Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2009

Developing a Unified Superset inQuantifying Ambiguities Among TropicalCyclone Best Track Data for the WesternNorth PacificMichael Robert Lowry

Follow this and additional works at the FSU Digital Library. For more information, please contact lib-ir@fsu.edu

FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

DEVELOPING A UNIFIED SUPERSET IN QUANTIFYING AMBIGUITIES AMONG

TROPICAL CYCLONE BEST TRACK DATA FOR THE WESTERN NORTH PACIFIC

By

MICHAEL ROBERT LOWRY

A Thesis submitted to the

Department of Meteorology

in partial fulfillment of the

requirements for the degree of

Master of Science

Degree Awarded:

Spring Semester, 2009

Copyright © 2009

Michael Robert Lowry

All Rights Reserved

ii

The members of the Committee approve the Thesis of Michael Robert Lowry defended on

November 26th, 2008.

___________________________________

James J. O’Brien

Professor Directing Thesis

___________________________________

Jon E. Ahlquist

Committee Member

___________________________________

Mark A. Bourassa

Committee Member

___________________________________

Robert E. Hart

Committee Member

The Office of Graduate Studies has verified and approved the above named committee members.

iii

ACKNOWLEDGMENTS

This research was funded by the NOAA ARC. Sincere gratitude is extended to Dr. James

J. O’Brien for opportunities provided professionally by him and growth experienced personally

from him. The latitude he granted to me in my research shaped the topic of this paper. Further

credit and appreciation is extended to Melissa Griffin and to the late Gary Watry of COAPS/FSU

for their invaluable contributions to this collaborative effort. Additional thanks to Buck Sampson

of NRL Monterey, John Knaff of NOAA/NESDIS, Brad Harris from JTWC, Chris Landsea of

NOAA/NHC, Hirotaka Kamahori and Akihiro Kikuchi of JMA, W.H. Lui from HKO and Li

Qingqing from STI for providing insight into the various best track data sets. I also wish to

acknowledge the keen guidance and intellectual counsel provided by my committee members:

Dr. Mark Bourassa, Dr. Robert Hart, and Dr. Jon Ahlquist. I owe a debt of gratitude to Ben

Nelson, Matthew Green, and Craig Fugate for creating a place for me at FDEM from my years in

college through my studies in graduate school. I will never forget the friends made and once in a

lifetime experiences had during my time with the division. I direct a special thanks to Jessica

Fieux, Michelle Stewart, Rebecca Smith, and my colleagues at COAPS/FSU for their continual

encouragement, suggestions, and support. Above all, I thank my family for the wonderful life

given to me.

iv

TABLE OF CONTENTS

List of Tables ......................................................................................................v

List of Figures.....................................................................................................vi

List of Abbreviations............................................................................................x

Abstract ..............................................................................................................xi

1. INTRODUCTION............................................................................................1

2. DATA AND METHODOLOGY......................................................................5

Data..........................................................................................................5

Basin Area of Responsibility ....................................................................5

Wind Speed Averaging Standards.............................................................6

Developing a Unified Western North Pacific Superset ..............................6

3. RESEARCH AND RESULTS........................................................................14

Position Ambiguities ..............................................................................14

Intensity Differences...............................................................................18

The Role of Dvorak’s Intensity Estimation Technique ............................19

The Reliability of Landfall Data and Extratropical Cyclones...................23

4. CONCLUSIONS............................................................................................43

APPENDIX A....................................................................................................46

APPENDIX B....................................................................................................50

REFERENCES ................................................................................................132

BIOGRAPHICAL SKETCH............................................................................137

v

LIST OF TABLES

2.1 AOR statistics from each of the four original WNP best track data sets

examined in this study. The percentage of points included in each center’s defined

AOR for the entire period of record is given in the final column. Because STI

provides no northern latitudinal or western longitudinal bounds on its data, AOR

statistics are supplemented by those when longitudes less than 100°E are

considered outside of its AOR......................................................................................... 9

3.1 The empirical relationship between current intensity (CI) number and

maximum sustained winds in knots. The Dvork, 1975, Dvorak, 1984, Koba, 1990,

and revised JMA CI/MSW relationships are shown....................................................... 30

vi

LIST OF FIGURES

1.1 A timeline of best track TC data implemented in this study ....................................... 4

2.1a TC area included in JTWC’s WNP AOR, indicated by shaded region.

Superimposed is JTWC’s original base of operations at Nimitz Hill, Guam. JTWC

moved to Pearl Harbor, Hawaii, on 1 January 1999 ......................................................... 9

2.1b TC area included in JMA’s AOR, indicated by shaded region. Superimposed

is JMA’s base of operations in Tokyo, Japan................................................................. 10

2.1c TC area included in STI’s AOR, indicated by shaded region. Superimposed

is STI’s location in Shanghai, China.............................................................................. 11

2.1d TC area included in HKO’s AOR, indicated by shaded region. Area enclosed

by solid lines is the AOR from 1961-1984. The AOR is expanded in 1985 to

include the area bounded by the dashed lines. Superimposed is HKO’s base of

operations in Hong Kong .............................................................................................. 12

2.2 Track for Storm 32 of 1989 from JTWC best track data, with intensity

indicated by shaded track. The system forms in the South China Sea and

continues into the Bay of Bengal, where it makes landfall along the east coast of

Andhra Pradesh, India, at Category 5 intensity. The full track appears in JTWC

WNP best track data, even though over two-thirds of the position estimates fall in

an area traditionally regarded as the North Indian Ocean (i.e., west of 100°E) ............... 13

3.1 Annual average absolute distance between position estimates in JTWC and

JMA best track data. Dashed line indicates raw annual averages and solid line

indicates smoothed averages after a 1-3-4-3-1 filter is applied. Overlaid are

significant developments in the TC observing system.................................................... 30

3.2 Annual mean number of WNP aircraft reconnaissance levied fixes and

investigations per TC, according to JTWC best track data. Red bars indicate

notable shifts in the annual mean: (1962-63) Advent of satellites allows

previously undetected open ocean TCs to be reconnoitered; (1972-73) SRP begins

using satellite to position TC fixes; (1987) Dedicated aircraft reconnaissance is

deactivated in the WNP................................................................................................. 31

3.3 Annually accumulated, unmodified PDI for the WNP from the unified WNP

Superset. Dashed lines indicate raw PDI and solid lines indicate smoothed PDI

after a 1-2-1 filter is applied. JTWC PDI is shown in red and JMA PDI in blue. R

for the entire period (1977-2006) is 0.83 ....................................................................... 31

vii

3.4 From Fig. 9 of Guard (1992), which shows the percent of warnings based

primarily on aircraft reconnaissance and satellite reconnaissance. Since some

estimates are based on terrestrial radar or extrapolation, the percent based on

satellite never reaches 100%.......................................................................................... 32

3.5 Annually accumulated, modified PDI for the WNP from the unified WNP

Superset. Dashed lines indicate raw PDI and solid lines indicate smoothed PDI

after a 1-2-1 filter is applied. JTWC PDI is shown in red and JMA PDI in blue.

MSW from JMA best track data for the period 1990-2006 are converted from the

1990 Koba et al. modified Dvorak table to the original 1984 Dvorak table for

MSW > 33 ms-1 according to equation (3.2). R for the entire period (1977-2006)

is 0.95 ........................................................................................................................... 32

3.6 Analysis diagram outlining the Dvorak intensity estimation technique

designed for use with visible satellite imagery from Dvorak, 1984 ................................ 33

3.7 Analysis diagram outlining the Dvorak intensity estimation technique

designed for use with enhanced infrared satellite imagery from Dvorak, 1984............... 34

3.8 Annual mean percent of points excluded from JTWC (red), JMA (blue), and

STI (green) Superset data from original best track data. Dashed lines indicate

unsmoothed data. Solid lines indicate smoothed data after a 1-2-1 filter is applied ........ 35

3.9 Regional Basic Synoptic Networks (RBSN) of global surface and upper air

stations from the World Weather Watch Programme of the World Meteorological

Organization. Station data are current as of 01 October 2008 (WMO, 2008).

Higher density coverage is found in mainland locations along the coast, with a

much sparser fixed surface network noted on islands and atolls dotting the ocean

basins............................................................................................................................ 35

3.10 Same as Fig 3.9 but cropped and enlarged to show networks across the WNP

basin ............................................................................................................................. 36

3.11 Spatial distribution of mean absolute distance between position estimates in

JTWC and JMA best track data from 1951 through 1969 from the WNP Superset.

Mean cell values are computed from all points contained within 2° x 2° cells and

binned within four classes set at approximately one and two standard deviations

from the sample mean ................................................................................................... 36

viii

3.12 JTWC (red) and JMA (blue) tracks for Storm 64 from the unified WNP

Superset (Storms 20 and 19 of 1953 from JTWC and JMA, respectively). Month

and day are labeled for 00Z estimates along with 1-minute average MSW in

parenthesis for JTWC data. Bullet points indicate 12-hour position estimates.

JTWC records three 6-hourly position estimates over Hainan, with one MSW

estimate as high as 110 kt. JMA track records no points over Hainan in its 6-

hourly best track data. Additionally, at 00Z on 2 November, position estimates

between the two agencies differ by nearly 250 km, such that at 00Z JTWC records

an 80 kt TC over the Gulf of Tonkin while JMA has a moderate TC inland over

Nghệ, a province along the north central Vietnam coast. Differences are also

found between landfall numbers in the Philippines. ....................................................... 37

3.13a Spatial distribution of mean absolute difference between intensity estimates

in JTWC and JMA best track data from 1977 through 1989 from the WNP

Superset. JMA MSW are converted to a 1-minute average using a 1.14 10-minute

to 1-minute conversion factor. Mean cell values are computed from all points

contained within 2° x 2° cells and binned within four classes set at approximately

one and two standard deviations from the sample mean................................................. 38

3.13b Spatial distribution of mean absolute difference between intensity estimates

in JTWC and JMA best track data from 1990 through 2006 from the WNP

Superset. JMA MSW are converted to a 1-minute average using a 1.14 10-minute

to 1-minute conversion factor. Mean cell values are computed from all points

contained within 2° x 2° cells and binned within four classes set at approximately

one and two standard deviations from the sample mean................................................. 39

3.14 Composite bar chart of annually accumulated unfiltered cubic MSW from

points over land in JTWC and JMA best track data from the WNP Superset.

JTWC aggregate is shown in red and JMA aggregate is shown in blue. Landfall

climatology includes only points captured over land in JTWC, JMA, and STI

Superset data. Annually accumulated cubic winds associated with points over

land from JTWC best track data exhibit an increasing trend significant at the 0.05

level. JMA data indicate a downward trend not significant at the 0.05 level................... 40

3.15 Spatial distribution of mean absolute distance between position estimates in

JTWC and JMA best track data from 1977 through 2006 from the WNP Superset.

Mean cell values are computed from all points contained within 2° x 2° cells and

binned within four classes set at approximately one and two standard deviations

from the sample mean ................................................................................................... 41

ix

3.16 Composite bar chart of annually accumulated unfiltered cubic MSW from

landfalling typhoons in JTWC and JMA best track data from the WNP Superset.

JTWC aggregate is shown in red and JMA aggregate is shown in blue.

Landfalling typhoon climatology includes only points captured over land of

typhoon intensity in JTWC, JMA, and STI Superset data. Annually accumulated

cubic winds associated with JTWC landfalling typhoons exhibit an upward trend

not significant at the 0.05 level. JMA data indicate a downward trend not

significant at the 0.05 level............................................................................................ 42

x

LIST OF ABBREVIATIONS

ACE Accumulated Cyclone Energy

AOR Area of Responsibility

APT Automated Picture Transmission

CMA China Meteorological Administration

DAPP Data Acquisition and Processing Program

DMSP Defense Meteorological Satellite Program

DOD Department of Defense

ESCAP Economic and Social Commission for Asia and the Pacific

GCM Global Climate Model

GMS Geostationary Meteorological Satellite

HKO Hong Kong Observatory

HURDAT Atlantic Basin Hurricane Database

JMA Japanese Meteorological Agency

JTWC Joint Typhoon Warning Center

MSLP Minimum Sea Level Pressure

MSW Maximum Sustained Winds

NOAA U.S. National Oceanic and Atmospheric Administration

NPMOCW Naval Pacific Meteorology and Oceanography Center West

PDI Power Dissipation Index

PMW Passive Microwave

QuikSCAT NASA’s Quick Scatterometer

RA Regional Association

RSMC Regional Specialized Meteorological Center

SRP Selective Reconnaissance Program

STI Shanghai Typhoon Institute

TC Tropical Cyclone

TRMM Tropical Rainfall Measuring Mission

TMI Tropical Rainfall Measuring Mission Microwave Imager

WMO World Meteorological Organization

WNP Western North Pacific

WWW World Weather Watch

xi

ABSTRACT

In the western North Pacific basin, several agencies archive “best track” data of tropical

cyclones. The Joint Typhoon Warning Center (JTWC) in Hawaii is responsible for the issuance

of tropical cyclone warnings for United States Department of Defense interests and has a record

of tropical cyclones extending back to 1945. The Japanese Meteorological Agency (JMA) is the

World Meteorological Organization (WMO) official Regional Specialized Meteorological Center

(RSMC) for the western North Pacific basin and has best track tropical cyclone data extending

back to 1951. The Shanghai Typhoon Institute (STI) of the Chinese Meteorological

Administration and the Hong Kong Observatory (HKO) of the Government of the Hong Kong

Special Administrative Region also have 6-hourly tropical cyclone data records from 1949 and

1961, respectively.

Western North Pacific (WNP) data sets are investigated in order to quantify ambiguities

in position and intensity estimates among the forecast institutions through the development of a

unified Superset. Ambiguities among the two primary warning centers (JMA and JTWC) are

presented in the context of a changing observation network, observational tools, and analysis

techniques since the beginning of tropical cyclone records. Mean differences in position

estimates are found between the two centers on the order of 60 km prior to the introduction of

meteorological satellites in 1961 and near 50 km following the deactivation of aircraft

reconnaissance in 1987. Results show a step function change among intensity in JTWC and JMA

best track data from 1989 to 1990 due to varying applications of the Dvorak intensity estimation

technique. Parsing best track data into landfall subsets does not ameliorate interagency

differences in position or intensity estimates. Additionally, analyses from Superset data call into

question the veracity of JTWC best track data during the period from 1995-1999. The

applicability of adopting an individual data set in discerning long term climate trends is

examined in light of these differences. Past efforts to analyze, assemble, and maintain a

complete, reliable best track tropical cyclone data set for the WNP are discussed among topical

methods of incorporating the Superset within a basin-wide re-analysis.

1

CHAPTER 1

INTRODUCTION

The occurrence of five landfalling hurricanes in the Atlantic basin in 2004, 10 landfalling

tropical cyclones in Japan in 2004, the most active Atlantic hurricane season on record in 2005,

the incidence of Hurricanes Katrina and Rita in 2005, and the formation of the most intense

hurricane on record in the Atlantic in 2005 have fueled a media frenzy surrounding the

influences of climate change upon global tropical cyclone (TC) activity. Empirical studies

(Trenberth, 2005; Emanuel, 2005) suggest an upward trend in TC accumulated energy and power

across the Atlantic and Pacific Ocean basins. Observations also indicate a sharp rise in the

number and percentage of intense TCs occurring globally since 1970 (Webster, 2005). More

recent papers (Klotzbach, 2006; Kamahori, 2006; Wu, 2006; Landsea, 2007) call into question

the veracity of historical TC best track data, citing inconsistencies within the TC observation

network and evolving analysis techniques for upward trends in intensity metrics. The present

state of modeling and theory remains irresolute to TC projections and is largely inconsistent with

empirical studies (Knutson and Tuleya, 2004; Vecchi and Soden, 2007). The observational data

has thus become a vital tool in discerning long-term variability of TCs, particularly within the

context of global climate change.

In the North Atlantic, historical TC best track data (HURDAT; Jarvinen et al., 1984;

Neumann et al., 1993) is available for the entire basin from the mid-19th Century through the

present. A rigorous re-analysis project to extend and correct HURDAT, led by the Hurricane

Research Division (HRD) of NOAA’s Atlantic Oceanographic and Meteorological Laboratory

(AOML), began in the late 1990s and continues as of this writing (Landsea et al., 2004). All

recommended changes to the original HURDAT data set are submitted by the re-analysis team to

the National Hurricane Center’s (NHC) Best Track Change Committee for approval. Thus far the

Best Track Change Committee has approved over 5000 revisions to HURDAT from the years

1851 to 1910, resulting in arguably the most accurate historical TC data set in the world

(Landsea et al.). Systematic errors and biases remain in HURDAT, however, especially in those

years prior to the advent of satellites (Landsea, 2007). Reconciling missed or undersampled

storms may prove impossible in areas of sparse or absent in-situ data. Nevertheless, HURDAT is

2

useful for a wide variety of purposes, including climate studies, once its proper caveats are

considered.

In the western North Pacific (WNP), generally the most active TC basin in the world, no

such organizational review of best track TC data has yet occurred. Several circumstances account

for this absence.

Four major historical TC data sets cover the basin. The Joint Typhoon Warning Center

(JTWC), the U.S. Department of Defense (DOD) agency responsible for the issuance of tropical

forecast warnings for the Pacific and Indian Oceans, houses an historical TC record dating back

to 1945. The Regional Specialized Meteorological Center (RSMC) Tokyo, one of six global

RSMCs created under the framework of the World Weather Watch (WWW) Program of the

World Meteorological Organization (WMO), was established in July 1989 at the Headquarters of

the Japanese Meteorological Agency (JMA) and provides TC advisories for the WNP. RSMC

Tokyo (hereinafter referred to as JMA) has best track TC data available for the basin from 1951

to the present. The Shanghai Typhoon Institute (STI) of the China Meteorological

Administration (CMA) and the Hong Kong Observatory (HKO) of the Government of the Hong

Kong Special Administrative Region also host 6-hourly historical TC data for the WNP from

1949 and 1961, respectively (see Fig 1.1).

In November 2001 a workshop co-sponsored by the WMO and U.S. National Weather

Service gathered representatives from nine countries to discuss the feasibility and desirability of

developing a WNP unified best track data. Participants, including researchers and insurers,

concluded that great value and a significant need exist for such a data set. A resolution was

drafted supporting the development of a unified best track data set and was later accepted by the

United Nations Economic and Social Commission for Asia and the Pacific (ESCAP)/WMO

Typhoon Committee (R. Murnane, personal communication, 2007). The Committee formed a

working group to develop a plan for assembling and maintaining the data set, but the working

group was dissolved shortly after its inception, as member countries were unable to find

adequate resources to fund the ambitious project (R. Murnane, personal communication). The

result is the continued use of multiple data sets in the WNP for verification, climate variability,

and extreme event risk studies, with conclusions shown to vary based upon the chosen set of

observations (e.g., Wu, 2006).

3

The goal of this study is to develop a unified TC historical record from three WNP best

track archives. HKO best track data are omitted from the initial run of the Superset because its 6-

hourly record is at least ten years shorter than any other set examined. In contrast to an integrated

best track set, the unified Superset contains three subsets of position and intensity data. The

Superset is not a re-analysis, but rather a repository of multiple position and intensity estimates

for TC points recorded across existing data sets.

An examination of historical TC data is conducted utilizing the newly developed WNP

Superset for the years 1951-2006. Ambiguities in track and intensity among the various TC

archives are assessed in the context of changing observational networks and analysis techniques

over the past 56 years. JTWC and JMA best track data, arguably the two most widely used

historical WNP data sets, manifest mean differences in track estimates on the order of 60 km

prior to the introduction of meteorological satellites in 1961 and near 50 km following the

deactivation of aircraft reconnaissance in 1987. Careful consideration to the applied Dvorak

technique after 1987 at WNP TC institutions must be taken when analyzing long term trends in

TC intensity. Results show a step function change among intensity in JTWC and JMA best track

data between 1989 and 1990. Revisions to the 1984 Dvorak table at RSMC Tokyo in 1990 likely

account for these differences. Remaining differences in intensity are attributed to

inhomogeneities in wind averaging standards and varying operational procedures. A cursory look

at regime shifts at JTWC during the 1990s reveals possible data quality issues in its best track

record during the period. Finally, Superset data parsed to landfall subsets do not ameliorate

discrepancies among best track records across the WNP.

The ultimate objective of this research is to reassert the salient need for a re-analysis of

historical WNP TCs. When examined together, best track sets across the basin yield significant

ambiguities in intensity and track data that rise and fall over time. Wind-pressure relationships,

aircraft reconnaissance, satellite tools, and operational practices are only part of a stew of

inconsistencies that will remain in empirical studies until they are fully addressed through a

thorough basin-wide re-analysis.

4

Fig 1.1: A timeline of best track TC data implemented in this study.

5

CHAPTER 2

DATA AND METHODOLOGY

Data

Four independent and comprehensive WNP TC best track data sets (JTWC, JMA, STI,

and HKO) are incorporated into this study. “Best track” is the best assessment of TC path and

intensity after all available post-storm data are reviewed. These data include 6-hourly estimates

(0000, 0600, 1200, 1800 UTC) of TC location (latitude/longitude) and intensity, and may include

other information such as storm name, wind radii, and development stage.

Intensity estimates are reported in terms of minimum sea level pressure (MSLP),

maximum sustained winds (MSW), or both. JTWC, STI, and HKO archive MSW for their entire

respective periods of record. JMA has a limited record of MSW for TCs from 1977 onward. In

addition to MSW, JMA, STI, and HKO also keep estimates of MSLP for all recorded systems

over their entire respective periods of record. JTWC begins its record of MSLP in 2001.

Data vary temporally and spatially among best track sets. Temporal variations are

discussed in the previous section (Fig 1.1). Differences in domain and wind averaging standards

are highlighted below.

Basin Area of Responsibility

Each international TC center in the WNP defines a domain, or basin area of responsibility

(AOR), across which historical TC data are recorded. The AOR is especially important at

warning agencies where TC forecasts are routinely issued. The AOR at JTWC and its

predecessors has always included the WNP, which encompasses the area north of the equator,

from 100°E to 180° (Fig 2.1a). Best track data from JTWC includes several instances of TCs

occurring outside of its prescribed WNP AOR (Table 2.1). In these cases, a TC either crossed

from (into) the central or eastern North Pacific into (from) the WNP or from (into) the WNP into

(from) the Bay of Bengal or Arabian Sea. Basin AOR discrepancies within best track data should

be carefully considered, as intense TCs outside of the AOR may appear in best track data. For

example, multiple typhoons of category four strength or higher on the regional Saffir-Simpson

hurricane classification scale are found in the JTWC WNP best track set at least 10°

longitudinally outside of its defined AOR (see Fig 2.2).

6

Like JTWC, STI best track TC data for the WNP cover a broad AOR north of the

equator, from 180° westward, with no defined western boundary (Fig 2.1c). A comparison of the

complete best track set, however, manifests nearly half as many points west of 100°E in STI data

than in JTWC best track basin data (Table 2.1). According to Table 2.1, of all four WNP best

track data sets examined, STI maintains the highest total percentage of data points within its

prescribed AOR, covering approximately 99.8% of all best track data. This percentage is

essentially unaltered when the longitudinal domain is reduced to match the AOR covered by

JTWC (i.e., excluding the area west of 100°E).

The JMA restricts its best track data to the region from the equator to 60°N, with a

longitudinal domain from 180° to 100°E (Fig 2.1b). HKO similarly defines a more restricted

AOR for the WNP. Its original AOR, from 100°E to 160°E and north of the equator to 45°N,

expands in 1985 to include the remaining portion of the WNP from 160°E to 180° (Fig 2.1d).

While HKO best track set has the lowest total percent of data points within its defined AORs,

comprising approximately 96.8% of all data, less than a tenth of those points found outside of the

AORs occur east of 180° or west of 100°E (Table 2.1). The remaining points outside of the

AORs fall poleward of 60°N, or east of 160°E prior to 1985.

Wind Speed Averaging Standards

WMO guidelines hold that all participating countries in Regional Associations (RA)

outside of RA IV (North America, Central America and the Caribbean) must average mean

surface wind over a 10-minute period. JMA, as an RSMC, and HKO, as a WMO designated TC

warning center, both adhere to this standard. JTWC, neither an official member nor participant of

the WMO, utilizes a 1-minute “sustained” wind in reporting TC intensity per the U.S. National

Hurricane Operations Plan. STI, not directly affiliated with the WMO, employs the CMA local

standard 2-minute averaging period for recording TC intensity. More specific issues related to

wind speed averaging standards are discussed in Section 3.

Developing a Unified Western North Pacific Superset

All four data sets are run through an initial quality control, consisting of conversion from

the original source format to a homogenous format, with a series of automated quality control

tests utilized to remove duplicate position/intensity data. As an example, beginning in 2004

7

JTWC includes, in addition to 34kt wind radii, 50kt, 64kt, and 100kt wind radii in its best track

files. These data are recorded as separate entries, with duplicate times, positions, and intensities.

The initial quality control removes from original JTWC best track data nearly 2000 of these

duplicates.

In order to quantify absolute differences in position and intensity estimates among the

four data sets, only 6-hourly points found in all best track sets need be included for this study. In

short, a 6-hourly point from one data set that has no match in one of the other three data sets does

not allow for direct interagency comparisons. Such discrepancies may stem from temporal

variations, AOR differences, forecaster discretion, evolving practices, conflicting observational

networks, or simply incomplete data. Employing this method systematically reduces original best

track data. This is especially apparent as new data sets are introduced, since the probability of

encountering an unsuccessful match in at least one other set is augmented with each inclusion.

Considerable data are lost when best track sets with relatively small AORs on fine time

scales are included. HKO houses a much shorter data record than the other three data sets

examined. Additionally, tangible differences remain between the area covered by HKO’s AOR

and those monitored by other WNP TC centers. For these reasons and for sake of brevity, HKO

best track data are omitted from the unified Superset. A perfunctory set is compiled to include all

four best track records; however, this is constructed only as a supplement to the Superset. Unlike

the Superset, these data are neither corrected nor robust.

A quick sweep through the remaining three best track sets (JTWC, JMA, and STI)

yielded a 3.2° latitude/longitude approximation for the maximum absolute difference between

corresponding position estimates. JTWC position estimates are objectively matched by date and

time to position estimates in JMA and STI data sets. If a match has a position difference greater

than 3.2° latitude or longitude, and less than 7° latitude/longitude, the match is flagged for

manual inspection. Multiple matches from individual data sets are also flagged for manual

inspection. All matches are binned to form a unified WNP Superset. Points for which no matches

are found or for which only one match is found are excluded.

Each storm match is assigned a total storm number for the Superset. The entire Superset

is further investigated to ensure no breaks within individual storm tracks. Those tracks with gaps

greater than 6 hours are manually reviewed for completion. In some cases, a system will

dissipate and re-form in one data set, while maintaining its circulation in another. In these

8

instances, the break is permitted, but a flag value is assigned to the 6-hourly storm entries

following the break. Total storm number remains unchanged before and after the break, with the

flag value indicating a potential discrepancy.

Difficulty in determining a purely objective method of matching points across all data

sets resulted in a laborious manual review. Close to 100 reviews, with upward of 800 manual

edits, were made to the Superset after all objective methods were exhausted. Given that the

Superset contains nearly 40,000 matches from 1951 through 2006, the objective technique

proved quite rigorous in its inception. All manual edits were logged and are supplied in

Appendix B.

The final unified WNP Superset includes position and intensity estimates at 6-hourly

(0000, 0600, 1200, 1800 UTC) steps from each of the three WNP best track sets included

(JTWC, JMA, and STI). Subsidiary information which vary according to data set, such as storm

name and wind radii, are not included in the Superset. JMA and STI best track data include a

column for development stage. This indicator is incorporated into the Superset for purposes of

interagency intensity comparisons across multiple wind averaging standards. MSW and MSLP

are included for each set for the periods over which those data are available. A complete

description and data format is provided with the Superset.

9

Table 2.1: AOR statistics from each of the four original WNP best track data sets examined in this study. The

percentage of points included in each center’s defined AOR for the entire period of record is given in the final

column. Because STI provides no northern latitudinal or western longitudinal bounds on its data, AOR

statistics are supplemented by those when longitudes less than 100°E are considered outside of its AOR.

Best Track

Data Set Total Points

Points within

AOR

Points

outside AOR

Points West

of 100°E

Points East of

180°

Ratio

(Points within AOR/Total Points) %

JTWC 51845 51343 502 173 329 99.03%

JMA 54777 53539 362 58 275 97.74%

STI 56474 56396* 78

** 24 78 99.86%

***

HKO 32487 31466 1021 79 13 96.86%

* 56372 when longitudes less than 100°E are considered outside the AOR

**102 when longitudes less than 100°E are considered outside the AOR

***99.82% when longitudes less than 100°E are considered outside the AOR

Fig 2.1a: TC area included in JTWC’s WNP AOR, indicated by shaded region. Superimposed is JTWC’s

original base of operations at Nimitz Hill, Guam. JTWC moved to Pearl Harbor, Hawaii, on 1 January 1999.

10

Fig 2.1b: TC area included in JMA’s AOR, indicated by shaded region. Superimposed is JMA’s base of

operations in Tokyo, Japan.

11

Fig 2.1c: TC area included in STI’s AOR, indicated by shaded region. Superimposed is STI’s location in

Shanghai, China.

12

Fig 2.1d: TC area included in HKO’s AOR, indicated by shaded region. Area enclosed by solid lines is the

AOR from 1961-1984. The AOR is expanded in 1985 to include the area bounded by the dashed lines.

Superimposed is HKO’s base of operations in Hong Kong.

13

Fig 2.2: Track for Storm 32 of 1989 from JTWC best track data, with intensity indicated by shaded track.

The system forms in the South China Sea and continues into the Bay of Bengal, where it makes landfall along

the east coast of Andhra Pradesh, India, at Category 5 intensity. The full track appears in JTWC WNP best

track data, even though over two-thirds of the position estimates fall in an area traditionally regarded as the

North Indian Ocean (i.e., west of 100°E).

14

CHAPTER 3

RESEARCH AND RESULTS

This section is broken up into several parts. First, a direct comparison of position

estimates from JTWC and JMA best track data of the WNP Superset are presented. Ambiguities

are ascertained and discussed as a diagnostic framework for latent issues within WNP best track

data. Next, an analysis of intensity data from the Superset is conducted. The gravity of assorted

wind speed averaging periods across the basin is reviewed, while a statistical examination of

intensity is administered in order to survey how temporal variations in distribution affect

intensity data. A chronology of the Dvorak technique at JTWC and JMA is summarized and

credible evidence of Dvorak’s role in purported TC intensity trends is highlighted. Finally, an ad

hoc filter is applied to landfalling typhoons in the Superset in the hopes of identifying a less

ambiguous historical record.

Position Ambiguities

Absolute differences in position estimates between fixes in JTWC best track data and

JMA best track data are computed utilizing the newly developed WNP Superset. The great circle

distance between two points to a first approximation is computed using the spherical Law of

Cosines:

d = acos(cos(φjtwc)*cos(φjma)*cos(λjma-λjtwc)+sin(φjtwc)*sin(φjma))*r, (3.1)

where d is the great circle distance, r is the earth’s radius, and λ and φ are longitude and latitude

in radians, respectively. While (3.1) is subject to small errors due to its spherical assumption

(i.e., neglecting ellipsoidal components), the formula provides computationally fast and accurate

estimates. Small distances (< approximately 1/60 of a degree, or a minute of arc) are not handled

well when using the Law of Cosines, but because best track position fixes are in 1/10 of a degree,

the use of this formula is appropriate.

The great circle distance is computed between the two data sets for each match, or point,

in the Superset. These distances are summed annually and the mean is calculated from each sum.

The average annual distances between position estimates at JTWC and JMA are plotted in Figure

3.1. To eliminate interannual variability, a 1-3-4-3-1 smoother is applied to the time series.

15

The average position ambiguities among the two agencies fluctuate between 50 km and

70 km in the late 1950s. In 1960 TIROS I, the first satellite solely dedicated to satellite

meteorology, was launched. By as early as 1962, JTWC began using satellite data to plan early

aircraft reconnaissance into suspect areas (ATCR, 1962). The introduction of Automated Picture

Transmission (APT) with TIROS VIII in December of 1963 allowed for the immediate broadcast

of satellite images to anyone on earth with proper receiving equipment. JTWC installed its first

APT system in 1964 (ATCR, 1964). The sharp decline in position ambiguities between JTWC

and JMA in the mid 1960s is a corollary of this much improved observation network. Not only

were the two centers seeing TCs for the first time, but the expanded coverage also had an

adjuvant type effect of reducing position ambiguities by increasing the relative number of

reconnaissance fixes and investigations per TC (Fig. 3.2).

The United States Air Force worked in tangent with the U.S. Navy during the early days

of aerial TC reconnaissance to provide dedicated weather reconnaissance for the WNP.

Typically the Air Force covered daytime investigations while the Navy covered nighttime

investigations (Guard et al., 1992). During the 1971 season, the Air Force squadron accounted

for roughly 60% of the fixes made, while the Navy provided the remaining 40% (ATCR, 1971).

In November of 1971, the Navy deactivated its weather reconnaissance mission in the WNP

(ATCR). In order to find additional data that might reconcile the loss of Navy reconnaissance,

the DoD made available to JTWC its Data Acquisition and Processing Program (DAPP) weather

satellite imagery in 1971 (Guard). Data from this program underwent an evaluation period to

determine whether or not it was a suitable alternative to aerial reconnaissance (Guard). The

Selective Reconnaissance Program (SRP) was initiated from this test bed, and by 1972 JTWC

began using satellite and radar to selectively position TC fixes (ATCR, 1972). By the end of

1973, DAPP was renamed the Defense Meteorological Satellite Program (DMSP). This marked a

new era, as the reliance on aircraft to position TCs steadily declined (Fig. 3.2). By 1974, over

40% of warnings in the WNP were based primarily on satellite data (Guard).

The time series of JTWC/JMA position ambiguities follows these developments

remarkably well. JMA had access to most of U.S. aerial reconnaissance data utilized by JTWC

(A. Kikuchi, personal communication, 2008). Not surprisingly, average position estimate

differences between the two agencies jump from 32 km in 1971 to nearly 50 km the following

year. As previously mentioned, WNP aircraft reconnaissance was cut substantially after 1971,

16

the final year of Navy weather reconnaissance. As SRP gradually reduced the number of

reconnaissance missions during the mid 1970s, average position ambiguities steadily rise.

In July of 1977 the first Japanese Geostationary Meteorological Satellite (GMS-1) for the

Pacific basin was launched and, a few months later, made operational by the JMA. While JTWC

DMSP sites initially lacked the equipment required to process or display GMS data, by 1979 the

center was receiving both GMS and U.S. Geostationary Operational Environmental Satellite

(GOES) data (ATCR, 1979; Guard et al., 1992). The addition of quasi-continuous satellite

coverage undoubtedly aided in reducing ambiguities in position estimates between warning

centers. Average distances between JTWC and JMA best track position estimates fall from

around 50 km in 1977 to 40 km in 1978. This is consistent with the earlier findings of Bell

(1979), which quantified best track uncertainties among warning centers from 1977-1978

(ATCR). By 1979 the ambiguities drop to levels seen only prior to the deactivation of Navy

aircraft reconnaissance. The addition of GMS-2 in 1981 and GMS-3 in 1984 by JMA, coupled

with next generation GOES data beginning in 1981, results in even better agreement among

position estimates at JTWC and JMA during the early 1980s.

A significant disruption to the TC observing network occurred in the WNP when U.S.

DoD dedicated aircraft reconnaissance was deactivated in August 1987. Previous studies have

examined the impact of this event upon position and intensity estimates as well as track and

intensity forecast errors (Martin and Gray, 1993; Mayfield et al., 1988; Sheets, 1989; Gray et al.,

1991). The most comprehensive investigation detailing position estimate differences in the WNP

was that of Martin and Gray (1993). The study examined TC position differences between

concurrent aircraft measurements and satellite estimates from 1979-1986. The authors found

mean position differences of 44 km between the two groups, with a 10% chance of encountering

differences greater than 93 km.

These findings are consistent with position differences ascertained between JTWC and

JMA from 1988-1998 in the ten years immediately following the loss of WNP aircraft

reconnaissance. Average position differences between the two centers during this period are 50

km, with the largest 10% of the differences being greater than 107 km. The larger differences of

the current study compared with the findings of Martin and Gray are not unexpected, as most

satellite estimates during the active reconnaissance period covered in the former study are not

independent of aircraft observations. Analysts making satellite position estimates often had

17

access to fix data from reconnaissance planes, which may have biased estimates to aircraft

observations (Martin and Gray, 1993). Removing these aircraft data from satellite estimates

would act to augment position estimate differences noted in the Martin study.

Martin and Gray also addressed differences in satellite position estimates between

independent, U.S. operational satellite analysis sites. This comparative analysis is perhaps most

analogous to the present study. JTWC and JMA position estimates prior to 1988 were based both

upon in situ and derived (i.e., satellite) data. Once aircraft reconnaissance ended, position fixes

became largely and often entirely tilted toward satellite analyses. Therefore, differences in

position estimates between warning centers post aircraft reconnaissance can be principally

attributed to varying satellite analyses. Martin and Gray found average position differences

between concurrent satellite estimates of 56 km, with an upper decile of the differences greater

than 110 km. These values agree remarkably well with the aforementioned position differences

between JTWC and JMA for the period 1988-1998. The satellite to satellite comparison in

Martin’s 1993 study proves to be an excellent predictor of inter-agency position ambiguities in

the absence of aircraft reconnaissance.

By the late 1990s, observations from passive microwave sensors aboard the Tropical

Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) greatly improved forecasters’

ability to remotely discern TC center positions. Passive microwave (PMW) frequencies are

insensitive to cirrus canopies, which often obscure low level cloud features and precipitation-free

TC circulation centers in visible and infrared images (Lee et al., 2002). PMW sensors are

especially valuable in locating the center position of weak tropical systems once strong winds

shear apart upper-level convection. The warmer low-level clouds are often absorbed by

background sea surface temperature in nighttime infrared imagery.

The TRMM satellite was launched in 1998, with TMI data made available operationally

beginning in late 1999 (Lee et al., 1999; Hawkins, et al., 2001). Additionally, data from the

Seawinds scatterometer aboard NASA’s Quick Scatterometer (QuikSCAT) launched in 1999

was made available by 2000. QuikSCAT provides TC surface winds, which are useful in locating

TC center positions, particularly with developing systems where a low pressure center forms

within a light wind, precipitation-free region (Edson, 2004). These advances spur a marked drop

in position ambiguities between JTWC and JMA from 1999 to 2000, with mean differences

falling to the lowest levels since the deactivation of aircraft reconnaissance in 1987.

18

Intensity Differences

A lack of homogeneity in wind speed averaging standards compounds existing issues of

interagency TC intensity comparisons (see Section 2). Seeking to conflate averaging periods

through derived wind conversion factors results in some uncertainty, particularly when

elucidating the mean wind versus the gust wind (Fujita, 1971; Simiu and Scanlon, 1978; Harper

et al., 2008; Krayer and Marshall, 1982). Moreover, the technical aspect of converting between

averaging periods is complicated by a raft of variant operational procedures, including the

absolute accuracy of maximum surface wind estimates from aircraft reconnaissance, ocean

buoys, land stations, and other in-situ observations. Nevertheless, seeking to address differences

in best track intensity between TC archives is fundamental to any discussion of observed long

term changes in WNP TCs. The Superset provides a lens through which to view passive issues in

the historical TC intensity record.

Previous studies of TC intensity have employed various integrals in defining storm

power, including Accumulated Cyclone Energy (ACE) and a TC Power Dissipation Index (PDI)

(Bell et al., 2000; Bister and Emanuel, 1998; Emanuel, 2005). The current study examines

intensity differences through PDI, defined as the sum of the maximum one-minute sustained

wind speed cubed, at six-hour intervals, for all periods when a tropical cyclone is of at least

tropical storm strength (i.e., MSW ≥ 17 ms-1

). Recent publications cite PDI in deconvolving long

term trends in TC intensity (Emanuel, 2005; Sriver and Huber, 2006; Kossin et al., 2007). By

considering the cube of the wind speed, PDI accentuates potential errors, notably in higher wind

speeds. Interagency intensity discrepancies are thereby emphasized through the ambit of PDI.

A contemporaneous comparison of annually accumulated PDI is conducted between

JTWC and JMA Superset MSW intensity data from 1977 through 2006. JMA MSW is first

converted from a 10-minute average to a 1-minute average using the de facto conversion factor

of 1.14 (Simiu and Scanlon, 1978). PDI for each data set is subsequently accumulated over each

year. A von Hann 1-2-1 smoother is applied to the time series of each set presented in Figure 3.3.

Annually accumulated PDI is a product of TC intensity, duration, and frequency

(Emanuel, 2007). Since duration and frequency remain constant among data sets within the WNP

unified Superset, interagency variations in annually accumulated PDI are a result only of

discrepancies in TC intensity estimates and from those sensitivities associated with wind speed

19

time averaging conversions. Applying the Mann-Kendall trend test to the 30-year raw time series

of JTWC annually accumulated PDI from the Superset shows an upward trend significant at the

0.05 level (Appendix A). Applying the same test to similar data from the JMA, however, results

in no significant trend. The implications of this analysis are especially disconcerting under the

purview of the Superset, whereby variant conclusions in 30-year trend analyses are solely an

outgrowth of differences in interagency intensity estimates.

The Role of Dvorak’s Intensity Estimation Technique

The evolution during the 1970s and 1980s of pattern recognition intensity estimates via

the Dvorak (1972, 1973, 1975, 1982, 1984) technique has proven essential to TC forecasters of

the satellite reconnaissance era. The percentage of warnings based on satellite data steadily rose

during this period across the WNP, as dedicated aircraft reconnaissance gradually diminished

(Fig. 3.4). The Dvorak technique has become the primary method of ascertaining TC intensity

across the region following the deactivation of aircraft reconnaissance in 1987 (Guard, 2004).

The original current intensity (CI) number to MSW index was developed by Dvorak from

empirical data across the WNP basin (Dvorak, 1973). Despite the complex nature of tropical

cyclones, the Dvorak technique is a rather perspicuous rubric of pattern recognition, whereby

combined kinematic (vorticity and vertical wind shear) and thermodynamic (convection and core

temperature) properties produce cloud patterns that relate to TC intensity (Velden et al., 2006).

The strength and distribution of the cyclonic winds, distorted by shear, are noted in these cloud

patterns. The degree of circular wind distortion is related back to MSW by means of Dvorak.

Furthermore, convective vigor in and around the eyewall in more intense TCs, derived remotely

from cloud temperature, is presently the dominant mechanism in assigning MSW estimates for

organized TCs (Velden et al.).

The process of evaluating TC intensity from satellite data is a relatively straightforward

procedure. A TC analyst first assigns a tropical (“T”) number (Tnum) based upon climatological

pattern recognition (Figures 3.6 and 3.7). This Tnum is converted into a CI number based upon

developmental stage. Finally a table is used to relate CI to MSW. MSLP estimates are derived

from MSW estimates through wind-pressure relationships (WPRs).

The 1-minute MSW averaging period assigned to the Dvorak table and implemented by

Regional Association (RA) IV (North and Central America) and JTWC is never explicitly

20

addressed by the technique’s architect (Velden et al., 2006). This tenuous relationship has since

been adopted and modified by other RAs to fit regional wind averaging standards. Australia’s

Tropical Cyclone Warning Centres apply a 0.871 1-minute to 10-minute conversion factor to all

MSW in their CI/MSW tables (Holland, 1993). HKO likewise multiples all 1-minute MSW by a

factor of 0.9 to convert to 10-minute winds (Wu, 2006). Meanwhile, the JMA, also exercising a

10-minute wind speed averaging standard, has solicited various conversion factors and

conversion equations since the late 1970s to adapt 1-minute MSW to a 10-minute averaging

period. These conversion techniques are nonuniform with respect to their application in Dvorak’s

CI/MSW table. Prior to 1990, JMA equated 1-minute winds to 10-minute winds from Dvorak,

1975 and Dvorak, 1984 for MSW less than or equal to 65 knots (33 ms-1

). For winds greater than

65 knots (33 ms-1

), the agency implemented the following conversion equation:

31.6 1.495

)(MSW MSW 1min

10min += , (3.2)

where MSW10min are 10-minute MSW and MSW1min are 1-minute MSW in knots (Japanese

Meteorological Agency, 1990; A. Kikuchi, personal communication, 2007).

By the mid-1970s, tangent with the decline of U.S. aircraft reconnaissance (Fig. 3.4), the

influence of the Dvorak, 1975 and later Dvorak, 1984 satellite interpretation models in intensity

estimation grew at JTWC. Several WPRs have been used by JTWC and incorporated within the

center’s Dvorak CI table over the years (ATCR, 1959-2004). The Dvorak, 1975 WPR was

ultimately replaced by Atkinson and Holliday (1977; hereinafter, AH77) at JTWC. The original

Dvorak CI/MSW relationship at JTWC, however, remains unaltered.

Correlations between PDI data from JTWC and JMA unfiltered Superset data are

examined for the period 1977-1989. A strong relationship is anticipated, as a significant

contribution to PDI variance is explained by storm duration and frequency (Maue and Hart,

2007); the Superset effectively eliminates variability from these components. The two time series

correlate at r > 0.95 for the period, even with variations in wind reporting practices. Aircraft

reconnaissance continued through most of the period, and data from these missions were shared

by JTWC and JMA (A. Kikuchi, personal communication, 2007). Additionally, both agencies

employed the Dvorak, 1975 and Dvorak, 1984 CI/MSW table, albeit modified at JMA to

21

accommodate a 10-minute averaging standard. This coordination complements the strong

correlation.

It should be emphasized that data quality issues related to reconnaissance estimation of

MSW from observed MSLP using AH77 are well documented in the literature (Knaff and Zehr,

2007). The purported underestimation from 1974-1987 of MSW in more intense TCs from

reconnaissance appears to be an artifact in both best track sets. The semblance is attributed to

interagency coordination of reconnaissance fix data (A. Kikuchi, personal communication,

2007).

The analysis is extended to include the entire period of MSW data (1977-2006). Although

the correlation of PDI from JTWC and JMA Superset data remains significant, the relationship

drops from r > 0.95 to r = 0.83. Bifurcation to the original analysis provides even further insight.

Examined individually, the periods 1977-1989 and 1990-2006 yield r > 0.95, which suggest a

critical shift within one or both of the best track archives.

JMA modified its CI/MSW relationship in 1990 based upon results from Koba et al.

(1990). The Dvorak, 1975, Dvorak, 1984, Koba, 1990, and revised JMA Dvorak CI/MSW tables

are summarized in Table 3.1. In order to investigate the effect of this change upon interagency

PDI correlation from the Superset, all MSW data from JMA Superset data post 1990 are

converted to their corresponding CI from the revised JMA CI/MSW table. These data are then

assigned a MSW from the original Dvorak, 1984 table from the CI. MSW are converted to a 10-

minute averaging period for MSW > 65 knots (33 ms-1

) per equation (3.2). The procedure

essentially debiases JMA intensity data post 1990.

A revised contemporaneous PDI comparison from debiased JMA Superset data and

JTWC Superset data from 1977-2006 is presented in Figure 3.5. This plot is shown as an

analogue to Figure 3.3. Similarly JMA MSW is converted from a 10-minute average to a

consistent 1-minute average and yearly accumulated PDI is smoothed to dampen interannual

variability. For the entire period (1977-2006) the corrected analysis substantially improves the

correlation between the two data sets. For the entire period (1977-2006) the two time series in

Figure 3.5 correlate at r = 0.95. This is contrasted to r = 0.83 for the uncorrected time series in

Figure 3.3. The strong correlation resulting from the debiasing algorithm follows that noted in

the Superset data for the period prior to integration of the Koba, 1990 study at JMA. The

22

debiasing technique also yields an upward trend in both the JTWC and JMA time series,

significant at the 0.05 level (Appendix A).

The debiasing algorithm did not rectify a notable difference in PDI between JTWC and

JMA Superset data around 1996. Unfiltered JTWC data in 1996 from Figure 3.5 tended toward a

higher PDI than JMA data by approximately one standard deviation from the 30-year sample

mean. In fact, time dependent differences in annually accumulated PDI for the years 1995-1997

are at least twice as large as in any other three year period examined. This aberration is also

found in Kossin, et al. (2007), in which PDI from a homogeneous intensity data set is compared

to PDI from JTWC original best track data.

The suspect period is considered through a cross check within the Superset of original

best track omissions. Time integrals of the annual number of points from each original best track

data set are computed along with their corresponding totals from the WNP unified Superset. The

annual mean percentage of points excluded from the Superset from original best track data from

the three archives set forth in this paper is reflected as a time series in Figure 3.8. These time

series exhibit storm intensity, location, and sample dependencies, with weaker, higher latitude

tropical storms during more active years tending toward greater a priori ambiguities among the

warning agencies. This observation is supported by JTWC: “The larger number of poorly

defined, or ‘low end’ systems contributed to some large initial position errors…” (ATCR, 1995).

The time series also show some influence from evolving detection and intensity estimate

procedures and technologies, as evident during the early 1970s when DMSP helped to refine

equivocal fixes.

A point is excluded from inclusion in the final WNP Superset if a match per criteria

defined in Section 2 is not made to a point in both of the other sets examined. It is impossible to

determine why specific points are excluded without inspecting each point and the role the above

mentioned dependencies may have played in interagency interpretations. Some systematic

elimination can be attributed to variation in basin AOR, although this parameter accounts for

only a fraction of those points excluded from the Superset. Nevertheless, Figure 3.8 provides an

important piece to the best track puzzle.

The percent of points excluded from JTWC and JMA Superset data remains generally

stable from the start of the record through the early 1980s, with some interdecadal variability

likely attributed to observational changes. The deactivation of dedicated aircraft reconnaissance

23

across the basin by the mid 1980s undoubtedly contributes to the rise in the JTWC time series

beginning during this era. Satellite estimates produce greater uncertainty compared with

empirical aircraft data shared across warning centers. It is also inferred from this analysis that

forecasters at JTWC during the late 1980s and early 1990s generally followed less restrictive

satellite interpretation guidelines than forecasters at JMA for early and late stages of TC

development. Even today satellite classification of certain systems, particularly “midget,” or very

small, TCs is non-standard among the warning agencies (Lander, 2008). The most recent

example of this is two “midget” TCs classified by JMA in 2007 but not warned on by JTWC nor

included in its best track data.

While Figure 3.8 highlights differences in the absence of aircraft reconnaissance, citing

classification procedures as the catalyst behind the sharp incline and subsequent peak in JTWC

data during the 1990s remains dubious. By 1996, the mean percentage of entries excluded from

JTWC original best track data via the WNP Superset climbs to 55%, or nearly 3σ above the 30-

year sample mean. U.S. legislation from the 1995 Base Realignment and Closure (BRAC) forced

JTWC to disestablish Naval Pacific Meteorology and Oceanography Center West/Joint Typhoon

Warning Center (NPMOCW/JTWC), Guam and transition operations to NPMOCW/JTWC,

Pearl Harbor, Hawaii. The transition period lasted until 1 January, 1999. During the transition,

JTWC conducted split operations as all tropical cyclone forecasting support equipment was

transferred to Hawaii (ATCR, 1998). The agency also experienced 90 percent personnel turnover

by the end of the transition (ATCR, 1998). These developments offer added insight into the

anomalously high percentages noted during this period.

The Reliability of Landfall Data and Extratropical Cyclones

In an effort to better understand changes in TCs against the backdrop of global climate

change and under the restrictions of inherently flawed best track data, research has turned to

global climate models (GCMs) to simulate the response of TCs in future environments (Knutson

and Tuleya, 2004). However, the horizontal resolution of the current family of GCMs and higher

resolution regional models to which downscaling techniques are applied is too low to reproduce

inner core processes which lead to intense TCs (Emanuel, 2006). Most recently, Emanuel (2008)

introduced a new downscaling approach in which nascent synthetic TC seedlings are produced

from GCM statistics, with 200-year projections drawn from seven GCMs run in support of the

24

Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4). While

Emanuel’s approach circumvents limitations associated with conventional techniques, his results

rest upon the credence of model simulations.

These theoretical simulations are not universally accepted by the general public, which

tend toward empirical evidence over numerical predictions. Consequently, despite well

documented issues across all global basin best track data, researchers continue to sift through and

compartmentalize observational data in order to find new truths and defang past assertions. Many

methods examine only portions of the data deemed reliable (Emanuel, 2005; Webster et al.,

2005; Klotzbach, 2006), others cite intensity proxies (Webster et al., 2005; Elsner et al., 2008) or

regression models (Kossin et al., 2007; Elsner et al., 2008), and some apply bias removal

schemes and quantifiable uncertainties to existing data (Emanuel, 2005; Landsea, 2007).

In the tropical Atlantic, landfall statistics have been presented against total storm

statistics in a search for symmetry (Landsea, 2005; Landsea, 2007). The predominant assumption

made with landfalling TCs is that detection and accuracy of near surface intensity estimates

increase along populated coastlines (see Figures 3.9 and 3.10). Since any trend from global

climate change should be more apparent over greater time scales, parsing entire best track data

by landfall also allows examination of a longer, “reliable” period of record. The WNP

experiences on average the formation of two to three times as many TCs than does the Atlantic

basin (Landsea, 2000). Implicit in this is a greater number of landfalling TCs within the

historical record. Therefore, it is unsurprising that TC/global climate change research across the

WNP has expanded into landfall subsets (Chan, 2008).

Multiple data sets offer a unique angle through which to test the validity of claims of

more reliable data near landfall. Reliable data include fewer inconsistencies and smaller

ambiguities. Reliable data also connote greater accuracy; accuracy, however, requires some

measure of ground truth, a parameter unavailable to every estimate. If the higher density network

of in situ observations over land increases a forecaster’s ability to estimate TC position and

intensity, it should be reflected in smaller ambiguities among independent estimates. The

Superset allows us to examine this issue through spatial distribution.

Figure 3.11 plots the spatial distribution of mean absolute distance between position

estimates in JTWC and JMA best track data from 1951 through 1969 from the WNP Superset.

Again the focus remains on the two most widely used data sets; STI data is generally of degraded

25

quality and consistency than the other data employed in this study. The years prior to 1970 are

chosen, since 1970 is widely regarded in the literature as the cut-off between reliable and

questionable data due to little or no satellite reconnaissance in the early period (Webster et al.,

2005; Landsea, 2007). Position errors ultimately affect landfall statistics, particularly among

low-end systems grazing the coastline. Position ambiguities among the warning agencies provide

some measure of the potential for error with such fixes.

Mean cell values are computed from all points contained within 2° x 2° cells and binned

within four classes set at approximately one and two standard deviations from the sample mean.

Figure 3.11 finds the smallest ambiguities over the deep tropics between roughly 10°N and

30°N. This is attributed to the presence of deeper, more frequent TCs at these latitudes. The mid-

latitudes exhibit sporadic distribution, with generally higher mean ambiguities, especially near

northern Japan, the Kuril Islands, and the Kamchatka Peninsula. This is expected with weaker,

transitioning TCs at higher latitudes where fewer land and ship observations exist. Moreover,

reconnaissance during the early to mid 1960s was routinely tasked for well developed TCs,

further impairing forecast estimates for weak or transitioning storms.

Position ambiguities actually tend toward mid range to high range values along coastal

regions and over locations inland. It is difficult to determine which terrestrial data were shared or

made available to each agency, but the symbiosis of land masses and higher ambiguities is

unlikely fortuitous. As a TC moves inland, it loses its energy source from the air-sea layer,

subsidence within the eye abates, and the structure decays, rendering storm center position

estimates problematic. Multiple observations from land stations or radar can also augment

uncertainty in synoptic positioning if the observations conflict. It is reiterated that minor

deviations in position estimates or tangential storm angles along coastlines may bring about

major changes to landfall subsets. Figure 3.12 illustrates one such discrepancy. The cyclone

passes over Hainan as a powerful typhoon in JTWC best track data, with its final position

estimate as an 80 kt typhoon over the Gulf of Tonkin. The corresponding estimate in JMA best

track data has the cyclone inland over Vietnam, over 700km west of the position fix in JTWC

data. Such discrepancies, particularly among more intense TCs, can greatly influence landfall

climatologies.

Position ambiguities underscore one obstacle with compartmentalizing TC data by

landfall. In the above section, temporal plots illustrate inconsistencies in reporting practices and

26

procedures between the TC warning agencies. A spatial analysis of intensity ambiguities is

desired in order to resolve the impact of land on interagency intensity estimates. If position

estimates suffer from ambiguities related to a lack of data or conflict therein, then uncertainties

in intensity estimates will also dilate across data sparse regions. The undersampled open ocean

areas prior to the satellite era and removed from major maritime routes make these data suspect.

While previous studies have reviewed sampling frequency associated with TC landfalls

(Landsea, 2007; Chang and Guo, 2007), none have sought to quantify errors or uncertainties

between cyclones over land and those over ocean.

Results shown in this paper are used to split the intensity spatial analysis among two

periods. A step function change in JTWC and JMA interagency intensity ambiguities occurs in

1990, with smaller overall ambiguities existing prior to 1990 when both agencies referenced

Dvorak, 1984 for satellite estimates. Differences in wind speed averages among the warning

centers are handled with an across-the-board 1.14 10-minute to 1-minute conversion factor.

Applying one such general conversion may seem imprudent in light of issues discussed above.

However, the spatial intensity analyses serve as a general layout of categorized ambiguities

rather than an exacting measure of uncertainty. The idea is to highlight those regions which may

account for the variability found in Figure 3.3, and to determine if land affects ambiguities

among intensity estimates, regardless of scale.

Mean MSW differences between JTWC and JMA data from the WNP Superset computed

within 2° x 2° cells for the period 1977 through 1989 are presented in Figure 3.13a. Cells

centered over land masses do not in general exhibit lower mean differences than those cells

centered over the open ocean. In fact, some of the largest differences occur near Tokyo and over

parts of northern Japan. This may result from a higher frequency of extratropical cyclones

passing near Japan. The prevalence of extratropical cyclones in best track data is discussed

further below.

The spatial plot otherwise reveals a similar pattern to that shown in the distance

composite, with smaller ambiguities in the deep tropics becoming larger across the mid-latitudes,

presumably an artifact of thermal asymmetries associated with weaker, transitioning systems.

Additionally, the continued presence of aircraft reconnaissance through the majority of this

period keeps ambiguities down over much of the open ocean. Decaying TCs inland from the

coast account for higher ambiguities across eastern Asia.

27

The absolute intensity differences of the latter period (1990 through 2006) compared with

the former period (1977 through 1989) manifest a higher mean and larger spread, as evident in

Figure 3.13b. Differences in the spatial distribution between the two periods are largely confined

to the deep tropics. In both analyses the mid-latitudes (poleward of 30°N) exhibit higher mean

differences; however, areas of the deep tropics in the latter period, especially open ocean regions,

are binned in a higher classification than the same cells in the former period. As previously

discussed the Dvorak scheme utilized at JMA and JTWC after 1990 differ only for intense TCs,

which generally form in the deep tropics. The absence of aircraft reconnaissance in the latter

period combined with variant Dvorak CI/MSW tables result in smaller mean ambiguities and a

lower spread in the intensity differences for estimates over land masses compared with estimates

over the open ocean.

If the various WNP TC warning agencies bias estimates to surface observations for storm

positions at landfall, the results should ameliorate the impact of nonstandard Dvorak schemes.

Spatial analyses of position ambiguities and intensity differences between JTWC and JMA

Superset data, however, offer little indication of growing congruence for points centered over or

near land. A composite bar chart of unfiltered cubic MSW from 1977 through 2006 for only

those points determined to be over land or within 50 km thereof based upon 5-minute elevation

data is presented in Figure 3.14. JMA wind data are converted from a 10-minute average to a 1-

minute average using a blanket 1.14 conversion factor. Only points found over land or within a

50 km land buffer in all JTWC, JMA, and STI Superset data are included in the subset. Figure

3.15 includes several instances of mean position estimate differences exceeding 50 km near the

coast, differences which can greatly affect landfall climatologies. Parallel data ensure deviations

are a consequence solely of intensity differences rather than differences in landfall frequency.

The method also employs the least ambiguous point data over land stemming from well defined

tropical systems.

With only one exception, annually accumulated cubic winds from JMA for points over

land tend toward higher aggregates than JTWC data for all years examined prior to the

deactivation of aircraft reconnaissance in 1987. From 1987 through 2006 the tendency reverses,

with JTWC aggregate data tending toward higher values than JMA data for all but two

consecutive years in the early 1990s. The largest absolute difference between consecutive years

in the two primary landfall subsets occurs between 1995 and 1996, a result consistent with earlier

28

findings from all WNP Superset data. Data from Figure 3.14 correlate at r = 0.74 for the entire

period (1977-2006), compared to r = 0.83 from Figure 3.3 for all WNP Superset data. Like the

time series in Figure 3.3, trend analyses of data in Figure 3.14 yield an upward trend significant

at the 0.05 level in JTWC aggregate data but a downward trend not significant at the 0.05 level in

JMA data. Bifurcation of the entire period to the 1990 Dvorak step change results in r = 0.75 for

the period prior to 1990 and r = 0.85 for the latter period (1990-2006).

The landfall subsets from the two primary warning agencies exhibit less agreement for

the entire period than intensity estimates from all WNP Superset data. In contrast to results from

all points in the WNP Superset, Figure 3.14 manifests a higher correlation in the post 1990

period than among over land points prior to 1990 when the two periods are examined

independently. Assuming JTWC and JMA apply the Dvorak intensity estimate technique

consistently throughout each period, with a step function change occurring at JMA in 1990, the

central change to the observation network between the two periods is the deactivation of

dedicated aircraft reconnaissance in 1987. The higher correlation in the period following aircraft

reconnaissance confirms less scatter within Dvorak, not necessarily greater reliance on surface

observations. A lower correlation for the entire period in the landfall subset compared to that for

all Superset data is indicative of surface observations which impugn independent estimates rather

than corroborate them.

As a final assessment of the veracity of intensity estimates from points over land, the

landfall subset is parsed to include only those points over land of typhoon intensity in all JTWC,

JMA, and STI Superset data. A composite bar chart of unfiltered cubic MSW from 1977 through

2006 from JTWC and JMA data is presented in Figure 3.16. As with Figure 3.13, JMA wind data

are converted from a 10-minute average to a 1-minute average using a 1.14 conversion factor.

Data for the entire period correlate at r = 0.81, similar to that found across all WNP Superset data

and higher than the coefficient computed from all points over land. Examination of the periods

prior to and subsequent to the 1990 modification of Dvorak at JMA yields correlations of r =

0.83 and r = 0.89, respectively.

Applying the Mann-Kendall trend test to the 30-year unfiltered data in Figure 3.16

produces an upward trend in JTWC aggregate data for points over land of typhoon intensity, but

a downward trend in JMA aggregate data. These trends follow trends computed from all WNP

Superset data and from the landfall subset; however, the trends in annually accumulated cubic

29

winds for points of typhoon intensity over land are not statistically significant at the 0.05 level in

either JTWC or JMA data. While compartmentalizing points over land of typhoon intensity does

not exonerate the influence of Dvorak on intensity estimates across the basin, it does appear to

dampen its effect. The progression from a significant upward trend in JTWC intensity data from

all points in the WNP Superset to no significant trend in JTWC aggregate data from only points

over land of typhoon intensity suggests that the detection of any trend may be an artifact of

satellite derived intensity techniques.

Best track data from each of the WNP warning agencies examined in this study contain

an indicator for the level of TC development. This is an important entry, for example, in

deciphering when a TC reaches typhoon status between data sets of irregular wind speed

averages. JMA records a grade column in its data to indicate development stage. Prior to 1977,

only one of two grade indicators is associated with each entry – either for the tropical depression

stage or for TCs of at least tropical storm intensity. JMA does not distinguish between the

tropical storm and typhoon stage until 1977 forward. The addition of tropical storm, severe

tropical storm, typhoon, and extra-tropical cyclone indicators to entries in JMA best track data

beginning in 1977 is utilized to sift out points over land of typhoon intensity for the above

analysis.

JTWC does not include a column for level of development in its best track data until

2000. Prior to 2000 it is impossible to determine whether an entry records estimates for a

tropical, subtropical, or extra-tropical cyclone. The extra-tropical stage of a cyclone’s lifecycle is

significant, as extra-tropical cyclones are generally devoid of tropical characteristics and

excluded from TC climatologies. JTWC and JMA Superset data are compared from 1977 to

1999 in order to uncover instances of extra-tropical entries from JMA that are included in but not

labeled as extra-tropical in JTWC best track data. Nearly 1000 points marked as extra-tropical in

JMA best track data are matched to JTWC best track data over the 23-year period. Because

inclusion of extra-tropical cyclones can substantially alter intensity trends and similar analyses,

proper disclaimers must be cited in the absence of complete metadata.

30

Table 3.1: The empirical relationship between current intensity (CI) number and maximum sustained winds

in knots. The Dvork, 1975, Dvorak, 1984, Koba, 1990, and revised JMA CI/MSW relationships are shown.

Maximum Sustained Winds (kt) CI #

Dvorak, 1975 (1-min) Dvorak, 1984 (1-min) Koba, 1990 (10-min) JMA (10-min)

1.0 25 25 22 25

1.5 25 25 29 25

2.0 30 30 36 30

2.5 35 35 43 35

3.0 45 45 50 45

3.5 55 55 57 55

4.0 65 65 64 65

4.5 77 77 71 70

5.0 90 90 78 77

5.5 102 102 85 85

6.0 115 115 93 93

6.5 127 127 100 100

7.0 140 140 107 107

7.5 155 155 115 115

8.0 170 170 120 122

Fig 3.1: Annual average absolute distance between position estimates in JTWC and JMA best track data.

Dashed line indicates raw annual averages and solid line indicates smoothed averages after a 1-3-4-3-1 filter is

applied. Overlaid are significant developments in the TC observing system.

31

Fig 3.2: Annual mean number of WNP aircraft reconnaissance levied fixes and investigations per TC,

according to JTWC best track data. Red bars indicate notable shifts in the annual mean: (1962-63) Advent of

satellites allows previously undetected open ocean TCs to be reconnoitered; (1972-73) SRP begins using

satellite to position TC fixes; (1987) Dedicated aircraft reconnaissance is deactivated in the WNP.

Fig 3.3: Annually accumulated, unmodified PDI for the WNP from the unified WNP Superset. Dashed lines

indicate raw PDI and solid lines indicate smoothed PDI after a 1-2-1 filter is applied. JTWC PDI is shown in

red and JMA PDI in blue. R for the entire period (1977-2006) is 0.83.

32

Fig 3.4: From Fig. 9 of Guard (1992), which shows the percent of warnings based primarily on aircraft

reconnaissance and satellite reconnaissance. Since some estimates are based on terrestrial radar or

extrapolation, the percent based on satellite never reaches 100%.

Fig 3.5: Annually accumulated, modified PDI for the WNP from the unified WNP Superset. Dashed lines

indicate raw PDI and solid lines indicate smoothed PDI after a 1-2-1 filter is applied. JTWC PDI is shown in

red and JMA PDI in blue. MSW from JMA best track data for the period 1990-2006 are converted from the

1990 Koba et al. modified Dvorak table to the original 1984 Dvorak table for MSW > 33 ms-1

according to

equation (3.2). R for the entire period (1977-2006) is 0.95.

33

Fig 3.6: Analysis diagram outlining the Dvorak intensity estimation technique designed for use with visible

satellite imagery from Dvorak, 1984.

34

Fig 3.7: Analysis diagram outlining the Dvorak intensity estimation technique designed for use with enhanced

infrared satellite imagery from Dvorak, 1984.

35

Fig 3.8: Annual mean percent of points excluded from JTWC (red), JMA (blue), and STI (green) Superset

data from original best track data. Dashed lines indicate unsmoothed data. Solid lines indicate smoothed data

after a 1-2-1 filter is applied.

Fig 3.9: Regional Basic Synoptic Networks (RBSN) of global surface and upper air stations from the World

Weather Watch Programme of the World Meteorological Organization. Station data are current as of 01

October 2008 (WMO, 2008). Higher density coverage is found in mainland locations along the coast, with a

much sparser fixed surface network noted on islands and atolls dotting the ocean basins.

36

Fig 3.10: Same as Fig 3.9 but cropped and enlarged to show networks across the WNP basin.

Fig 3.11: Spatial distribution of mean absolute distance between position estimates in JTWC and JMA best

track data from 1951 through 1969 from the WNP Superset. Mean cell values are computed from all points

contained within 2° x 2° cells and binned within four classes set at approximately one and two standard

deviations from the sample mean.

37

Fig 3.12: JTWC (red) and JMA (blue) tracks for Storm 64 from the unified WNP Superset (Storms 20 and 19

of 1953 from JTWC and JMA, respectively). Month and day are labeled for 00Z estimates along with 1-

minute average MSW in parenthesis for JTWC data. Bullet points indicate 12-hour position estimates. JTWC

records three 6-hourly position estimates over Hainan, with one MSW estimate as high as 110 kt. JMA track

records no points over Hainan in its 6-hourly best track data. Additionally, at 00Z on 2 November, position

estimates between the two agencies differ by nearly 250 km, such that at 00Z JTWC records an 80 kt TC over

the Gulf of Tonkin while JMA has a moderate TC inland over Nghệ, a province along the north central

Vietnam coast. Differences are also found between landfall numbers in the Philippines.

38

Fig 3.13a: Spatial distribution of mean absolute difference between intensity estimates in JTWC and JMA

best track data from 1977 through 1989 from the WNP Superset. JMA MSW are converted to a 1-minute

average using a 1.14 10-minute to 1-minute conversion factor. Mean cell values are computed from all points

contained within 2° x 2° cells and binned within four classes set at approximately one and two standard

deviations from the sample mean.

39

Fig 3.13b: Spatial distribution of mean absolute difference between intensity estimates in JTWC and JMA

best track data from 1990 through 2006 from the WNP Superset. JMA MSW are converted to a 1-minute

average using a 1.14 10-minute to 1-minute conversion factor. Mean cell values are computed from all points

contained within 2° x 2° cells and binned within four classes set at approximately one and two standard

deviations from the sample mean.

40

Fig 3.14: Composite bar chart of annually accumulated unfiltered cubic MSW from points over land in

JTWC and JMA best track data from the WNP Superset. JTWC aggregate is shown in red and JMA

aggregate is shown in blue. Landfall climatology includes only points captured over land in JTWC, JMA, and

STI Superset data. Annually accumulated cubic winds associated with points over land from JTWC best

track data exhibit an increasing trend significant at the 0.05 level. JMA data indicate a downward trend not

significant at the 0.05 level.

41

Fig 3.15: Spatial distribution of mean absolute distance between position estimates in JTWC and JMA best

track data from 1977 through 2006 from the WNP Superset. Mean cell values are computed from all points

contained within 2° x 2° cells and binned within four classes set at approximately one and two standard

deviations from the sample mean.

42

Fig 3.16: Composite bar chart of annually accumulated unfiltered cubic MSW from landfalling typhoons in

JTWC and JMA best track data from the WNP Superset. JTWC aggregate is shown in red and JMA

aggregate is shown in blue. Landfalling typhoon climatology includes only points captured over land of

typhoon intensity in JTWC, JMA, and STI Superset data. Annually accumulated cubic winds associated with

JTWC landfalling typhoons exhibit an upward trend not significant at the 0.05 level. JMA data indicate a

downward trend not significant at the 0.05 level.

43

CHAPTER 5

CONCLUSIONS

A combined WNP Superset is developed from existing best track data across multiple TC

agencies. Data from three of four comprehensive best track records are compared over 56 years,

and concurrent, quasi-independent TC estimates matched amongst JTWC, JMA, and STI best

track archives are merged into a unified best track Superset. WNP Superset data include

concurrent position and intensity estimates from all three WNP best track data sets. Differences

between data at the two primary warning agencies are quantified and presented as ambiguities.

Position and intensity ambiguities are utilized as a tool for elucidating unresolved issues in

historical TC data across the basin.

Position ambiguities between JTWC and JMA Superset data closely follow changes in

the WNP TC observation network. Ambiguities on the order of 50 km to 70 km fall sharply upon

the introduction of the first meteorological satellites in the early 1960s. Ambiguities between the

two agencies continue to decline until the deactivation of Navy aircraft reconnaissance in 1971.

With the launch of GMS satellite in 1977, satellite reconnaissance advanced from sporadic

transmission to continuous surveillance. The presence of aircraft reconnaissance paired with

emerging satellite technology perpetuates the steady decline in mean ambiguities among the two

forecast agencies through the mid-1980s. Ambiguities jump by 25% in the years immediately

following decommission of dedicated aircraft reconnaissance in 1987. Mean ambiguities remain

steady near pre-satellite levels in the post reconnaissance era until microwave imagers and

satellite derived ocean wind data became available by the turn of the century, after which time

position estimates between JTWC and JMA begin to coalesce.

Non standard wind speed averaging periods across the WNP complicate direct

comparisons among TC best track intensity data. Intensity differences from JTWC and JMA

Superset data are examined through PDI, a metric sensitive to small variations in intensity

estimates, particularly among intense TCs. Results show a fraction between the warning agencies

after 1990 once JMA revised its Dvorak CI/MSW relationship based upon an earlier study by

Koba et al. (1990). The step function change at JMA affects intensity trends in the absence of

aircraft reconnaissance. Without validation of the two Dvorak schemes applied across the basin

since 1990, however, it is inconclusive which relationship exhibits greater accuracy. Recent

44

work by Knaff and Zehr (2007) find a more valid WPR within Koba et al. (1990) compared to

the AH77 WPR incorporated within JTWC Dvorak tables. WPRs do not affect MSW estimates

from Dvorak, as the derived empirical relationship relates CI to MSW rather than MSLP.

Reconnaissance data in the early period may bias MSW estimates too high in cases where

aircraft measured pressure is converted to MSW via WPRs (Knaff and Zehr).

JMA Superset MSW data from 1990 forward are paired with their corresponding CI

values from the revised JMA Dvorak table. CI values are subsequently matched to MSW

estimates from Dvorak, 1984. The debiasing algorithm corrects an apparent discontinuity around

1990, with 30-year post adjustment upward trends significant at the 0.05 level detected in both

sets. A notable discrepancy between PDI analyses remains in the mid-1990s. The study attributes

these differences to data issues related to equipment transfers and personnel turnover at JTWC

during its phased relocation from Guam to Pearl Harbor, Hawaii, from 1995-1999. The percent

of points excluded from JTWC Superset data from original best track data exceeds 50% during

this period. The difference is also noted in contemporaneous PDI comparisons between the

global homogeneous UW/NCDC tropical cyclone intensity record and the full JTWC best track

archive (Kossin et al., 2007). The findings call into question the quality of JTWC best track data

during the 1990s. Users should exercise extreme caution when establishing climatologies or

using such uncorrected records to assess risk. Until deficiencies are properly addressed, metadata

contained within the present study and other similar work will be the only form of quality control

behind historical WNP TC records.

Parsing Superset data into points over land estimated as such in all sets or

compartmentalizing landfall subsets into landfalls of typhoon intensity does not increase

agreement in position or intensity estimates between JTWC and JMA archives. Lower mean

ambiguities are concentrated in the tropical belt –between roughly 10°N and 30°N – rather than

along the coast or over land. Better defined, more intense TCs form in the deep tropics compared

with regions in higher latitudes. Mean ambiguities poleward of 30°N are subject to the greatest

scatter due to increased frequency of weakening, transitioning, and extra-tropical TCs.

Decommission of aircraft reconnaissance in the late 1980s created a significant void in available

surface data. Forecasters likely looked to land observations to confirm derived estimates in the

absence of aircraft reconnaissance. This is supported by lower mean intensity ambiguities over

land in the post reconnaissance period.

45

Exploring available resources to incorporate other germane data into the WNP Superset

would aid in future studies. Infusing aircraft reconnaissance data into the Superset for example

may shed new light on interagency differences. A comparison method of analysis similar to that

developed by Martin and Gray (1993) would allow researchers to quantify how forecaster

estimates and ambiguities among warning agencies change with and without available center fix

data from reconnaissance missions. Additionally, matching nearby global marine surface

observations from the International Comprehensive Ocean-Atmosphere Data Set (ICOADS) to

existing entries from the three WNP best track sets included in the first iteration of the unified

Superset may assist in resolving the largest discrepancies noted between best track archives.

Pending permission from the Government of the Hong Kong Special Administrative Region,

including HKO best track data from 1961 forward would provide additional information and,

since both HKO and JMA record 10 minute average MSW, may permit direct comparisons of

wind intensity data. Because the AOR at HKO (Fig 2.1d) is 25% to 44% smaller than the

smallest domain of the other three WNP sets examined, it is more advantageous to add the data

as supplementary information rather than excluding thousands of points from the other data sets

solely due to AOR differences.

Ultimately publicly available data from all sources related to position and intensity

interpretation for TCs in the WNP, including post event analyses like the homogeneous

UW/NCDC tropical cyclone intensity record, can be compiled into a Superset database.

Documentation replete with metadata is the primary focus of any future extension to the

Superset. Without the data behind the data deep deficiencies will remain in best track records,

deficiencies clouded by illusions of scrupulous consideration absent of new analyses. This

research and the unified WNP Superset outlined herein may give impetus to WMO Members and

scientific organizations across the region to fortify a plan to adequately address these issues

either through complete metadata or by means of a rigorous and systematic organizational

review of all best track archives.

46

APPENDIX A

STATISTICAL ANALYSIS

To evaluate trends in intensity among best track data, the Mann-Kendall statistical trend

test is performed on WNP Superset data. The Mann-Kendall test is a non-parametric (i.e.,

distribution independent) test which compares relative magnitudes of data rather than actual

values (Gilbert, 1987). No prior assumptions are made regarding the particular distribution of the

sample data.

In this study, unsmoothed annually accumulated PDI is the metric. The null hypothesis,

H0, is that the data are randomly ordered in time and exhibit no trend, with the alternative

hypothesis, H1, being that the data follow some monotonic trend at a determined level of

significance. The Mann-Kendall test is consistent with previous TC trend detection studies

(Webster et al., 2005; Kossin et al., 2007). The analysis procedure for calculating the Mann-

Kendall test statistic and associated probability is described below.

Mann-Kendall Test Statistic

Under H0, let x1, x2, xj, …xn denote a series of n data points, with xj as the jth datum in the

series. The Mann-Kendall test statistic, S, is calculated as

)(sgn1

1 1

kj

n

k

n

kj

xxS −=∑ ∑−

= +=

where

0,1

0,0

0,1

)sgn(

<−−

=−

>−+

=−

kj

kj

kj

kj

xx

xx

xx

xx

The initial value of S is assumed to be zero, implying no trend. The sample data are listed as an

ordered time series, with positive values and negative values computed between each data point

and its preceding values across a triangular table. The Table illustrates this method from the first

10 years of the 30-year annually accumulated PDI of the JTWC Superset examined in this study.

The test statistic S is derived from the sum of the totals computed along the bottommost row. A

large positive or negative value of S indicates the presence of an increasing or decreasing trend,

47

respectively. In order to statistically quantify whether S is significantly different from zero,

probability is determined.

Standard Normal Variate and Probability

The test statistic, S, is found by Kendall (1975) to be asymptotically normally distributed.

For n < 10, the theoretical distribution of S derived by both Mann (1945) and Kendall (1975) can

be compared to the absolute value of S (Gilbert, 1987). However, for n ≥ 10, as is the case in the

current study, a normal approximation test is used. Tied groups (i.e., two datum of equal value)

in time series with a sample size near 10 may invalidate the normal approximation, but since the

30-year PDI data have neither tied groups nor small n, the normal approximation holds.

The variance of the test statistic is described by

18/)52)(1()52)(1()(1

+−−+−= ∑=

p

j

jjj tttnnnSVar

where n is the number of data points, p is the number of tied groups, and tj is the number of

datum in the jth

tied group. According to Kendall (1975), a normal approximation can be used

once the standard normal variate, or normalized test statistic, Z is determined by

<+

=

>−

=

0,)]([

1

0,0

0,)]([

1

Z

2/1

2/1

SifSVar

S

Sif

SifSVar

S

Therefore, from the Table, S = 13, Var(S) = 269, and Z = 0.7 for the 10-year sample data.

Once the normalized test statistic Z is known, the standard normal distribution probability

density function, which has probability density

2

2

2

1)(

Z

eZf−

=

π

is employed to calculate probability. This probability is then compared to a level of significance

to resolve the null hypothesis. The significance level chosen for this study is 0.05, or

alternatively the 0.95 confidence level. This requirement of 0.95 confidence for statistical

significance of a trend is consistent with the literature.

48

For the sample data in the Table, a p-value of 0.23, or 0.77 confidence level, is computed.

Although S is positive, indicating the presence of an upward trend, the probability calculated is

less than the chosen confidence level, so H0 is accepted for the 10-year sample.

49

Table: 10-year sample JTWC PDI data from the WNP Superset, with positive and negative values listed for

each year. The Mann-Kendall test statistic S is computed from the summation of all total values.

Year 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986

PDI (m/s)3 18047116 25702890 34954720 26866012 25597972 40829528 28658616 32833536 24417980 38046284

1977 18047116 1 1 1 1 1 1 1 1 1

1978 25702890 1 1 -1 1 1 1 -1 1

1979 34954720 -1 -1 1 -1 -1 -1 1

1980 26866012 -1 1 1 1 -1 1

1981 25597972 1 1 1 -1 1

1982 40829528 -1 -1 -1 -1

1983 28658616 1 -1 1

1984 32833536 -1 1

1985 24417980 1

1986 38046284

Total 1 2 1 -2 5 2 3 -6 7

50

APPENDIX B

MANUAL TRACK CHANGES TO THE WNP SUPERSET

1. Storm 5 of the full track set is missing position and intensity estimates at 12Z on 7/2/51.

Storm 5 of the full track set matches storms 5,6,7 of 1951 from the JTWC, JMA, and STI

sets, respectively. Position and intensity estimates are included for storm 5 at 6Z and 18Z on

7/2/51. The 12Z point was not originally included in the full track set because the position

match falls outside of the prescribed 3.2° buffer. In order not to have breaks within the storm

track, this missing point is manually ingested back into all final uniform tracks. A flag value

of 1 is listed in each final uniform set for this point since at least two of the three position

matches fall outside of the prescribed 3.2° buffer.

2. Storm 24 of the full track set has duplicate matches for 8/7/52 at both 0Z and 6Z. The

duplicate match is found in Storm 10 of the STI set, which has two different serial numbers

for the same catalog number (Storm 10 of 1952 – Jeanne). The track with serial number

195210s appears to be an addendum to the track of serial number 195210. However, the two

tracks overlap at times 0Z and 6Z on 8/7/52 (i.e., two instances at each time). The latitude

and longitude values as well as the intensity values are similar for both instances at each

time, but differ slightly. It is unknown which instance for each time is correct, so the

addendum position and intensity values (195210s) are retained. The positions at 0Z and 6Z

on 8/7/52 from serial number 195210 are completely removed from all final uniform tracks.

3. Storm 27 of the full track set has duplicate matches for 9/1/52 at both 6Z and 12Z. The

duplicate match is found in Storm 16 of the STI set, which has two different serial numbers

for the same catalog number (Storm 16 of 1952 – Mary). The track with serial number

195216s appears to be an addendum to the track of serial number 195216. However, the two

51

tracks overlap at times 6Z and 12Z on 9/1/52 (i.e., two instances at each time). The latitude

and longitude values as well as the intensity values are similar for both instances at each

time, but differ slightly. It is unknown which instance for each time is correct, so the original

position and intensity values (195216) are retained, as they are more in line with the

surrounding position values and closer to the JMA position values than the addendum

values. The positions at 6Z and 12Z on 9/1/52 from serial number 195216s are completely

removed from all final uniform tracks.

4. Storm 41 of the full track set has duplicate matches for 11/27/52 at both 6Z and 12Z. The

duplicate match is found in Storm 31 of the STI set, which has two different serial numbers

for the same catalog number (Storm 31 of 1952 – Della). The track with serial number

195231s appears to be an addendum to the track of serial number 195231. However, the two

tracks overlap at times 6Z and 12Z on 11/27/52 (i.e., two instances at each time). The

latitude and longitude values as well as the intensity values are similar for both instances at

each time, but differ slightly. It is unknown which instance for each time is correct, so the

addendum position and intensity values (195231s) are retained, as they are more in line with

the surrounding position values and closer to the JMA and JTWC position values than the

original values. The positions at 6Z and 12Z on 11/27/52 from serial number 195231 are

completely removed from all final uniform tracks.

5. Storm 64 of the full track set is missing position and intensity estimates at 18Z on 11/1/53

and at 0Z on 11/2/53. Storm 64 of the full track set matches storms 20,19,26 of 1953 from

the JTWC, JMA, and STI sets, respectively. Position and intensity estimates are included for

storm 64 at 12Z on 11/1/53. The 18Z point on 11/1/53 and 0Z point on 11/2/53 were not

originally included in the full track set because the position matches between at least two of

52

the three sets fall outside of the prescribed 3.2° buffer. In order to avoid breaks within the

storm track, these missing points are manually ingested back into all final uniform tracks. A

flag value of 1 is listed in each final uniform set for these points since at least two of the

three position matches fall outside of the prescribed 3.2° buffer.

6. Storm 66 of the full track set is missing position estimates at 6Z on 11/28/53. Storms

22,21,29 from the JTWC, JMA, and STI sets, respectively, match to yield storm 66 of the

full track set. Storm 29 from the STI set falls out of the prescribed 3.2° buffer at 6Z on

11/28/53 when compared to the JTWC set (storm 22 of 1953). Therefore, while this point in

the STI set matches position estimates from storms 22 and 21 of 1953 in the JTWC and

JMA sets, respectively, it was not originally included in the full track set. The missing point

is manually added back into each final uniform set.

7. Storm 72 of the full track set has duplicate matches for 8/19/54 at both 0Z and 6Z. The

duplicate match is found in Storm 10 of the STI set, which has two different serial numbers

for the same catalog number (Storm 10 of 1954 – Grace). The track with serial number

195410s appears to be an addendum to the track of serial number 195410. However, the two

tracks overlap at times 0Z and 6Z on 8/19/54 (i.e., two instances at each time). The latitude

and longitude values as well as the intensity values are similar for both instances at each

time, but differ slightly. It is unknown which instance for each time is correct, so the original

position and intensity values (195410) are retained, as they are more in line with the

surrounding position values and closer to the JMA and JTWC position values than the

addendum values. The positions at 0Z and 6Z on 8/19/54 from serial number 195410s are

completely removed from all final uniform tracks.

53

8. Storm 96 of the full track set changed from 9,9,10 (JTWC, JMA, STI) to match the correct

storms 9,10,11.

9. Storm 99 of the full track set has duplicate matches for 8/24/55 at 0Z, 6Z, and 12Z. The

duplicate match is found in Storm 19 of the STI set, which has two different serial numbers

for the same catalog number (Storm 19 of 1955 – Iris). The track with serial number

195519s appears to be an addendum to the track of serial number 195519. However, the two

tracks overlap at times 0Z, 6Z, and 12Z on 8/24/55 (i.e., two instances at each time). The

latitude and longitude values as well as the intensity values are similar for both instances at

each time, but differ slightly. It is unknown which instance for each time is correct, so the

addendum position and intensity values (195519s) are retained, as they are closer to the

JMA and JTWC position values than the original values. The values at 0Z, 6Z, and 12Z on

8/24/55 from serial number 195519 are completely removed from all final uniform tracks.

10. Storm 102 of the full track set is missing position and intensity estimates at 12Z on 10/1/55.

Storm 102 of the full track set matches storms 15,22,27 of 1955 from the JTWC, JMA, and

STI sets, respectively. Position and intensity estimates are included for storm 102 at 6Z and

18Z on 10/1/55. The 12Z point was not originally included in the full track set because the

position match falls outside of the prescribed 3.2° buffer. In order not to have breaks within

the storm track, this missing point is manually ingested back into all final uniform tracks. A

flag value of 1 is listed in each final uniform set for this point since at least two of the three

position matches fall outside of the prescribed 3.2° buffer.

11. Storm 125 of the full track set is missing position and intensity estimates at 0Z on 9/25/56.

Storm 125 of the full track set matches storms 16,16,29 of 1956 from the JTWC, JMA, and

STI sets, respectively. Position and intensity estimates are included for storm 125 at 18Z on

54

9/24/56. The 0Z point on 9/25/56 was not originally included in the full track set because the

position match between at least two of the three sets falls outside of the prescribed 3.2°

buffer. In order to avoid breaks within the storm track, this missing point is manually

ingested back into all final uniform tracks. A flag value of 1 is listed in each final uniform

set for this point since at least two of the three position matches fall outside of the prescribed

3.2° buffer.

12. Storm 125 of the full track set reads 16,16,29 (JTWC, JMA, STI) then switches to 16,17,29.

The remove line program removes the 16,17,29 match since storm 125 is matched first to

storms 16,16,29. Storm 16 in the JMA ends on 9/25/56 at 0Z. The same storm appears to

reform as Storm 17 on 9/26/56 at 0Z. So data from 9/25/56 at 6Z to 9/26/56 at 18Z is

missing. It’s apparent from the JTWC and STI sets (whose tracks do not end or change) that

Storm 17 in the JMA is just a continuation of Storm 16. The 16,17,29 match originally taken

out of the full track set by the remove line program is added back to all final uniform tracks

(5 lines – 9/26/56 at 0Z through 9/27/56 at 0Z). Since the JMA data set records this system

as two separate storms and the JTWC and STI data sets keep it as one continuous storm, a

flag value of 2 is listed in each final uniform set for the 16,17,29 match of storm 125. This

flag value also indicates a break within storm 125 of the full track set between 0Z on 9/25/56

and 0Z on 9/26/59.

13. Storm 129 of the full track set reads 21,21,34 (JTWC, JMA, STI). This is an incorrect

match. It should read 21,22,35.

14. Storm 130 of the full track set reads 22,21,24 (JTWC, JMA, STI). This storm was removed

from the final track, since the incorrect Storm 129 read 21,21,34 (JTWC,JMA,STI). The 21

55

and 34 storms from the JMA and STI, respectively, were the same as the incorrect JMA and

STI for Storm 129. These six lines were added back to all three data sets.

15. Storm 143 of the full track set is missing position and intensity estimates at 6Z, 12Z, and

18Z on 9/19/57. Storm 143 of the full track set matches storms 12,13,18 of 1957 from the

JTWC, JMA, and STI sets, respectively. Position and intensity estimates are included for

storm 143 at 0Z on 9/19/57 and 0Z on 9/20/57. The 6Z, 12Z, and 18Z points were not

originally included in the full track set because the position matches fall outside of the

prescribed 3.2° buffer. In order not to have breaks within the storm track, the missing points

are manually ingested back into all final uniform tracks. A flag value of 1 is listed in each

final uniform set for these points since at least two of the three position matches fall outside

of the prescribed 3.2° buffer.

16. Storm 168 of the full track set (“June”) only has matches for 0Z and 6Z on 9/20/58. The

track of the storm in all three sets (16,23,26 of 1958 from the JTWC, JMA and STI,

respectively) extends beyond 6Z on 9/20/58; however, the match between at least two of the

sets falls outside of the prescribed 3.2° buffer through 18Z on 9/21/58 (end point of the STI

track for “June”). While track differences are large at times (e.g.,12° latitude difference at

18Z on 9/21/58 between the JMA and STI sets), it is clearly the same cyclone in all sets.

Because the system is classified as a strong tropical storm in all three data sets (65kt msw in

the JTWC and 985mb mslp in the STI) during the period after 6Z on 9/20/58, the matches

thereafter that fall outside of the 3.2° buffer (from 12Z on 9/20/58 through 18Z on 9/21/58)

are included in the final, unified Superset.

17. Storm 175 of the full track set reads 23,29,34 (JTWC, JMA, STI). Storms 29 and 34 from

the JMA and STI, respectively, are already matched to storm 21 of 1958 from the JTWC

56

(storm 173 of the full track set). Storm 23 of 1958 from the JTWC appears to be a duplicate,

correction, or addendum to storm 21 of the same year. The original match is retained (storm

173) and storm 175 of the full track set is completely removed from all final uniform tracks.

18. Storm 176 of the full track set reads 24,23,26 (JTWC, JMA, STI). Storms 23 and 26 from

the JMA and STI, respectively, are already matched to storm 16 of 1958 from the JTWC

(storm 168 of the full track set). Storm 24 of 1958 from the JTWC may be a duplicate or

addendum to storm 16 of the same year. JTWC storm notes from 1958 comment storm 24 is

a central North Pacific cyclone. For these reasons, storm 176 of the full track set is

completely removed from all final uniform tracks.

19. Storm 185 of the full track set has duplicate matches for 8/31/59 at 12Z. The duplicate

match is found in Storm 14 of the STI set, which has two different serial numbers for the

same catalog number (Storm 14 of 1959 – Joan). The track with serial number 195914s

appears to be an addendum to the track of serial number 195914. However, the two tracks

overlap at 12Z on 8/31/59 (i.e., two instances). The latitude and longitude values as well as

the intensity values are similar for both instances, but differ slightly. It is unknown which

instance is correct, so the original position and intensity values (195914) are retained, as

they are more in line with the surrounding position values and closer to the JMA position

values than the addendum values. The positions at 12Z on 8/31/59 from serial number

195914s are completely removed from all final uniform tracks.

20. Storm 208 of the full track set has duplicate matches for 8/8/60 at 6Z. The duplicate match is

found in Storm 13 of the STI set, which has two different serial numbers for the same

catalog number (Storm 13 of 1960 – Trix). The track with serial number 196013s appears to

be an addendum to the track of serial number 196013. However, the two tracks overlap at

57

6Z on 8/8/60 (i.e., two instances). The latitude and longitude values as well as the intensity

values are similar for both instances, but differ slightly. It is unknown which instance is

correct, so the original position and intensity values (196013) are retained, as they are more

in line with the surrounding position values and closer to the JMA and JTWC position

values than the addendum values. The values at 6Z on 8/8/60 from serial number 196013s

are completely removed from all final uniform tracks.

21. Storm 232 of the full track set has duplicate matches for 5/26/61 at 18Z and 5/27/61 at 0Z.

The duplicate match is found in Storm 5 of the STI set, which has two different serial

numbers for the same catalog number (Storm 5 of 1961 – Betty). The track with serial

number 196105s appears to be an addendum to the track of serial number 196105. However,

the two tracks overlap at times 18Z on 5/26/61 and 0Z on 5/27/61 (i.e., two instances at each

time). The latitude and longitude values as well as the intensity values are similar for both

instances at each time, but differ slightly. It is unknown which instance for each time is

correct, so the original position and intensity values (196105) are retained, as they are more

in line with the surrounding position values and closer to the JMA and JTWC position

values than the addendum values. The values at 18Z on 5/26/61 and 0Z on 5/27/61 from

serial number 196105s are completely removed from all final uniform tracks.

22. Storm 234 of the full track set begins with 8,6,11 (JTWC, JMA, STI). This is an incorrect

match. Storm 6 and 11 from the JMA and STI, respectively, are not found in the JTWC data

set. Storm 8 of the JTWC from 1961 correctly matches storms 7 and 12 of the JMA and STI.

Lines from Storm 234 with match 8,7,12 were originally removed. Since they are correct,

the 20 lines are added back to the final tracks of each of the three sets. The 8,6,11 match

originally included is removed from the final sets.

58

23. Storm 234 of the full track set has duplicate matches at 12Z and 18Z on 6/30/61 and at 0Z

on 7/1/61. The duplicate matches are found in Storm 12 of the STI set, which has two

different serial numbers for the same catalog number (Storm 12 of 1961 – Doris). The track

with serial number 196112s appears to be an addendum to the track of serial number

196112. However, the two tracks overlap at 12Z and 18Z on 6/30/61 and at 0Z on 7/1/61

(i.e., two instances at each time). The latitude and longitude values as well as the intensity

values are similar for both instances at each time, but differ slightly. It is unknown which

instance is correct, so the original position and intensity values (196112) are retained, as

they are more in line with the surrounding position values and closer to the JMA and JTWC

position values than the addendum values. The values at 12Z and 18Z on 6/30/61 and at 0Z

on 7/1/61 from serial number 196112s are completely removed from all final uniform tracks.

24. Storm 302 of the full track set reads 59,27,33 (JTWC, JMA, STI). Storms 27 and 33 of 1962

from the JMA and STI sets, respectively, are also assigned to storm 289 of the full track set

(storm 26 of 1962 from the JTWC set). Storm 302 of the full track set is retained; however,

since the JTWC data set records this system as two separate storms and the JMA and STI

sets keep it as one continuous storm, a flag value of 2 is listed in each final uniform set for

the 59,27,33 match of storm 302.

25. Storm 309 of the full track set has duplicate matches for 7/18/63 at 18Z. The duplicate

match is found in Storm 8 of the STI set, which has two different serial numbers for the

same catalog number (Storm 8 of 1963 – Wendy). The track with serial number 196308s

appears to be an addendum to the track of serial number 196308. However, the two tracks

overlap at 18Z on 7/18/63 (i.e., two instances). The latitude and longitude values as well as

the intensity values are similar for both instances, but differ slightly. It is unknown which

59

instance is correct, so the original position and intensity values (196308) are retained, as

they are more in line with the surrounding position values and closer to the JTWC position

values than the addendum values. The values at 18Z on 7/18/63 from serial number 196308s

are completely removed from all final uniform tracks.

26. Storm 327 of the full track set is missing position and intensity estimates at 18Z on 12/27/63

through 12Z on 12/28/63. Storm 327 of the full track set matches storms 25,24,33 of 1963

from the JTWC, JMA, and STI sets, respectively. Position and intensity estimates are

included for storm 327 at 12Z on 12/27/63 and at 18Z on 12/28/63. The 18Z point on

12/27/63 and 0Z, 6Z, and 12Z points on 12/28/63 were not originally included in the full

track set because the position matches between the JTWC and JMA sets and STI and JMA

sets fall outside of the prescribed 3.2° buffer. In order to avoid breaks within the storm track,

these missing points are manually ingested back into all final uniform tracks. A flag value of

1 is listed in each final uniform set for these points since at least two of the three position

matches fall outside of the prescribed 3.2° buffer. A comparison of position estimates in all

three sets suggests that the JMA longitude position estimates from the original JMA data set

for this period (18Z on 12/27/63 through 12Z on 12/28/63) may be in error.

27. Storm 373 of the full track set reads 54,27,33 (JTWC, JMA, STI). This is an incorrect

match. Storms 27 and 33 from the JMA and STI sets, respectively, are correctly matched to

storm 362 of the full track set (storm 32 of 1964 from the JTWC set). Storm 54 of 1964

from the JTWC set does not have a match in either the JMA or STI data sets. As such, storm

373 is completely removed from all final uniform tracks (1 line).

28. Storm 388 of the full track set has duplicate matches for 7/26/65 at both 0Z and 6Z. The

duplicate match is found in Storm 21 of the STI set, which has two different serial numbers

60

for the same catalog number (Storm 21 of 1965 – Harriet). The track with serial number

196521s appears to be an addendum to the track of serial number 196521. However, the two

tracks overlap at times 0Z and 6Z on 7/26/65 (i.e., two instances at each time). The latitude

and longitude values as well as the intensity values are similar for both instances at each

time, but differ slightly. It is unknown which instance for each time is correct, so the original

position and intensity values (196521) are retained, as they are more in line with the

surrounding position values and closer to the JMA and JTWC position values than the

addendum values. The values at 0Z and 6Z on 7/26/65 from serial number 196521s are

completely removed from all final uniform tracks.

29. Storm 408 of the full track set is completely removed from all final tracks. Storm 408 reads

34,32,40 (JTWC, JMA, STI). Storm 407 of the full track set reads 33,32,40 (JTWC, JMA,

STI). Because the JMA and JTWC storm numbers are the same for Storms 407 and 408, all

storms in storm 408 are removed by the remove line program. Storms 33 and 34 from the

JTWC appear to be the same storm, with Storm 34 being a continuation of Storm 33. Both

storms, however, have points for 11/19/65 at 12Z (11/19/65 being the last point for Storm 33

and the first point for Storm 34). However, it is left as two storms and the 27 removed lines

from Storm 408 are added back to all the final tracks.

30. Storms 409-411 of the full track set missing

31. Storm 412 completely removed from full track set; Storms 20 and 28 from 1965 from the

JMA and STI data sets, respectively, had already been used for Storm 395 of the full track

set.

32. Storms 413-415 of the full track set missing

61

33. Storm 416 of the full track set begins with 1,1,1 (JTWC, JMA, STI). This is an incorrect

match. Storm 1 of the STI is not found in the JTWC or JMA data sets. As a result lines from

Storm 416 with match 1,1,2 (JTWC, JMA, STI), the correct match, were originally

removed. Since they are correct, the 35 lines are added back to the final tracks of each of the

three sets. The 1,1,1 matches (two lines) originally included are removed from the final sets.

34. Storm 446 completely removed from full track set (4 lines); Storms 12 and 15 from 1966

from the JMA and STI data sets, respectively, had already been used for Storm 426 of the

full track set. Storm 51 from 1966 in the JTWC data set appears to be a continuation (and

addendum) of Storm 11 from 1966. Storm 11 from the JTWC end on 8/16/66 at 6Z. Storm

51 from the JTWC picks up at 8/17/66 at 12Z.

35. Storm 447 of the full track set missing

36. Storm 449 completely removed from full track set (4 lines); Storms 22 and 26 from 1966

from the JMA and STI data sets, respectively, had already been used for Storm 435 of the

full track set. Storm 54 from 1966 in the JTWC data set does not appear in the JMA or STI

sets.

37. Storm 450 of the full track set missing

38. Storm 451 of the full track set missing

39. Storm 463 of the full track set reads 9,9,13 (JTWC, JMA, STI). This is an incorrect match. It

should read 9,10,14. All 9,9,13 lines are removed (10 lines) and the 9,10,14 lines taken out

by the remove line program are added back in (31 lines).

40. Storm 473 has two points for the same date/time from 8/29/67 at 18Z to 8/30/67 at 12Z

(supposed to be 4 points, but 8 points in the final track files, as each point is listed twice).

This is happening because the STI set has two instances of “Nora”, Storm 31 of 1967. The

62

first “Nora” track is taken, since this is the longer and seemingly correct track. The second

“Nora” track only has the 4 points toward the end (8/29/67 at 18Z to 8/30/67 at 12Z). 4

instances of repeat lines are removed from each final set.

41. Storm 487 of the full track set has duplicate matches for 11/18/67 at both 12Z and 18Z. The

duplicate match is found in Storm 51 of the STI set, which has two different serial numbers

for the same catalog number (Storm 51 of 1967 – Gilda). The track with serial number

196751s appears to be an addendum to the track of serial number 196751. However, the two

tracks overlap at times 12Z and 18Z on 11/18/67 (i.e., two instances at each time). The

latitude and longitude values as well as the intensity values are similar for both instances at

each time, but differ slightly. It is unknown which instance for each time is correct, so the

original position and intensity values (196751) are retained, as they are more in line with the

surrounding position values and closer to the JMA and JTWC position values than the

addendum values. The values at 12Z and 18Z on 11/18/67 from serial number 196751s are

completely removed from all final uniform tracks.

42. Storm 490 completely removed from the full track set (five lines); Storms 10 and 14 from

1967 from the JMA and STI data sets, respectively, had already been used for Storm 463 of

the full track set. The first four or five lines of Storm 9 of the JTWC of 1967 look

problematic. They seem to align more closely to storm 9 and 13 of the JMA and STI,

respectively. However, from about 7/22/67 at 12Z onward, it’s obvious that Storm 9 of the

JTWC of 1967 is the same as storms 10 and 14 from the JMA and STI, respectively. Storm

51 of the JTWC of 1967 should probably be the beginning point of Storm 9 of 1967. Since

this is uncertain, Storm 51 of the JTWC, which corresponds to Storm 490 from the full track

set, is completely removed.

63

43. Storm 491 completely removed from the full track set (2 lines); Storm 11 from 1967 from

the JMA set was already used for Storm 464 of the full track set. While Storm 15 from the

STI matches Storm 52 from the JTWC from 1967, there is no match from the JMA set.

44. Storm 500 of the full track set has duplicate matches for 7/25/68 at both 12Z and 18Z. The

duplicate match is found in Storm 9 of the STI set, which has two different serial numbers

for the same catalog number (Storm 9 of 1968 – Nadine). The track with serial number

196809s appears to be an addendum to the track of serial number 196809. However, the two

tracks overlap at times 12Z and 18Z on 7/25/68 (i.e., two instances at each time). The

latitude and longitude values as well as the intensity values are similar for both instances at

each time, but differ slightly. It is unknown which instance for each time is correct, so the

original position and intensity values (196809s) are retained, as they are more in line with

the surrounding position values and closer to the JMA and JTWC position values than the

addendum values. The positions at 12Z and 18Z on 7/25/68 from serial number 196809s are

completely removed from all final uniform tracks.

45. Storm 507 of the full track set has duplicate matches for 9/6/68 at 0Z, 6Z, and 12Z. The

duplicate matches are found in Storm 21 of the STI set, which has two different serial

numbers for the same catalog number (Storm 21 of 1968 – Wendy). The track with serial

number 196821s appears to be an addendum to the track of serial number 196821. However,

the two tracks overlap at 0Z, 6Z, and 12Z on 9/6/68 (i.e., two instances). The latitude and

longitude values as well as the intensity values are similar for both instances, but differ

slightly. It is unknown which instance is correct, so the original position and intensity values

(196821) are retained, as they are more in line with the surrounding position values and

closer to the JMA and JTWC position values than the addendum values. The values at 0Z,

64

6Z, and 12Z on 9/6/68 from serial number 196821s are completely removed from all final

uniform tracks.

46. Storm 607 of the full track set has duplicate matches for 10/10/71 at 12Z and 18Z and for 0Z

on 10/11/71. The duplicate matches are found in Storm 40 of the STI set, which has two

different serial numbers for the same catalog number (Storm 40 of 1971 – Faye). The track

with serial number 197140s appears to be an addendum to the track of serial number

197140. However, the two tracks overlap at 12Z and 18Z on 10/10/71 and at 0Z on 10/11/71

(i.e., two instances at each time). The latitude and longitude values as well as the intensity

values are similar for both instances at each time, but differ slightly. It is unknown which

instance is correct, so the original position and intensity values (197140) are retained for

12Z and 18Z on 10/10/71 and the addendum position and intensity values are retained for 0Z

on 10/11/71. These values are more in line with the surrounding position values and closer

to the JMA position values. The values at 12Z and 18Z on 10/10/71 from serial number

197140s and the values at 0Z on 10/11/71 from serial number 197140 are completely

removed from all final uniform tracks.

47. Storm 608 completely removed by the remove line program, as no instance of Storm 32

from 1971 is found in the JMA or STI sets. Storms 32 and 40 from the JMA and STI sets,

respectively, were already used for Storm 607 of the full track set.

48. Storm 625 of the full track set changed from 12,11,15 (JTWC, JMA, STI) to 12,12,16. The

remove line program kept an incorrect line and removed the correct line. The one incorrect

line was removed and the 21 correct lines are added back to each final set.

49. Storm 657 of the full track set has duplicate matches for 8/20/73 at 0Z through 8/21/73 at

18Z. The duplicate matches are found in Storm 12 of the STI set, which has two different

65

serial numbers for the same catalog number (Storm 12 of 1973 – Joan). The track with serial

number 197312s appears to be an addendum to the track with serial number 197312.

However, the two tracks overlap at 0Z, 6Z, 12Z, and 18Z on 8/20/73 and at 0Z, 6Z, 12Z, and

18Z on 8/21/73 (i.e., two instances at each time). It is unknown which instance is correct, so

the original position and intensity values (197312) are retained, as they are more in line with

the surrounding position values and closer to the JMA and JTWC position values than the

addendum values. The values at 0Z, 6Z, 12Z, and 18Z on 8/20/73 and at 0Z, 6Z, 12Z, and

18Z on 8/21/73 from serial number 197312s are completely removed from all final uniform

tracks.

50. Storm 683 of the full track set should begin on 8/11/74 at 0Z; all data sets have that

date/time. However, it begins on 8/12/74 at 0Z. Added points from 8/11/74 at 0Z to 8/11/74

at 18Z to all final sets. The STI set has three separate tracks for Storm 18 of 1974 (“Mary”).

As a result points 8/20/74 at 0Z, 8/20/74 at 6Z, and 8/20/74 at 12Z appear twice in the final

set while points 8/21/74 at 18Z and 8/22/74 at 0Z appear three times in the final set. The first

track of “Mary” is used until it ends on 8/21/74 at 0Z. The third instance of track “Mary”

(Storm 18 of 1974 from STI) is used from point 8/21/74 at 6Z onward. The second instance

of track “Mary” (5 lines) is discarded from the STI set.

51. Storm 687 completely removed from the full track set (4 lines); Storm 17 and 24 from the

JMA and STI, respectively, belong with Storm 20 of 1974 from the JTWC (Storm 688 of the

full track set). A matching storm from the JMA set cannot be found for Storm 19 of 1974

from the JTWC set. Storm 25 from the STI appears to match Storm 19 of 1974 from the

JTWC set.

66

52. Storm 688 of the full track set changed from 20,16,23 (JTWC, JMA, STI) to 20,17,24.

Storms 17 and 24 of 1974 from the JMA and STI, respectively, were incorrectly matched

with Storm 687. The remove line program saw these two storms already matching with

Storm 687 of the full track set, so incorrectly kept Storms 16 and 23 from the JMA and STI

as the correct matches to Storm 20 from the JTWC. All three sets (20,17,24) also included

the point 8/30/74 at 18Z, but for some reason this was not included in the int or check files.

This point was manually added to each final set.

53. Storm 711 of the full track set changed from 8,4,9 (JTWC, JMA, STI) to 8,6,12. Storms 4

and 9 of 1975 from the JMA and STI sets, respectively, had already been matched with

Storm 6 of 1975 from the JTWC (Storm 709 of the full track set). All three sets (8,6,12) also

included the point 8/17/75 at 0Z, but for some reason this was not included in the int or

check files. This point was manually added to each final set.

54. Storm 721 completely removed from the full track set (1 line); Storms 16 and 25 of 1975

from the JMA and STI sets, respectively, match with Storm 19 of 1975 from the JTWC.

Storm 18 of 1975 from the JTWC appears to be the beginnings of Storm 19. However,

Storm 18 (Storm 721 from the full track set) is thrown out since there are no complete

matches from the JMA and STI sets.

55. Storm 723 of the full track set is missing points at 12Z and 18Z on 10/27/75. Storms

20,17,26 from the JTWC, JMA, and STI sets, respectively, match to yield storm 723 of the

full track set. However, at 12Z on 10/27/75 the longitude of storm 26 in the STI set falls

outside of the prescribed 3.2° buffer when compared to the JMA set. Storm 26 of the STI set

also falls outside of the prescribed buffer on 10/27/75 at 18Z when compared to both the

JTWC and JMA sets (storms 20 and 17 of 1975). Therefore, while these points in the STI set

67

match position estimates from storms 20 and 17 of 1975 in the JTWC and JMA sets,

respectively, they were not originally included in the full track set. The two missing points

are manually added back into each final uniform set, making storm 723 of the full track set

continuous (i.e., no breaks within the storm track).

56. Storm 727 completely removed from the full track set (3 lines); Storm 24 of 1975 from the

JTWC is not found in either the JMA or STI sets. The program incorrectly matches Storms

21 and 30 from the JMA and STI sets, respectively, to Storm 24 of the JTWC set. They

should be correctly paired with Storm 25 of 1975 from the JTWC (see below).

57. Storm 728 from the full track set is added back to the final tracks. The remove line program

incorrectly removed these 17 lines because Storms 21 and 30 from the JMA and STI sets,

respectively, had already been matched (incorrectly) to Storm 727.

58. Storm 746 of the full track set has duplicate matches for 9/13/76 at 0Z and 6Z. The duplicate

matches are found in Storm 22 of the STI set, which has two different serial numbers for the

same catalog number (Storm 22 of 1976 – Georgia). The track with serial number 197622s

appears to be an addendum to the track of serial number 197622. However, the two tracks

overlap at 0Z and 6Z on 9/13/76 (i.e., two instances at each time). The latitude and longitude

values as well as the intensity values are similar for both instances at each time, but differ

slightly. It is unknown which instance is correct, so the addendum position and intensity

values (197622s) are retained, as they are more in line with the surrounding position values

and closer to the JMA position values than the original values. The values at 0Z and 6Z on

9/13/76 from serial number 197622 are completely removed from all final uniform tracks.

59. Storm 762 of the full track set has duplicate matches for 8/20/77 at 12Z through 8/22/77 at

12Z. The duplicate matches are found in Storm 12 of the STI set, which has two different

68

serial numbers for the same catalog number (Storm 12 of 1977 – Amy). The tracks with

serial numbers 197712s, 197712s-1, and 197712s-2 appear to be addendums to the track of

serial number 197712. However, the two tracks overlap between 12Z on 8/20/77 and 12Z on

8/22/77 (i.e., two instances at each time). In fact, multiple duplicates are included (i.e., three

instances at one time) at 6Z on 8/22/77. The latitude and longitude values as well as the

intensity values are similar for all instances at each time, but differ slightly. It is unknown

which instance is correct for each time, so the original position and intensity values

(197712) are retained, as they are more in line with the surrounding position values and

closer to the JMA position values than the addendum values. The values at 12Z and 18Z on

8/20/77, at 0Z, 6Z, 12Z, and 18Z on 8/21/77, and at 0Z, 6Z, and 12Z on 8/22/77 from serial

numbers 197712s, 197712s-1, and 197712s-2 are completely removed from all final uniform

tracks.

60. Storm 794 of the full track set is missing position and intensity estimates at 12Z on 9/22/78.

Storm 794 of the full track set matches storms 20,20,26 of 1978 from the JTWC, JMA, and

STI sets, respectively. Position and intensity estimates are included for storm 794 at 6Z and

18Z on 9/22/78. The 12Z point was not originally included in the full track set because the

position match falls outside of the prescribed 3.2° buffer. In order not to have breaks within

the storm track, this missing point is manually ingested back into all final uniform tracks. A

flag value of 1 is listed in each final uniform set for this point since at least two of the three

position matches fall outside of the prescribed 3.2° buffer.

61. Storm 807 of the full track set is missing position and intensity estimates at 18Z on 1/9/79

and at 6Z on 1/10/79. Storm 807 of the full track set matches storms 1,1,1 of 1979 from the

JTWC, JMA, and STI sets, respectively. Position and intensity estimates are included for

69

storm 807 at 12Z on 1/9/79 and 0Z on 1/10/79; position and intensity estimates are also

included at 12Z on 1/10/79. The 18Z and 6Z points were not originally included in the full

track set because the position matches fall outside of the prescribed 3.2° buffer. In order not

to have breaks within the storm track, these missing points are manually ingested back into

all final uniform tracks. A flag value of 1 is listed in each final uniform set for these points

since at least two of the three position matches fall outside of the prescribed 3.2° buffer.

62. Storm 824 of the full track set is missing position and intensity estimates at 18Z on 9/17/79.

Storm 824 of the full track set matches storms 18,15,24 of 1979 from the JTWC, JMA, and

STI sets, respectively. Position and intensity estimates are included for storm 824 at 12Z on

9/17/79 and at 0Z on 9/10/79. The 18Z point was not originally included in the full track set

because the position match falls outside of the prescribed 3.2° buffer. In order not to have

breaks within the storm track, this missing point is manually ingested back into all final

uniform tracks. A flag value of 1 is listed in each final uniform set for this point since at

least two of the three position matches fall outside of the prescribed 3.2° buffer.

63. Storm 830 of the full track set has duplicate matches for 11/7/79 at 0Z and 6Z. The duplicate

matches are found in Storm 31 of the STI set, which has two different serial numbers for the

same catalog number (Storm 31 of 1979 – Vera). The track with serial number 197931s

appears to be an addendum to the track of serial number 197931. However, the two tracks

overlap at 0Z and 6Z on 11/7/79 (i.e., two instances at each time). The latitude and longitude

values as well as the intensity values are similar for both instances at each time, but differ

slightly. It is unknown which instance is correct, so the original position and intensity values

(197931) are retained, as they are more in line with the surrounding position values and

70

closer to the JMA position values than the addendum values. The values at 0Z and 6Z on

11/7/79 from serial number 197931s are completely removed from all final uniform tracks.

64. Storm 850 of the full track set reads 16,13,19 (JTWC, JMA, STI). Storm 851 of the full

track set reads 17,13,21 (JTWC, JMA, STI). Both sets are correct matches; however, JMA

keeps the system as one storm. The JTWC and STI record two separate storms for the one

storm in the JMA set. The system is kept as two separate storms in the unified set, since

position estimates are missing from 9/5/80 at 12Z through 9/5/80 at 18Z from the JTWC set.

Storm 19 and Storm 21 of 1981 in the original STI set could be combined into one system,

but it is unclear if the two storms are from the same parent circulation, as indicated by the

JMA data. Since the JTWC and STI data sets record this system as two separate storms and

the JMA data set keeps it as one continuous storm, a flag value of 2 is listed in each final

unified set for the 17,13,21 match of storm 851.

65. Storm 864 of the full track set is missing position and intensity estimates at 18Z on 4/19/81.

Storm 864 of the full track set matches storms 2,2,2 of 1981 from the JTWC, JMA, and STI

sets, respectively. Position and intensity estimates are included for storm 864 at 12Z on

4/19/81 and at 0Z on 4/20/81. The 18Z point was not originally included in the full track set

because the position match falls outside of the prescribed 3.2° buffer. In order not to have

breaks within the storm track, this missing point is manually ingested back into all final

uniform tracks. A flag value of 1 is listed in each final uniform set for this point since at

least two of the three position matches fall outside of the prescribed 3.2° buffer.

66. Storm 873 of the full track set reads 11,11,12 (JTWC, JMA, STI). It is uncertain whether or

not storm 12 of the STI set is the correct match to storms 11 and 11 of 1981 from the JTWC

and JMA sets. Storm 12 of the STI set is correctly matched to storm 874 of the full track set

71

(storms 12 and 12 from the JTWC and JMA sets). It is possible that storms 11 and 11 of

1981 from the JTWC and JMA sets are the incipient disturbance of storm 874 of the full

track set. If this is true, the beginning of storm 12 from the STI set may be the correct match

for storms 11 and 11 from the JTWC and JMA sets. Due to these uncertainties, storm 873 is

retained in the full track set; however, a flag value of 2 is listed in each final uniform set for

the 11,11,12 match of storm 873, since storm 12 of the STI set is already correctly matched

to storm 874.

67. Storm 873 of the full track set is missing position and intensity estimates at 0Z on 8/2/81.

Storm 873 of the full track set matches storms 11,11,12 of 1981 from the JTWC, JMA, and

STI sets, respectively. Position and intensity estimates are included for storm 873 at 18Z on

8/1/81. The 0Z point was not originally included in the full track set because the position

match between the JMA and STI sets falls outside of the prescribed 3.2° buffer. This

missing point should be included since it is found in all three data sets. The point is

manually ingested back into all final uniform tracks; however, a flag value of 1, indicating

track ambiguities greater than 3.2°, is not listed in each final uniform set since a flag value

of 2 is already assigned to this point (see discussion for track change #65).

68. Storm 890 of the full track set has duplicate matches for 12/20/81 at 6Z. The duplicate

match is found in Storm 32 of the STI set, which has two different serial numbers for the

same catalog number (Storm 32 of 1981 – Kit). The track with serial number 198132s

appears to be an addendum to the track of serial number 198132. However, the two tracks

overlap at 6Z on 12/20/81 (i.e., two instances). The latitude and longitude values as well as

the intensity values are similar for both instances, but differ slightly. It is unknown which

instance is correct, so the addendum position and intensity values (198132s) are retained, as

72

they are more in line with the surrounding position values and closer to the JMA and JTWC

position values than the original values. The values at 6Z on 12/20/81 from serial number

198132 are completely removed from all final uniform tracks.

69. Storm 899 completely removed from the full track set (1 line); Storm 8 of 1982 from the

JTWC set looks to be a continuation of Storm 6 of 1982 from the JTWC. It is not certain if

this is the case, however. The STI set has a system (Storm 8, “Val”) that matches with the

JTWC Storm 8 of 1982. Storm 6 of 1982 from the STI (“Tess”) matches Storm 6 of 1982

from the JTWC. The JMA appears to have combined “Tess” and “Val” into one system in

their best track set (Storm 6 of 1982). Since the points after 7/2/82 at 0Z are called into

question, the JMA track is cut at that point. Storm 8 of 1982 from the JTWC and “Val”

(Storm 8) of 1982 from the STI are thrown out.

70. Storm 901 of the full track set has duplicate matches for 7/29/82 at 0Z, 6Z, and 12Z. The

duplicate match is found in Storm 10 of the STI set, which has two different serial numbers

for the same catalog number (Storm 10 of 1982 – Andy). The track with serial number

198210s appears to be an addendum to the track of serial number 198210. However, the two

tracks overlap at 0Z, 6Z, and 12Z on 7/29/82 (i.e., two instances at each time). The latitude

and longitude values as well as the intensity values are similar for both instances at each

time, but differ slightly. It is unknown which instance is correct for each time, so the original

position and intensity values (198210) are retained, as they are more in line with the

surrounding position values and closer to the JMA and JTWC position values than the

addendum values. The values at 0Z, 6Z, and 12Z on 7/29/82 from serial number 198210s are

completely removed from all final uniform tracks.

73

71. Storm 941 from the full track set changed from 22,21,23 (JTWC, JMA, STI) to 22,21,24.

Storm 24 of 1983 from the STI appears to be a continuation of Storm 23. Four points are

missing between Storm 23 and 24 of 1983 in the STI, however. While Storm 23 from the

STI matches with Storms 22 and 21 of 1983 from the JTWC and JMA, respectively, it is

very weak during this period. Storm 24 of 1983 from the STI, on the other hand, becomes a

tropical storm (“Ruth”). In order not to inflate storm numbers, only one of the STI tracks

was chosen (the latter track, Storm 24). Storm 23 from the STI and the corresponding points

from the JTWC and JMA sets were disregarded.

72. Storm 946 of the full track set has duplicate matches for 7/3/84 at 6Z. The duplicate match is

found in Storm 3 of the STI set, which has two different serial numbers for the same catalog

number (Storm 3 of 1984 – Alex). The track with serial number 198403s appears to be an

addendum to the track of serial number 198403. However, the two tracks overlap at 6Z on

7/3/84 (i.e., two instances). The latitude and longitude values as well as the intensity values

are similar for both instances, but differ slightly. It is unknown which instance is correct, so

the original position and intensity values (198403) are retained, as they are more in line with

the surrounding position values and closer to the JMA and JTWC position values than the

addendum values. The values at 6Z on 7/3/84 from serial number 198403s are completely

removed from all final uniform tracks.

73. Storm 954 of the full track set reads 11,10,11 (JTWC, JMA, STI). This is an incorrect

match. It should read 11,10,13. While storm 11 of 1984 from the STI may be the beginnings

of storm 954 of the full track set, it is uncertain, especially since the two storms (11 and 13

from the STI) overlap from 12Z on 8/15/84 through 0Z on 8/17/84. As a result, all 11,10,11

74

lines are removed (3 lines) and the 11,10,13 lines taken out by the remove line program are

added back in (27 lines).

74. Storm 987 from the full track set changed from 14,12,18 (JTWC, JMA, STI) to 14,14,20.

Storms 12 and 18 from the JMA and STI, respectively, had already been matched with

Storm 985. Storms 14 and 20 of 1985 from the JMA and STI are the correct matches for

Storm 14 of 1985 from the JTWC.

75. Storm 1009 of the full track set has duplicate matches for 8/1/86 at 6Z. The duplicate match

is found in Storm 13 of the STI set, which has two different serial numbers for the same

catalog number (Storm 13 of 1986 – Sarah). The track with serial number 198613s appears

to be an addendum to the track of serial number 198613. However, the two tracks overlap at

6Z on 8/1/86 (i.e., two instances). The latitude and longitude values as well as the intensity

values are similar for both instances, but differ slightly. It is unknown which instance is

correct, so the original position and intensity values (198613) are retained, as they are more

in line with the surrounding position values and closer to the JMA and JTWC position

values than the addendum values. The values at 6Z on 8/1/86 from serial number 198613s

are completely removed from all final uniform tracks.

76. Storm 1051 of the full track set is missing position and intensity estimates at 18Z on

11/20/87. Storm 1051 of the full track set matches storms 23,22,25 of 1987 from the JTWC,

JMA, and STI sets, respectively. Position and intensity estimates are included for storm

1051 at 12Z on 11/20/87 and at 0Z on 11/21/87. The 18Z point was not originally included

in the full track set because the position match falls outside of the prescribed 3.2° buffer. In

order not to have breaks within the storm track, this missing point is manually ingested back

75

into all final uniform tracks. A flag value of 1 is listed in each final uniform set for this point

since at least two of the three position matches fall outside of the prescribed 3.2° buffer.

77. Storm 1091 completely removed from the full track set (5 lines); Storm 12 of 1989 from the

JTWC set cannot be accurately matched with any storm from the JMA set. Storm 15 of 1989

from the STI, however, does appear to be the same system as Storm 12 from the JTWC. The

matching system for Storm 1091 from the JMA set (Storm 12) matches with Storm 13 of

1989 from the JTWC, not Storm 12.

78. Storm 1092 is added back to the final sets. The remove line program incorrectly removed

this system from the final sets, since Storms 12 and 14 of 1989 from the JMA and STI sets,

respectively, had already been matched (incorrectly) with Storm 1091.

79. Storm 1100 of the full track set has duplicate matches for 9/11/89 at 12Z and 18Z and

9/12/89 at 0Z. The duplicate match is found in Storm 26 of the STI set, which has two

different serial numbers for the same catalog number (Storm 26 of 1989 – Sarah). The track

with serial number 198926s appears to be an addendum to the track of serial number

198926. However, the two tracks overlap at 12Z and 18Z on 9/11/89 and at 0Z on 9/12/89

(i.e., two instances at each time). The latitude and longitude values as well as the intensity

values are similar for both instances at each time, but differ slightly. It is unknown which

instance is correct for each time, so the original position and intensity values (198926) are

retained for 12Z and 18Z on 9/11/89 and the addendum position and intensity values

(198926s) are retained for 0Z on 9/12/89, as these values are most in line with the

surrounding position values and closest to the JMA and JTWC position values. The values at

12Z and 18Z on 9/11/89 and 0Z on 9/12/89 from serial numbers 198926s and 198926,

respectively, are completely removed from all final uniform tracks.

76

80. Storm 1118 completely removed from the full track set (6 lines); Storm 4 of 1990 from the

JTWC cannot be accurately matched with any storm from the JMA set. Storm 7 of 1990

from the STI may match Storm 4 from the JTWC, but it is difficult to say with certainty.

81. Storm 1119 is added back to the final sets. The remove line program incorrectly removed

this system from the final sets, since Storms 4 and 6 of 1990 from the JMA and STI sets,

respectively, had already been assigned (incorrectly) to Storm 1118 (removed).

82. Storm 1131 of the full track set has duplicate matches for 9/7/90 at 12Z and 18Z. The

duplicate match is found in Storm 24 of the STI set, which has two different serial numbers

for the same catalog number (Storm 24 of 1990 – Dot). The track with serial number

199024s appears to be an addendum to the track of serial number 199024. However, the two

tracks overlap at 12Z and 18Z on 9/7/90 (i.e., two instances at each time). The latitude and

longitude values as well as the intensity values are similar for both instances at each time,

but differ slightly. It is unknown which instance is correct, so the original position and

intensity values (199024) are retained, as they are more in line with the surrounding position

values and closer to the JMA and JTWC position values than the addendum values. The

values at 12Z and 18Z on 9/7/90 from serial number 199024s are completely removed from

all final uniform tracks.

83. Storm 1158 completely removed from the full track set (11 lines). Storm 13 of 1991 from

the JTWC set cannot be accurately matched with any storm from the JMA or STI sets.

Storm 10 of 1991 from the JMA is the same system as storm 12 from the STI; however,

these storms match storm 11 from the JTWC, not storm 13 from the JTWC, as the match for

storm 1158 yields. Storm 1156 is the correct match for storms 10 and 12 of 1991 from the

JMA and STI sets.

77

84. Storm 1192 of the full track set has duplicate matches for 8/31/92 at 18Z. The duplicate

match is found in Storm 17 of the STI set, which has two different serial numbers for the

same catalog number (Storm 17 of 1992 – Polly). The track with serial number 199217s

appears to be an addendum to the track of serial number 199217. However, the two tracks

overlap at 18Z on 8/31/92 (i.e., two instances). The latitude and longitude values as well as

the intensity values are similar for both instances, but differ slightly. It is unknown which

instance is correct, so the original position and intensity values (199217) are retained, as

they are more in line with the surrounding position values and closer to the JMA and JTWC

position values than the addendum values. The values at 18Z on 8/31/92 from serial number

199217s are completely removed from all final uniform tracks.

85. Storm 1257 from the full track set changed from 12,7,8 (JTWC, JMA, STI) to 12,8,10;

Storms 7 and 8 of 1994 from the JMA and STI sets, respectively, had already been assigned

to Storm 1255.

86. Storm 1290 completely removed from the full track set (6 lines); Storm 6 of 1995 from the

JTWC cannot be matched with any storm from the JMA or STI sets. The program

incorrectly assigns Storms 4 and 4 of 1995 from the JMA and STI sets, respectively, to

Storm 6 from the JTWC.

87. Storm 1291 from the full track set added back to the final sets. It was incorrectly discarded

by the remove line program, since Storms 4 and 4 of 1995 from the JMA and STI sets,

respectively, were already assigned to Storm 1290. This earlier match was incorrect (see

above).

88. Storm 1339 completely removed from the full track set (1 line); Storm 21 of 1996 from the

JTWC cannot be matched with any storm from the JMA or STI sets.

78

89. Storm 1355 is problematic. While Storm 37 of 1996 from the JTWC agrees well with Storm

29 of 1996 from the STI (“Ernie”) it does not agree completely with Storm 25 of 1996 from

the JMA (also “Ernie”). Instead, Storm 39 of 1996 from the JTWC agrees with the points

from 11/06/96 at 12Z to 11/10/96 at 0Z of Storm 25 in the JMA set. Thereafter, Storm 37 of

the JTWC matches with the points in Storm 25 of the JMA set. Because of this discrepancy

in the earlier points, only points from 11/10/96 at 0Z to 11/16/96 at 0Z are included in Storm

1355 (definite matches from all three data sets). Storm 39 of 1996 from the JTWC may or

may not be part of Storm 25 from the JMA. As a result, Storm 39 from the JTWC is not

considered.

90. Storm 1375 of the full track set has duplicate matches for 8/20/97 at 0Z. The duplicate

match is found in Storm 14 of the STI set, which has two different serial numbers for the

same catalog number (Storm 14 of 1997 – Winnie). The track with serial number 199714s

appears to be an addendum to the track of serial number 199714. However, the two tracks

overlap at 0Z on 8/20/97 (i.e., two instances). The latitude and longitude values are similar

for both instances, but differ slightly. It is unknown which instance is correct, so the original

position values (199714) are retained, as they are more in line with the surrounding position

values and closer to the JMA and JTWC position values than the addendum values. The

values at 0Z on 8/20/97 from serial number 199714s are completely removed from all final

uniform tracks.

91. Storm 1393 completely removed from the full track set (4 lines); Storm 1 of 1998 from the

JTWC is found in the STI set (Storm 2 of 1998) but not in the JMA set.

79

92. Storm 1394 from the full track set added back to the final sets. This Storm was discarded by

the remove line program because Storms 1 and 1 of 1998 from the JMA and STI sets,

respectively, had already been assigned (incorrectly) to Storm 1 from the JTWC set.

93. Storm 1440 of the full track set is missing position and intensity estimates at 12Z on

9/12/99. Storm 1440 of the full track set matches storms 21,15,18 of 1999 from the JTWC,

JMA, and STI sets, respectively. Position and intensity estimates are included for storm

1440 at 6Z and 18Z on 9/12/99. The 12Z point was not originally included in the full track

set because the position match falls outside of the prescribed 3.2° buffer. In order not to

have breaks within the storm track, this missing point is manually ingested back into all final

uniform tracks. A flag value of 1 is listed in each final uniform set for this point since at

least two of the three position matches fall outside of the prescribed 3.2° buffer.

94. Storm 1552 of the full track set has a duplicate match for 4/12/03 at 6Z. The duplicate match

is found in Storm 2 of the JTWC set, which includes two entries for the same date and time.

The position and intensity information are the same for both entries; however, the wind radii

data (wind intensity and radius code) are different. The first entry includes 34kt wind radii

while the second entry includes 50kt wind radii. Because the Superset does not include wind

radii data, only one instance of Storm 2 at 6Z on 4/12/1997 is retained. The duplicate entry

is removed from all final uniform tracks.

JTWC began including 35kt wind radii information within its best track data beginning in

2001. By 2004 34kt, 50kt, 64kt, and 100kt wind radii data were included within individual storm

files in the best track data. Instead of incorporating all wind radii information in one entry for

each date and time within an individual storm file, duplicate entries with the same date, time,

80

position, and intensity information are included, but with different wind radii data (wind

intensity and radius code) for all available wind radii. For example, Super Typhoon Sudal in

2004 (Storm 3 of 2004 in the JTWC set) has four entries at 0Z on 4/8/2004 – the first entry

containing wind radii data for 34kt wind radii, the second entry for 50kt wind radii, the third

entry for 64kt wind radii, and the fourth entry for 100kt wind radii. The table below contains all

duplicate entries from 2004, originally included because of multiple wind radii data, which are

removed from the final Superset:

Change

Number

JTWC

Storm

Number

JMA

Storm

Number

STI

Storm

Number

Year Month Day Hour

Superset

Storm

Number

95 3 1 3 2004 4 5 6 1580

96 3 1 3 2004 4 5 12 1580

97 3 1 3 2004 4 5 18 1580

98 3 1 3 2004 4 6 0 1580

99 3 1 3 2004 4 6 6 1580

100 3 1 3 2004 4 6 6 1580

101 3 1 3 2004 4 6 12 1580

102 3 1 3 2004 4 6 12 1580

103 3 1 3 2004 4 6 18 1580

104 3 1 3 2004 4 6 18 1580

105 3 1 3 2004 4 7 0 1580

106 3 1 3 2004 4 7 0 1580

107 3 1 3 2004 4 7 6 1580

108 3 1 3 2004 4 7 6 1580

109 3 1 3 2004 4 7 12 1580

110 3 1 3 2004 4 7 12 1580

111 3 1 3 2004 4 7 18 1580

112 3 1 3 2004 4 7 18 1580

113 3 1 3 2004 4 8 0 1580

114 3 1 3 2004 4 8 0 1580

115 3 1 3 2004 4 8 0 1580

116 3 1 3 2004 4 8 6 1580

117 3 1 3 2004 4 8 6 1580

81

118 3 1 3 2004 4 8 12 1580

119 3 1 3 2004 4 8 12 1580

120 3 1 3 2004 4 8 12 1580

121 3 1 3 2004 4 8 18 1580

122 3 1 3 2004 4 8 18 1580

123 3 1 3 2004 4 9 0 1580

124 3 1 3 2004 4 9 0 1580

125 3 1 3 2004 4 9 0 1580

126 3 1 3 2004 4 9 6 1580

127 3 1 3 2004 4 9 6 1580

128 3 1 3 2004 4 9 6 1580

129 3 1 3 2004 4 9 12 1580

130 3 1 3 2004 4 9 12 1580

131 3 1 3 2004 4 9 12 1580

132 3 1 3 2004 4 9 18 1580

133 3 1 3 2004 4 9 18 1580

134 3 1 3 2004 4 10 0 1580

135 3 1 3 2004 4 10 0 1580

136 3 1 3 2004 4 10 6 1580

137 3 1 3 2004 4 10 6 1580

138 3 1 3 2004 4 10 12 1580

139 3 1 3 2004 4 10 12 1580

140 3 1 3 2004 4 10 18 1580

141 3 1 3 2004 4 10 18 1580

142 3 1 3 2004 4 10 18 1580

143 3 1 3 2004 4 11 0 1580

144 3 1 3 2004 4 11 0 1580

145 3 1 3 2004 4 11 0 1580

146 3 1 3 2004 4 11 6 1580

147 3 1 3 2004 4 11 6 1580

148 3 1 3 2004 4 11 6 1580

149 3 1 3 2004 4 11 12 1580

150 3 1 3 2004 4 11 12 1580

151 3 1 3 2004 4 11 12 1580

152 3 1 3 2004 4 11 18 1580

153 3 1 3 2004 4 11 18 1580

154 3 1 3 2004 4 12 0 1580

155 3 1 3 2004 4 12 0 1580

156 3 1 3 2004 4 12 0 1580

82

157 3 1 3 2004 4 12 6 1580

158 3 1 3 2004 4 12 6 1580

159 3 1 3 2004 4 12 6 1580

160 3 1 3 2004 4 12 12 1580

161 3 1 3 2004 4 12 12 1580

162 3 1 3 2004 4 12 12 1580

163 3 1 3 2004 4 12 18 1580

164 3 1 3 2004 4 12 18 1580

165 3 1 3 2004 4 13 0 1580

166 3 1 3 2004 4 13 0 1580

167 3 1 3 2004 4 13 0 1580

168 3 1 3 2004 4 13 6 1580

169 3 1 3 2004 4 13 6 1580

170 3 1 3 2004 4 13 12 1580

171 3 1 3 2004 4 13 12 1580

172 3 1 3 2004 4 13 12 1580

173 3 1 3 2004 4 13 18 1580

174 3 1 3 2004 4 13 18 1580

175 3 1 3 2004 4 13 18 1580

176 3 1 3 2004 4 14 0 1580

177 3 1 3 2004 4 14 0 1580

178 3 1 3 2004 4 14 0 1580

179 3 1 3 2004 4 14 6 1580

180 3 1 3 2004 4 14 6 1580

181 3 1 3 2004 4 14 6 1580

182 3 1 3 2004 4 14 12 1580

183 3 1 3 2004 4 14 12 1580

184 3 1 3 2004 4 14 18 1580

185 3 1 3 2004 4 14 18 1580

186 3 1 3 2004 4 15 0 1580

187 3 1 3 2004 4 15 0 1580

188 3 1 3 2004 4 15 6 1580

189 3 1 3 2004 4 15 12 1580

190 4 2 4 2004 5 14 6 1581

191 4 2 4 2004 5 14 12 1581

192 4 2 4 2004 5 14 18 1581

193 4 2 4 2004 5 15 0 1581

194 4 2 4 2004 5 15 6 1581

195 4 2 4 2004 5 15 12 1581

83

196 4 2 4 2004 5 15 18 1581

197 4 2 4 2004 5 16 0 1581

198 4 2 4 2004 5 16 0 1581

199 4 2 4 2004 5 16 6 1581

200 4 2 4 2004 5 16 6 1581

201 4 2 4 2004 5 16 12 1581

202 4 2 4 2004 5 16 12 1581

203 4 2 4 2004 5 16 18 1581

204 4 2 4 2004 5 17 0 1581

205 4 2 4 2004 5 17 6 1581

206 4 2 4 2004 5 17 6 1581

207 4 2 4 2004 5 17 12 1581

208 4 2 4 2004 5 17 12 1581

209 4 2 4 2004 5 17 18 1581

210 4 2 4 2004 5 17 18 1581

211 4 2 4 2004 5 18 0 1581

212 4 2 4 2004 5 18 0 1581

213 4 2 4 2004 5 18 6 1581

214 4 2 4 2004 5 18 6 1581

215 4 2 4 2004 5 18 12 1581

216 4 2 4 2004 5 18 12 1581

217 4 2 4 2004 5 18 18 1581

218 4 2 4 2004 5 19 0 1581

219 4 2 4 2004 5 19 0 1581

220 4 2 4 2004 5 19 6 1581

221 4 2 4 2004 5 19 12 1581

222 4 2 4 2004 5 19 18 1581

223 4 2 4 2004 5 20 0 1581

224 4 2 4 2004 5 20 6 1581

225 4 2 4 2004 5 20 12 1581

226 4 2 4 2004 5 20 18 1581

227 4 2 4 2004 5 21 0 1581

228 6 3 6 2004 5 17 18 1583

229 6 3 6 2004 5 18 0 1583

230 6 3 6 2004 5 18 6 1583

231 6 3 6 2004 5 18 12 1583

232 6 3 6 2004 5 18 18 1583

233 6 3 6 2004 5 19 0 1583

234 6 3 6 2004 5 19 6 1583

84

235 6 3 6 2004 5 19 12 1583

236 6 3 6 2004 5 19 18 1583

237 6 3 6 2004 5 20 0 1583

238 6 3 6 2004 5 20 6 1583

239 6 3 6 2004 5 20 12 1583

240 7 4 7 2004 6 7 6 1584

241 7 4 7 2004 6 7 12 1584

242 7 4 7 2004 6 7 18 1584

243 7 4 7 2004 6 8 0 1584

244 7 4 7 2004 6 8 6 1584

245 7 4 7 2004 6 8 6 1584

246 7 4 7 2004 6 8 12 1584

247 7 4 7 2004 6 8 12 1584

248 7 4 7 2004 6 8 18 1584

249 7 4 7 2004 6 8 18 1584

250 7 4 7 2004 6 9 0 1584

251 7 4 7 2004 6 9 0 1584

252 7 4 7 2004 6 9 6 1584

253 7 4 7 2004 6 9 6 1584

254 7 4 7 2004 6 9 12 1584

255 7 4 7 2004 6 9 12 1584

256 7 4 7 2004 6 9 18 1584

257 7 4 7 2004 6 9 18 1584

258 7 4 7 2004 6 10 0 1584

259 7 4 7 2004 6 10 0 1584

260 7 4 7 2004 6 10 6 1584

261 7 4 7 2004 6 10 6 1584

262 7 4 7 2004 6 10 12 1584

263 7 4 7 2004 6 10 18 1584

264 7 4 7 2004 6 11 0 1584

265 8 5 8 2004 6 11 12 1585

266 8 5 8 2004 6 11 18 1585

267 8 5 8 2004 6 12 0 1585

268 8 5 8 2004 6 12 0 1585

269 8 5 8 2004 6 12 6 1585

270 8 5 8 2004 6 12 6 1585

271 8 5 8 2004 6 12 12 1585

272 8 5 8 2004 6 12 12 1585

273 8 5 8 2004 6 12 18 1585

85

274 9 6 9 2004 6 14 6 1586

275 9 6 9 2004 6 14 12 1586

276 9 6 9 2004 6 14 18 1586

277 9 6 9 2004 6 14 18 1586

278 9 6 9 2004 6 15 0 1586

279 9 6 9 2004 6 15 0 1586

280 9 6 9 2004 6 15 6 1586

281 9 6 9 2004 6 15 6 1586

282 9 6 9 2004 6 15 12 1586

283 9 6 9 2004 6 15 12 1586

284 9 6 9 2004 6 15 18 1586

285 9 6 9 2004 6 15 18 1586

286 9 6 9 2004 6 16 0 1586

287 9 6 9 2004 6 16 0 1586

288 9 6 9 2004 6 16 6 1586

289 9 6 9 2004 6 16 6 1586

290 9 6 9 2004 6 16 12 1586

291 9 6 9 2004 6 16 12 1586

292 9 6 9 2004 6 16 18 1586

293 9 6 9 2004 6 16 18 1586

294 9 6 9 2004 6 17 0 1586

295 9 6 9 2004 6 17 0 1586

296 9 6 9 2004 6 17 6 1586

297 9 6 9 2004 6 17 6 1586

298 9 6 9 2004 6 17 12 1586

299 9 6 9 2004 6 17 12 1586

300 9 6 9 2004 6 17 18 1586

301 9 6 9 2004 6 17 18 1586

302 9 6 9 2004 6 18 0 1586

303 9 6 9 2004 6 18 0 1586

304 9 6 9 2004 6 18 6 1586

305 9 6 9 2004 6 18 6 1586

306 9 6 9 2004 6 18 12 1586

307 9 6 9 2004 6 18 12 1586

308 9 6 9 2004 6 18 18 1586

309 9 6 9 2004 6 18 18 1586

310 9 6 9 2004 6 19 0 1586

311 9 6 9 2004 6 19 0 1586

312 9 6 9 2004 6 19 6 1586

86

313 9 6 9 2004 6 19 6 1586

314 9 6 9 2004 6 19 12 1586

315 9 6 9 2004 6 19 12 1586

316 9 6 9 2004 6 19 18 1586

317 9 6 9 2004 6 19 18 1586

318 9 6 9 2004 6 20 0 1586

319 9 6 9 2004 6 20 0 1586

320 9 6 9 2004 6 20 6 1586

321 9 6 9 2004 6 20 6 1586

322 9 6 9 2004 6 20 12 1586

323 9 6 9 2004 6 20 12 1586

324 9 6 9 2004 6 20 18 1586

325 9 6 9 2004 6 21 0 1586

326 9 6 9 2004 6 21 6 1586

327 10 7 10 2004 6 24 18 1587

328 10 7 10 2004 6 25 0 1587

329 10 7 10 2004 6 25 6 1587

330 10 7 10 2004 6 25 12 1587

331 10 7 10 2004 6 26 18 1587

332 10 7 10 2004 6 27 0 1587

333 10 7 10 2004 6 27 6 1587

334 10 7 10 2004 6 27 6 1587

335 10 7 10 2004 6 27 12 1587

336 10 7 10 2004 6 27 12 1587

337 10 7 10 2004 6 27 18 1587

338 10 7 10 2004 6 27 18 1587

339 10 7 10 2004 6 28 0 1587

340 10 7 10 2004 6 28 0 1587

341 10 7 10 2004 6 28 6 1587

342 10 7 10 2004 6 28 6 1587

343 10 7 10 2004 6 28 12 1587

344 10 7 10 2004 6 28 12 1587

345 10 7 10 2004 6 28 18 1587

346 10 7 10 2004 6 28 18 1587

347 10 7 10 2004 6 29 0 1587

348 10 7 10 2004 6 29 0 1587

349 10 7 10 2004 6 29 6 1587

350 10 7 10 2004 6 29 6 1587

351 10 7 10 2004 6 29 12 1587

87

352 10 7 10 2004 6 29 12 1587

353 10 7 10 2004 6 29 18 1587

354 10 7 10 2004 6 29 18 1587

355 10 7 10 2004 6 30 0 1587

356 10 7 10 2004 6 30 0 1587

357 10 7 10 2004 6 30 6 1587

358 10 7 10 2004 6 30 6 1587

359 10 7 10 2004 6 30 12 1587

360 10 7 10 2004 6 30 12 1587

361 10 7 10 2004 6 30 18 1587

362 10 7 10 2004 6 30 18 1587

363 10 7 10 2004 7 1 0 1587

364 10 7 10 2004 7 1 0 1587

365 10 7 10 2004 7 1 6 1587

366 10 7 10 2004 7 1 6 1587

367 10 7 10 2004 7 1 12 1587

368 10 7 10 2004 7 1 12 1587

369 10 7 10 2004 7 1 18 1587

370 11 8 11 2004 6 27 0 1588

371 11 8 11 2004 6 27 6 1588

372 11 8 11 2004 6 27 12 1588

373 11 8 11 2004 6 27 18 1588

374 11 8 11 2004 6 28 0 1588

375 11 8 11 2004 6 28 6 1588

376 11 8 11 2004 6 28 6 1588

377 11 8 11 2004 6 28 12 1588

378 11 8 11 2004 6 28 12 1588

379 11 8 11 2004 6 28 18 1588

380 11 8 11 2004 6 28 18 1588

381 11 8 11 2004 6 29 0 1588

382 11 8 11 2004 6 29 0 1588

383 11 8 11 2004 6 29 6 1588

384 11 8 11 2004 6 29 6 1588

385 11 8 11 2004 6 29 12 1588

386 11 8 11 2004 6 29 12 1588

387 11 8 11 2004 6 29 18 1588

388 11 8 11 2004 6 29 18 1588

389 11 8 11 2004 6 30 0 1588

390 11 8 11 2004 6 30 0 1588

88

391 11 8 11 2004 6 30 6 1588

392 11 8 11 2004 6 30 6 1588

393 11 8 11 2004 6 30 12 1588

394 11 8 11 2004 6 30 12 1588

395 11 8 11 2004 6 30 18 1588

396 11 8 11 2004 6 30 18 1588

397 11 8 11 2004 7 1 0 1588

398 11 8 11 2004 7 1 6 1588

399 13 10 13 2004 7 26 0 1590

400 13 10 13 2004 7 26 6 1590

401 13 10 13 2004 7 26 6 1590

402 13 10 13 2004 7 26 12 1590

403 13 10 13 2004 7 26 12 1590

404 13 10 13 2004 7 26 18 1590

405 13 10 13 2004 7 26 18 1590

406 13 10 13 2004 7 27 0 1590

407 13 10 13 2004 7 27 0 1590

408 13 10 13 2004 7 27 6 1590

409 13 10 13 2004 7 27 6 1590

410 13 10 13 2004 7 27 12 1590

411 13 10 13 2004 7 27 12 1590

412 13 10 13 2004 7 27 18 1590

413 13 10 13 2004 7 27 18 1590

414 13 10 13 2004 7 28 0 1590

415 13 10 13 2004 7 28 0 1590

416 13 10 13 2004 7 28 6 1590

417 13 10 13 2004 7 28 6 1590

418 13 10 13 2004 7 28 12 1590

419 13 10 13 2004 7 28 12 1590

420 13 10 13 2004 7 28 18 1590

421 13 10 13 2004 7 28 18 1590

422 13 10 13 2004 7 29 0 1590

423 13 10 13 2004 7 29 0 1590

424 13 10 13 2004 7 29 6 1590

425 13 10 13 2004 7 29 6 1590

426 13 10 13 2004 7 29 12 1590

427 13 10 13 2004 7 29 18 1590

428 13 10 13 2004 7 30 0 1590

429 13 10 13 2004 7 30 6 1590

89

430 13 10 13 2004 7 30 12 1590

431 13 10 13 2004 7 30 18 1590

432 13 10 13 2004 7 31 0 1590

433 13 10 13 2004 7 31 6 1590

434 14 12 16 2004 8 5 6 1591

435 14 12 16 2004 8 5 12 1591

436 14 12 16 2004 8 5 12 1591

437 14 12 16 2004 8 5 18 1591

438 14 12 16 2004 8 5 18 1591

439 14 12 16 2004 8 6 0 1591

440 14 12 16 2004 8 6 0 1591

441 14 12 16 2004 8 6 6 1591

442 14 12 16 2004 8 6 6 1591

443 14 12 16 2004 8 6 12 1591

444 14 12 16 2004 8 6 12 1591

445 14 12 16 2004 8 6 18 1591

446 14 12 16 2004 8 7 0 1591

447 14 12 16 2004 8 7 6 1591

448 16 13 17 2004 8 9 18 1593

449 16 13 17 2004 8 10 0 1593

450 16 13 17 2004 8 10 6 1593

451 16 13 17 2004 8 10 12 1593

452 16 13 17 2004 8 10 18 1593

453 16 13 17 2004 8 10 18 1593

454 16 13 17 2004 8 11 0 1593

455 16 13 17 2004 8 11 0 1593

456 16 13 17 2004 8 11 6 1593

457 16 13 17 2004 8 11 6 1593

458 16 13 17 2004 8 11 12 1593

459 16 13 17 2004 8 11 12 1593

460 16 13 17 2004 8 11 18 1593

461 16 13 17 2004 8 11 18 1593

462 16 13 17 2004 8 12 0 1593

463 16 13 17 2004 8 12 0 1593

464 16 13 17 2004 8 12 6 1593

465 16 13 17 2004 8 12 6 1593

466 16 13 17 2004 8 12 12 1593

467 16 13 17 2004 8 12 12 1593

468 16 13 17 2004 8 12 18 1593

90

469 16 13 17 2004 8 13 0 1593

470 18 15 19 2004 8 17 12 1595

471 18 15 19 2004 8 17 18 1595

472 18 15 19 2004 8 18 0 1595

473 18 15 19 2004 8 18 0 1595

474 18 15 19 2004 8 18 6 1595

475 18 15 19 2004 8 18 6 1595

476 18 15 19 2004 8 18 12 1595

477 18 15 19 2004 8 18 12 1595

478 18 15 19 2004 8 18 18 1595

479 18 15 19 2004 8 18 18 1595

480 18 15 19 2004 8 19 0 1595

481 18 15 19 2004 8 19 6 1595

482 18 15 19 2004 8 19 18 1595

483 19 16 21 2004 8 20 0 1596

484 19 16 21 2004 8 20 6 1596

485 19 16 21 2004 8 20 12 1596

486 19 16 21 2004 8 20 18 1596

487 19 16 21 2004 8 21 0 1596

488 19 16 21 2004 8 21 0 1596

489 19 16 21 2004 8 21 6 1596

490 19 16 21 2004 8 21 6 1596

491 19 16 21 2004 8 21 12 1596

492 19 16 21 2004 8 21 12 1596

493 19 16 21 2004 8 21 18 1596

494 19 16 21 2004 8 21 18 1596

495 19 16 21 2004 8 22 0 1596

496 19 16 21 2004 8 22 0 1596

497 19 16 21 2004 8 22 6 1596

498 19 16 21 2004 8 22 6 1596

499 19 16 21 2004 8 22 12 1596

500 19 16 21 2004 8 22 12 1596

501 19 16 21 2004 8 22 18 1596

502 19 16 21 2004 8 22 18 1596

503 19 16 21 2004 8 23 0 1596

504 19 16 21 2004 8 23 0 1596

505 19 16 21 2004 8 23 6 1596

506 19 16 21 2004 8 23 6 1596

507 19 16 21 2004 8 23 12 1596

91

508 19 16 21 2004 8 23 12 1596

509 19 16 21 2004 8 23 18 1596

510 19 16 21 2004 8 23 18 1596

511 19 16 21 2004 8 24 0 1596

512 19 16 21 2004 8 24 0 1596

513 19 16 21 2004 8 24 6 1596

514 19 16 21 2004 8 24 6 1596

515 19 16 21 2004 8 24 12 1596

516 19 16 21 2004 8 24 12 1596

517 19 16 21 2004 8 24 18 1596

518 19 16 21 2004 8 24 18 1596

519 19 16 21 2004 8 25 0 1596

520 19 16 21 2004 8 25 0 1596

521 19 16 21 2004 8 25 6 1596

522 19 16 21 2004 8 25 6 1596

523 19 16 21 2004 8 25 12 1596

524 19 16 21 2004 8 25 12 1596

525 19 16 21 2004 8 25 18 1596

526 19 16 21 2004 8 25 18 1596

527 19 16 21 2004 8 26 0 1596

528 19 16 21 2004 8 26 0 1596

529 19 16 21 2004 8 26 6 1596

530 19 16 21 2004 8 26 6 1596

531 19 16 21 2004 8 26 12 1596

532 19 16 21 2004 8 26 12 1596

533 19 16 21 2004 8 26 18 1596

534 19 16 21 2004 8 26 18 1596

535 19 16 21 2004 8 27 0 1596

536 19 16 21 2004 8 27 0 1596

537 19 16 21 2004 8 27 6 1596

538 19 16 21 2004 8 27 6 1596

539 19 16 21 2004 8 27 12 1596

540 19 16 21 2004 8 27 12 1596

541 19 16 21 2004 8 27 18 1596

542 19 16 21 2004 8 27 18 1596

543 19 16 21 2004 8 28 0 1596

544 19 16 21 2004 8 28 0 1596

545 19 16 21 2004 8 28 6 1596

546 19 16 21 2004 8 28 6 1596

92

547 19 16 21 2004 8 28 12 1596

548 19 16 21 2004 8 28 12 1596

549 19 16 21 2004 8 28 18 1596

550 19 16 21 2004 8 28 18 1596

551 19 16 21 2004 8 29 0 1596

552 19 16 21 2004 8 29 0 1596

553 19 16 21 2004 8 29 6 1596

554 19 16 21 2004 8 29 6 1596

555 19 16 21 2004 8 29 12 1596

556 19 16 21 2004 8 29 12 1596

557 19 16 21 2004 8 29 18 1596

558 19 16 21 2004 8 29 18 1596

559 19 16 21 2004 8 30 0 1596

560 19 16 21 2004 8 30 0 1596

561 19 16 21 2004 8 30 6 1596

562 19 16 21 2004 8 30 6 1596

563 19 16 21 2004 8 30 12 1596

564 19 16 21 2004 8 30 12 1596

565 19 16 21 2004 8 30 18 1596

566 19 16 21 2004 8 31 0 1596

567 19 16 21 2004 8 31 6 1596

568 20 17 20 2004 8 20 18 1597

569 20 17 20 2004 8 21 0 1597

570 20 17 20 2004 8 21 6 1597

571 20 17 20 2004 8 21 12 1597

572 20 17 20 2004 8 21 18 1597

573 20 17 20 2004 8 22 0 1597

574 20 17 20 2004 8 22 6 1597

575 20 17 20 2004 8 22 12 1597

576 20 17 20 2004 8 22 18 1597

577 20 17 20 2004 8 23 0 1597

578 20 17 20 2004 8 23 6 1597

579 20 17 20 2004 8 23 12 1597

580 20 17 20 2004 8 23 18 1597

581 20 17 20 2004 8 23 18 1597

582 20 17 20 2004 8 24 0 1597

583 20 17 20 2004 8 24 0 1597

584 20 17 20 2004 8 24 6 1597

585 20 17 20 2004 8 24 6 1597

93

586 20 17 20 2004 8 24 12 1597

587 20 17 20 2004 8 24 12 1597

588 20 17 20 2004 8 24 18 1597

589 20 17 20 2004 8 24 18 1597

590 20 17 20 2004 8 25 0 1597

591 20 17 20 2004 8 25 0 1597

592 20 17 20 2004 8 25 6 1597

593 20 17 20 2004 8 25 6 1597

594 20 17 20 2004 8 25 12 1597

595 20 17 20 2004 8 25 12 1597

596 20 17 20 2004 8 25 18 1597

597 20 17 20 2004 8 25 18 1597

598 20 17 20 2004 8 26 0 1597

599 22 18 23 2004 8 28 6 1599

600 22 18 23 2004 8 28 12 1599

601 22 18 23 2004 8 28 18 1599

602 22 18 23 2004 8 29 0 1599

603 22 18 23 2004 8 29 6 1599

604 22 18 23 2004 8 29 12 1599

605 22 18 23 2004 8 29 18 1599

606 22 18 23 2004 8 29 18 1599

607 22 18 23 2004 8 30 0 1599

608 22 18 23 2004 8 30 0 1599

609 22 18 23 2004 8 30 6 1599

610 22 18 23 2004 8 30 6 1599

611 22 18 23 2004 8 30 12 1599

612 22 18 23 2004 8 30 12 1599

613 22 18 23 2004 8 30 18 1599

614 22 18 23 2004 8 30 18 1599

615 22 18 23 2004 8 31 0 1599

616 22 18 23 2004 8 31 0 1599

617 22 18 23 2004 8 31 6 1599

618 22 18 23 2004 8 31 6 1599

619 22 18 23 2004 8 31 12 1599

620 22 18 23 2004 8 31 12 1599

621 22 18 23 2004 8 31 18 1599

622 22 18 23 2004 8 31 18 1599

623 22 18 23 2004 9 1 0 1599

624 22 18 23 2004 9 1 0 1599

94

625 22 18 23 2004 9 1 6 1599

626 22 18 23 2004 9 1 6 1599

627 22 18 23 2004 9 1 12 1599

628 22 18 23 2004 9 1 12 1599

629 22 18 23 2004 9 1 18 1599

630 22 18 23 2004 9 1 18 1599

631 22 18 23 2004 9 2 0 1599

632 22 18 23 2004 9 2 0 1599

633 22 18 23 2004 9 2 6 1599

634 22 18 23 2004 9 2 6 1599

635 22 18 23 2004 9 2 12 1599

636 22 18 23 2004 9 2 12 1599

637 22 18 23 2004 9 2 18 1599

638 22 18 23 2004 9 2 18 1599

639 22 18 23 2004 9 3 0 1599

640 22 18 23 2004 9 3 0 1599

641 22 18 23 2004 9 3 6 1599

642 22 18 23 2004 9 3 6 1599

643 22 18 23 2004 9 3 12 1599

644 22 18 23 2004 9 3 12 1599

645 22 18 23 2004 9 3 18 1599

646 22 18 23 2004 9 3 18 1599

647 22 18 23 2004 9 4 0 1599

648 22 18 23 2004 9 4 0 1599

649 22 18 23 2004 9 4 6 1599

650 22 18 23 2004 9 4 6 1599

651 22 18 23 2004 9 4 12 1599

652 22 18 23 2004 9 4 12 1599

653 22 18 23 2004 9 4 18 1599

654 22 18 23 2004 9 4 18 1599

655 22 18 23 2004 9 5 0 1599

656 22 18 23 2004 9 5 0 1599

657 22 18 23 2004 9 5 6 1599

658 22 18 23 2004 9 5 6 1599

659 22 18 23 2004 9 5 12 1599

660 22 18 23 2004 9 5 12 1599

661 22 18 23 2004 9 5 18 1599

662 22 18 23 2004 9 5 18 1599

663 22 18 23 2004 9 6 0 1599

95

664 22 18 23 2004 9 6 0 1599

665 22 18 23 2004 9 6 6 1599

666 22 18 23 2004 9 6 6 1599

667 22 18 23 2004 9 6 12 1599

668 22 18 23 2004 9 6 12 1599

669 22 18 23 2004 9 6 18 1599

670 22 18 23 2004 9 6 18 1599

671 22 18 23 2004 9 7 0 1599

672 22 18 23 2004 9 7 0 1599

673 22 18 23 2004 9 7 6 1599

674 22 18 23 2004 9 7 6 1599

675 22 18 23 2004 9 7 12 1599

676 22 18 23 2004 9 7 12 1599

677 23 19 24 2004 9 5 6 1600

678 23 19 24 2004 9 5 12 1600

679 23 19 24 2004 9 5 18 1600

680 23 19 24 2004 9 6 0 1600

681 23 19 24 2004 9 6 6 1600

682 23 19 24 2004 9 6 12 1600

683 25 21 26 2004 9 21 18 1602

684 25 21 26 2004 9 22 0 1602

685 25 21 26 2004 9 22 6 1602

686 25 21 26 2004 9 22 12 1602

687 25 21 26 2004 9 22 18 1602

688 25 21 26 2004 9 22 18 1602

689 25 21 26 2004 9 23 0 1602

690 25 21 26 2004 9 23 0 1602

691 25 21 26 2004 9 23 6 1602

692 25 21 26 2004 9 23 6 1602

693 25 21 26 2004 9 23 12 1602

694 25 21 26 2004 9 23 12 1602

695 25 21 26 2004 9 23 18 1602

696 25 21 26 2004 9 23 18 1602

697 25 21 26 2004 9 24 0 1602

698 25 21 26 2004 9 24 0 1602

699 25 21 26 2004 9 24 6 1602

700 25 21 26 2004 9 24 6 1602

701 25 21 26 2004 9 24 12 1602

702 25 21 26 2004 9 24 12 1602

96

703 25 21 26 2004 9 24 18 1602

704 25 21 26 2004 9 24 18 1602

705 25 21 26 2004 9 25 0 1602

706 25 21 26 2004 9 25 0 1602

707 25 21 26 2004 9 25 6 1602

708 25 21 26 2004 9 25 6 1602

709 25 21 26 2004 9 25 12 1602

710 25 21 26 2004 9 25 12 1602

711 25 21 26 2004 9 25 18 1602

712 25 21 26 2004 9 25 18 1602

713 25 21 26 2004 9 26 0 1602

714 25 21 26 2004 9 26 0 1602

715 25 21 26 2004 9 26 6 1602

716 25 21 26 2004 9 26 6 1602

717 25 21 26 2004 9 26 12 1602

718 25 21 26 2004 9 26 12 1602

719 25 21 26 2004 9 26 18 1602

720 25 21 26 2004 9 26 18 1602

721 25 21 26 2004 9 27 0 1602

722 25 21 26 2004 9 27 0 1602

723 25 21 26 2004 9 27 6 1602

724 25 21 26 2004 9 27 6 1602

725 25 21 26 2004 9 27 12 1602

726 25 21 26 2004 9 27 12 1602

727 25 21 26 2004 9 27 18 1602

728 25 21 26 2004 9 27 18 1602

729 25 21 26 2004 9 28 0 1602

730 25 21 26 2004 9 28 0 1602

731 25 21 26 2004 9 28 6 1602

732 25 21 26 2004 9 28 6 1602

733 25 21 26 2004 9 28 12 1602

734 25 21 26 2004 9 28 12 1602

735 25 21 26 2004 9 28 18 1602

736 25 21 26 2004 9 28 18 1602

737 25 21 26 2004 9 29 0 1602

738 25 21 26 2004 9 29 0 1602

739 25 21 26 2004 9 29 6 1602

740 25 21 26 2004 9 29 12 1602

741 26 22 27 2004 10 6 0 1603

97

742 26 22 27 2004 10 6 6 1603

743 26 22 27 2004 10 6 12 1603

744 26 22 27 2004 10 6 12 1603

745 26 22 27 2004 10 6 18 1603

746 26 22 27 2004 10 6 18 1603

747 26 22 27 2004 10 7 0 1603

748 26 22 27 2004 10 7 0 1603

749 26 22 27 2004 10 7 6 1603

750 26 22 27 2004 10 7 6 1603

751 26 22 27 2004 10 7 12 1603

752 26 22 27 2004 10 7 12 1603

753 26 22 27 2004 10 7 18 1603

754 26 22 27 2004 10 7 18 1603

755 26 22 27 2004 10 8 0 1603

756 26 22 27 2004 10 8 0 1603

757 26 22 27 2004 10 8 6 1603

758 26 22 27 2004 10 8 6 1603

759 26 22 27 2004 10 8 12 1603

760 26 22 27 2004 10 8 12 1603

761 26 22 27 2004 10 8 18 1603

762 26 22 27 2004 10 8 18 1603

763 26 22 27 2004 10 9 0 1603

764 26 22 27 2004 10 9 0 1603

765 26 22 27 2004 10 9 6 1603

766 26 22 27 2004 10 9 6 1603

767 26 22 27 2004 10 9 12 1603

768 26 22 27 2004 10 9 12 1603

769 26 22 27 2004 10 9 18 1603

770 27 23 28 2004 10 13 12 1604

771 27 23 28 2004 10 13 18 1604

772 27 23 28 2004 10 13 18 1604

773 27 23 28 2004 10 14 0 1604

774 27 23 28 2004 10 14 0 1604

775 27 23 28 2004 10 14 6 1604

776 27 23 28 2004 10 14 6 1604

777 27 23 28 2004 10 14 12 1604

778 27 23 28 2004 10 14 12 1604

779 27 23 28 2004 10 14 18 1604

780 27 23 28 2004 10 14 18 1604

98

781 27 23 28 2004 10 15 0 1604

782 27 23 28 2004 10 15 0 1604

783 27 23 28 2004 10 15 6 1604

784 27 23 28 2004 10 15 6 1604

785 27 23 28 2004 10 15 12 1604

786 27 23 28 2004 10 15 12 1604

787 27 23 28 2004 10 15 18 1604

788 27 23 28 2004 10 15 18 1604

789 27 23 28 2004 10 16 0 1604

790 27 23 28 2004 10 16 0 1604

791 27 23 28 2004 10 16 6 1604

792 27 23 28 2004 10 16 6 1604

793 27 23 28 2004 10 16 12 1604

794 27 23 28 2004 10 16 12 1604

795 27 23 28 2004 10 16 18 1604

796 27 23 28 2004 10 16 18 1604

797 27 23 28 2004 10 17 0 1604

798 27 23 28 2004 10 17 0 1604

799 27 23 28 2004 10 17 6 1604

800 27 23 28 2004 10 17 6 1604

801 27 23 28 2004 10 17 12 1604

802 27 23 28 2004 10 17 12 1604

803 27 23 28 2004 10 17 18 1604

804 27 23 28 2004 10 17 18 1604

805 27 23 28 2004 10 18 0 1604

806 27 23 28 2004 10 18 0 1604

807 27 23 28 2004 10 18 6 1604

808 27 23 28 2004 10 18 6 1604

809 27 23 28 2004 10 18 12 1604

810 27 23 28 2004 10 18 12 1604

811 27 23 28 2004 10 18 18 1604

812 27 23 28 2004 10 18 18 1604

813 27 23 28 2004 10 19 0 1604

814 27 23 28 2004 10 19 0 1604

815 27 23 28 2004 10 19 6 1604

816 27 23 28 2004 10 19 6 1604

817 27 23 28 2004 10 19 12 1604

818 27 23 28 2004 10 19 12 1604

819 27 23 28 2004 10 19 18 1604

99

820 27 23 28 2004 10 19 18 1604

821 27 23 28 2004 10 20 0 1604

822 27 23 28 2004 10 20 6 1604

823 28 24 29 2004 10 17 0 1605

824 28 24 29 2004 10 17 6 1605

825 28 24 29 2004 10 17 12 1605

826 28 24 29 2004 10 17 18 1605

827 28 24 29 2004 10 18 0 1605

828 28 24 29 2004 10 18 0 1605

829 28 24 29 2004 10 18 6 1605

830 28 24 29 2004 10 18 6 1605

831 28 24 29 2004 10 18 12 1605

832 28 24 29 2004 10 18 12 1605

833 28 24 29 2004 10 18 18 1605

834 28 24 29 2004 10 18 18 1605

835 28 24 29 2004 10 19 0 1605

836 28 24 29 2004 10 19 0 1605

837 28 24 29 2004 10 19 6 1605

838 28 24 29 2004 10 19 6 1605

839 28 24 29 2004 10 19 12 1605

840 28 24 29 2004 10 19 12 1605

841 28 24 29 2004 10 19 18 1605

842 28 24 29 2004 10 19 18 1605

843 28 24 29 2004 10 20 0 1605

844 28 24 29 2004 10 20 0 1605

845 28 24 29 2004 10 20 6 1605

846 28 24 29 2004 10 20 6 1605

847 28 24 29 2004 10 20 12 1605

848 28 24 29 2004 10 20 12 1605

849 28 24 29 2004 10 20 18 1605

850 28 24 29 2004 10 20 18 1605

851 28 24 29 2004 10 21 0 1605

852 28 24 29 2004 10 21 0 1605

853 28 24 29 2004 10 21 6 1605

854 28 24 29 2004 10 21 6 1605

855 28 24 29 2004 10 21 12 1605

856 28 24 29 2004 10 21 12 1605

857 28 24 29 2004 10 21 18 1605

858 28 24 29 2004 10 21 18 1605

100

859 28 24 29 2004 10 22 0 1605

860 28 24 29 2004 10 22 0 1605

861 28 24 29 2004 10 22 6 1605

862 28 24 29 2004 10 22 6 1605

863 28 24 29 2004 10 22 12 1605

864 28 24 29 2004 10 22 12 1605

865 28 24 29 2004 10 22 18 1605

866 28 24 29 2004 10 22 18 1605

867 28 24 29 2004 10 23 0 1605

868 28 24 29 2004 10 23 0 1605

869 28 24 29 2004 10 23 6 1605

870 28 24 29 2004 10 23 6 1605

871 28 24 29 2004 10 23 12 1605

872 28 24 29 2004 10 23 12 1605

873 28 24 29 2004 10 23 18 1605

874 28 24 29 2004 10 23 18 1605

875 28 24 29 2004 10 24 0 1605

876 28 24 29 2004 10 24 0 1605

877 28 24 29 2004 10 24 6 1605

878 28 24 29 2004 10 24 6 1605

879 28 24 29 2004 10 24 12 1605

880 28 24 29 2004 10 24 12 1605

881 28 24 29 2004 10 24 18 1605

882 28 24 29 2004 10 24 18 1605

883 28 24 29 2004 10 25 0 1605

884 28 24 29 2004 10 25 0 1605

885 28 24 29 2004 10 25 6 1605

886 28 24 29 2004 10 25 6 1605

887 28 24 29 2004 10 25 12 1605

888 28 24 29 2004 10 25 18 1605

889 28 24 29 2004 10 26 0 1605

890 29 25 30 2004 11 16 12 1606

891 29 25 30 2004 11 16 18 1606

892 29 25 30 2004 11 17 0 1606

893 29 25 30 2004 11 17 6 1606

894 29 25 30 2004 11 17 6 1606

895 29 25 30 2004 11 17 12 1606

896 29 25 30 2004 11 17 12 1606

897 29 25 30 2004 11 17 18 1606

101

898 29 25 30 2004 11 17 18 1606

899 29 25 30 2004 11 18 0 1606

900 29 25 30 2004 11 18 0 1606

901 29 25 30 2004 11 18 6 1606

902 29 25 30 2004 11 18 6 1606

903 29 25 30 2004 11 18 12 1606

904 29 25 30 2004 11 18 12 1606

905 29 25 30 2004 11 18 18 1606

906 29 25 30 2004 11 18 18 1606

907 29 25 30 2004 11 19 0 1606

908 29 25 30 2004 11 19 0 1606

909 29 25 30 2004 11 19 6 1606

910 29 25 30 2004 11 19 6 1606

911 29 25 30 2004 11 19 12 1606

912 29 25 30 2004 11 19 12 1606

913 29 25 30 2004 11 19 18 1606

914 29 25 30 2004 11 19 18 1606

915 29 25 30 2004 11 20 0 1606

916 29 25 30 2004 11 20 0 1606

917 29 25 30 2004 11 20 6 1606

918 29 25 30 2004 11 20 12 1606

919 29 25 30 2004 11 20 18 1606

920 29 25 30 2004 11 21 0 1606

921 29 25 30 2004 11 21 6 1606

922 29 25 30 2004 11 21 6 1606

923 29 25 30 2004 11 21 12 1606

924 29 25 30 2004 11 21 12 1606

925 29 25 30 2004 11 21 18 1606

926 29 25 30 2004 11 21 18 1606

927 29 25 30 2004 11 22 0 1606

928 29 25 30 2004 11 22 0 1606

929 29 25 30 2004 11 22 6 1606

930 29 25 30 2004 11 22 6 1606

931 29 25 30 2004 11 22 12 1606

932 29 25 30 2004 11 22 12 1606

933 29 25 30 2004 11 22 18 1606

934 29 25 30 2004 11 22 18 1606

935 29 25 30 2004 11 23 0 1606

936 29 25 30 2004 11 23 6 1606

102

937 29 25 30 2004 11 23 12 1606

938 29 25 30 2004 11 23 18 1606

939 29 25 30 2004 11 24 0 1606

940 29 25 30 2004 11 24 6 1606

941 30 27 32 2004 11 29 6 1607

942 30 27 32 2004 11 29 12 1607

943 30 27 32 2004 11 29 18 1607

944 30 27 32 2004 11 29 18 1607

945 30 27 32 2004 11 30 0 1607

946 30 27 32 2004 11 30 0 1607

947 30 27 32 2004 11 30 6 1607

948 30 27 32 2004 11 30 6 1607

949 30 27 32 2004 11 30 12 1607

950 30 27 32 2004 11 30 12 1607

951 30 27 32 2004 11 30 18 1607

952 30 27 32 2004 11 30 18 1607

953 30 27 32 2004 12 1 0 1607

954 30 27 32 2004 12 1 0 1607

955 30 27 32 2004 12 1 6 1607

956 30 27 32 2004 12 1 6 1607

957 30 27 32 2004 12 1 12 1607

958 30 27 32 2004 12 1 12 1607

959 30 27 32 2004 12 1 18 1607

960 30 27 32 2004 12 1 18 1607

961 30 27 32 2004 12 2 0 1607

962 30 27 32 2004 12 2 0 1607

963 30 27 32 2004 12 2 6 1607

964 30 27 32 2004 12 2 6 1607

965 30 27 32 2004 12 2 12 1607

966 30 27 32 2004 12 2 12 1607

967 30 27 32 2004 12 2 18 1607

968 30 27 32 2004 12 2 18 1607

969 30 27 32 2004 12 3 0 1607

970 30 27 32 2004 12 3 0 1607

971 30 27 32 2004 12 3 6 1607

972 30 27 32 2004 12 3 6 1607

973 30 27 32 2004 12 3 12 1607

974 30 27 32 2004 12 3 18 1607

975 32 29 34 2004 12 20 0 1609

103

976 32 29 34 2004 12 20 6 1609

977 32 29 34 2004 12 20 12 1609

978 32 29 34 2004 12 20 18 1609

979 1 1 1 2005 1 17 6 1610

980 1 1 1 2005 1 17 12 1610

981 1 1 1 2005 1 17 18 1610

982 1 1 1 2005 1 18 0 1610

983 1 1 1 2005 1 18 6 1610

984 1 1 1 2005 1 18 12 1610

985 2 2 2 2005 3 15 6 1611

986 2 2 2 2005 3 15 12 1611

987 2 2 2 2005 3 15 18 1611

988 2 2 2 2005 3 15 18 1611

989 2 2 2 2005 3 16 0 1611

990 2 2 2 2005 3 16 0 1611

991 2 2 2 2005 3 16 6 1611

992 2 2 2 2005 3 16 12 1611

993 2 2 2 2005 3 16 18 1611

994 2 2 2 2005 3 17 0 1611

995 3 3 3 2005 4 23 12 1612

996 3 3 3 2005 4 23 18 1612

997 3 3 3 2005 4 23 18 1612

998 3 3 3 2005 4 24 0 1612

999 3 3 3 2005 4 24 0 1612

1000 3 3 3 2005 4 24 6 1612

1001 3 3 3 2005 4 24 6 1612

1002 3 3 3 2005 4 24 12 1612

1003 3 3 3 2005 4 24 12 1612

1004 3 3 3 2005 4 24 18 1612

1005 3 3 3 2005 4 24 18 1612

1006 3 3 3 2005 4 25 0 1612

1007 3 3 3 2005 4 25 0 1612

1008 3 3 3 2005 4 25 6 1612

1009 3 3 3 2005 4 25 6 1612

1010 3 3 3 2005 4 25 12 1612

1011 3 3 3 2005 4 25 12 1612

1012 3 3 3 2005 4 25 18 1612

1013 3 3 3 2005 4 25 18 1612

1014 3 3 3 2005 4 26 0 1612

104

1015 3 3 3 2005 4 26 0 1612

1016 3 3 3 2005 4 26 6 1612

1017 3 3 3 2005 4 26 6 1612

1018 3 3 3 2005 4 26 12 1612

1019 3 3 3 2005 4 26 12 1612

1020 3 3 3 2005 4 26 18 1612

1021 3 3 3 2005 4 27 0 1612

1022 4 4 4 2005 6 1 6 1613

1023 4 4 4 2005 6 1 12 1613

1024 4 4 4 2005 6 1 12 1613

1025 4 4 4 2005 6 1 18 1613

1026 4 4 4 2005 6 1 18 1613

1027 4 4 4 2005 6 2 0 1613

1028 4 4 4 2005 6 2 0 1613

1029 4 4 4 2005 6 2 6 1613

1030 4 4 4 2005 6 2 6 1613

1031 4 4 4 2005 6 2 12 1613

1032 4 4 4 2005 6 2 12 1613

1033 4 4 4 2005 6 2 18 1613

1034 4 4 4 2005 6 2 18 1613

1035 4 4 4 2005 6 3 0 1613

1036 4 4 4 2005 6 3 0 1613

1037 4 4 4 2005 6 3 6 1613

1038 4 4 4 2005 6 3 6 1613

1039 4 4 4 2005 6 3 12 1613

1040 4 4 4 2005 6 3 12 1613

1041 4 4 4 2005 6 3 18 1613

1042 4 4 4 2005 6 3 18 1613

1043 4 4 4 2005 6 4 0 1613

1044 4 4 4 2005 6 4 0 1613

1045 4 4 4 2005 6 4 6 1613

1046 4 4 4 2005 6 4 6 1613

1047 4 4 4 2005 6 4 12 1613

1048 4 4 4 2005 6 4 12 1613

1049 4 4 4 2005 6 4 18 1613

1050 4 4 4 2005 6 4 18 1613

1051 4 4 4 2005 6 5 0 1613

1052 4 4 4 2005 6 5 0 1613

1053 4 4 4 2005 6 5 6 1613

105

1054 4 4 4 2005 6 5 6 1613

1055 4 4 4 2005 6 5 12 1613

1056 4 4 4 2005 6 5 12 1613

1057 4 4 4 2005 6 5 18 1613

1058 4 4 4 2005 6 5 18 1613

1059 4 4 4 2005 6 6 0 1613

1060 4 4 4 2005 6 6 0 1613

1061 4 4 4 2005 6 6 6 1613

1062 4 4 4 2005 6 6 6 1613

1063 4 4 4 2005 6 6 12 1613

1064 4 4 4 2005 6 6 12 1613

1065 4 4 4 2005 6 6 18 1613

1066 4 4 4 2005 6 6 18 1613

1067 4 4 4 2005 6 7 0 1613

1068 4 4 4 2005 6 7 0 1613

1069 4 4 4 2005 6 7 6 1613

1070 4 4 4 2005 6 7 6 1613

1071 4 4 4 2005 6 7 12 1613

1072 4 4 4 2005 6 7 12 1613

1073 4 4 4 2005 6 7 18 1613

1074 4 4 4 2005 6 7 18 1613

1075 4 4 4 2005 6 8 0 1613

1076 4 4 4 2005 6 8 0 1613

1077 4 4 4 2005 6 8 6 1613

1078 4 4 4 2005 6 8 6 1613

1079 4 4 4 2005 6 8 12 1613

1080 4 4 4 2005 6 8 18 1613

1081 4 4 4 2005 6 9 0 1613

1082 4 4 4 2005 6 9 6 1613

1083 5 5 5 2005 7 13 12 1614

1084 5 5 5 2005 7 13 18 1614

1085 5 5 5 2005 7 13 18 1614

1086 5 5 5 2005 7 14 0 1614

1087 5 5 5 2005 7 14 0 1614

1088 5 5 5 2005 7 14 6 1614

1089 5 5 5 2005 7 14 6 1614

1090 5 5 5 2005 7 14 12 1614

1091 5 5 5 2005 7 14 12 1614

1092 5 5 5 2005 7 14 18 1614

106

1093 5 5 5 2005 7 14 18 1614

1094 5 5 5 2005 7 15 0 1614

1095 5 5 5 2005 7 15 0 1614

1096 5 5 5 2005 7 15 6 1614

1097 5 5 5 2005 7 15 6 1614

1098 5 5 5 2005 7 15 12 1614

1099 5 5 5 2005 7 15 12 1614

1100 5 5 5 2005 7 15 18 1614

1101 5 5 5 2005 7 15 18 1614

1102 5 5 5 2005 7 16 0 1614

1103 5 5 5 2005 7 16 0 1614

1104 5 5 5 2005 7 16 6 1614

1105 5 5 5 2005 7 16 6 1614

1106 5 5 5 2005 7 16 12 1614

1107 5 5 5 2005 7 16 12 1614

1108 5 5 5 2005 7 16 18 1614

1109 5 5 5 2005 7 16 18 1614

1110 5 5 5 2005 7 17 0 1614

1111 5 5 5 2005 7 17 0 1614

1112 5 5 5 2005 7 17 6 1614

1113 5 5 5 2005 7 17 6 1614

1114 5 5 5 2005 7 17 12 1614

1115 5 5 5 2005 7 17 12 1614

1116 5 5 5 2005 7 17 18 1614

1117 5 5 5 2005 7 17 18 1614

1118 5 5 5 2005 7 18 0 1614

1119 5 5 5 2005 7 18 0 1614

1120 5 5 5 2005 7 18 6 1614

1121 5 5 5 2005 7 18 6 1614

1122 5 5 5 2005 7 18 12 1614

1123 5 5 5 2005 7 18 12 1614

1124 5 5 5 2005 7 18 18 1614

1125 5 5 5 2005 7 19 0 1614

1126 5 5 5 2005 7 19 6 1614

1127 7 7 7 2005 7 23 18 1616

1128 7 7 7 2005 7 24 0 1616

1129 7 7 7 2005 7 24 6 1616

1130 7 7 7 2005 7 24 12 1616

1131 7 7 7 2005 7 24 18 1616

107

1132 7 7 7 2005 7 25 0 1616

1133 9 9 9 2005 8 1 18 1618

1134 9 9 9 2005 8 2 0 1618

1135 9 9 9 2005 8 2 6 1618

1136 9 9 9 2005 8 2 12 1618

1137 9 9 9 2005 8 2 12 1618

1138 9 9 9 2005 8 2 18 1618

1139 9 9 9 2005 8 2 18 1618

1140 9 9 9 2005 8 3 0 1618

1141 9 9 9 2005 8 3 0 1618

1142 9 9 9 2005 8 3 6 1618

1143 9 9 9 2005 8 3 6 1618

1144 9 9 9 2005 8 3 12 1618

1145 9 9 9 2005 8 3 12 1618

1146 9 9 9 2005 8 3 18 1618

1147 9 9 9 2005 8 3 18 1618

1148 9 9 9 2005 8 4 0 1618

1149 9 9 9 2005 8 4 0 1618

1150 9 9 9 2005 8 4 6 1618

1151 9 9 9 2005 8 4 6 1618

1152 9 9 9 2005 8 4 12 1618

1153 9 9 9 2005 8 4 12 1618

1154 9 9 9 2005 8 4 18 1618

1155 9 9 9 2005 8 4 18 1618

1156 9 9 9 2005 8 5 0 1618

1157 9 9 9 2005 8 5 0 1618

1158 9 9 9 2005 8 5 6 1618

1159 9 9 9 2005 8 5 6 1618

1160 9 9 9 2005 8 5 12 1618

1161 9 9 9 2005 8 5 12 1618

1162 9 9 9 2005 8 5 18 1618

1163 9 9 9 2005 8 6 0 1618

1164 9 9 9 2005 8 6 6 1618

1165 10 10 10 2005 8 12 12 1619

1166 10 10 10 2005 8 12 18 1619

1167 10 10 10 2005 8 13 0 1619

1168 11 11 11 2005 8 20 12 1620

1169 11 11 11 2005 8 20 18 1620

1170 11 11 11 2005 8 21 0 1620

108

1171 11 11 11 2005 8 21 0 1620

1172 11 11 11 2005 8 21 6 1620

1173 11 11 11 2005 8 21 6 1620

1174 11 11 11 2005 8 21 12 1620

1175 11 11 11 2005 8 21 12 1620

1176 11 11 11 2005 8 21 18 1620

1177 11 11 11 2005 8 21 18 1620

1178 11 11 11 2005 8 22 0 1620

1179 11 11 11 2005 8 22 0 1620

1180 11 11 11 2005 8 22 6 1620

1181 11 11 11 2005 8 22 6 1620

1182 11 11 11 2005 8 22 12 1620

1183 11 11 11 2005 8 22 12 1620

1184 11 11 11 2005 8 22 18 1620

1185 11 11 11 2005 8 22 18 1620

1186 11 11 11 2005 8 23 0 1620

1187 11 11 11 2005 8 23 0 1620

1188 11 11 11 2005 8 23 6 1620

1189 11 11 11 2005 8 23 6 1620

1190 11 11 11 2005 8 23 12 1620

1191 11 11 11 2005 8 23 12 1620

1192 11 11 11 2005 8 23 18 1620

1193 11 11 11 2005 8 23 18 1620

1194 11 11 11 2005 8 24 0 1620

1195 11 11 11 2005 8 24 0 1620

1196 11 11 11 2005 8 24 6 1620

1197 11 11 11 2005 8 24 6 1620

1198 11 11 11 2005 8 24 12 1620

1199 11 11 11 2005 8 24 12 1620

1200 11 11 11 2005 8 24 18 1620

1201 11 11 11 2005 8 24 18 1620

1202 11 11 11 2005 8 25 0 1620

1203 11 11 11 2005 8 25 0 1620

1204 11 11 11 2005 8 25 6 1620

1205 11 11 11 2005 8 25 6 1620

1206 11 11 11 2005 8 25 12 1620

1207 11 11 11 2005 8 25 12 1620

1208 11 11 11 2005 8 25 18 1620

1209 11 11 11 2005 8 25 18 1620

109

1210 11 11 11 2005 8 26 0 1620

1211 11 11 11 2005 8 26 0 1620

1212 11 11 11 2005 8 26 6 1620

1213 11 11 11 2005 8 26 12 1620

1214 11 11 11 2005 8 27 0 1620

1215 12 12 12 2005 8 22 0 1621

1216 12 12 12 2005 8 22 6 1621

1217 12 12 12 2005 8 22 12 1621

1218 12 12 12 2005 8 23 6 1621

1219 12 12 12 2005 8 23 12 1621

1220 12 12 12 2005 8 23 18 1621

1221 12 12 12 2005 8 24 0 1621

1222 12 12 12 2005 8 24 6 1621

1223 12 12 12 2005 8 24 12 1621

1224 13 13 13 2005 8 27 18 1622

1225 13 13 13 2005 8 28 0 1622

1226 13 13 13 2005 8 28 6 1622

1227 13 13 13 2005 8 28 12 1622

1228 13 13 13 2005 8 28 12 1622

1229 13 13 13 2005 8 28 18 1622

1230 13 13 13 2005 8 28 18 1622

1231 13 13 13 2005 8 29 0 1622

1232 13 13 13 2005 8 29 0 1622

1233 13 13 13 2005 8 29 6 1622

1234 13 13 13 2005 8 29 6 1622

1235 13 13 13 2005 8 29 12 1622

1236 13 13 13 2005 8 29 12 1622

1237 13 13 13 2005 8 29 18 1622

1238 13 13 13 2005 8 29 18 1622

1239 13 13 13 2005 8 30 0 1622

1240 13 13 13 2005 8 30 0 1622

1241 13 13 13 2005 8 30 6 1622

1242 13 13 13 2005 8 30 6 1622

1243 13 13 13 2005 8 30 12 1622

1244 13 13 13 2005 8 30 12 1622

1245 13 13 13 2005 8 30 18 1622

1246 13 13 13 2005 8 30 18 1622

1247 13 13 13 2005 8 31 0 1622

1248 13 13 13 2005 8 31 0 1622

110

1249 13 13 13 2005 8 31 6 1622

1250 13 13 13 2005 8 31 6 1622

1251 13 13 13 2005 8 31 12 1622

1252 13 13 13 2005 8 31 12 1622

1253 13 13 13 2005 8 31 18 1622

1254 13 13 13 2005 8 31 18 1622

1255 13 13 13 2005 9 1 0 1622

1256 13 13 13 2005 9 1 0 1622

1257 13 13 13 2005 9 1 6 1622

1258 14 14 14 2005 8 30 6 1623

1259 14 14 14 2005 8 30 12 1623

1260 14 14 14 2005 8 30 18 1623

1261 14 14 14 2005 8 30 18 1623

1262 14 14 14 2005 8 31 0 1623

1263 14 14 14 2005 8 31 0 1623

1264 14 14 14 2005 8 31 6 1623

1265 14 14 14 2005 8 31 6 1623

1266 14 14 14 2005 8 31 12 1623

1267 14 14 14 2005 8 31 12 1623

1268 14 14 14 2005 8 31 18 1623

1269 14 14 14 2005 8 31 18 1623

1270 14 14 14 2005 9 1 0 1623

1271 14 14 14 2005 9 1 0 1623

1272 14 14 14 2005 9 1 6 1623

1273 14 14 14 2005 9 1 6 1623

1274 14 14 14 2005 9 1 12 1623

1275 14 14 14 2005 9 1 12 1623

1276 14 14 14 2005 9 1 18 1623

1277 14 14 14 2005 9 1 18 1623

1278 14 14 14 2005 9 2 0 1623

1279 14 14 14 2005 9 2 0 1623

1280 14 14 14 2005 9 2 6 1623

1281 14 14 14 2005 9 2 6 1623

1282 14 14 14 2005 9 2 12 1623

1283 14 14 14 2005 9 2 12 1623

1284 14 14 14 2005 9 2 18 1623

1285 14 14 14 2005 9 2 18 1623

1286 14 14 14 2005 9 3 0 1623

1287 14 14 14 2005 9 3 0 1623

111

1288 14 14 14 2005 9 3 6 1623

1289 14 14 14 2005 9 3 6 1623

1290 14 14 14 2005 9 3 12 1623

1291 14 14 14 2005 9 3 12 1623

1292 14 14 14 2005 9 3 18 1623

1293 14 14 14 2005 9 3 18 1623

1294 14 14 14 2005 9 4 0 1623

1295 14 14 14 2005 9 4 0 1623

1296 14 14 14 2005 9 4 6 1623

1297 14 14 14 2005 9 4 6 1623

1298 14 14 14 2005 9 4 12 1623

1299 14 14 14 2005 9 4 12 1623

1300 14 14 14 2005 9 4 18 1623

1301 14 14 14 2005 9 4 18 1623

1302 14 14 14 2005 9 5 0 1623

1303 14 14 14 2005 9 5 0 1623

1304 14 14 14 2005 9 5 6 1623

1305 14 14 14 2005 9 5 6 1623

1306 14 14 14 2005 9 5 12 1623

1307 14 14 14 2005 9 5 12 1623

1308 14 14 14 2005 9 5 18 1623

1309 14 14 14 2005 9 5 18 1623

1310 14 14 14 2005 9 6 0 1623

1311 14 14 14 2005 9 6 0 1623

1312 14 14 14 2005 9 6 6 1623

1313 14 14 14 2005 9 6 6 1623

1314 14 14 14 2005 9 6 12 1623

1315 14 14 14 2005 9 6 12 1623

1316 15 15 15 2005 9 6 18 1624

1317 15 15 15 2005 9 7 6 1624

1318 15 15 15 2005 9 7 12 1624

1319 15 15 15 2005 9 8 0 1624

1320 15 15 15 2005 9 8 6 1624

1321 15 15 15 2005 9 8 12 1624

1322 15 15 15 2005 9 8 18 1624

1323 15 15 15 2005 9 9 0 1624

1324 15 15 15 2005 9 9 6 1624

1325 15 15 15 2005 9 9 6 1624

1326 15 15 15 2005 9 9 12 1624

112

1327 15 15 15 2005 9 9 12 1624

1328 15 15 15 2005 9 9 18 1624

1329 15 15 15 2005 9 9 18 1624

1330 15 15 15 2005 9 10 0 1624

1331 15 15 15 2005 9 10 0 1624

1332 15 15 15 2005 9 10 6 1624

1333 15 15 15 2005 9 10 6 1624

1334 15 15 15 2005 9 10 12 1624

1335 15 15 15 2005 9 10 12 1624

1336 15 15 15 2005 9 10 18 1624

1337 15 15 15 2005 9 10 18 1624

1338 15 15 15 2005 9 11 0 1624

1339 15 15 15 2005 9 11 0 1624

1340 15 15 15 2005 9 11 6 1624

1341 15 15 15 2005 9 11 6 1624

1342 15 15 15 2005 9 11 12 1624

1343 15 15 15 2005 9 11 12 1624

1344 15 15 15 2005 9 11 18 1624

1345 17 18 18 2005 9 22 6 1626

1346 17 18 18 2005 9 23 0 1626

1347 17 18 18 2005 9 23 6 1626

1348 17 18 18 2005 9 24 12 1626

1349 17 18 18 2005 9 24 18 1626

1350 17 18 18 2005 9 25 0 1626

1351 17 18 18 2005 9 25 0 1626

1352 17 18 18 2005 9 25 6 1626

1353 17 18 18 2005 9 25 6 1626

1354 17 18 18 2005 9 25 12 1626

1355 17 18 18 2005 9 25 12 1626

1356 17 18 18 2005 9 25 18 1626

1357 17 18 18 2005 9 25 18 1626

1358 17 18 18 2005 9 26 0 1626

1359 17 18 18 2005 9 26 0 1626

1360 17 18 18 2005 9 26 6 1626

1361 17 18 18 2005 9 26 12 1626

1362 17 18 18 2005 9 26 18 1626

1363 17 18 18 2005 9 27 0 1626

1364 18 17 17 2005 9 21 12 1627

1365 18 17 17 2005 9 21 18 1627

113

1366 18 17 17 2005 9 22 0 1627

1367 18 17 17 2005 9 22 0 1627

1368 18 17 17 2005 9 22 6 1627

1369 18 17 17 2005 9 22 6 1627

1370 18 17 17 2005 9 22 12 1627

1371 18 17 17 2005 9 22 12 1627

1372 18 17 17 2005 9 22 18 1627

1373 18 17 17 2005 9 22 18 1627

1374 18 17 17 2005 9 23 0 1627

1375 18 17 17 2005 9 23 0 1627

1376 18 17 17 2005 9 23 6 1627

1377 18 17 17 2005 9 23 6 1627

1378 18 17 17 2005 9 23 12 1627

1379 18 17 17 2005 9 23 12 1627

1380 18 17 17 2005 9 23 18 1627

1381 18 17 17 2005 9 23 18 1627

1382 18 17 17 2005 9 24 0 1627

1383 18 17 17 2005 9 24 0 1627

1384 18 17 17 2005 9 24 6 1627

1385 18 17 17 2005 9 24 6 1627

1386 18 17 17 2005 9 24 12 1627

1387 18 17 17 2005 9 24 12 1627

1388 18 17 17 2005 9 24 18 1627

1389 18 17 17 2005 9 24 18 1627

1390 18 17 17 2005 9 25 0 1627

1391 18 17 17 2005 9 25 0 1627

1392 18 17 17 2005 9 25 6 1627

1393 18 17 17 2005 9 25 6 1627

1394 18 17 17 2005 9 25 12 1627

1395 19 19 19 2005 9 26 18 1628

1396 19 19 19 2005 9 27 0 1628

1397 19 19 19 2005 9 27 0 1628

1398 19 19 19 2005 9 27 6 1628

1399 19 19 19 2005 9 27 6 1628

1400 19 19 19 2005 9 27 12 1628

1401 19 19 19 2005 9 27 12 1628

1402 19 19 19 2005 9 27 18 1628

1403 19 19 19 2005 9 27 18 1628

1404 19 19 19 2005 9 28 0 1628

114

1405 19 19 19 2005 9 28 0 1628

1406 19 19 19 2005 9 28 6 1628

1407 19 19 19 2005 9 28 6 1628

1408 19 19 19 2005 9 28 12 1628

1409 19 19 19 2005 9 28 12 1628

1410 19 19 19 2005 9 28 18 1628

1411 19 19 19 2005 9 28 18 1628

1412 19 19 19 2005 9 29 0 1628

1413 19 19 19 2005 9 29 0 1628

1414 19 19 19 2005 9 29 6 1628

1415 19 19 19 2005 9 29 6 1628

1416 19 19 19 2005 9 29 12 1628

1417 19 19 19 2005 9 29 12 1628

1418 19 19 19 2005 9 29 18 1628

1419 19 19 19 2005 9 29 18 1628

1420 19 19 19 2005 9 30 0 1628

1421 19 19 19 2005 9 30 0 1628

1422 19 19 19 2005 9 30 6 1628

1423 19 19 19 2005 9 30 6 1628

1424 19 19 19 2005 9 30 12 1628

1425 19 19 19 2005 9 30 12 1628

1426 19 19 19 2005 9 30 18 1628

1427 19 19 19 2005 9 30 18 1628

1428 19 19 19 2005 10 1 0 1628

1429 19 19 19 2005 10 1 0 1628

1430 19 19 19 2005 10 1 6 1628

1431 19 19 19 2005 10 1 6 1628

1432 19 19 19 2005 10 1 12 1628

1433 19 19 19 2005 10 1 12 1628

1434 19 19 19 2005 10 1 18 1628

1435 19 19 19 2005 10 1 18 1628

1436 19 19 19 2005 10 2 0 1628

1437 19 19 19 2005 10 2 0 1628

1438 19 19 19 2005 10 2 6 1628

1439 19 19 19 2005 10 2 6 1628

1440 19 19 19 2005 10 2 12 1628

1441 19 19 19 2005 10 2 12 1628

1442 19 19 19 2005 10 2 18 1628

1443 21 20 20 2005 10 11 6 1630

115

1444 21 20 20 2005 10 11 12 1630

1445 21 20 20 2005 10 11 12 1630

1446 21 20 20 2005 10 11 18 1630

1447 21 20 20 2005 10 11 18 1630

1448 21 20 20 2005 10 12 0 1630

1449 21 20 20 2005 10 12 0 1630

1450 21 20 20 2005 10 12 6 1630

1451 21 20 20 2005 10 12 6 1630

1452 21 20 20 2005 10 12 12 1630

1453 21 20 20 2005 10 12 12 1630

1454 21 20 20 2005 10 12 18 1630

1455 21 20 20 2005 10 12 18 1630

1456 21 20 20 2005 10 13 0 1630

1457 21 20 20 2005 10 13 0 1630

1458 21 20 20 2005 10 13 6 1630

1459 21 20 20 2005 10 13 6 1630

1460 21 20 20 2005 10 13 12 1630

1461 21 20 20 2005 10 13 12 1630

1462 21 20 20 2005 10 13 18 1630

1463 21 20 20 2005 10 13 18 1630

1464 21 20 20 2005 10 14 0 1630

1465 21 20 20 2005 10 14 0 1630

1466 21 20 20 2005 10 14 6 1630

1467 21 20 20 2005 10 14 6 1630

1468 21 20 20 2005 10 14 12 1630

1469 21 20 20 2005 10 14 12 1630

1470 21 20 20 2005 10 14 18 1630

1471 21 20 20 2005 10 14 18 1630

1472 21 20 20 2005 10 15 0 1630

1473 21 20 20 2005 10 15 0 1630

1474 21 20 20 2005 10 15 6 1630

1475 21 20 20 2005 10 15 6 1630

1476 21 20 20 2005 10 15 12 1630

1477 21 20 20 2005 10 15 12 1630

1478 21 20 20 2005 10 15 18 1630

1479 21 20 20 2005 10 15 18 1630

1480 21 20 20 2005 10 16 0 1630

1481 21 20 20 2005 10 16 0 1630

1482 21 20 20 2005 10 16 6 1630

116

1483 21 20 20 2005 10 16 6 1630

1484 21 20 20 2005 10 16 12 1630

1485 21 20 20 2005 10 16 12 1630

1486 21 20 20 2005 10 16 18 1630

1487 21 20 20 2005 10 16 18 1630

1488 21 20 20 2005 10 17 0 1630

1489 21 20 20 2005 10 17 0 1630

1490 21 20 20 2005 10 17 6 1630

1491 21 20 20 2005 10 17 6 1630

1492 21 20 20 2005 10 17 12 1630

1493 21 20 20 2005 10 17 12 1630

1494 21 20 20 2005 10 17 18 1630

1495 21 20 20 2005 10 17 18 1630

1496 21 20 20 2005 10 18 0 1630

1497 21 20 20 2005 10 18 0 1630

1498 21 20 20 2005 10 18 6 1630

1499 21 20 20 2005 10 18 6 1630

1500 21 20 20 2005 10 18 12 1630

1501 21 20 20 2005 10 18 12 1630

1502 21 20 20 2005 10 18 18 1630

1503 22 21 21 2005 10 29 18 1631

1504 22 21 21 2005 10 30 0 1631

1505 22 21 21 2005 10 30 6 1631

1506 22 21 21 2005 10 30 6 1631

1507 22 21 21 2005 10 30 12 1631

1508 22 21 21 2005 10 30 12 1631

1509 22 21 21 2005 10 30 18 1631

1510 22 21 21 2005 10 30 18 1631

1511 22 21 21 2005 10 31 0 1631

1512 22 21 21 2005 10 31 0 1631

1513 22 21 21 2005 10 31 6 1631

1514 22 21 21 2005 10 31 6 1631

1515 22 21 21 2005 10 31 12 1631

1516 22 21 21 2005 10 31 12 1631

1517 22 21 21 2005 10 31 18 1631

1518 22 21 21 2005 10 31 18 1631

1519 22 21 21 2005 11 1 0 1631

1520 22 21 21 2005 11 1 0 1631

1521 22 21 21 2005 11 1 6 1631

117

1522 22 21 21 2005 11 1 12 1631

1523 22 21 21 2005 11 1 18 1631

1524 22 21 21 2005 11 2 0 1631

1525 24 23 23 2005 11 16 6 1633

1526 24 23 23 2005 11 16 12 1633

1527 24 23 23 2005 11 16 18 1633

1528 24 23 23 2005 11 17 0 1633

1529 24 23 23 2005 11 17 6 1633

1530 24 23 23 2005 11 17 6 1633

1531 24 23 23 2005 11 17 12 1633

1532 24 23 23 2005 11 17 12 1633

1533 24 23 23 2005 11 17 18 1633

1534 24 23 23 2005 11 17 18 1633

1535 24 23 23 2005 11 18 0 1633

1536 24 23 23 2005 11 18 6 1633

1537 24 23 23 2005 11 18 12 1633

1538 24 23 23 2005 11 18 18 1633

1539 24 23 23 2005 11 19 0 1633

1540 24 23 23 2005 11 19 6 1633

1541 2 1 1 2006 5 10 0 1636

1542 2 1 1 2006 5 10 6 1636

1543 2 1 1 2006 5 10 12 1636

1544 2 1 1 2006 5 10 18 1636

1545 2 1 1 2006 5 10 18 1636

1546 2 1 1 2006 5 11 0 1636

1547 2 1 1 2006 5 11 0 1636

1548 2 1 1 2006 5 11 6 1636

1549 2 1 1 2006 5 11 6 1636

1550 2 1 1 2006 5 11 12 1636

1551 2 1 1 2006 5 11 12 1636

1552 2 1 1 2006 5 11 18 1636

1553 2 1 1 2006 5 12 0 1636

1554 2 1 1 2006 5 12 6 1636

1555 2 1 1 2006 5 12 6 1636

1556 2 1 1 2006 5 12 12 1636

1557 2 1 1 2006 5 12 12 1636

1558 2 1 1 2006 5 12 18 1636

1559 2 1 1 2006 5 12 18 1636

1560 2 1 1 2006 5 13 0 1636

118

1561 2 1 1 2006 5 13 0 1636

1562 2 1 1 2006 5 13 6 1636

1563 2 1 1 2006 5 13 6 1636

1564 2 1 1 2006 5 13 12 1636

1565 2 1 1 2006 5 13 12 1636

1566 2 1 1 2006 5 13 18 1636

1567 2 1 1 2006 5 13 18 1636

1568 2 1 1 2006 5 14 0 1636

1569 2 1 1 2006 5 14 0 1636

1570 2 1 1 2006 5 14 6 1636

1571 2 1 1 2006 5 14 6 1636

1572 2 1 1 2006 5 14 12 1636

1573 2 1 1 2006 5 14 12 1636

1574 2 1 1 2006 5 14 18 1636

1575 2 1 1 2006 5 14 18 1636

1576 2 1 1 2006 5 15 0 1636

1577 2 1 1 2006 5 15 0 1636

1578 2 1 1 2006 5 15 6 1636

1579 2 1 1 2006 5 15 6 1636

1580 2 1 1 2006 5 15 12 1636

1581 2 1 1 2006 5 15 12 1636

1582 2 1 1 2006 5 15 18 1636

1583 2 1 1 2006 5 15 18 1636

1584 2 1 1 2006 5 16 0 1636

1585 2 1 1 2006 5 16 0 1636

1586 2 1 1 2006 5 16 6 1636

1587 2 1 1 2006 5 16 6 1636

1588 2 1 1 2006 5 16 12 1636

1589 2 1 1 2006 5 16 12 1636

1590 2 1 1 2006 5 16 18 1636

1591 2 1 1 2006 5 16 18 1636

1592 2 1 1 2006 5 17 0 1636

1593 2 1 1 2006 5 17 0 1636

1594 2 1 1 2006 5 17 6 1636

1595 2 1 1 2006 5 17 6 1636

1596 2 1 1 2006 5 17 12 1636

1597 2 1 1 2006 5 17 12 1636

1598 2 1 1 2006 5 17 18 1636

1599 2 1 1 2006 5 17 18 1636

119

1600 4 3 3 2006 7 1 12 1638

1601 4 3 3 2006 7 1 18 1638

1602 4 3 3 2006 7 2 0 1638

1603 4 3 3 2006 7 2 6 1638

1604 4 3 3 2006 7 2 12 1638

1605 4 3 3 2006 7 2 18 1638

1606 4 3 3 2006 7 3 0 1638

1607 4 3 3 2006 7 3 6 1638

1608 4 3 3 2006 7 3 6 1638

1609 4 3 3 2006 7 3 12 1638

1610 4 3 3 2006 7 3 12 1638

1611 4 3 3 2006 7 3 18 1638

1612 4 3 3 2006 7 3 18 1638

1613 4 3 3 2006 7 4 0 1638

1614 4 3 3 2006 7 4 0 1638

1615 4 3 3 2006 7 4 6 1638

1616 4 3 3 2006 7 4 6 1638

1617 4 3 3 2006 7 4 12 1638

1618 4 3 3 2006 7 4 12 1638

1619 4 3 3 2006 7 4 18 1638

1620 4 3 3 2006 7 4 18 1638

1621 4 3 3 2006 7 5 0 1638

1622 4 3 3 2006 7 5 0 1638

1623 4 3 3 2006 7 5 6 1638

1624 4 3 3 2006 7 5 6 1638

1625 4 3 3 2006 7 5 12 1638

1626 4 3 3 2006 7 5 12 1638

1627 4 3 3 2006 7 5 18 1638

1628 4 3 3 2006 7 5 18 1638

1629 4 3 3 2006 7 6 0 1638

1630 4 3 3 2006 7 6 0 1638

1631 4 3 3 2006 7 6 6 1638

1632 4 3 3 2006 7 6 6 1638

1633 4 3 3 2006 7 6 12 1638

1634 4 3 3 2006 7 6 12 1638

1635 4 3 3 2006 7 6 18 1638

1636 4 3 3 2006 7 6 18 1638

1637 4 3 3 2006 7 7 0 1638

1638 4 3 3 2006 7 7 0 1638

120

1639 4 3 3 2006 7 7 6 1638

1640 4 3 3 2006 7 7 6 1638

1641 4 3 3 2006 7 7 12 1638

1642 4 3 3 2006 7 7 12 1638

1643 4 3 3 2006 7 7 18 1638

1644 4 3 3 2006 7 7 18 1638

1645 4 3 3 2006 7 8 0 1638

1646 4 3 3 2006 7 8 0 1638

1647 4 3 3 2006 7 8 6 1638

1648 4 3 3 2006 7 8 6 1638

1649 4 3 3 2006 7 8 12 1638

1650 4 3 3 2006 7 8 12 1638

1651 4 3 3 2006 7 8 18 1638

1652 4 3 3 2006 7 8 18 1638

1653 4 3 3 2006 7 9 0 1638

1654 4 3 3 2006 7 9 0 1638

1655 4 3 3 2006 7 9 6 1638

1656 4 3 3 2006 7 9 12 1638

1657 6 5 5 2006 7 20 0 1640

1658 6 5 5 2006 7 20 6 1640

1659 6 5 5 2006 7 20 12 1640

1660 6 5 5 2006 7 20 18 1640

1661 6 5 5 2006 7 20 18 1640

1662 6 5 5 2006 7 21 0 1640

1663 6 5 5 2006 7 21 0 1640

1664 6 5 5 2006 7 21 6 1640

1665 6 5 5 2006 7 21 6 1640

1666 6 5 5 2006 7 21 12 1640

1667 6 5 5 2006 7 21 12 1640

1668 6 5 5 2006 7 21 18 1640

1669 6 5 5 2006 7 21 18 1640

1670 6 5 5 2006 7 22 0 1640

1671 6 5 5 2006 7 22 0 1640

1672 6 5 5 2006 7 22 6 1640

1673 6 5 5 2006 7 22 6 1640

1674 6 5 5 2006 7 22 12 1640

1675 6 5 5 2006 7 22 12 1640

1676 6 5 5 2006 7 22 18 1640

1677 6 5 5 2006 7 22 18 1640

121

1678 6 5 5 2006 7 23 0 1640

1679 6 5 5 2006 7 23 0 1640

1680 6 5 5 2006 7 23 6 1640

1681 6 5 5 2006 7 23 6 1640

1682 6 5 5 2006 7 23 12 1640

1683 6 5 5 2006 7 23 12 1640

1684 6 5 5 2006 7 23 18 1640

1685 6 5 5 2006 7 24 0 1640

1686 6 5 5 2006 7 24 6 1640

1687 6 5 5 2006 7 24 12 1640

1688 6 5 5 2006 7 24 18 1640

1689 6 5 5 2006 7 25 0 1640

1690 6 5 5 2006 7 25 6 1640

1691 7 6 6 2006 8 1 18 1641

1692 7 6 6 2006 8 2 0 1641

1693 7 6 6 2006 8 2 6 1641

1694 7 6 6 2006 8 2 12 1641

1695 7 6 6 2006 8 2 18 1641

1696 7 6 6 2006 8 3 0 1641

1697 7 6 6 2006 8 3 0 1641

1698 7 6 6 2006 8 3 6 1641

1699 7 6 6 2006 8 3 6 1641

1700 7 6 6 2006 8 3 12 1641

1701 8 8 8 2006 8 6 6 1642

1702 8 8 8 2006 8 6 12 1642

1703 8 8 8 2006 8 6 18 1642

1704 8 8 8 2006 8 7 0 1642

1705 8 8 8 2006 8 7 0 1642

1706 8 8 8 2006 8 7 6 1642

1707 8 8 8 2006 8 7 6 1642

1708 8 8 8 2006 8 7 12 1642

1709 8 8 8 2006 8 7 12 1642

1710 8 8 8 2006 8 7 18 1642

1711 8 8 8 2006 8 7 18 1642

1712 8 8 8 2006 8 8 0 1642

1713 8 8 8 2006 8 8 0 1642

1714 8 8 8 2006 8 8 6 1642

1715 8 8 8 2006 8 8 6 1642

1716 8 8 8 2006 8 8 12 1642

122

1717 8 8 8 2006 8 8 12 1642

1718 8 8 8 2006 8 8 18 1642

1719 8 8 8 2006 8 8 18 1642

1720 8 8 8 2006 8 9 0 1642

1721 8 8 8 2006 8 9 0 1642

1722 8 8 8 2006 8 9 6 1642

1723 8 8 8 2006 8 9 6 1642

1724 8 8 8 2006 8 9 12 1642

1725 8 8 8 2006 8 9 12 1642

1726 8 8 8 2006 8 9 18 1642

1727 8 8 8 2006 8 9 18 1642

1728 8 8 8 2006 8 10 0 1642

1729 8 8 8 2006 8 10 0 1642

1730 8 8 8 2006 8 10 6 1642

1731 8 8 8 2006 8 10 6 1642

1732 8 8 8 2006 8 10 12 1642

1733 8 8 8 2006 8 10 12 1642

1734 8 8 8 2006 8 10 18 1642

1735 8 8 8 2006 8 10 18 1642

1736 8 8 8 2006 8 11 0 1642

1737 8 8 8 2006 8 11 6 1642

1738 9 7 7 2006 8 6 6 1643

1739 9 7 7 2006 8 6 12 1643

1740 9 7 7 2006 8 6 18 1643

1741 9 7 7 2006 8 7 0 1643

1742 9 7 7 2006 8 7 6 1643

1743 9 7 7 2006 8 7 12 1643

1744 9 7 7 2006 8 7 18 1643

1745 9 7 7 2006 8 8 0 1643

1746 11 10 10 2006 8 17 18 1645

1747 14 13 13 2006 9 10 18 1648

1748 14 13 13 2006 9 11 0 1648

1749 14 13 13 2006 9 11 6 1648

1750 14 13 13 2006 9 11 12 1648

1751 14 13 13 2006 9 11 12 1648

1752 14 13 13 2006 9 11 18 1648

1753 14 13 13 2006 9 11 18 1648

1754 14 13 13 2006 9 12 0 1648

1755 14 13 13 2006 9 12 0 1648

123

1756 14 13 13 2006 9 12 6 1648

1757 14 13 13 2006 9 12 6 1648

1758 14 13 13 2006 9 12 12 1648

1759 14 13 13 2006 9 12 12 1648

1760 14 13 13 2006 9 12 18 1648

1761 14 13 13 2006 9 12 18 1648

1762 14 13 13 2006 9 13 0 1648

1763 14 13 13 2006 9 13 0 1648

1764 14 13 13 2006 9 13 6 1648

1765 14 13 13 2006 9 13 6 1648

1766 14 13 13 2006 9 13 12 1648

1767 14 13 13 2006 9 13 12 1648

1768 14 13 13 2006 9 13 18 1648

1769 14 13 13 2006 9 13 18 1648

1770 14 13 13 2006 9 14 0 1648

1771 14 13 13 2006 9 14 0 1648

1772 14 13 13 2006 9 14 6 1648

1773 14 13 13 2006 9 14 6 1648

1774 14 13 13 2006 9 14 12 1648

1775 14 13 13 2006 9 14 12 1648

1776 14 13 13 2006 9 14 18 1648

1777 14 13 13 2006 9 14 18 1648

1778 14 13 13 2006 9 15 0 1648

1779 14 13 13 2006 9 15 0 1648

1780 14 13 13 2006 9 15 6 1648

1781 14 13 13 2006 9 15 6 1648

1782 14 13 13 2006 9 15 12 1648

1783 14 13 13 2006 9 15 12 1648

1784 14 13 13 2006 9 15 18 1648

1785 14 13 13 2006 9 15 18 1648

1786 14 13 13 2006 9 16 0 1648

1787 14 13 13 2006 9 16 0 1648

1788 14 13 13 2006 9 16 6 1648

1789 14 13 13 2006 9 16 6 1648

1790 14 13 13 2006 9 16 12 1648

1791 14 13 13 2006 9 16 12 1648

1792 14 13 13 2006 9 16 18 1648

1793 14 13 13 2006 9 16 18 1648

1794 14 13 13 2006 9 17 0 1648

124

1795 14 13 13 2006 9 17 0 1648

1796 14 13 13 2006 9 17 6 1648

1797 14 13 13 2006 9 17 6 1648

1798 14 13 13 2006 9 17 12 1648

1799 16 14 14 2006 9 18 0 1650

1800 16 14 14 2006 9 18 6 1650

1801 16 14 14 2006 9 18 12 1650

1802 16 14 14 2006 9 18 18 1650

1803 16 14 14 2006 9 19 0 1650

1804 16 14 14 2006 9 19 0 1650

1805 16 14 14 2006 9 19 6 1650

1806 16 14 14 2006 9 19 6 1650

1807 16 14 14 2006 9 19 12 1650

1808 16 14 14 2006 9 19 12 1650

1809 16 14 14 2006 9 19 18 1650

1810 16 14 14 2006 9 19 18 1650

1811 16 14 14 2006 9 20 0 1650

1812 16 14 14 2006 9 20 0 1650

1813 16 14 14 2006 9 20 6 1650

1814 16 14 14 2006 9 20 6 1650

1815 16 14 14 2006 9 20 12 1650

1816 16 14 14 2006 9 20 12 1650

1817 16 14 14 2006 9 20 18 1650

1818 16 14 14 2006 9 20 18 1650

1819 16 14 14 2006 9 21 0 1650

1820 16 14 14 2006 9 21 0 1650

1821 16 14 14 2006 9 21 6 1650

1822 16 14 14 2006 9 21 6 1650

1823 16 14 14 2006 9 21 12 1650

1824 16 14 14 2006 9 21 12 1650

1825 16 14 14 2006 9 21 18 1650

1826 16 14 14 2006 9 21 18 1650

1827 16 14 14 2006 9 22 0 1650

1828 16 14 14 2006 9 22 0 1650

1829 16 14 14 2006 9 22 6 1650

1830 16 14 14 2006 9 22 6 1650

1831 16 14 14 2006 9 22 12 1650

1832 16 14 14 2006 9 22 12 1650

1833 16 14 14 2006 9 22 18 1650

125

1834 16 14 14 2006 9 22 18 1650

1835 16 14 14 2006 9 23 0 1650

1836 16 14 14 2006 9 23 0 1650

1837 16 14 14 2006 9 23 6 1650

1838 16 14 14 2006 9 23 6 1650

1839 16 14 14 2006 9 23 12 1650

1840 16 14 14 2006 9 23 12 1650

1841 16 14 14 2006 9 23 18 1650

1842 16 14 14 2006 9 23 18 1650

1843 16 14 14 2006 9 24 0 1650

1844 16 14 14 2006 9 24 0 1650

1845 16 14 14 2006 9 24 6 1650

1846 16 14 14 2006 9 24 6 1650

1847 16 14 14 2006 9 24 12 1650

1848 16 14 14 2006 9 24 12 1650

1849 16 14 14 2006 9 24 18 1650

1850 16 14 14 2006 9 25 0 1650

1851 18 15 16 2006 9 26 12 1652

1852 18 15 16 2006 9 26 18 1652

1853 18 15 16 2006 9 27 0 1652

1854 18 15 16 2006 9 27 0 1652

1855 18 15 16 2006 9 27 6 1652

1856 18 15 16 2006 9 27 6 1652

1857 18 15 16 2006 9 27 12 1652

1858 18 15 16 2006 9 27 12 1652

1859 18 15 16 2006 9 27 18 1652

1860 18 15 16 2006 9 27 18 1652

1861 18 15 16 2006 9 28 0 1652

1862 18 15 16 2006 9 28 0 1652

1863 18 15 16 2006 9 28 6 1652

1864 18 15 16 2006 9 28 6 1652

1865 18 15 16 2006 9 28 12 1652

1866 18 15 16 2006 9 28 12 1652

1867 18 15 16 2006 9 28 18 1652

1868 18 15 16 2006 9 28 18 1652

1869 18 15 16 2006 9 29 0 1652

1870 18 15 16 2006 9 29 0 1652

1871 18 15 16 2006 9 29 6 1652

1872 18 15 16 2006 9 29 6 1652

126

1873 18 15 16 2006 9 29 12 1652

1874 18 15 16 2006 9 29 12 1652

1875 18 15 16 2006 9 29 18 1652

1876 18 15 16 2006 9 29 18 1652

1877 18 15 16 2006 9 30 0 1652

1878 18 15 16 2006 9 30 0 1652

1879 18 15 16 2006 9 30 6 1652

1880 18 15 16 2006 9 30 6 1652

1881 18 15 16 2006 9 30 12 1652

1882 18 15 16 2006 9 30 12 1652

1883 18 15 16 2006 9 30 18 1652

1884 18 15 16 2006 9 30 18 1652

1885 18 15 16 2006 10 1 0 1652

1886 18 15 16 2006 10 1 0 1652

1887 18 15 16 2006 10 1 6 1652

1888 18 15 16 2006 10 1 6 1652

1889 18 15 16 2006 10 1 12 1652

1890 21 18 19 2006 10 10 12 1655

1891 21 18 19 2006 10 10 18 1655

1892 21 18 19 2006 10 11 0 1655

1893 21 18 19 2006 10 11 6 1655

1894 21 18 19 2006 10 11 12 1655

1895 21 18 19 2006 10 11 18 1655

1896 21 18 19 2006 10 12 0 1655

1897 21 18 19 2006 10 12 6 1655

1898 21 18 19 2006 10 12 12 1655

1899 21 18 19 2006 10 12 18 1655

1900 21 18 19 2006 10 12 18 1655

1901 21 18 19 2006 10 13 0 1655

1902 21 18 19 2006 10 13 0 1655

1903 21 18 19 2006 10 13 6 1655

1904 21 18 19 2006 10 13 6 1655

1905 21 18 19 2006 10 13 12 1655

1906 21 18 19 2006 10 13 12 1655

1907 21 18 19 2006 10 13 18 1655

1908 21 18 19 2006 10 13 18 1655

1909 21 18 19 2006 10 14 0 1655

1910 21 18 19 2006 10 14 0 1655

1911 21 18 19 2006 10 14 6 1655

127

1912 21 18 19 2006 10 14 6 1655

1913 21 18 19 2006 10 14 12 1655

1914 21 18 19 2006 10 14 12 1655

1915 21 18 19 2006 10 14 18 1655

1916 21 18 19 2006 10 14 18 1655

1917 21 18 19 2006 10 15 0 1655

1918 21 18 19 2006 10 15 0 1655

1919 21 18 19 2006 10 15 6 1655

1920 21 18 19 2006 10 15 12 1655

1921 22 19 20 2006 10 27 18 1656

1922 22 19 20 2006 10 28 0 1656

1923 22 19 20 2006 10 28 6 1656

1924 22 19 20 2006 10 28 6 1656

1925 22 19 20 2006 10 28 12 1656

1926 22 19 20 2006 10 28 12 1656

1927 22 19 20 2006 10 28 18 1656

1928 22 19 20 2006 10 28 18 1656

1929 22 19 20 2006 10 29 0 1656

1930 22 19 20 2006 10 29 0 1656

1931 22 19 20 2006 10 29 6 1656

1932 22 19 20 2006 10 29 6 1656

1933 22 19 20 2006 10 29 12 1656

1934 22 19 20 2006 10 29 12 1656

1935 22 19 20 2006 10 29 18 1656

1936 22 19 20 2006 10 29 18 1656

1937 22 19 20 2006 10 30 0 1656

1938 22 19 20 2006 10 30 0 1656

1939 22 19 20 2006 10 30 6 1656

1940 22 19 20 2006 10 30 6 1656

1941 22 19 20 2006 10 30 12 1656

1942 22 19 20 2006 10 30 12 1656

1943 22 19 20 2006 10 30 18 1656

1944 22 19 20 2006 10 30 18 1656

1945 22 19 20 2006 10 31 0 1656

1946 22 19 20 2006 10 31 0 1656

1947 22 19 20 2006 10 31 6 1656

1948 22 19 20 2006 10 31 6 1656

1949 22 19 20 2006 10 31 12 1656

1950 22 19 20 2006 10 31 12 1656

128

1951 22 19 20 2006 10 31 18 1656

1952 22 19 20 2006 10 31 18 1656

1953 22 19 20 2006 11 1 0 1656

1954 22 19 20 2006 11 1 0 1656

1955 22 19 20 2006 11 1 6 1656

1956 22 19 20 2006 11 1 6 1656

1957 22 19 20 2006 11 1 12 1656

1958 22 19 20 2006 11 1 12 1656

1959 22 19 20 2006 11 1 18 1656

1960 22 19 20 2006 11 1 18 1656

1961 22 19 20 2006 11 2 0 1656

1962 22 19 20 2006 11 2 0 1656

1963 22 19 20 2006 11 2 6 1656

1964 22 19 20 2006 11 2 6 1656

1965 22 19 20 2006 11 2 12 1656

1966 22 19 20 2006 11 2 18 1656

1967 22 19 20 2006 11 3 0 1656

1968 23 20 21 2006 11 10 0 1657

1969 23 20 21 2006 11 10 6 1657

1970 23 20 21 2006 11 10 6 1657

1971 23 20 21 2006 11 10 12 1657

1972 23 20 21 2006 11 10 12 1657

1973 23 20 21 2006 11 10 18 1657

1974 23 20 21 2006 11 10 18 1657

1975 23 20 21 2006 11 11 0 1657

1976 23 20 21 2006 11 11 0 1657

1977 23 20 21 2006 11 11 6 1657

1978 23 20 21 2006 11 11 6 1657

1979 23 20 21 2006 11 11 12 1657

1980 23 20 21 2006 11 11 12 1657

1981 23 20 21 2006 11 11 18 1657

1982 23 20 21 2006 11 11 18 1657

1983 23 20 21 2006 11 12 0 1657

1984 23 20 21 2006 11 12 0 1657

1985 23 20 21 2006 11 12 6 1657

1986 23 20 21 2006 11 12 6 1657

1987 23 20 21 2006 11 12 12 1657

1988 23 20 21 2006 11 12 18 1657

1989 23 20 21 2006 11 13 0 1657

129

1990 23 20 21 2006 11 13 6 1657

1991 23 20 21 2006 11 13 12 1657

1992 24 21 22 2006 11 27 18 1658

1993 24 21 22 2006 11 28 0 1658

1994 24 21 22 2006 11 28 6 1658

1995 24 21 22 2006 11 28 12 1658

1996 24 21 22 2006 11 28 18 1658

1997 24 21 22 2006 11 28 18 1658

1998 24 21 22 2006 11 29 0 1658

1999 24 21 22 2006 11 29 0 1658

2000 24 21 22 2006 11 29 6 1658

2001 24 21 22 2006 11 29 6 1658

2002 24 21 22 2006 11 29 12 1658

2003 24 21 22 2006 11 29 12 1658

2004 24 21 22 2006 11 29 18 1658

2005 24 21 22 2006 11 29 18 1658

2006 24 21 22 2006 11 30 0 1658

2007 24 21 22 2006 11 30 0 1658

2008 24 21 22 2006 11 30 6 1658

2009 24 21 22 2006 11 30 6 1658

2010 24 21 22 2006 11 30 12 1658

2011 24 21 22 2006 11 30 12 1658

2012 24 21 22 2006 11 30 18 1658

2013 24 21 22 2006 11 30 18 1658

2014 24 21 22 2006 12 1 0 1658

2015 24 21 22 2006 12 1 0 1658

2016 24 21 22 2006 12 1 6 1658

2017 24 21 22 2006 12 1 6 1658

2018 24 21 22 2006 12 1 12 1658

2019 24 21 22 2006 12 1 12 1658

2020 24 21 22 2006 12 1 18 1658

2021 24 21 22 2006 12 1 18 1658

2022 24 21 22 2006 12 2 0 1658

2023 24 21 22 2006 12 2 0 1658

2024 24 21 22 2006 12 2 6 1658

2025 24 21 22 2006 12 2 6 1658

2026 24 21 22 2006 12 2 12 1658

2027 24 21 22 2006 12 2 12 1658

2028 24 21 22 2006 12 2 18 1658

130

2029 24 21 22 2006 12 2 18 1658

2030 24 21 22 2006 12 3 0 1658

2031 24 21 22 2006 12 3 0 1658

2032 24 21 22 2006 12 3 6 1658

2033 24 21 22 2006 12 3 6 1658

2034 24 21 22 2006 12 3 12 1658

2035 24 21 22 2006 12 3 12 1658

2036 24 21 22 2006 12 3 18 1658

2037 24 21 22 2006 12 3 18 1658

2038 24 21 22 2006 12 4 0 1658

2039 24 21 22 2006 12 4 0 1658

2040 24 21 22 2006 12 4 12 1658

2041 24 21 22 2006 12 4 18 1658

2042 24 21 22 2006 12 5 0 1658

2043 25 22 23 2006 12 8 18 1659

2044 25 22 23 2006 12 9 0 1659

2045 25 22 23 2006 12 9 6 1659

2046 25 22 23 2006 12 9 12 1659

2047 25 22 23 2006 12 9 18 1659

2048 25 22 23 2006 12 9 18 1659

2049 25 22 23 2006 12 10 0 1659

2050 25 22 23 2006 12 10 0 1659

2051 25 22 23 2006 12 10 6 1659

2052 25 22 23 2006 12 10 6 1659

2053 25 22 23 2006 12 10 12 1659

2054 25 22 23 2006 12 10 12 1659

2055 25 22 23 2006 12 10 18 1659

2056 25 22 23 2006 12 10 18 1659

2057 25 22 23 2006 12 11 0 1659

2058 25 22 23 2006 12 11 0 1659

2059 25 22 23 2006 12 11 6 1659

2060 25 22 23 2006 12 11 6 1659

2061 25 22 23 2006 12 11 12 1659

2062 25 22 23 2006 12 11 12 1659

2063 25 22 23 2006 12 11 18 1659

2064 25 22 23 2006 12 11 18 1659

2065 25 22 23 2006 12 12 0 1659

2066 25 22 23 2006 12 12 0 1659

2067 25 22 23 2006 12 12 6 1659

131

2068 25 22 23 2006 12 12 6 1659

2069 25 22 23 2006 12 12 12 1659

2070 25 22 23 2006 12 12 12 1659

2071 25 22 23 2006 12 12 18 1659

2072 25 22 23 2006 12 12 18 1659

2073 25 22 23 2006 12 13 0 1659

2074 25 22 23 2006 12 13 0 1659

2075 25 22 23 2006 12 13 6 1659

2076 25 22 23 2006 12 13 6 1659

2077 25 22 23 2006 12 13 12 1659

2078 25 22 23 2006 12 13 12 1659

2079 25 22 23 2006 12 13 18 1659

132

REFERENCES

Annual Tropical Cyclone Reports (ATCR), 1959-2004, Joint Typhoon Warning Center.

Atkinson, G. D. and C. R. Holliday, 1977: Tropical cyclone minimum sea level

pressure/maximum sustained wind relationship for the western North Pacific. Mon. Wea.

Rev., 105, 421-427.

Bell, G. D. and Coauthors, 2000: Climate Assessment for 1999. Bull. Amer. Meteor. Soc., 81,

1149-1165.

Bell, G. J., 1979: Operational forecasting of tropical cyclones: Past, present, and future. Aust.

Met. Mag., 27, 249-258.

Bister, M., and K. A. Emanuel, 1998: Dissipative heating and hurricane intensity. Meteor. Atm.

Phys., 52, 233-240.

Chan, J. C., 2008: Interannual and Interdecadal Variations of Tropical Cyclone Landfall in East

Asia. Eos Trans. AGU, 89(23), West. Pac. Geophy. Meet. Suppl., Abstract U33A-01.

Chang, E. K. M., and Y. Guo, 2007: Is the number of North Atlantic tropical cyclones

significantly underestimated prior to the availability of satellite observations? Geophys.

Res. Lett., 34, L14801, doi:10.1029/2007GL030169.

Dvorak, V. F., 1972: A technique for the analysis and forecasting of tropical cyclone intensities

from satellite pictures. NOAA Tech. Memo. NESS 36, 15 pp.

——, V. F., 1973: A technique for the analysis and forecasting of tropical cyclone intensities

from satellite pictures. NOAA Tech. Memo. NESS 45, 19 pp.

——, V. F., 1975: Tropical cyclone intensity analysis and forecasting from satellite imagery.

Mon. Wea. Rev., 103, 420–430.

——, V. F., 1982: Tropical cyclone intensity analysis and forecasting from satellite visible or

enhanced infrared imagery. NOAA National Environmental Satellite Service,

Applications Laboratory Training Notes, 42 pp.

——, V. F., 1984: Tropical cyclone intensity analysis using satellite data. NOAA Tech. Rep. 11,

45 pp.

Edson, R. T., 2004: Tropical cyclone analysis techniques from QuikSCAT NRCS, wind and

ambiguity data and microwave imagery. 26th Conf. on Hurricanes and Tropical

Meteorology, Miami, FL, Amer. Meteor. Soc.

Elsner, J. B., J. P. Kossin, and T. H. Jagger, 2008: The increasing intensity of the strongest

tropical cyclones, Nature, 455, 92-95.

133

Emanuel, K. A., 2005: Increasing destructiveness of tropical cyclones over the past 30 years.

Nature, 436, 686-688.

——, K. A., 2006: Climate and tropical cyclone activity: A new model downscaling approach.

J. Climate, 19, 4797-4802.

——, K. A., 2007: Environmental factors affecting tropical cyclone power dissipation. J.

Climate, 20, 5497-5509.

——, K. A., R. Sundararajan, and J. Williams, 2008: Hurricanes and global warming: Results

from downscaling IPCC AR4 simulations. Bull. Amer. Meteor. Soc, 89, 347-367.

Fujita, T. T., 1971: Proposed Characterization of Tornadoes and Hurricanes by Area and

Intensity. SMRP Research Paper No. 91, Department of the Geophy. Sci., The University

of Chicago, 42 pp.

Gilbert, R. O., 1987: Statistical methods for environmental pollution monitoring. Wiley and

Sons, 336 pp.

Gray, W. M., C. Neumann and T. L. Tsui, 1991: Assessment of the role of aircraft

reconnaissance on tropical cyclone analysis and forecasting. Bull. Amer. Meteor. Soc., 72,

1867-1883.

Guard, C. P., L. E. Carr, F. H. Wells, R. A. Jeffries, N. D. Gural, and D. K. Edson, 1992: Joint

Typhoon Warning Center and the challenges of multi-basin tropical cyclone forecasting.

Wea. Forecasting, 7, 328-352.

——, C. P., 2004: The longevity of the Dvorak TC. intensity technique in the Pacific and Indian

Ocean regions. Preprints, 26th

Conf. on Hurricanes and Tropical Meteorology, Miami,

FL, Amer. Meteor. Soc., 210-211.

Hawkins, J. D., T. F. Lee, J. Turk, C. Sampson, J. Kent, and K. Richardson. 2001. Real-time

Internet distribution of satellite products for tropical cyclone reconnaissance. Bull. Am.

Meteorol. Soc., 82, 567-578.

Harper, B. A., J. Kepert and J. Ginger, 2008: Wind speed time averaging conversions for tropical

cyclone conditions. 28th Conf. on Hurricanes and Tropical Meteorology, Orlando, FL,

Amer. Meteor. Soc.

Holland, G. J., 1993: "Ready Reckoner" - Chapter 9, Global Guide to Tropical Cyclone

Forecasting, WMO/TC-No. 560, Report No. TCP-31, World Meteorological

Organization, Geneva.

——, G. J., 1997: The maximum Potential Intensity of Tropical Cyclones., J. Atmos. Sci., 54,

2519-2541.

134

Japanese Meteorological Agency, 1990: Guidelines for Forecast Operations – Typhoon forecast.

Forecast Department, 150pp. (in Japanese)

Jarvinen, B. R., C. J. Neumann, M. A. S. Davis, 1984: "A tropical cyclone data tape for the

North Atlantic basin, 1886-1983: Contents, limitations, and uses", NOAA Technical

Memorandum NWS NHC 22.

Kamahori, H., Yamazaki, N., Mannoji, N., and K.Takahashi, 2006: Variability in Intense

Tropical Cyclone Days in the Western North Pacific. SOLA 2, 104-107.

Kendall, M. G., 1975: Rank Correlation Methods. 4th ed. Charles Griffin and Company, 282 pp.

Klotzbach, P. J., 2006: Trends in global tropical cyclone activity over the past twenty years

(1986–2005). Geophys. Res. Lett., 33, L10805, doi:1029/2006GL025881.

Knaff, J. A. and R. M. Zehr, 2007: Reexamination of tropical cyclone pressure wind

relationships. Wea. Forecasting, 22:1, 71-88.

Knutson, T. K., and R. E. Tuleya, 2004: Impact of CO2-induced warming on simulated hurricane

intensity and precipitation: Sensitivity to the choice of climate model and convective

parameterization. J. Climate, 17(18), 3477-3495.

Koba, H., T. Hagiwara, S. Asano, and S. Akashi, 1990: Relationships between CI number from

Dvorak’s technique and minimum sea level pressure of maximum wind speed of tropical

cyclone (in Japanese). J. Meteor. Res., 42, 59-67.

Kossin, J. P., K. R. Knapp, D. J. Vimont, R. J. Murnane, and B. A. Harper, 2007: A globally

consistent reanalysis of hurricane variability and trends. Geophys. Res. Lett., 34, L04815,

doi:10.1029/2006GL028836.

Krayer, W. R. and R. D. Marshall, 1992: Gust factors applied to hurricane winds. Bull. Amer.

Met. Soc., 73, 613-617.

Lander, M., 2008: A comparison of typhoon best-track data in the western North Pacific:

irreconcilable differences. Extended Abstracts, 28th

Conf. on Hurricanes and Tropical

Meteorology, Orlando, FL, Amer. Meteor. Soc.

Landsea, C. W., 2000: "Climate variability of tropical cyclones: Past, Present and Future" Storms

edited by R. A. Pielke, Sr. and R. A. Pielke, Jr, Routledge, New York, p.220-241

——, C. W., C. Anderson, N. Charles, G. Clark, J. Dunion, J. Fernandez-Partagas, P.

Hungerford, C. Neumann, and M. Zimmer, 2004: The Atlantic hurricane database re-

analysis project: Documentation for the 1851-1910 alterations and additions to the

HURDAT database. Hurricanes and Typhoons: Past, Present and Future, R. J. Murname

and K.-B. Liu, Eds., Columbia University Press, 177-221.

135

——, C. W., 2005: Hurricanes and global warming. Nature, 438, E11-13,

doi:10.1038/nature04477.

——, C. W., 2007: Counting Atlantic Tropical Cyclones back to 1900. EOS, 88, 197 & 2002.

Lee, T. F., J. D. Hawkins, F. J. Turk, K. Richardson, C. Sampson, and J. Kent, 1999: Tropical

cyclone images now can be viewed “live” on the Web. EOS, 80, 612–614.

——, T. F., F. J. Turk, J. Hawkins, and K. Richardson, 2002: Interpretation of TRMM TMI

Images of Tropical Cyclones. Earth Interactions, 6, 1-17.

Mann, H. B., 1945: Non-parametric tests against trend. Econometrica, 13, 245–259.

Martin, J. D. and W. M. Gray, 1993: Tropical cyclone observation and forecasting with and

without aircraft reconnaissance. Wea. Forecasting, 8, 519–532.

Mayfield, M. C. McAdie, and A. Pike, 1988: A preliminary evaluation of the dispersion of

tropical cyclone position and intensity estimates determined from satellite imagery.

NOAA NHC research paper, 16 pp.

Maue, R. and R. Hart, 2007: Comment on: “Low frequency variability in globally integrated

tropical cyclone power dissipation.” Geophys. Res. Lett., 34, L11703,

doi:1029/2006GL028283.

Neumann, C. J., B. R. Jarvinen, C. J. McAdie, and J. D. Elms, 1993: Tropical cyclones of the

North Atlantic Ocean, 1871-1992. NOAA NESDIS, Historical Climatology Series 6-2,

193 pp. (NHC homepage for most recent track information: http://www.nhc.noaa.gov)

Sheets, R. C., 1989: Impacts of aircraft reconnaissance “fixes” on tropical cyclone track

predictions. Minutes, 43rd

Interdepartmental Hurricane Conference, 10 – 13 January

1989, Homestead AFB, (office of the Federal Coordinator for Meteorological Services

and Supporting Research).

Simiu, E. and R. H. Scanlon, 1978: Wind Effects on Structures. Wiley-Interscience, 458 pp.

Sriver, R. and M. Huber, 2006: Low frequency variability in globally integrated tropical cyclone

power dissipation. Geophys. Res. Lett., 33, L11705, doi: 10.1029/2006GL026167

Trenberth, K. E., 2005: Uncertainty in Hurricanes and Global Warming. Science, 308, 1753-

1754.

Vecchi, G. A. and B. J. Soden, 2007. Increased tropical Atlantic wind shear in model projections

of global warming. Geophys. Res. Lett., 34, L08702, doi:10.1029/2006GL028905.

136

Velden, C., B. Harper, F. Wells, J. L. Beven, R. Zehr, T. Olander, M. Mayfield, C. P. Guard, M.

Lander, R. Edson, L. Avila, A. Burton, M. Turk, A. Kikuchi, A. Christian, P. Caroff, and

P. McCrone, 2006: The Dvorak Tropical Cyclone Intensity Estimation Technique: A

Satellite-Based Method that Has Endured for over 30 Years. Bull. Am. Meteorol. Soc, 87,

1195–1210.

Webster, P. J., G. J. Holland, J. A. Curry, and H. R. Chang, 2005: Changes in tropical cyclone

number, duration, and intensity in a warming environment. Science, 309, 1844-1846.

Wu, M-. C., K-. H. Yeung, and W-. L. Chang, 2006: Trends in Western North Pacific Tropical

Cyclone Intensity. EOS, 87, 537-538.

137

BIOGRAPHICAL SKETCH

Michael Robert Lowry was born and raised in Destrehan, Louisiana, a small town west of

New Orleans along the banks of the Mississippi River. As a young boy, weather events were the

highlights of his years, from the 1989 December snow to the mass evacuation ahead of Hurricane

Andrew in 1992. His father a lab technician and his mother a school teacher, science and

education were instilled in him from a very young age. Upon graduation from Destrehan High

School as class salutatorian in 2001, Michael attended the Florida State University in

Tallahassee, Florida, from 2001 to 2005. There he graduated summa cum laude with honors,

obtaining a Bachelor of Science degree in meteorology, with minors in mathematics and physics,

in April 2005. As an undergraduate he was given the opportunity to intern under the State

Meteorologist of Florida at the Florida Division of Emergency Management, where he worked

part time from June of 2004 through his time in graduate school. Michael began his graduate

studies in meteorology in 2005 at the Center for Ocean-Atmospheric Prediction Studies

(COAPS) at the Florida State University under former State Climatologist of Florida and Robert

O. Lawton Distinguished Professor of Meteorology and Oceanography Dr. James J. O’Brien.

Michael was offered a full-time position as a meteorologist with Northrop Grumman Corporation

two years into his Masters of Science degree. He relocated to Alexandria, Virginia, in March of

2008, to begin his career while finishing his MS thesis remotely.