere inclu d ed in the stud y. The prim ary m etrics of com p arison are the arc and arc-sector annu al average exp osure and d ep osition d erived from a set of 610 rep resentative releases d uring the year 2000 at the DOE ARM SGP site.

To ad d valid ity to the stud y the resu lts w ere obtained u sing the m od els in their norm al m od es. Each m od el w as run by personnel w ho norm ally u tilize these m od els, and all sim u lations w ere d one ind epend ently and w ithout ad ju stment. In ord er to make the best p ossible com parison of the ATD com p onents, w e chose a d ry d ep osition velocity and sp ecified w ashout coefficients, based on norm al p rop erties of the m od els, that gave the sam e or very sim ilar removal rates, and these choices w ere mad e before the mod el ru ns started . The sam e set of 610 release tim es, d erived from MACCS2's norm al selection criteria, w as used by all mod els. Each cod e u sed hou rly m eteorological d ata and ran each case until all the released m aterial exited the 160.9-km (100-m ile) rad ius d om ain, 80.5-km (50-mile) rad ius for RASCAL/ RATCH ET. The characteristics of the release - location, start time, d uration, am ou nt of d ep ositing and non-d ep ositing species, height, and heat energy of release - w ere id entical for all m od els.

10.1 Arc Averages The annu al average exposures for the d ep ositing and non-d ep ositing sp ecies and the annu al average d eposition for arcs at d istances of 14.5-16.1, 30.6-32.2, 78.8-80.5, and 159.3-160.9 km (9-10, 19-20, 49-50, and 99-100 m iles) from the sou rce are given in Tables 14-16. In all cases the arc average exp osu res and d ep ositions for MACCS2, RASCAL, and RATCH ET d iffer from LODI by less than a factor of tw o (ratio betw een 0.5 and Table 14. N on-Depositing Species Arc Average Exp osu re (Bq-s/ m 3) and Ratio to LODI Mod el 16.1 km (10 m i) 32.2 km (20 m i) 80.5 km (50 m i) 161 km (100 m i)

MACCS2 8.02e7 (1.58) 2.39e7 (1.01) 4.77e6 (0.64) 1.80e6 (0.65)

RASCAL 7.32e7 (1.45) 3.09e7 (1.30) 8.41e6 (1.12)

RATCH ET 3.24e7 (0.64) 1.33e7 (0.56) 3.59e6 (0.48)

LODI 5.06e7 (1.00) 2.36e7 (1.00) 7.49e6 (1.00) 2.75e6 (1.00) 53

Table 15. Depositing Species Arc Average Exp osu re (Bq-s/ m 3) and Ratio to LODI Mod el 16.1 km (10 m i) 32.2 km (20 m i) 80.5 km (50 m i) 161 km (100 m i)

MACCS2 5.18e7 (1.41) 1.40e7 (1.05) 2.49e6 (0.81) 7.86e5 (0.89)

RASCAL 5.91e7 (1.61) 2.01e7 (1.50) 3.94e6 (1.28)

RATCH ET 2.89e7 (0.79) 1.09e7 (0.81) 2.69e6 (0.88)

LODI 3.68e7 (1.00) 1.34e7 (1.00) 3.07e6 (1.00) 8.86e5 (1.00)

Table 16. Arc Average Dep osition (Bq/ m 2) and Ratio to LODI Mod el 16.1 km (10 m i) 32.2 km (20 m i) 80.5 km (50 m i) 161 km (100 m i)

MACCS2 5.57e5 (1.21) 1.53e5 (0.96) 2.87e4 (0.78) 8.96e3 (0.83)

RASCAL 7.20e5 (1.56) 2.34e5 (1.46) 4.71e4 (1.29)

RATCH ET 3.10e5 (0.67) 1.06e5 (0.66) 2.63e4 (0.71)

LODI 4.62e5 (1.00) 1.60e5 (1.00) 3.67e4 (1.00) 1.08e4 (1.00) 2.00), excep t RATCH ET at 80.5 km (50 m iles) for the non-d ep ositing sp ecies. MACCS2's exp osu re and d eposition values have a tend ency to be higher close to the sou rce and low er at d istances of 80.5 and 160.9 km (50 and 100 m iles). RASCAL and RATCH ET have the sam e tend ency, but w ith sm aller m agnitu d e. RASCAL consistently has higher and RATCH ET low er exposures and d ep ositions than LODI. Ratios of RASCAL to RATCH ET exposures and d epositions are often larger than tw o; this is attributed to faster vertical d ispersion in RATCH ETs new and m ore com p lex d isp ersion m od el.

The agreement am ong these mod els is gratifying. The exp lanation seems to be related to the fact that arc averaging m inim izes the im p ortance of transp ort since released m aterial m ust m ove aw ay from the sou rce and after som e transit tim e cross the arcs at 16.1, 32.2., 80.5, 160.9 km (10, 20, 50, and 100 m iles). Disp ersion and d ep osition are both related to travel time. The non-d epositing sp ecies is su bject only to vertical and horizontal d ispersion and transport, and , after an initial p eriod , is w ell m ixed throughou t the bou nd ary layer. Large d ifferences in exp osu re w ou ld requ ire large d ifferences in m ixing heights am ong the m od els, bu t MACCS2 u sed the seasonal average m ixing heights d erived from the hourly m ixing heights u sed by LODI so the d ifferences are not large. In ad d ition the d ry and w et d ep osition rates w ere chosen to be sim ilar; therefore, as long as the transit tim es are sim ilar, there shou ld not be large d ifferences in d epositing species exp osu res either. The fact that the largest d ifferences 54

are betw een RASCAL and RATCH ET w hich have the sam e plu m e trajectory and d iffer only in the d ispersion param eterization tend s to su p p ort this argu m ent.

10.2 Arc-Sector Averages The second m etric chosen for com parison is the annu al average arc-sector exp osu re and d ep osition. The sam e 1.6 km (one m ile) w id e arcs at 16.1, 32.2, 80.5, and 160.9 km (10, 20, 50, and 100 m iles) and the sixteen 221/2 d egree d irectional sectors from north clockw ise through north-northw est p rovid e 64 valu es for MACCS2 and ADAPT/ LODI (48 for RASCAL/ RATCH ET) of exp osu re for the non-d ep ositing material and 64 (48) valu es of exposure and d eposition for the d ep ositing m aterial. Exp osu re for the non-d ep ositing species is plotted and com p ared w ith the bar grap hs in Figu res 26-29.

Sim ilar p lots of exposure for the d ep ositing material are show n in Figu res 30-33 and of d ep osition in Figu res 34-37. In all these p lots the sector to the north is rep eated on both sid es of the plot d enoted N and N 2. These p lots show the angu lar d istribu tions of the released m aterial in ad d ition to the d ecrease of exp osu re and d eposition w ith d istance.

All mod els prod uce similar angular d istribu tions that reflect the m ean annu al w ind cycle. The largest concentrations are to the north, w ith a second ary m axim u m to the south; a relatively sm all am ount of m aterial goes w est and esp ecially east.

The arc-sector exposures and d epositions for MACCS2 are generally w ithin a factor of tw o of the correspond ing values for the state-of-the-art m od el, LODI. Of the 192 exp osu res and d epositions (4 arcs, 16 sectors, 2 exp osu res and 1 d ep osition), only nine are m ore than tw ice as large - all in the 16.1-km (10-m ile) arc - and 12 are less than half as large, - four in the 80.5-km (50-mile) and eight in the 160.9-km (100-mile arc) - and these are usually in sectors w here the exp osu re or d eposition is sm aller. The higher valu es close in and low er values at greater d istances for MACCS2 corresp ond to the sam e trend noted for the arc average exp osu re and d eposition. Differences greater than a factor of three are seen only tw ice, both for the non-d ep ositing m aterial; these are in the WSW sector of the 80.5-km (50-mile) arc (ratio = 0.31) and in the N N E sector of the 160.9-km (100-mile) arc (ratio = 0.33).

RASCAL and RATCH ET arc-sector exp osu res and d epositions have many m ore valu es d iffering by m ore than a factor of tw o from LODI, bu t this is p artly related to the fact that RASCAL tend s to consistently p rod u ce higher and RATCH ET low er valu es than LODI. RASCAL has 33 of 144 exposu res and d epositions (3 arcs, 16 sectors, 2 exp osu res and 1 d eposition) m ore than tw ice as large as LODI, and none less than half as large.

Ten of these are m ore than three tim es as large as LODI. RATCH ET has three of 144 exp osu res and d epositions m ore than tw ice as large as LODI and 33 less than half as large. Of these 33, ten are less than one-third as large as LODI.

55

Figure 26. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 16.9-km (10-mile) Arc.

Figure 27. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 32.2-km (20-mile) Arc.

56

Figure 28. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 80.5-km (50-mile) Arc.

Figure 29. Arc-Sector Exposure for N on-Dep ositing Sp ecies on 160.9-km (100-mile) Arc.

57

Figure 30. Arc-Sector Exposure for Dep ositing Sp ecies on 16.1-km (10-mile) Arc.

Figure 31. Arc-Sector Exposure for Dep ositing Sp ecies on 32.2-km (20-mile) Arc.

58

Figure 32. Arc-Sector Exposure for Dep ositing Sp ecies on 80.5-km (50-mile) Arc.

Figure 33. Arc-Sector Exposure for Dep ositing Sp ecies on 160.9-km (100-mile) Arc.

59

Figure 34. Arc-Sector Deposition on 16.1-km (10-mile) Arc.

Figure 35. Arc-Sector Deposition on 32.2-km (20-mile) Arc.

60

Figure 36. Arc-Sector Deposition on 80.5-km (50-mile) Arc.

Figure 37. Arc-Sector Deposition on 160.9-km (100-mile) Arc.

61

Becau se each m od el has a d ifferent arc average exp osu re and d ep osition, it is d ifficu lt to p ortray how closely the angular d istribu tions actu ally agree. Figu res 38-49 show the angu lar d istributions norm alized by each m od els arc average concentration. The ord inate in these plots is logarithm ic so that m u ltiplicative changes are p rop ortional to d istance, i.e., a value tw ice the average is ju st as far above the average line as a value half the average is below , and a valu e four tim es the average is tw ice as far above the average line as a value tw ice the average. These figu res also inclu d e the north sector on both sid es.

The angu lar d istributions of exp osure and d ep osition are qu ite similar for all mod els and again reflect the d istribution of the w ind . The highest valu es are to the north w here the exposu re or d eposition is 2-3 tim es the average; interm ed iate valu es, near the average, occur in southerly d irections; low er valu es, from one-half to three-qu arters of the average, occur to the w est of the sou rce; and the low est valu es, often less the one-half the average, occur to the east, corresp ond ing to infrequ ent w esterly w ind s. The largest d ifferences in norm alized exp osure and d ep osition occu r in sectors to the east and w est w here the values of exposu re and d eposition are sm aller.

In general, the angular d istribution from MACCS2 seem s to corresp ond m ore closely w ith LODI than RASCAL or RATCH ET. This is a bit su rp rising since RASCAL and RATCH ET follow ind ivid ual plum es m ore closely than MACCS2, and the annu al d istribu tions are averages of ind ivid u al p lu m es from the 610 releases ju st like LODI.

Where local m axim a (m inim a) of the cu rves are d isp laced , it is often by only one sector; that could be a result of ind ivid ual plum es taking slightly d ifferent tracks and show ing u p in neighboring sectors. LODI also m akes u se of u p p er-level w ind d ata; therefore, w ind d irection shear w ith height w ou ld be represented in LODI bu t not in the other m od els. For m ost plu m es from ind ivid u al releases, exp osu re and d ep osition are confined to tw o or three sectors. The d ifferences in norm alized d istribu tions d o not increase w ith d istance, in fact they m ay even d ecrease. Larger d ifferences in d ep osition are probably d ue to relatively infrequ ent large rain events occu rring at d ifferent locations. H eavy rain over a period of an hou r can d ep osit m ost of the d ep ositing m aterial in a local area and largely d eplete the plu m e.

10.3 Tw o-D imensional Exposure and D eposition While not a p rim ary m etric of com p arison, it is interesting to exam ine the tw o-d imensional exp osure and d eposition p lots from each m od el; these are show n in Figu res 50-52. The d ifferences in these plots are only p artly d u e to d ifferences in results; they also d epend on the location and sp acing of the d ata u sed to constru ct them and to p articu lar featu res of the m od els. The MACCS2 p lots are based on rad ial/ sector exp osu re (d eposition) d ata, specifically, 29 not very evenly sp aced rad ii from 0.16 to 320.8 km (0.1 to 200 miles) and 16 sectors. In these figu res the d ata are p lotted for rad ii 62

Figure 38. N orm alized Exposure for N on-Dep ositing Material on the 16.1-km (10-mile) Arc.

Figure 39. N orm alized Exposure for N on-Dep ositing Material on the 32.2-km (20-mile) Arc.

63

Figure 40. N orm alized Exposure for N on-Dep ositing Material on the 80.5-km (50-mile) Arc.

Figure 41. N orm alized Exposure for N on-Dep ositing Material on the 160.9-km (100-mile) Arc.

64

Figure 42. N orm alized Exposure for Dep ositing Material on the 16.1-km (10-m ile)

Arc.

Figure 43. N orm alized Exposure for Dep ositing Material on the 32.2-km (20-m ile)

Arc.

65

Figure 44. N orm alized Exposure for Dep ositing Material on the 80.5-km (50-m ile)

Arc.

Figure 45. N orm alized Exposure for Dep ositing Material on the 160.9-km (100-mile) Arc.

66

Figure 46. N orm alized Deposition on the 16.9-km (10-mile) Arc.

Figure 47. N orm alized Deposition on the 32.2-km (20-mile) Arc.

67

Figure 48. N orm alized Deposition on the 80.5-km (50-mile) Arc.

Figure 49. N orm alized Deposition on the 160.9-km (100-mile) Arc.

68

Figure 50. Annual Average Exposure for N on-Dep ositing Sp ecies. Contou r levels are 108, 5x107, 2x107, 107, 5x106, 2x106, 106, 5x105, 2x105, 105, and 5x104 Bq-s/ m 3. Resu lts for each m od el are as ind icated . N ote that RASCAL and RATCH ET only p rovid e d ata w ithin 80.5 km (50 m iles) of the sou rce.

from 0 to 160.9 km (100 m iles), 25 rad ii. RASCAL and RATCH ET d ata are in three grid s:

an inner grid w ith a spacing of 0.8 km (0.5 miles) over a range from -16.1 to +16.1 km

(-10 to +10 m iles), an interm ed iate grid w ith a sp acing of 2 km (1.25 m iles) over a range from -40.2 to +40.2 km (-25 to +25 m iles), and an ou ter grid w ith a sp acing of 4 km 69

Figure 51. Annual Average Exposu re for Dep ositing Sp ecies. Contou r levels are 108, 5x107, 2x107, 107, 5x106, 2x106, 106, 5x105, 2x105, 105, and 5x104 Bq-s/ m 3. Resu lts for each m od el are as ind icated . N ote that RASCAL and RATCH ET only p rovid e d ata w ithin 80.5 km (50 m iles) of the sou rce.

(2.5 m iles) over a range from -80.5 to +80.5 km (-50 to +50 m iles). The d ata for LODI are from the concentration grid w ith a sp acing of 1 km over a range from -175 to +175 km

(-109 to +109 m iles) from the source. Since w e are prim arily interested in the 70

Figure 52. Annual Deposition. Contou r levels are 106, 5x105, 2x105, 105, 5x104, 2x104, 104, 5x103, 2x103, 103, and 5x102 Bq-s/ m 3. Results for each m od el are as ind icated .

N ote that RASCAL and RATCH ET only p rovid e d ata w ithin 80.5 km (50 miles) of the source.

d istribu tion over the 8-160.9 km (5-100 m ile) range, the highest valu es close to the sou rce are not contoured . Also note that RASCAL and RATCH ET p lots cover only a 80.5-km (50-mile) square about the sou rce w hile LODI and MACCS2 go ou t 160.9 km 71

(100 m iles). The sm ooth contou rs in the p lots for MACCS2 are a resu lt of the solu tion techniqu e, the assum ed straight line transp ort, and the w id e sp acing of the d ata points (400 p oints). The LODI figures inclu d e som e high frequ ency noise that is a featu re of m apping parcels to a grid , especially a high-d ensity (closely-sp aced ) concentration grid (122,500 ind ivid ual exposures or d ep ositions are u sed in constru cting the contou r p lots).

RASCAL, RATCH ET, and LODI all show featu res in these annu al averages that app ear to preserve ind ivid ual plum es, and there seem s to be general agreem ent abou t the d irection of these plum es. The RASCAL and RATCH ET d ata are in qu ite close agreement except for the magnitud e of the exp osu re or d eposition. This is exp ected since these mod els are very closely related and the main d ifference is the tu rbu lent d iffusion form u lation. RASCAL and RATCH ET also have isolated d ow nw ind high d eposition contours that are not present in MACCS2 or LODI p lots. These are p resu m ably d ue to rap id w et d eposition w hen rain occu rs several hours after the release. The closer spacing of the contou rs for MACCS2 com p ared w ith LODI, as one m oves aw ay from the release location, is evid ence of the m ore rap id d ecrease of exp osu re and d eposition w ith d istance for MACCS2. In general, the similarities in the d istributions of exposure and d eposition show n by these p lots are greater than the d ifferences, particularly w hen consid eration is given to the d ifferent d ensity (closeness of sp acing) of the und erlying d ata. The m ore com p lex m od els certainly show m ore d etail in structure; how ever, the sm oothed d istribu tion still show the com m on featu res that w e noted in the p revious sections on arc and arc-sector averages.

10.4 Summary of Results All of the arc average and the great m ajority of the arc-sector average exp osures and d ep ositions are w ithin a factor of tw o w hen com p aring MACCS2 to the state-of-the-art m od el, LODI. Sim ilar com parisons of RASCAL and RATCH ET to LODI also have m ost exp osu res and d epositions w ithin a factor of tw o of LODI. In fact the largest d ifferences in results are betw een the closely related RASCAL and RATCH ET m od els.

We can id entify at least tw o caveats to the d iscu ssion of m od el d ifferences. First, this stu d y w as perform ed in an area w ith sm ooth or favorable terrain and p ersistent w ind s althou gh w ith structure in the form of low -level noctu rnal jets and severe storm s. In regions w ith com plex terrain, particu larly if the su rface w ind d irection changes w ith height, cau tion should be used . Second , MACCS2 p red icts a too rap id d ecrease of exp osu re w ith d istance; this should be consid ered w hen MACCS2 is u sed to estim ate consequ ences at d istances greater than 321.8 km (200 m iles). H ow ever, this second caveat is tem pered by the fact that the m ajority of the d ep osition (and exp osu re to d epositing m aterial) is w ithin this 321.8-km (200-m ile) d istance.

72

11. REFEREN CES Band er, T. J. 1982. PA V A N : A n A tmospheric Dispersion Program for Evaluating Design Basis A ccidental Releases of Radioactive M aterials for N uclear Power Stations, N UREG/ CR-2858, U.S. N u clear Regulatory Com m ission, Washington, D.C.

Bonner, W. D., 1968. Clim atology of the Low Level Jet, M on. W ea. Rev., 96, 833-850.

Briggs, G. A. 1984. Plum e Rise and Bu oyancy Effects in A tmospheric Science and Power Production, Ed . D. Rand erson, DOE/ TIC-27601, U.S. Dep artm ent of Energy.

Chanin, D., M. L. Young, and J. Rand all, 1998. Code M anual for M A CCS2, N UREG/ CR-6613.

Delle Monache, L. 2002. A erosol Property Comparison within and above the A BL at the A RM Program SGP Site, M.S. Thesis, San Jose State University, UCRL-LR-146711.

(http :/ / w w w .llnl.gov/ tid / lof/ d ocum ents/ p d f/ 240689.p d f)

EPA, 2000. M eteorological M onitoring Guidance for Regulatory M odeling A pplications, EPA-454/ R-99-005, U.S. Environm ental Protection Agency, Research Triangle Park, N C.

(http :/ / w w w .w ebm et.com / m et_m onitoring/ toc.htm l)

Erm ak, D.L., and J.S. N asstrom , 2000. A Lagrangian Stochastic Diffu sion Method for Inhom ogeneous Turbulence, A tmos. Environ., 34, 7, 1059-1068.

Foster, K.T., G. Sugiyam a, J. S. N asstrom , J. M. Leone, Jr., S. T. Chan, and B. M. Bow en, 2000. The Use of an Operational Mod el Evalu ation System for Mod el Intercom p arison, International Journal of Environment and Pollution, 14, 1-6, 77-88.

Gifford , F. A. 1961. Use of Routine Meteorological Observations for Estimating Atm osp heric Dispersion, N uclear Safety, 2, 4, 47-51.

Gold er, D. 1972. Relations Am ong Stability Param eters in the Su rface layer, Boundary-Layer M eteorology, 3, 1, 47-58.

H anna, S. R., G. A. Briggs, and R. P. H osker. 1982. Handbook on A tmospheric Diffusion.

DOE/ TIC-11223, U.S. Departm ent of Energy, Washington, D.C.

H effter, J. L. 1980. Transport-layer Dep th Calculations, Second Joint Conference on A pplications of A ir-Pollution M eteorology, N ew Orleans, LA.

73

Monin, A. S. and A. M. Obukhov, 1954. Basic Law s of Tu rbu lent Mixing in the Grou nd Layer of the Atm osphere, Trans. Geophys. Inst. A kad. N auk, USSR, 151, 1, 63-87.

N asstrom , J.S., G. Sugiyam a, J.M. Leone, Jr., and D.L. Erm ak, 2000. A Real-time Atm osp heric Dispersion Mod eling System , Preprint, Eleventh Joint Conference on the A pplications of A ir Pollution M eteorology, Long Beach, CA, Jan. 9-14, 2000, Am erican Meteorological Society, Boston, MA.

(http :/ / w w w .llnl.gov/ tid / lof/ d ocum ents/ p d f/ 237149.p d f)

N RC, "Onsite Meteorological Program s," Regu latory Gu id e 1.23, Febru ary 1972.

(Single cop ies of printed regulatory gu id es are available free of charge by w riting the Rep rod u ction and Distribution Services Section, U.S. N u clear Regu latory Comm ission, Washington, DC 20555-0001; by send ing an e-m ail requ est to d istribu tion@nrc.gov; or by faxing to (301) 415-2289.)

Panofsky, H .A., H . Tennekes, D. H . Lenschow , and J. C. Wyngaard , 1977. The Characteristics of Turbulent Velocity Com p onents in the Su rface Layer Und er Convective Cond itions, Boundary-Layer M eteorology 11, 3, 55-61.

Pasqu ill, F. 1961. The Estim ation of the Disp ersion of Wind borne Material, The M eteorological M agazine, 90, 33-49.

Ram sd ell, Jr., J.V., C. A. Sim onen, and K. W. Bu rk, 1994. Regional A tmospheric Transport Code for Hanford Emission Tracking (RA TCHET), PN WD-2224 H EDR, Battelle, Pacific N orthw est Laboratories, Richland , Washington. Available throu gh the N ational Technical Inform ation Service, (http :/ / w w w .ntis.gov), ord er nu m ber: DE94006449.

Ram sd ell, J. V., G. F. Athey, T. J. Band er, R. I. Sherp elz, 1988. The M esorad Dose A ssessment M odel Computer Code, N UREG/ CR-4000 Vol. 2, U.S. N u clear Regu latory Com m ission, Washington, D.C.

Sagend orf, J. F., J. T. Goll, and W. F. Sand u sky, 1982. X OQDOQ: Computer Program for the Routine M eteorological Evaluation of Routine Effluent Releases an N uclear Power Stations, N UREG/ CR-2919, U.S. N uclear Regu latory Com m ission, Washington, D.C.

Seinfeld , J. H . 1986. A tmospheric Chemistry and Physics of A ir Pollution, John Wiley &

Sons, N ew York.

Sherp elz, R. I., T. J. Band er, G. F. Athey, J. V. Ram sd ell, 1986. The M esorad Dose A ssessment M odel V ol. 1: Technical Basis, N UREG/ CR-4000 Vol. 1, U.S. N u clear Regu latory Com m ission, Washington, D.C 74

Ship ler, D.B., B.A. N ap ier, W. T. Farris, and M.D. Freshley. 1996. H anford Environm ental Dose Reconstruction Project-An Overview . Health Physics 71, 4, 532-544.

Slinn, W. G. N . 1984. Precipitation Scavenging, in A tmospheric Science and Power Production, Ed . D. Rand erson, DOE/ TIC-27601, U.S. Dep artm ent of Energy.

Sjoreen, A. L., J. V. Ram sd ell, Jr., T. J. McKenna, S. A. McGuire, C. Fosm ire, G. F. Athey.

2001. RA SCA L 3.0: Description of M odels and M ethods, N UREG-1741. U.S. N u clear Regulatory Com m ission, Washington, D.C.

Su giyam a, G., and S. T. Chan, 1998. A N ew Meteorological Data Assim ilation Mod el for Real-Tim e Em ergency Response, Preprint, 10th Joint Conference on the A pplications of A ir Pollution M eteorology, Phoenix, AZ (11-16 Janu ary, 1998), Am . Met. Soc., Boston, MA.

285-289. ( http:/ / w w w .llnl.gov/ tid / lof/ d ocum ents/ p d f/ 232515.p d f)

Turner,. D. B. 1964. A Diffusion Mod el for an Urban Area, J. of A pplied M eteorology, 3, 1, 83-91.

Zilitinkevich, S.S. 1972. On the Determ ination of the H eight of the Ekm an Bou nd ary Layer, Boundary-Layer M eteorol., 3, 141-145.

75

76 APPEN D IX A - CON TEN TS OF D ATA CD The CD includ ed w ith this report includ es the follow ing:

1) FORTRAN source cod es used to read ARM d ata from the netCDF files and provid e input files of meteorological observations for u se by MACCS2, RASCAL/ RATCH ET, and ADAPT/ LODI,

2) FORTRAN source cod es u sed to convert LODI grid d ed d ata to arc-sector exposure and d eposition and to calcu late the annu al w eighted average,

3) C shell scripts that cycled throu gh the 610 cases, ru nning ADAPT to p rovid e the w ind field s and LODI to p rod u ce the integrated exp osu re and d eposition,

4) IDL proced ures used for qu ality control analysis and m od ification of ADAPT input m eteorological observation files,

5) som e sam ple ADAPT and LODI inp u t and ou tp u t files,

6) som e sam ple ind ivid ual case and annu al average arc-sector concentration files,

7) meteorological d ata files p rovid ed to MACCS2 and RASCAL/ RATCH ET,

8) hourly m ixing heights and scavenging rates u sed by ADAPT/ LODI,

9) the list of 610 w eather trials and their associated w eights,

10) MATLAB m -files used for 2D exp osure and d ep osition p lots,

11) C shell script used to cycle throu gh the smoothing of LODI exp osures and d epositions and a sam p le inp u t file,

12) sample sm oothed exposure and d ep osition d ata,

13) file lists for each type of arm d ata.

Many of the source cod es and scrip ts locate files w ithin the sp ecific d irectory stru ctu re that w as u sed at LLN L for this project. The CD u ses this sam e d irectory stru ctu re to organize the files. On the com puter u sed for this p roject the d irectory stru ctu re given below is a subd irectory in m y hom e d irectory, / u / cm ole/ nrc. The files w ere transferred d irectly from a UN IX system so the files m ay not be comp letely comp atible w ith a Wind ow s PC. The subd irectory structu re is as follow s:

ad ap td ata - observed m eteorology ASCII inp u t files for ADAPT ad ap tru n - ADAPT ru n d irectory (ad ap t_ru n.scr, inp u t nam elist file, ADAPT ou tp u t log file arm d ata - ARM d ata d irectory (no files) 60w pd n - N OAA w eather p rofile netCDF d ata files 915rw p - 915 MH z ARM p rofiler netCDF d ata files aeri - AERI d ata d irectory (no files)

B1 - AERI netCDF d ata files for H illsboro, KA B4 - AERI netCDF d ata files for Vici, OK B5 - AERI netCDF d ata files for Morris, OK 77

B6 - AERI netCDF d ata files for Pu rcell, OK C1 - AERI netCDF d ata files for Central Facility okm - Oklahom a Mesonet netCDF d ata files sm os01 - SMOS netCDF d ata files for Larned , KA sm os03 - SMOS netCDF d ata files for Le Roy, KA sm os04 - SMOS netCDF d ata files for Plevna, KA sm os05 - SMOS netCDF d ata files for H alsted , KA sm os06 - SMOS netCDF d ata files for Tow and a, KA sm os07 - SMOS netCDF d ata files for Elk Falls, KA sm os08 - SMOS netCDF d ata files for Cold w ater, KA sm os09 - SMOS netCDF d ata files for Ashton, KA sm os11 - SMOS netCDF d ata files for Byron, OK sm os13 - SMOS netCDF d ata files for Central Facility (Lam ont, OK) sm os15 - SMOS netCDF d ata files for Ringw ood , OK sm os20 - SMOS netCDF d ata files for Meeker, OK sm os21 - SMOS netCDF d ata files for Okm u lgee, OK sm os24 - SMOS netCDF d ata files for Cyril, OK sm os25 - SMOS netCDF d ata files for Sem inole, OK sond e - Sond e netCDF d ata files ascond ata - Arc-sector concentration d ata for each of 610 cases cod es - Directory for m ost of the cod es (no files) arac2m accs - Fortran sou rce cod es for converting resu lts from rectangu lar grid to arc-sectors and averaging resu lts over 610 cases arm 2arac - Fortran sou rce cod es for read ing ARM d ata and p rep aring ADAPT observed m eteorology d ata files qc - IDL sou rce cod es for d isp laying and correcting observed m eteorology d ata files grid gen - ADAPT/ LODI grid netCDF files lod iou t - LODI inp u t nam elist files and ou tp u t files (ru n log, p article p osition, exp osu re, d eposition for each case) lod iru n - LODI ru n d irectory (ru n_cases.scr, lod i_run.scr, incrd at8, jd 2m d , WeatherTrials.txt, scav_rate.txt) resu lts - MATLAB sou rce cod es for 2D concentration p lots sm oothd ata - Slightly sm oothed grid d ed exp osu re and d ep osition netCDF files sm oothing - Sm oothing ru n d irectory (sm ooth_cases.scr, inp u t file) su rfaced ata - Surface d ata files sent to PN L for RASCAL/ RATCH ET and Sand ia for MACCS2 78