Impacts of Climate Variability and Change on Eastern North Pacific Tropical Cyclones

Jerry Yu Jien

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Physical and Environmental Sciences University of Toronto Scarborough

Jerry Yu Jien

Doctor of Philosophy

Department of Physical and Environmental Sciences University of Toronto Scarborough

2015

Abstract

Damages inflicted by tropical cyclones (TCs) worldwide have increased in recent decades with climate change and variability playing key roles in altering TC characteristics. In this thesis, the impact of natural variability is explored, using ENSO conditions, and climate change on the nature of eastern North Pacific (ENP) TCs. The first research objective of the thesis focused on a spatial-temporal separation of ENP storms based on El Niño-Southern Oscillation (ENSO) phase (temporal variability, El

Niño, La Niña, neutral) and regional storm stratification (spatial variability, east and west). The western development region (WDR) storms were found to be more sensitive to influences of ENSO. In particular, during El Niño years, there were more WDR storms. The second research objective explored the ENSO impact on the trajectory of

ENP storm tracks by examining the locations for genesis and downgradation points and storm track movements. The storm tracks were strongly influenced by ENSO phases, with significant differences detected for many ENSO pairings. However, when storm data are regionally separated the latitudinal movement of WDR storms tend to be more extensive during El Niño conditions and as a result there are more landfalling TCs. The third research objective of the dissertation explored the importance of near-time sea

ii surface temperatures (SSTs) on storm intensities. SST thresholds were found that were critical for the sustenance of stronger ENP storms that achieve hurricane and major hurricane status. Significantly, the minimum SST threshold varied between the MDR subdivisions. For major hurricanes, the SST requirements for EDR and WDR are substantially lower than that found in the North Atlantic basin at 28.25°C. Although

SSTs appear to contribute little in determining the ultimate maximum storm intensity for

ENP storms in general, when ENP storms are regionally divided, SSTs are found to be highly associated with the WDR major hurricanes. Evidently the recent warming shown in the distribution of storm-bounded SSTs has led to the rise of maximum potential intensity for ENP storms. Overall, the common theme that emerged from these three studies is that ENP storm characteristics associated with WDR are inherently more sensitive to climate variability and change.

iii Acknowledgements

The completion of my thesis would not have been possible without the scientific expertise and moral support from faculty, staff and friends/students within and outside of the University of Toronto Scarborough (UTSC) community.

I would like to express my foremost gratitude to my thesis supervisor, Dr. William

Gough, who is fully committed to the well-being of his students. I would like to thank members of my PhD Examination Committee: Dr. George Arhonditsis, Dr. Mathew

Wells, Dr. Ken Butler and Dr. Monirul Mirza for their feedback which improved my thesis. In addition, to my internal committee members, the completion of my thesis would not have been possible without the thorough review provided by Dr. Jennifer

Collins from the University of South Florida. I am particularly thankful to Dr. Adam

Fenech for his encouragement and making my graduate study a fulfilling experience. In addition, I would like to thank Dr. Carl Mitchell and Dr. Tanzina Mohsin whom I had the privilege to work for many years as their teaching assistants. For some of the figures that require GIS, I would like to recognize Dr. Mike Doughty and Dr. Vincent Cheng for their generous support.

Within the Department of Physical and Environmental Science, Office of the Dean and

Vice-Principal and Department of International Academic Programs and Initiatives at

UTSC, we have a great team of wonderfully cheerful administrative staff who are deeply committed to student experience. For that, I thank you all dearly and encourage you to keep up this amazing work!

Outside the UTSC community, I would also like to express my gratitude to colleagues associated with Agriculture and Agri-Food Canada in Ottawa for their tremendous support during the latter end of my graduate student career through their mentorships and funding on relevant research projects.

It has been a great experience knowing and working with each and every member of the

UTSC Climate Lab. Finally, friends and families I cannot thank you enough for propelling me forward towards the completion of my PhD program.

v Table of Contents

Contents Acknowledgements……………………………………………………………………… iv

Table of Contents ……………………………………………………………………….. vi

List of Tables ………………………………………………………………………….…ix

List of Figures …………………………………………………………………………..xiii

Chapter 1: Introduction ···························································· 1

1.1 Background ·································································· 1 1.2 Tropical cyclone activity (frequency, intensity and duration) ······· 4 1.3 Impact of El-Niño Southern Oscillation on tropical cyclones ········ 6 1.4 Sea surface temperature effect on tropical cyclone intensity ······· 10 1.5 Research Gap ····························································· 12 1.6 Study Area ································································· 14 1.7 Objectives ································································· 18 1.7.1 Chapter 2 ··········································································· 21 1.7.2 Chapter 3 ··········································································· 22 1.7.3 Chapter 4 ··········································································· 23 1.8 References: ································································ 25

Chapter 2: The Influence of El Niño-Southern Oscillation on Tropical Cyclone Activity in the Eastern North Pacific Basin ······················· 32

2.1 Abstract ···································································· 32 2.2 Introduction ······························································· 32 Deleted: 33 2.3 Methods ···································································· 36 2.3.1 Classification of ENSO events ·················································· 36 2.3.2 TC activity and TC intensity ···················································· 37 2.3.3 Statistical analyses ································································ 38 2.4 Results ····································································· 41 2.4.1 Classification of ENSO events ·················································· 41

vi 2.4.2 Time series of storm activity and intensity ···································· 43 2.4.3 Statistical analyses ································································ 50 2.4.3.a Two-way analysis of variance ··············································· 50 2.4.3.b Correlation analysis ··························································· 54 2.5 Discussion ································································· 63 2.6 Conclusion ································································· 69 2.7 References ································································· 71

Chapter 3: The Impact of El Niño-Southern Oscillation on Eastern North Pacific Tropical Cyclone Tracks ······································· 75

3.1 Abstract ···································································· 75 3.2 Introduction ······························································· 76 3.3 Data ········································································· 80 3.4 Methods ···································································· 80 3.5 Results ····································································· 83 3.5.1 Impacts of El Niño-Southern Oscillation and regional divisions ··········· 84 3.5.1.a El Niño-Southern Oscillation differences ·································· 84 3.5.1.b Regional differences ··························································· 96 3.5.1.c Combined El Niño-Southern Oscillation and regional differences ··· 103 3.5.2 Primary Genesis Region ······················································· 107 3.6 Storm Impact ····························································· 115 3.7 Summary and Conclusion ·············································· 118 3.8 References: ······························································· 122

Chapter 4: Near-time Sea Surface Temperature and Tropical Cyclone Intensity in the Eastern North Pacific basin ································ 127

4.1 Abstract ··································································· 127 4.2 Introduction ······························································ 128 4.3 Data and Methods ······················································· 132 4.4 Longitudinal Division of ENP basin ·································· 136 4.5 Displacements of maximum TC intensity and maximum SST and initial genesis point ····················································· 140 4.6 Correlation between SST and TC intensity ·························· 147

vii 4.6.1 All Observations ································································ 147 4.6.2 MDR Subdivisions ····························································· 150 4.7 Upper bound of TC Intensity by SST Groups ······················· 159 4.8 Conclusion ································································ 169 4.9 References ································································ 173

Chapter 5: Conclusion ··························································· 176

5.1 Summary and Discussion: ·············································· 176 5.2 Future Research ·························································· 182 5.3 References ································································ 184

Appendix I – Conditions for tropical cyclone development ················· 186

Appendix II – Saffir-Simpson classification of tropical cyclone intensity · 187

viii List of Tables Table 1-1: Strength of correlations between six measures of TC activity and SSTs in MDR, EDR and WDR averaged over 1971-2012 from May to November...20

Table 2-1: Years from 1971-2012 are classified by ENSO phases based on the averages of the eight sliding bimonthly MEI values* from April-May to November- December ……….………………………………………………………..…42

Table 2-2: TC activity in the eastern North Pacific basin in the main (eastern and western) development regions from 1971-2012; while seasonal averages from 1981-2010 are incorporated into the derivation of NTC ……………..…..…46

Table 2-3: Mean seasonal (1971-2012) values of six NTC measures subdivided into EDR and WDR during three ENSO conditions ………………..…………………50

Table 2-4: A two-way analysis of variance is tested for effects of ENSO and regional division on the NTC activity index (%) ……………………………………...51

Table 2-5: A two-way analysis of variance is tested for effects of ENSO and regional division on the PDI (scale of 1.0 × 107 m3 s-3) ……………………………...52

Table 2-6: Correlations of NTC and PDI with MEI are determined as the Pearson coefficient of correlation (r) and the correlation of determination (r2) for Figures 2-8 and 2-9. Bolded values are statistically significant at p< 0.05 ...……………………………………………………………………………..62

Table 2-7: Correlations of PDI1, PDI2 and PDI3 with bimonthly (May-November) MEI values in MDR, EDR and WDR from 1971-2012. Bolded values are statistically significant at p<0.05.…..……………………………………….67

ix Table 3-1: A two-way analysis of variance is tested for effects of ENSO and regional divisions on latitudes of storm genesis locations …………………..……….88

Table 3-2: A two-way analysis of variance is tested for effects of ENSO and regional divisions on longitudes of storm genesis locations ..…………………………89

Table 3-3: A two-way analysis of variance is tested for effects of ENSO and regional divisions on latitudes of storm downgradation locations .…………………....90

Table 3-4: A two-way analysis of variance is tested for effects of ENSO and regional divisions on longitudes of storm downgradation locations .…...……………..91

Table 3-5: A two-way analysis of variance is tested for effects of ENSO and regional divisions on latitudinal movement of storm tracks ..………….……………...93

Table 3-6: A two-way analysis of variance is tested for effects of ENSO and regional divisions on the longitudinal movement of storm tracks …………….……....94

Table 3-7: A two-way analysis of variance is tested for effects of ENSO and regional divisions on total storm duration (days) ...……..…………………………..…95

Table 3-8: A two-way analysis of variance is tested for effects of ENSO and regional divisions on maximum storm intensity ...…………………...………………97

Table 3-9: A two-way analysis of variance is tested for effects of ENSO and regional divisions on latitudes of the locations of maximum storm strength ……...... 98

Table 3-10: A two-way analysis of variance is tested for effects of ENSO and regional divisions on longitudes of the locations of maximum storm strength ...…....99

x Table 3-11: A two-way analysis of variance is tested for effects of ENSO and regional divisions on total storm duration (days) between locations of storm genesis and maximum storm strength ...……………………………………………100

Table 3-12: A two-way analysis of variance is tested for effects of ENSO and regional divisions on the longitudinal shift between locations of storm genesis and maximum storm strength ...……...………………………………………...109

Table 3-13: Comparison of the group percentage of EDR and WDR storms in crossing the northern-most boundary. To reject the null hypothesis of similarity between three ENSO phases at p < 0.05, the χ2 value must exceeds 5.99………………………………………………………………..111

Table 4-1: Annual Number of storm count (TS, H and MH) stratified between EDR and WDR averaged from 1982-2013 based on the location where the peak intensity is established. Total storm count is also sorted according to where storms were initially detected ...……………………………………….…..136

Table 4-2: Regression analysis of storm intensity of all EDR and WDR observations with daily SST at seven days prior to storm passage with Day 0 being the arrival day of storm passage ...……………………..……………………………..151

Table 4-3: Summary of TC intensities stratified into SST bins ……………...………..160

Table 4-4: Details on storms that attained the highest relative winds at each SST bin..162

Table 4-5: Summary of EDR storm intensities stratified into SST bins …………...….165

Table 4-6: Summary of WDR storm intensities stratified into SST bins ……...………165

xi Table A-1: Classification of TC strength for storms that originated in the eastern North Pacific and North Atlantic basins. TCs greater than 33m/s are categorized as hurricanes on the Saffir-Simpson Scale. ...……………………..………….187

xii List of Figures

Figure 1-1 Global distribution of tropical cyclones at seven major development basins. The numbers provided are the annual averages of tropical cyclones observed at each basin. Image provided by USA TODAY from its website at http://usatoday30.usatoday.com/weather/hurricane/tropical-cyclone- basins.htm, modified from Williams (2009)..……………………..………….3

Figure 1-2 a) Sea surface temperature anomalies between strong El Niño years (1972, 1982, 1987, 1993 and 1997) and strong La Niña years (1971, 1973, 1975, 1988 and 2010) and b) same as in a) but for surface pressure differences. Image provided by the NOAA-ESRL Physical Sciences Division, Boulder Colorado from their Web site at http://www.esrl.noaa.gov/psd/; data were described in Kalnay et al. (1996)..……………………..……………………..8

Figure 1-3 Monthly TC frequency and TC duration averaged over 1971-2012 from May to November in the main development region (MDR)……………………...16

Figure 1-4 Monthly sea surface temperature averaged over 1971-2012 at the main development region (MDR) and its subdivisions of EDR and WDR……………………………………………………………………...…16

Figure 1-5 same as in Figure 1-3 but for the western development region (WDR).……19

Figure 1-6 Same as in Figure 1.2 but for the eastern development region (EDR)………19

Figure 2-1 The main development region is longitudinally divided into eastern (EDR) and western (WDR) development regions for eastern North Pacific tropical cyclones. Dots represent areas of major urban centers ………………...…..39

Figure 2-2 Time series for the seasonal a) frequency (TS, H and MH) and b) duration (TSD, HD and MHD) at EDR from 1971-2012 …………………………....44

xiii

Figure 2-3 Time series for the seasonal a) frequency (TS, H and MH) and b) duration (TSD, HD and MHD) at WDR from 1971-2012 …………………………...45

Figure 2-4 Time series charts for NTC at a) MDR with horizontal dashed lines showing values at the 25th and 75th percentiles and when TC development region is subdivided into b) EDR/WDR during 1971-2012. Dashed vertical lines are colored to distinguish El Niño (red), La Niña (blue) and neutral (grey) years, in consistent with a) ………………………………………………………...47

Figure 2-6 a) The net tropical cyclone (NTC) activity index and b) power dissipation index (PDI; scale of 1.0 × 107m3 s-3) under ten El Niño, La Niña and neutral years are summarized as boxplots at regions of main (MDR), western (WDR) and eastern (EDR) development regions. Green diamonds represent NTC and PDI averages at each ENSO condition. Results of statistical evaluations of NTC and PDI differences are in Table 2-4 …………………………...…….55

Figure 2-7 Quantile-quantile (QQ) plots for a) NTC and b) PDI (scale of 1.0 × 107.m3 s-3) values at WDR have lines of normality passing through the first and third quartiles. Data are transformed and presented in c) and d). The Shapiro-Wilk test statistic (W), with its p-value, is provided before and after data transformations …………………………………………………………..…57

xiv Figure 2-8 Linear regressions of MEI values with data transformed for NTC at a) MDR, b) EDR and (c) WDR under three ENSO conditions during 1971-2012. Pearson coefficient of correlation (r) and correlation of variation (r2) are provided in Table 4 …………………………………………………………60

Figure 2-9 Linear regressions of MEI values with data transformed for PDI (1.0 × 107 m3 s-3) at a) MDR, b) EDR and c) WDR under three ENSO conditions during 1971-2012. Pearson coefficient of correlation (r) and correlation of variation (r2) are provided in Table 4 …………………………………………………61

Figure 2-10 Time series of residuals (observed-predicted) NTC values and correlograms (autocorrelation plots) are evaluated with Ljung-Box test statistic (chi- squared value), with its p-value, based on the first 20 lags at WDR for a) NTC and b) PDI (scale of 1.0 × 107 m3 s-3) during 1971-2012 ………….. 63

Figure 2-11 Boxplot summaries for data distributions of a) PDI1, b) PDI2 and c) PDI3 (scale of 1.0 × 107 m3 s-3) at MDR, EDR and WDR. Grey diamonds represent the regional averages ……………………………………………66

Figure 3-1 Bimonthly MEI averages from April-May to November-December with the ten highest (lowest) marked as El Niño (La Niña) years ………………...…81

Figure 3-2 Map of EDR and WDR boundary divisions of the eastern North Pacific basin with surrounding North American major cities …………………………….83

Figure 3-3 Distributions of a) latitudes and b) longitudes of genesis locations for all ENP storms. Green diamonds represent averages within each ENSO phase ……85

Figure 3-4 Distributions of a) latitudes and b) longitudes of downgradation locations for all ENP storms. Green diamonds represent averages within each ENSO phase ……………………………………………………………………..…86

Figure 3-5 Distributions of a) northward and b) westward storm track movements all ENP storms. Green diamonds represent averages within each ENSO phase ……………………………………..………………………………....92

Figure 3-6 Distributions of latitudes and longitudes of ENP storm downgradation locations based on regional division. Hollow diamonds represent the regional averages ……...…………………………..………………………………..101

Figure 3-7 Distributions of latitudes and longitudes of ENP storm genesis locations based on regional division. Hollow diamonds represent the regional averages ……………………………………..…………………………….101

Figure 3-8 Distributions of latitudinal and longitudinal movements of ENP storm genesis locations based on regional division. Hollow diamonds represent the regional averages ………………………………..…………………………………..102

Figure 3-9 Comparison of the total and annual/seasonal average storm frequency in MDR subdivisions during all three ENSO phases ………………………...104

Figure 3-10 Distributions of latitudes and longitudes for EDR and WDR storm genesis locations within each ENSO phase. Green diamonds represent individual averages ………………………………..………………………………...105

Figure 3-11 Distributions of latitudes and longitudes for EDR and WDR storm downgradation locations within each ENSO phase. Green diamonds represent individual averages ………..…………………………………..106

Figure 3-12 Distributions of latitudinal and longitudinal movements for EDR and WDR storm downgradation locations within each ENSO phase. Green diamonds represent individual averages ………..…………………………………..108

xvi Figure 3-13 a) EDR and b) WDR storm tracks that had crossed 20°N and made landfalls ………..…………………………………………………...112

Figure 3-14 Locations where maximum relative winds are achieved for EDR and WDR storms under three ENSO phases ………………………………………..114

Figure 3-15 Storm locations at the a) genesis, and b) downgradation stages under three ENSO phases …..………………………………………………………...116

Figure 3-16 Storm locations at the a) genesis, and b) downgradation locations for EDR and WDR storms ………………………………………………………...117

Figure 3-17 Locations of EDR and WDR storms at the a) genesis, and b) downgradation under three ENSO phases …..……………………………………………119

Figure 4-1 Annual proportions of WDR storms to all ENP storms from 1982-2013 …138

Figure 4-2 Annual proportions of EDR-derived WDR storms to the total WDR storm count from 1982-2013 …..…………………………………………………139

Figure 4-3 a) Distribution of latitudinal differences between the locations of storm genesis and maximum lifetime intensity for all ENP storms from 1982-2013 and b) same as a), but for longitudinal differences …………………….….142

Figure 4-4 a) Distribution of latitudinal differences of locations between storm genesis and maximum lifetime intensity for EDR and WDR storms from 1982-2013 and b) same as a), but for longitudinal differences ………………………..143

Figure 4-5 a) Distribution of latitudinal differences between the locations of maximum SST and maximum lifetime intensity for all ENP storms from 1982-2013 and b) same as a), but for longitudinal differences …………………………….144

xvii Figure 4-6 a) Distribution of latitudinal differences between locations of maximum SST and maximum lifetime intensity for EDR and WDR storms from 1982-2013 and b) same as a), but for longitudinal differences ……………………..…146

Figure 4-7 Scatterplot and regression fit for all observations of TC intensity against the weekly SSTs averaged over seven days prior to storm passage ………...... 148

Figure 4-8 Daily SST response (with standard error) to the influence of storm passage over a 15-day time series ………………………………………………….149

Figure 4-9 Same as in Figure 4-8, but for storm genesis and maximum lifetime intensity. Dashed lines represent the daily range of the highest rate of SST reduction .….………………………………………………………………151

Figure 4-10 a) Relative wind speeds for all EDR storm observations and b) same as a) but for all WDR storm observations …………………………………..…152

Figure 4-11 Same as Figure 4-8, but separated for storm passages of EDR and WDR observations. Dashed lines represent the reference SSTs of local minima over for EDR and WDR …………………………………………………154

Figure 4-12 a) Same as Figure 4-11, but for the average SST conditions during storm genesis and b) same as a) but during the maximum lifetime intensity…...155

Figure 4-13 a) Scatterplot and linear regression of maximum lifetime intensity of relative velocity for EDR storms (N=213) and SST a week prior to passage of storm center and b) same as a) but for WDR storms (N=283). The dashed lines with cooler (warmer) SST indicates the thresholds for hurricane (major hurricane) genesis ………………………………………………………..156

xviii Figure 4-14 Scatterplot and linear regression of maximum lifetime intensity of relative velocity for WDR storms that reached the strength as major hurricanes with SST at the locations where storm passages occur three days ago ……….158

Figure 4-15 Anomaly of TC observations in each SST bin during 1982-2013 compared to 1963-1993 …………………………………………………………..…161

Figure 4-16 Storm relative winds at the maximum intensity and 99th, 95th, 90th and 50th intensity percentiles at all SST bins after translational speeds have been accounted ………………………………………………………………...163

Figure 4-17 Scatterplot and regression line drawn for the maximum storm intensity attained by each SST bin with and without the maximum intensity at the 31°C bin ………………………………………………………………….163

Figure 4-18 a) Same as Figure 4-16., but only for EDR storm observations and b) Same as Figure 4-16., but only for WDR storm observations .…………………166

Figure 4-19 Scatterplot for maximum storm intensity attained by each SST bin of all EDR observations. A locally weighted scatterplot smoothing (LOWESS) is applied …………………………………………………………………...168

Figure 4-20 Non-linear least-squares (NLS) fitting of maximum storm intensities bounded by SST bins at WDR, with the LOWESS curve added. …….…170

xix 1 Chapter 1: Introduction

1.1 Background

The recent fifth assessment report (AR5) of Intergovernmental Panel on Climate

Change (IPCC, 2013) states the combined land and ocean temperature globally has increased 0.89°C since the beginning of the 20th century. Evidence of a warmer environment such as the melting of ice sheets and mountain glacier (Mitrovica et al.,

2001; Pritchard et al., 2009; Chen et al., 2013) and the thermal expansion of seawater

(Cazenave and Nerem, 2004; Solomon et al., 209) all contribute to the global sea level rise, threatening small islands and low-lying coastal areas. Many of these changes due to a warmer climate could have long-lasting effects on the availability of water resources even if anthropogenic contributions of radiative forcing from greenhouse gases are immediately eradicated (Solomon et al., 2009). The general consensus within the scientific community is that a warmer climate will increase the frequency and severity of weather extremes. Overall, a warmer environment generally elevates the atmosphere’s water-holding capacity, translating to an accelerated hydrological cycle (Trenberth, 1998;

Held and Soden, 2006).

In consequence of the increased atmospheric moisture due to the observed warming, the number of hydro-meteorological disasters has been on the rise (Auld,

2008). While these extreme weather events are becoming the new norms, many have significantly intensified (Meehl et al., 2000; IPCC, 2013), placing more inland populations at risk and significantly affecting those who reside in coastal regions. Since

1 all storm-related impacts are experienced on private and public assets, socioeconomic changes and reforms are often triggered in the aftermath of extreme events. For example, wide-spreading US storm damages in the 1990s had led to the adjustments of insurance policy following escalations of insurance payouts (Changnon et al., 1997). Among the most dangerous and costliest type of storms (Emanuel, 2001; Sriver and Huber, 2007), tropical cyclones (TC) characteristics are expected to shift in accordance to changes in climate and resulting weather extremes (Meehl et al., 2000; Cai et al., 2014).

Compared to other natural disasters and weather extremes, our improved understanding of TC interactions with the surrounding environment has led to more accurate forecasts of TC tracks and characteristics. TCs are distinguished from other storm systems through their persistence over tropical oceans where non-frontal, low- pressure systems develop. This warm core system is self-sustained through the enthalpy heat exchange over oceans, while maintaining converging surface winds and diverging winds aloft. TC origins are distributed in seven primary development basins over tropical oceans (Figure 1-1). However, the zone of TC influence expands beyond the tropics as it often extends westward along its dominant storm path and progresses poleward or recurves eastward when entering the mid-latitude region. Although until 2004, no TCs have formed in either the eastern South Pacific nor South Atlantic basins due to their climatologically high vertical wind shear (> 10 m s-1) and low sea surface temperature

(SST; < 26.5°C), Catarina (2004) was the first TC that reached hurricane strength in the

South Atlantic Ocean (Pezza and Simmonds, 2005).

Though the number of TC-related deaths in the U.S. has decreased over the years

(Shultz et al., 2005), insured losses due to TC-inflicted damages have increased and been projected to rise (Mendelsohn et al., 2012). Previously, the recent upsurge in TC damage has been accounted for within the context of natural variability. In the U.S. where North

Atlantic TCs are considered to be seasonal threats along the eastern seaboard and the

Gulf coast, TC damages have been adjusted to account for inflation, changes of coastal population and wealth distributions (Pielke and Landsea, 1998). TC damage is a strong indicator of the interannual fluctuation of El-Niño Southern Oscillation (ENSO).

Overall, TC damage is statistically greater during La Niña conditions than during El Niño years (Pielke and Landsea, 1999). However, in the case of Hurricane Andrew (1992), such a linkage did not hold. For example, even though only six North Atlantic storms

3 formed during a strong El-Niño year in 1992, Hurricane Andrew brought the costliest damage until Katrina in 2005. As rising potential storm damage is highly likely in a warming environment (Emanuel, 2005), how the recent climatic warming impacts TC intensity has become an important topic to investigate (Webster et al., 2005).

1.2 Tropical cyclone activity (frequency, intensity and duration)

Though global warming exerts a direct impact on conditions (Appendix I) for TC formation and development, how warming regulates seasonal TC frequency remains uncertain. Though the total TC count varies regionally on a year-to-year basis, there remains an absence of a climatological trend on both a global and regional level

(Henderson-Sellers et al., 1998; Webster et al., 2005; Frank and Young, 2007). Such seasonal TC variability is often linked to ENSO phenomenon that is characterized by shifts in convective processes in the equatorial Pacific Ocean. While TC observation shows the annual TC frequency has not changed in recent decades, the annual TC number observed globally is largely maintained between 80-90 events (Frank and Young, 2007).

Even results from the model projected TC frequency due to future greenhouse gas input remain inconclusive (Henderson-Sellers et al., 1998; Bengtsson et al., 2007).

Evidence has already shown TC intensity (as measured by maximum surface wind speed) has increased noticeably since the 1970s when satellite imagery began monitoring changes in storm strength (Dvorak, 1975). Associated SST warming provides one environmental factor that has been positively correlated with TC intensity (Emanuel,

2005; Steenhof and Gough, 2008; Ralph and Gough, 2009) and storms equivalent to category 4 and 5 hurricanes (Webster et al., 2005). In fact, the poleward migration of

4 maximum TC intensity is observed globally and hypothesized to be linked to the latitudinal expansion of the tropical Hadley circulation due to environmental warming

(Kossin et al., 2014). Should the trend of rising TC intensity continue, some have suggested revisiting the current classification system (Appendix II) for TC intensity to better reflect its increased numbers and realistically anticipate potential hurricane impacts.

Although some TCs of long duration were observed in recent years, there has been much debate whether climate change will impact future storm duration. Even under the likelihood of an expanded tropical region and a greater TC poleward movement

(Kossin et al., 2014), climate models do not respond with a global increase in TC duration (Knutson et al., 2010). However, a global rise in storms of Saffir-Simpson category 4 or 5 hurricanes (Webster et al., 2005) may have extended TC duration for some of the strongest storms. In the North Atlantic basin where TC data are most reliable and extensive, the increase in the number of short-duration (< 2 days) TCs appears to contribute to the rise in TC frequency (Knutson et al., 2010). However, it is uncertain if changes in TC observation practices may have led to step-wise increases in the climatological record. Based on an effort to reconstruct past time series of TC count, the trend in short term duration TCs is balanced by insignificant increases of the medium- and long-lived TCs (Landsea et al., 2010).

When TC parameters, frequency, intensity and duration, are collectively integrated into a single TC index, a particular storm season can be characterized in terms of its net tropical cyclone activity (NTC; Gray et al., 1994) and its power dissipation index (PDI; Emanuel, 2005), with the latter being similar to the definition of accumulated

5 cyclone energy (Bell et al., 2000). These two metrics are more holistic ways to quantify seasonal differences of TC activity (NTC) and TC intensity (PDI) and their relationships to environmental influences. Although much of the research effort has focused on the

North Atlantic storms, they only contribute less than 15% of the annual global total. The

Eastern North Pacific (ENP) in contrast is the most active TC development basin. Yet, impacts of a changing environment and natural variability on ENP storm climatology have received relatively little attention.

1.3 Impact of El-Niño Southern Oscillation on tropical cyclones

Low-frequency atmosphere-ocean variations such as El-Niño Southern

Oscillation (ENSO) have a profound impact on regional variations of weather patterns

(Timmerman et al., 1999; Collins et al., 2010). Extreme weather events such as TCs are largely controlled by the regional variability of equatorial Pacific SSTs driven by the variability of ENSO events. As part of this natural cycle, El Niño events occur when abnormally warm SSTs are detected in the east-central to eastern region of equatorial

Pacific Ocean. During El Niño events, westerly transport of surface water in the equatorial Pacific Ocean is often coupled with weak upwelling of deep cold waters near the shores of Peru and Ecuador (Schwing et al., 2002; Amador et al., 2006). In contrast,

La Niña episodes often induce a shallower thermocline in the ENP basin resulting in colder SST anomalies (Schwing et al., 2002; Amador et al., 2006). Depending on the spatial distribution of SST variability at the equatorial Pacific Ocean, the Ocean Niño

Index considers El Niño events when five consecutive three month-running mean of SST anomaly exceeds 5°C at the Niño 3.4 region (170°W - 120°W and 5°S - 5°N; Hurrell and

6 Trenberth, 1999). Alternatively, additional ocean-atmospheric components can be incorporated when ENSO events are defined through the Multivariate ENSO Index (MEI;

Wolter and Timlin, 2011).

As ENSO signatures of warmer SST and lower surface pressure in the eastern equatorial Pacific Ocean during strong El Niño years are mostly confined to the southern hemisphere (Figure 1-2a and b), south Pacific TCs are affected through ENSO-induced migration of the south Pacific convergence zone (SPCZ) generating a favourable environment for convective activity (Cai et al., 2012). This SPCZ movement off the

Australian coast could explain the seasonal distribution of intense TC occurrences generated and landfall as a result of an ENSO phase shift (Dowdy et al., 2012). Other regionally defined methods of ENSO detection, such as the Southern Oscillation Index that rely on the monthly difference of atmospheric pressure difference between Tahiti and

Darwin, are also linked to differences in the seasonal fluctuation of TC occurrences

(Nicholls, 1979; Dowdy et al., 2012) and TC days (Nicholls, 1985). At the northwestern

Australian basin, though the impact of ENSO on local TCs is not strong (Goebbert and

Leslie, 2010), rainfall events associated with the paleo-climate record of TC landfalls are strongly affected by ENSO events (Denniston et al., 2015).

Although ENSO development peaks during the months of the boreal autumn and winter seasons, the teleconnection behaviour of ENSO still plays an important role on

TCs in the northern hemisphere (Ashok et al., 2007). ENSO-induced effects are predominantly detected in the North Atlantic and Northwestern Pacific basins characterized by seesaw patterns of TC activity on an interannual temporal scale consistent with ENSO variations. A particularly important subset of TC activity are those

Figure 1-2 a) Sea surface temperature anomalies between strong El Niño years (1972, 1982, 1987, 1993 and 1997) and strong La Niña years (1971, 1973, 1975, 1988 and 2010) and b) same as in a) but for surface pressure differences. Image is provided by the NOAA-ESRL Physical Sciences Division, Boulder Colorado from their Web site at http://www.esrl.noaa.gov/psd/; data from Kalnay et al. (1996).

8 that landfall. This has been explored for typhoons (Chan, 1985; Wang and Chan, 2002) and hurricanes (Saunders et al., 2000). Variations in the storm genesis location resulting from the changing ENSO phases also affect the length of the storm development stage during which storms are most sensitive to environmental conditions for intensification

(Pan 1982; Chan 1985; Chan, 2000; Wang and Chan 2002; Chia and Ropelewski 2002).

Such a constraint regulates the period a storm is sustained prior to encountering either a frictional surface upon landfall or colder mid-latitude waters (Wang and Chan, 2002).

Camargo and Sobel (2005) have suggested that the longer a storm experiences conducive conditions for storm development, the more likely it is capable of reaching maximum potential intensity.

While recent changes in climate have not lead to a detectable increase in the number of TCs, natural variability imposes a profound impact on the regional distribution of TC occurrence and its eventual track. At the sub-basin level, both longitudinal and latitudinal locations of TC genesis are observed to oscillate with ENSO events. For both the western Pacific basin (Chan, 1985; Wu and Lau, 1992; Wang and Chan, 2002) and the Australian basin (Basher and Zheng, 1995; Goebbert and Leslie, 2010; Dowdy et al.,

2012), the average storm genesis locations tend to shift towards the equator and eastward towards the International Date Line during El-Niño events. During La Niña events, a westward drift of storm track often leads to greater landfall frequency. However, the total TC count during a given season appears to be independent of an ENSO-induced shift within an individual basin’s main development region (Lander, 1993; Wang and

Chan, 2002).

9 Though the interannual variability of ENP storm frequency is not entirely driven by ENSO, it has been shown the spatial representation of ENP storm tracks is critically depended on ENSO variation (Camargo et al., 2008). Previously, the mean location of

TC genesis during El Niño years is found to experience a westward shift and begin closer to the equator (Irwin and Davis, 1999). As a result, ENP storm genesis region generally expands westward (eastward) during the El Niño (La Niña) phase (Irwin and Davis,

1999). However, storm track distances between ENSO phases were not found to be statistically different even though ENSO-induced differences in the downgradation locations were detected.

1.4 Sea surface temperature effect on tropical cyclone intensity

Arguably, the most widely discussed environmental factor that contributes to the

TC formation and intensification is SST. Although TCs are clearly inhibited by cooler ocean water (Emanuel, 2006; Dare and McBride, 2011b), a fully matured storm may decouple somewhat from sources of warm water and yet still develop (Holland, 1997).

Since warm ocean water tends to support higher TC wind speed through promotion of atmospheric instability (Palmén, 1948; Gray, 1968) and ranges of thresholds of minimum

SSTs have been proposed to provide the required thermodynamic energy for cyclogenesis

(Holland, 1997; Dare and McBride, 2011a). Hence, not only an abnormally warm SST is generally required for above-normal TC occurrences, it is also a critical factor in determining the maximum potential TC intensity.

Following the back-to-back above-normal North Atlantic hurricane seasons in

2004 and 2005, there has been increased public interest in the connection between

10 hurricanes and global warming. While Webster et al. (2005) have shown the recent record of SST warming and the combined number of tropical storms and hurricanes in the

North Atlantic basin have increased between 1995 and 2005, Emanuel (2005) points out the recent SST warming relative to temperature increases in the tropospheric profile is highly critical to storm intensity. Instead of using basin-wide SST averages, Michaels et al. (2006) suggested contemporaneous SSTs that are measured immediately below storm centers should be used when examining the SST-TC intensity relationship. This alternative approach in applying near-time SSTs showed that SSTs do not have a dominant effect on storm intensity and that other factors may be more critical in elevating and/or suppressing the maximum sustained TC winds.

Models have been used to assess the impact of projected SST warming on TC intensity. In response to increased SSTs, moist convective instability is generally enhanced thus favouring more intense storms in the future. Through the continuous improvement of model resolution in simulating the surrounding environment for storm genesis and development, dynamically downscaled hurricane models are now more capable at simulating the internal structure of storm vortices (Bengtsson et al., 1996). It is generally agreed that the maximum potential intensity will rise by 10%-20% under a warming scenario of a doubled carbon dioxide forcing (Henderson-Sellers et al., 1998).

More recent evidence has also shown that higher radiative forcing (global warming) will generate a higher storm frequency (Emanuel, 2013).

Even though seasonal SST variability is characterized by both low-frequency and high-frequency climate variability, a consistent warming of ocean water temperature is observed across all major TC development basins (Webster et al., 2005; Elsner et al.,

11 2008). As moisture is evaporated from the warm ocean surface, moist entropy is transported and contributes to storm intensification, despite other hindrances such as the storm-induced cooling of surface water, ocean spray and interference of surrounding dry air. Using PDI as an indicator for seasonal storm intensity and potential societal damage, the recent surge in North Atlantic storm intensity is highly correlated with SST warming

(Emanuel, 2005; Wu et al., 2008). Elsewhere, although there seems to be a lack of PDI trends in both western Pacific and ENP basins, PDIs in both regions are still found to correlate to their respective SSTs (Wu et al., 2008).

1.5 Research Gap

Although there had been numerous attempts to explore the relationship between the phases of ENSO and individual TC parameters (frequency, duration and intensity), their connections remain unclear. It is hypothesized that the consolidation of all three measures of TC activity into a single storm metric will provide a more representative way to evaluate the ENSO impact on seasonal storm activity and intensity. Part of the reason for the absence of a clear understanding could be a lack of reliable TC data and the choice of ENSO index. Because many ENSO indicators, such as Ocean Niño Indices and

Southern Oscillation Index, rely only on a single environmental field, the Multivarite

ENSO Index (MEI; Wolter and Timlin, 1998), which incorporates conditions of multiple environmental factors, is likely a more suitable metric for ENSO. Since the warm (El

Niño) phase of ENSO is often associated with abnormally high tropical convection in the eastern Pacific region, increase in ENP storm activity and intensity are expected.

12 ENSO-induced effect is not only limited to storm activity and intensity but is also critical to locations of TC genesis and downgradation. Potential shifts of storm location might contribute to changes in the distance a storm travels. Earlier work from Irwin and

Davis (1999) linked ENSO fluctuation with the initial and final geographic positions of

ENP storms. Statistically significant differences were found when mean locations of storm origin and downgradation and the ENSO phases. Although the easterly and southerly movements were statistically identical between any ENSO pairing (i.e. El Niño vs La Niña, El Niño vs neutral or La Niña vs neutral), such a conclusion may change as the local storm data archive is updated. It is possible that as a more poleward or longitudinal movement of storm track is realized under climate change, TC duration will also be lengthened in response. Understanding the regional variation of ENSO-modified storm track points and movements may help to pinpoint areas that are more vulnerable to

TC risks.

The recent response of TC intensity to SST warming has been well researched in the field of climate change impact assessment. Globally, the maximum TC winds had dramatically increased with SSTs (Elsner et al., 2008). Although their result shows significant SST control is linked to only the most intense storms, SST impact on lifetime maximum storm strength has not been investigated. A recent observational study in the

North Atlantic basin concludes that near-time SSTs did not contribute significantly to the lifetime-maximum TC winds (Michaels et al., 2006). In general, although the development of major hurricane strength in the North Atlantic basin requires a minimum of 28.25°C, a continuous function between SSTs and maximum intensities of major hurricanes was not detected. While an SST threshold is required to achieve certain levels

13 of storm intensity, Michaels et al. (2006) also noted other factors may be more critical in determining the ultimate TC intensity. However, similar work on the transient change in

TC intensity due to recent SST warming has not been investigated in the ENP basin.

1.6 Study Area

The study area encompasses the prime area for ENP tropical cyclogenesis, which will be referred as the main development region (MDR; 10-20°N and 85-140°W) hereafter. Though storms may form outside of MDR, they would still spend the majority of their lifetime within the MDR. Seasonal monitoring of storm activity in the ENP and

North Atlantic basins is done by the U.S. National Hurricane Center. However, environmental conditions usually become ripe for ENP cyclogenesis (Appendix I), earlier than its neighbouring North Atlantic basin. In practice, NHC’s official hurricane season for the ENP basin starts on May 15 and ends on November 30, with the peak season occurring from July to September (Wu and Chu, 2007).

Given that most TCs form over the tropics, storms are steered by easterly winds

(Trade winds) and move predominantly westward. As ENP storms encounter strong mid- latitudinal wind flows, they would often track poleward and recurve to landfall in Mexico

(Englehart et al., 2008). Though rare, some of them may enter into the Gulf of Mexico as

North Atlantic storms. Rarely does an ENP storm landfall in the U.S. By the time ENP storm influence could be experienced in the U.S., most would be downgraded to TC remnants. However, they can still bring in a large amount of rainfall, relieving some of the driest areas in southwest U.S. (Corbosiero et al., 2009; Ritchie et al., 2011). Through

14 the paleotempestological record, an 1858 TC was discovered to have reached southern

California as a hurricane (Chenoweth and Landsea, 2004).

ENP storm systems are highly interactive with its immediate geography and environment. Many of the local TCs were derived from tropical disturbances that can be traced as far away as the North African easterly waves (Avila and Pasch, 1992). In fact,

Collins (2010) noted many of these storm precursors were once identified as disturbances in the North Atlantic basin. However, they were able to intensify and upgrade to a TC status only when reaching the ENP basin, suggesting an inter-basin, yet negative, relationship. A more conducive environment for TC development at ENP basin triggers many of these convective systems, while some are strengthened by an orographic effect when passing through the mountainous area over Central America (Zehnder, 1991). In cases where tropical depressions had already intensified into tropical storms or hurricanes before crossing into the adjacent basin, at present, their North Atlantic storm names are maintained unless they had been downgraded to disturbances. Many of these westward- moving ENP storms often track westward towards islands in the Central Pacific Ocean and beyond. In fact, the longest-lasting TC ever recorded, Hurricane John (1994), had reached as far as Northwestern Pacific basin before recurvature.

The monthly distribution of TC activity generally follows the monthly SST averages in MDR where, consistent with other TC basins, ENP storm activity is most active during the warmest SST conditions (Figure 1-3). Consistent with Wu and Chu

(2007), ENP storm activity peaks during July-September when SSTs peak (Figure 1-4), accounting for 70% of total storms in a season. Of the less active months, June has a higher number of total storms and hurricanes than October and November. In

15 8 20 7 18 16 6 14 5 12 4 10 3 8 6 2

TC Frequency 4 1 2 TC Duration (Days) 0 0 May June July Aug Sept Oct Nov TS H HD MH MHD TSD

Figure 1-3 Monthly TC frequency (TS, H, MH) and TC duration (TSD, HD, MHD) averaged over 1971-2012 from May to November in the main development region of the eastern North Pacific basin.

29 28 27

C) 26 ° ( 25 Sea Surface Temperature MDR EDR WDR

Figure 1-4 Monthly sea surface temperature averaged over 1971-2012 at the main development region (MDR) of the eastern North Pacific basin and its subdivisions of EDR and WDR.

16 contrast to the century-long storm record at the North Atlantic basin, a less active TC month in June is compared with the tail-end of the local hurricane season (June-

November).

Much advancement in the current understanding of environmental influences on

ENP storm activity has been achieved by the geospatial division of local storm activity, breaking the ENP MDR into two subregions (Collins and Mason, 2000; Ralph and

Gough, 2009). Although many ENP storms are initiated closer to the coastlines of

Mexico and other countries of Central America, most reach their maximum intensities far from shore. This subdivision analysis led to the detection of regional variation of TC sensitivity to various environmental conditions (Collins and Mason, 2000; Collins, 2010;

Collins and Roache, 2011). Compared to the North Atlantic basin, where seasonal variability of dynamical influences such as vertical wind shear imposes a dominant influence on TC characteristics, ENP storms are more responsive to changes in thermodynamic limitations such as relative humidity in the mid-troposphere (Collins,

2007; Collins, 2010; Collins and Roache, 2011).

Other thermodynamic controls on ENP storm activity were also explored by

Ralph and Gough (2009), in which they used a different longitudinal line (112°W) from

Collins and Mason (2000) to determine boundaries of eastern (EDR) and western (WDR) development regions. Despite a climatologically warmer SST in EDR (10-20°N and 85-

112°W), where stronger TCs are more prevalent, Ralph and Gough (2009) found parameters of storm frequency and duration are significantly correlated with SST only in

WDR (10-20°N and 112-140°W). Similar conclusion was arrived earlier in Collins and

17 Mason (2000), which used 116°W as the geophysical division of MDR storms east of the

International Date Line (118°). For instance, while TC frequencies and durations, except for major hurricanes, are observed to peak when the warmest monthly SST co-occurs in

WDR (Figure 1-5), EDR storm activity tends to be equally active between July to

September (Figure 1-6). As such, correlation strengths between all monthly SSTs and

TC parameters are stronger for WDR storms than EDR storms (Table 1-1).

1.7 Objectives

As noted previously, the relative SST measurements from the climatological average may be important at triggering or suppressing local TC activity. Ralph and

Gough (2009) reasoned that deviations from the 26.5°C criterion for TC formation

(Appendix I) had substantially contributed to a regional difference of SST-TC activity relationship. Because the seasonal condition in EDR is already above this SST threshold

(Gray, 1968), deviations from EDR’s seasonal SST climatology may not induce the same impact on TC activity as SST variations in WDR, where its long-term average of seasonal SST is closer to the threshold. It is expected any abnormal SST warming

(cooling) in the WDR will induce a greater upswing (downswing) of storm activity than the EDR.

The goal of the thesis is to explore the influences of natural variation and climate change to the most recent ENP storm climatology where storms are longitudinally stratified. Previous work (Collins and Mason, 2000) indicated that WDR storm frequency is more sensitive to mid-level relative humidity; while, counts of WDR hurricanes and major hurricanes are more responsive to differences in vertical wind shear

2 6

5 1.5 4

1 3

2 0.5 TC Frequency 1 TC Duration (days) 0 0 May June July Aug Sept Oct Nov TS H HD MH MHD TSD

Figure 1-5 same as in Figure 1-3 but for the western development region of the eastern North Pacific basin.

6 14

5 12 10 4 8 3 6 2 4

TC Frequency 1 2 TC Duration (days) 0 0 May June July Aug Sept Oct Nov TS H HD MH MHD TSD

Figure 1-6 Same as in Figure 1.2 but for the eastern development region of the eastern North Pacific basin.

19 Table 1-1: Strength of correlations between six measures of TC activity and SSTs in MDR, EDR and WDR averaged over 1971-2012 from May to November

TS TSD H HD MH MHD MDR 0.69 0.78 0.71 0.80 0.70 0.78 EDR 0.59 0.59 0.59 0.54 0.54 0.46 WDR 0.84 0.85 0.88 0.84 0.77 0.80

between the 850mb and 200mb. As such, it is hypothesized that aspects of WDR storm characteristics will demonstrate a greater degree of association with ENSO fluctuation.

Although Ralph and Gough (2009) found individual parameters of WDR storm activity are more sensitive to SST variability, their study was conducted using basin-wide SST averages, rather than the near-time SST conditions measured at the storm location. Given that the underlying SSTs recorded at each storm center have become warmer than the past analysis (Whitney and Hobgood, 1997), the maximum potential intensity is also expected to rise. Hence, the following research questions emerge from the works of

Collins and Mason (2000) and Ralph and Gough (2009) pertaining to the ENP storm basin:

1. How do storm metrics that measure seasonal storm activity and intensity vary as a

function of ENSO conditions in the main development region (MDR)?

2. How do storm metrics that measure seasonal storm activity and intensity vary as a

function of ENSO conditions compare between the MDR subdivisions?

3. How do locations of storm genesis, maximum lifetime intensity and

downgradation and track movement vary as a function of ENSO conditions in the

MDR?

20 4. How do locations of storm genesis, maximum lifetime intensity and

downgradation and track movement vary as a function of ENSO conditions

compare between MDR subdivisions?

5. How do near-time SSTs link to the observations of TC intensities in the MDR

subdivisions?

6. How has the storm’s maximum potential intensity been limited by SSTs using the

most recent SST record?

An outline of the following sections illustrates a more detailed overview of the thesis. Chapter 2, which has been published in the Journal of Climate (Jien et al., 2015), statistically addresses the significance of ENSO influence on metrics of TC activity and intensity within MDR and its subdivisions and addresses research questions 1 and 2.

Chapter 3 highlights the ENSO impact on the geographic distributions of EDR and WDR storm tracks at different storm stages of its lifecycle and addresses research questions 3 and 4. Chapter 4 investigates the effect of near-time SSTs on the maximum potential TC intensity and the maximum lifetime TC winds and addresses research questions 5 and 6.

Chapter 5 concludes the thesis by suggesting future work.

1.7.1 Chapter 2

This chapter presents an overview of TC climatology in the ENP basin and statistically compares storm activity and storm intensity through a combination of ENSO classifications and regional divisions. Storm data from 1971-2012 is binned according to

ENSO phases of El Niño, La Niña and neutral conditions. Since locations of ENP storm

21 genesis mostly remain closer to the continental shoreline, storms are longitudinally stratified. As the relationship between ENSO and TC characteristics in the ENP basin is not as definitive as other storm basins, the importance of ENSO at modulating seasonal

TC activity and intensity at both the basin and sub-basin levels is tested. Based on previous work of Collins and Mason (2000), we hypothesize a longitudinal difference of

TC sensitivity to ENSO phases. Specifically, linkages between indices of TC activity and TC intensity with ENSO are addressed using correlation measures and compared between spatial divisions of ENP basin.

1.7.2 Chapter 3

Chapter 3 is an extended work from Irwin and Davis (1999) in which associations of points of genesis and downgradation of ENP storms and their eventual tracks with

ENSO are investigated. Statistical comparisons of characteristics of storm tracks are made under different ENSO pairings at MDR and its subdivisions. Since it was hypothesized previously that TC features in WDR are more sensitive to ENSO influences, significant differences of ENSO-induced longitudinal and/or latitudinal displacement of locations of genesis and dissipation are expected to occur only in WDR.

ENSO modulation of storm genesis and downgradation points will likely affect the length of storm track. As such it is investigated if the length of storm track is modulated by

ENSO influences and whether such variation of storm movement is maintained in both the EDR and WDR. Although most storms form or endure the intensification process within boundaries of the MDR, and its subdivisions, by the time that some dissipate, they are likely to traverse these boundary divisions. As such this study also addresses how

22 likely storms would track outside these boundaries between different years of ENSO conditions.

1.7.3 Chapter 4

Based on the previous investigations of historical relationship of SST and TC wind speed during a 20-year (1967-1986) timespan, we expand the connection between contemporaneous SST and TC winds by updating the analysis from 1971-2013. It is hypothesized that although SSTs provides the upper limit for TC intensification process, it is not the sole factor in determining the observed maximum sustained wind speeds.

Since the relationship of storm intensity with monthly and weekly SST datasets have been explored elsewhere (Evans, 1993; Michaels et al., 2006), the temporal resolution is refined to a daily SST dataset. In addition, the spatial resolution of the current SST dataset has also been substantially upgraded to better differentiate gridded SST differences. The relationship between more updated SST observations and maximum sustained wind speed will generate a better representation of transient changes of TC strength in response to the underlying SST.

With the availability of a near-time SST record and corresponding storm tracks, the SST threshold for TC genesis also is assessed. Since the importance of a SST threshold in contributing to the seasonal TC genesis index has already been determined

(Gray, 1968), the existence of a SST threshold is hypothesized in determining the number of storms generated in the ENP basin. Based on the previous analyses that 26.5°C is the

SST threshold for TC genesis, we determine if this value differs for ENP storms between, and within, sub-basins. Specifically, measures of SST thresholds for storms that attained

23 hurricane and major hurricane strengths are compared with those that occurred in the

North Atlantic basin (Michaels et al., 2006).

24 1.8 References:

Amador, J. A., E. J. Alfaro, O. G. Lizano, & V. O. Magaña, 2006: Atmospheric forcing of the eastern tropical Pacific: A review. Progress in Oceanography, 69(2-4), 101-142.

Ashok, K., S. K. Behera, S. A. Rao, S. A., Weng, H., and Yamagata, T. (2007). El Niño modoki and its possible teleconnection. Journal of Geophysical Research: Oceans, 112(11), C11007.

Auld, H., 2008: Disaster Risk Reduction under Current and Changing Climate Conditions: Important Roles for the National Meteorological and Hydrological Services. World Meteorological Organization, 57(1), 118-125.

Avila, L.A., and R. J. Pasch, 1992: Atlantic tropical systems of 1991. Monthly Weather Review, 120, 2688–2696.

Basher, R. E. and X. Zheng, 1995: Tropical cyclones in the Southwest Pacific: Spatial patterns and relationships to Southern Oscillation and sea surface temperature. Journal of Climate, 8, 1249-1260.

Bell, G. D., and Coauthors, 2000: Climate assessment for 1999. Bulletin of the American Meteorological Society., 81, S1–S50.

Bengtsson, L., M. Botzet, and M. Esch, 1996: Will greenhouse gas-induced warming over the next 50 years lead to higher frequency and greater intensity of hurricanes? Tellus, Series A: Dynamic Meteorology and Oceanography, 48(1), 57-73.

Bengtsson, L., K. I. Hodges, M. Esch, N. Keenlyside, L. Kornblueh, J-J. Luo, et al., 2007: How may tropical cyclones change in a warmer climate? Tellus, 59A, 539-561.

Cai, W., M. Lengaigne, S. Borlace, M. Collins, T. Cowan, M. J. McPhaden, . . . M. J. Widlansky, 2012: More extreme swings of the south Pacific convergence zone due to greenhouse warming. Nature, 488(7411), 365-369.

Cai, W., S. Borlace, M. Lengaigne, P. Van Rensch, M. Collins, G. Vecchi, A. Timmermann, A. Santoso, M. McPhaden, L. Wu, M. H. England, G. Wang, E. Guilyardi, and F.-. Jin, 2014: Increasing frequency of extreme El Niño events due to greenhouse warming. Nature Climate Change, 4(2), 111-116.

Camargo, S. J., and A. H. Sobel, 2005: Western north Pacific tropical cyclone intensity and ENSO. Journal of Climate, 18(15), 2996-3006.

Camargo, S. J., K. A. Emanuel, and A. H. Sobel, 2007: Use of a genesis potential index to diagnose ENSO effects on tropical cyclone genesis. Journal of Climate, 20(19), 4819- 4843. doi:10.1175/JCLI4282.1.

25 Camargo, S. J., A. W. Robertson, A. G. Barnston, and M. Ghil, 2008: Clustering of eastern north Pacific tropical cyclone tracks: ENSO and MJO effects. Geochemistry, Geophysics, Geosystems, 9(6), Q06V05

Cazenave, A., and R. S. Nerem, 2004: Present-day sea level change: Observations and causes. Reviews of Geophysics, 42(3), RG3001 1-20.

Chan, J. C. L., 1985: Tropical cyclone activity in the northwest Pacific in relation to the El Nino/Southern oscillation phenomenon. Monthly Weather Review, 113(4), 599-606.

Chan, J. C. L., 2000: Tropical cyclone activity over the western north Pacific associated with El Niño and La ninña events. Journal of Climate, 13(16), 2960-2972.

Changnon, S. A., D. Changnon, E. Ray Fosse, D. C. Hoganson, R. J. Roth Sr., and J. M. Totsch, 1997: Effects of recent weather extremes on the insurance industry: Major implications for the atmospheric sciences. Bulletin of the American Meteorological Society, 78(3), 425-435.

Chen, J. L., C. R. Wilson, and B. D. Tapley, 2013: Contribution of ice sheet and mountain glacier melt to recent sea level rise. Nature Geoscience, 6(7), 549-552.

Chenoweth, M., and C. Landsea, 2004: The San Diego hurricane of 2 October 1858. Bulletin of the American Meteorological Society, 85(11), 1689-1697.

Chia, H. H., and C. F. Ropelewski, 2002: The interannual variability in the genesis location of tropical cyclones in the northwest Pacific. Journal of Climate, 15(20), 2934- 2944.

Collins, J. M., 2007: The relationship of ENSO and relative humidity to interannual variations of hurricane frequency in the North-East Pacific Ocean. Papers of the Applied Geography Conferences, 30, 324-333.

Collins, J., 2010: Contrasting High North-East Pacific Tropical Cyclone Activity. Southeastern Geographer, 50, 83-98.

Collins, J. M., and I. M. Mason, 2000: Local environmental conditions related to seasonal tropical cyclone activity in the Northeast Pacific basin. Geophysical Research Letters, 27, 3881-3884.

Collins, J. M., and D. R. Roache, 2011: The 2009 Hurricane Season in the Eastern North Pacific Basin: An Analysis of Environmental Conditions. Monthly Weather Review, 139(6), 1673-1681.

26 Collins, M., S.-. An, W. Cai, A. Ganachaud, E. Guilyardi, F.-. Jin, M. Jochum, M. Lengaigne, S. Power, A. Timmermann, G. Vecchi, and A. Wittenberg, 2010: The impact of global warming on the tropical Pacific Ocean and El Nĩno. Nature Geoscience, 3(6), 391-397.

Corbosiero, K. L., M. J. Dickinson, and L. F. Bosart, 2009: The contribution of eastern north Pacific tropical cyclones to the rainfall climatology of the southwest united states. Monthly Weather Review, 137(8), 2415-2435.

Dare, R. A., and J. L. Mcbride, 2011a: Sea surface temperature response to tropical cyclones. Monthly Weather Review, 139(12), 3798-3808.

Dare, R. A., and J. L. Mcbride, 2011b: The threshold sea surface temperature condition for tropical cyclogenesis. Journal of Climate, 24(17), 4570-4576.

Denniston, R. F., G. Villarini, A. N. Gonzales, K. -. Wyrwoll, V. J. Polyak, C. C. Ummenhofer, M. S. Lachniet, A. D. Wanamaker Jr., W. F. Humphreys, D. Woods, and J. Cugley, 2015: Extreme rainfall activity in the Australian tropics reflects changes in the el Niño/Southern oscillation over the last two millennia. Proceedings of the National Academy of Sciences of the United States of America, 112(15), 4576-4581.

Dowdy, A. J., L. Qi, D. Jones, H. Ramsay, R. Fawcett, and Y. Kuleshov, 2012: Tropical cyclone climatology of the south Pacific Ocean and its relationship to El Niño-southern oscillation. Journal of Climate, 25(18), 6108-6122.

Dvorak, V. F., 1975: Tropical cyclone intensity analysis and forecasting from satellite imagery. Monthly Weather Review, 103, 420–430.

Elsner, J. B., J. P. Kossin, and T. H. Jagger, 2008: The increasing intensity of the strongest tropical cyclones. Nature, 455(7209), 92-95.

Emanuel, K. A., 2001: Contribution of tropical cyclones to meridional heat transport by the oceans. Journal of Geophysical Research: Atmospheres, 106(D14), 14771-14781.

Emanuel, K. A., 2005: Increasing Destructiveness of Tropical Cyclones over the Past 30 years. Nature , 436, 686-688.

Emanuel, K., 2006: Hurricanes: Tempests in a greenhouse. Physics Today, 59(8), 74-75.

Emanuel, K. A., 2013: Downscaling CMIP5 climate models shows increased tropical cyclone activity over the 21st century. Proceedings of the National Academy of Sciences of the United States of America, 110(30), 12219-12224.

Englehart, P. J., M. D. Lewis, and A. V. Douglas, 2008: Defining the frequency of near- shore tropical cyclone activity in the eastern North Pacific from historical surface observations (1921-2005). Geophysical Research Letters, 35, L03706.

27 Evans, J. E., 1993: Sensitivity of tropical cyclone intensity to sea surface temperature, Journal of Climate, 6, 1133– 1140.

Frank, W. M., and G. S. Young, 2007: The interannual variability of tropical cyclones. Monthly Weather Review, 135(10), 3587-3598.

Goebbert, K. H., and L. M. Leslie, 2010: Interannual variability of northwest Australian tropical cyclones. Journal of Climate, 23(17), 4538-4555.

Gray, W. M., 1968: The Origin of Tropical Disturbances and Storms. Monthly Weather Review, 96, 669-700.

Gray, W. M., C. W. Landsea, P. W. Mielke Jr, and K. J. Berry, 1994: Predicting Atlantic basin seasonal tropical cyclone activity by 1 June. Weather and Forecasting, 9(1), 103- 115.

Held, I. M., and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. Journal of Climate, 19(21), 5686-5699.

Henderson-Sellers, A., H. Zhang, G. Berz, K. Emanuel, W. Gray, C. Landsea, et al., 1998: Tropical Cyclone and Global Climate Change: A Post-IPCC Assessment. Bulletin of the American Meteorological Society , 79, 19-37.

Holland, G. J., 1997: The maximum potential intensity of tropical cyclones. Journal of the Atmospheric Sciences, 54(21), 2519-2541.

Hurrell, J. W., and K. E. Trenberth, 1999: Global sea surface temperature analyses: Multiple problems and their implications for climate analysis, modeling, and reanalysis. Bulletin of the American Meteorological Society, 80, 2661-2678.

IPCC, 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 1535 pp, doi:10.1017/CBO9781107415324.

Irwin III, R. P., and R. E. Davis, 1999: The relationship between the Southern Oscillation Index and tropical cyclone tracks in the eastern North Pacific. Geophysical Research Letters, 26(15), 2251-2254.

Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bulletin of the American Meteorological Society., 77, 437–471.

28 Korty, R. L., K. A. Emanuel, and J. R. Scott, 2008: Tropical cyclone-induced upper- ocean mixing and climate: Application to equable climates. Journal of Climate, 21(4), 638-654. Kossin, J. P., K. A. Emanuel, and G. A. Vecchi, 2014: The poleward migration of the location of tropical cyclone maximum intensity. Nature, 509(7500), 349-352.

Knutson, T. R., J. L. McBride, J. Chan, K. Emanuel, G. Holland, C. Landsea, . . . M. Sugi,, 2010: Tropical cyclones and climate change. Nature Geoscience, 3(3), 157-163.

Lander, M. A., 1994: An exploratory analysis of the relationship between tropical storm formation in the western north Pacific and ENSO. Monthly Weather Review, 122(4), 636- 651.

Landsea, C. W., G. A. Vecchi, L. Bengtsson, and T. R. Knutson, 2010: Impact of duration thresholds on Atlantic tropical cyclone counts. Journal of Climate, 23(10), 2508- 2519.

Meehl, G. A., F. Zwiers, J. Evans, T. Knutson, L. Mearns, and P. Whetton, 2000: Trends in extreme weather and climate events: Issues related to modeling extremes in projections of future climate change. Bulletin of the American Meteorological Society, 81(3), 427-436.

Mendelsohn, R., K. Emanuel, S. Chonabayashi, and L. Bakkensen, 2012: The impact of climate change on global tropical cyclone damage. Nature Climate Change, 2(3), 205- 209.

Michaels, P. J., P. C. Knappenberger, and R. E. Davis, 2006: Sea-surface temperatures and tropical cyclones in the Atlantic basin. Geophysical Research Letters, 33(9)， L09708, doi:10.1029/2006GL025757

Mitrovica, J. X., M. E. Tamisiea, J. L. Davis, and G. A. Milne, 2001: Recent mass balance of polar ice sheets inferred from patterns of global sea-level change. Nature, 409(6823), 1026-1029.

Nicholls, N., 1979: A possible method for predicting seasonal tropical cyclone activity in the Australian region. Monthly Weather Review, 107(9), 1221-1224.

Nicholls, N., 1984: The southern oscillation, sea-surface-temperature, and interannual fluctuations in Australian tropical cyclone activity. Journal of Climatology, 4(6), 661- 670.

Nicholls, N., 1985: Predictability of interannual variations of Australian seasonal tropical cyclone activity. Monthly Weather Review, 113(7), 1144-1149.

Palmén, E., 1948: On the formation and structure of tropical hurricanes. Geophysica, 3, 26–38. Pan, Y. H., 1982: The effect of the thermal state of equatorial eastern Pacific on the

29 frequency of typhoon over the western Pacific (in Chinese with English abstract). Acta Meteorlogical Sinica,40, 24–34.

Pezza, A. B., and I. Simmonds, 2005: The first south Atlantic hurricane: Unprecedented blocking, low shear and climate change. Geophysical Research Letters, 32(15), L15712.

Pielke Jr., R. A., and C. W. Landsea, 1998: Normalized hurricane damages in the united states: 1925-95. Weather and Forecasting, 13(3 PART2), 621-631.

Pielke Jr., R. A., and Landsea, C. N. (1999). La Niña, El Niño, and Atlantic hurricane damages in the United States. Bulletin of the American Meteorological Society, 80(10), 2027-2033.

Pritchard, H. D., R. J. Arthern, D. G. Vaughan, and L. A. Edwards, 2009: Extensive dynamic thinning on the margins of the Greenland and Antarctic ice sheets. Nature, 461(7266), 971-975.

Ralph, T. U., and W. A. Gough, 2009: The influence of sea-surface temperatures on Eastern North Pacific tropical cyclone activity. Theoretical and Applied Climatology, 95, 257-264.

Ritchie, E. A., K. M. Wood, D. S. Gutzler, and S. R. White, 2011: The influence of eastern Pacific tropical cyclone remnants on the southwestern United States. Monthly Weather Review, 139(1), 192-210.

Saunders, M. A., R. E. Chandler, C. J. Merchant, and F. P. Roberts, 2000: Atlantic hurricanes and NW Pacific typhoons: ENSO spatial impacts on occurrence and landfall. Geophysical Research Letters, 27(8), 1147-1150.

Schwing, F. B., T. Murphree, L. DeWitt, and P. M. Green, 2002: The evolution of oceanic and atmospheric anomalies in the northeast Pacific during the El Niño and La Niña events of 1995-2001. Progress in Oceanography, 54(1-4), 459-491.

Shultz, J.M., J. Russell, and Z. Espinel, 2005: Epidemiology of tropical cyclones: The dynamics of disaster, disease, and development. Epidemiologic Reviews, 27, 21-35.

Solomon, S., G. -., Plattner, R. Knutti, and P. Friedlingstein, 2009: Irreversible climate change due to carbon dioxide emissions. Proceedings of the National Academy of Sciences of the United States of America, 106(6), 1704-1709.

Sriver, R. L., and M. Huber, 2007: Observational evidence for an ocean heat pump induced by tropical cyclones. Nature, 447(7144), 577-580.

Steenhof, P.A. and W.A. Gough, 2008: The impact of tropical sea surface temperatures on various measures of Atlantic tropical cyclone activity. Theoretical and Applied Climatology 92, 249-255.

30 Timmermann, A., J. Oberhuber, A. Bacher, M. Esch, M. Latif, and E. Roeckner, 1999: Increased El Nino frequency in a climate model forced by future greenhouse warming. Nature, 398(6729), 694-697.

Trenberth, K. E., 1998: Atmospheric moisture residence times and cycling: Implications for rainfall rates and climate change. Climatic Change, 39(4), 667-694.

Wang, B., and J. C. L. Chan, 2002: How strong ENSO events affect tropical storm activity over the western north Pacific. Journal of Climate, 15(13), 1643-1658.

Wang, C., S. -. Lee, and D. B. Enfield, 2008: Atlantic warm pool acting as a link between Atlantic Multidecadal Oscillation and Atlantic tropical cyclone activity. Geochemistry, Geophysics, Geosystems, 9(5), Q05V03.

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.

Whitney, L. D., and J. S. Hobgood, 1997: The relationship between sea surface temperatures and maximum intensities of tropical cyclones in the eastern north Pacific Ocean. Journal of Climate, 10(11), 2921-2930.

Williams, J., 2009: Tropical cyclones form in global ‘basins’ [Image]. Retrieved 9 July, 2015, from USA TODAY, http://usatoday30.usatoday.com/weather/hurricane/tropical- cyclone-basins.htm

Wolter, K., and M. S. Timlin, 2011: El Niño/Southern oscillation behaviour since 1871 as diagnosed in an extended multivariate ENSO index (MEI.ext). International Journal of Climatology, 31(7), 1074-1087.

Wu, L., B. Wang, and S. A. Braun, 2008: Implications of tropical cyclone power dissipation index. International Journal of Climatology, 28, 727-731.

Wu, P. and P. -. Chu, 2007: Characteristics of tropical cyclone activity over the eastern north Pacific: The extremely active 1992 and the inactive 1977. Tellus, Series A: Dynamic Meteorology and Oceanography, 59A(4), 444-454.

Wu, G., and N. C. Lau, 1992: A GCM simulation of the relationship between tropical- storm formation and ENSO. Monthly Weather Review, 120(6), 958-977.

Zehnder, J. A., 1991: The interaction of planetary-scale tropical easterly waves with topography: A mechanism for the initiation of tropical cyclones. Journal of the Atmospheric Sciences, 48(10), 1217-1230.

31 2 Chapter 2: The Influence of El Niño-Southern Oscillation on Tropical Cyclone Activity in the Eastern North Pacific Basin

2.1 Abstract

The interannual variability of tropical cyclone (TC) activity due to El Niño–

Southern Oscillation (ENSO) in the main development region of the eastern North Pacific basin has received scant attention. Herein the authors classify years of El Niño, La Niña, and neutral conditions using the multivariate ENSO index (MEI). Storm measurements of the net tropical cyclone activity index and power dissipation index are used to summarize the overall seasonal TC activity and TC intensity between 1971 and 2012.

Both measures are found to be statistically dependent on the ENSO phases in the basin’s main development region. However, when the area is longitudinally divided, only the western portion of the development region experienced a significant difference (p <

0.05). Specifically, El Niño years are characterized by more frequent, more intense events compared to La Niña conditions for this subregion. Correlation analyses on the relationships between the MEI and both TC indices demonstrate correlations between

ENSO and TC activity and intensity that are statistically significant (p < 0.05) only in the western region. These relationships have the potential to improve the short-term forecast of the local TC activity and intensity on a seasonal basis for public awareness and disaster preparation.

2.2 Introduction

The El Niño–Southern Oscillation (ENSO) phenomenon is arguably the most dominant control of the interannual climate variability on a global scale (Rasmusson and

Carpenter 1982). Atmospheric variations in the ENSO cycle can be monitored by the mean sea level pressure difference between Tahiti and Darwin, Australia, known as the

Southern Oscillation index (Ropelewski and Jones 1987). Oceanic anomalies at the

Niño-3 (5°N-5°S and 160°E-150°W) and Niño-3.4 (5°N-5°S and 170°-120°W) regions of the equatorial Pacific Ocean during periods of warmer (colder) sea surface temperature are also applied to quantify the extent of El Niño (La Niña) events (Trenberth 1997;

Wolter and Timlin 1998). Overall, the El Niño–induced atmospheric instability in the eastern equatorial Pacific Ocean shifts westward when transitioning to the La Niña phase.

Given its feedback on the global atmospheric circulation and its effect on worldwide climate anomalies (Pielke and Landsea 1999; Landsea 2000), ENSO is also critical to the short-term prediction for a range of extreme weather events, including seasonal forecasts of tropical cyclones (TCs) in multiple ocean basins (Camargo et al.

2007). In additional to its role as an important contributor to seasonal TC activity

(frequency, intensity, and duration), the impact of ENSO may vary depending on the strength of its signal and the location of the TC development region (Pielke and

Landsea, 1999). In general, during the La Niña phase, an above-normal TC activity is observed in the southwestern Pacific (Nicholls 1979), the northwestern Pacific

(Chan 1985), and North Atlantic (Gray 1984a) TC development basins, while a greater number of storms form during the El Niño phase in the Central Pacific basin, west of

140°W to 180oW (Chu and Wang 1997).

The interannual fluctuation of the ENSO cycle acts as a crucial control to regional

33 storm intensity (Gray 1984b; Bove et al. 1998). For instance, La Niña (El Niño) years are linked to a greater (lower) number of intense TC landfalls in the U.S. during seasons when more (less) North Atlantic storms develop (Klotzbach 2011). Since TC potential destruction power can be empirically related and annually aggregated to its intensity

(Emanuel 2005), the ENSO cycle plays an important role in determining the potential

TC-afflicted societal damages (Chu and Wang 1997; Pielke and Landsea 1999). Because more landfalling storms mature into hurricanes on the eastern U. S., the annual average economic damage (Hebert et al. 1997) caused by North Atlantic TCs is exacerbated during La Niña years (Pielke and Landsea 1999). However, for the world’s most active tropical storm genesis region when considered on a per unit area and time bases at the eastern North Pacific (ENP) basin (Molinari et al. 2000), the relationship of TC intensity with ENSO has not been explored. Typically, ENP storms dissipate over the open sea.

However, when they landfall, they not only affect continental North America but also possess the ability to inflict their strengths on the Hawaiian Islands in the Central Pacific basin. For instance, the formation of Iniki (1992) in the ENP basin later reached the category-4 hurricane strength and was accounted for the financial loss of USD 2.5 billion in Hawaii, during the 1991/92 El Niño year (Chu and Wang 1997).

In the Pacific Ocean, where local signatures of El Niño/La Niña conditions are evident, (Nicholls 1992; Trenberth 1997; Wolter and Timlin 1998), strong northwestern

Pacific TCs, or typhoons, are suppressed immediately after El Niño years (Chan 1985;

Wang and Chan 2002; Chu 2004). Moreover, there is a general consensus that the distribution of TC genesis area in the northwestern Pacific basin has experienced a southeastward drift during the peak typhoon season of El Niño years relative to La Niña

34 years (Chan 1985; Wang and Chan 2002). Thus, El Niño storms tend to travel a greater distance (both westward and northward) prior to experiencing unfavorable conditions to dissipate. Continuing eastward in the North Pacific, it is anticipated that the interannual variation of the ENSO cycle would also influence TC activity (frequency, intensity, and duration) in the ENP basin. Specifically, it is hypothesized that ENP storm seasons during the El Niño phase will be reflected by a heightened TC activity with more intense

TCs; while, the La Niña phase will be more subdued.

In the ENP basin where the influence of ENSO on seasonal TC formation and development has remained largely undetermined (Lupo et al. 2008), there is, however, strong evidence of ENSO modification on the longitudinal shift in locations of TC origin and dissipation (Irwin and Davis 1999). Specifically, during El Niño events, more TCs tend to form west of the primary region of TC genesis, while more storms are distributed in the eastern region during the La Niña phase. Though Irwin and Davis (1999) conclude the seasonal total storm count does not deviate between the ENSO phases, Collins (2007) suggests TC frequency of certain storm categories is sensitive to the alternation of El

Niño and La Niña events only in the western division of the ENP main development region (MDR; 10–20°N and 85–140°W). Thus, this unequal east–west distribution of storm genesis contributes to such a difference of spatial sensitivity to environmental conditions (Collins and Mason 2000; Ralph and Gough 2009). In particular, thermodynamic influences at the western subdivision of MDR are argued to enforce such an ENSO-induced difference of TC frequency (Collins 2007, 2010; Klotzbach and Blake

2013).

To date, there remains a gap in the literature that statistically evaluates the

35 relationship of TC climatology in the ENP basin with ENSO. Instead of relying on seasonal storm count alone, an alternative option is to characterize metrics of seasonal TC activity and TC intensity through combining individual measures of storm frequency, intensity, and duration. Though Collins and Roache (2011) integrated these measurements into metrics that better define and quantify seasonal storm activity and intensity, a long-term record of these metrics is needed to establish their relationships with ENSO events. Following the works of Ralph and Gough (2009) and Collins and

Mason (2000), our study region will encompass MDR subdivisions to statistically illustrate the spatial sensitivity of regional storm activity as well as the interannual variation due to ENSO influences. Because globally stronger TCs tend to show a stronger relationship with ENSO (Frank and Young 2007), seasonal TC intensity will also be examined as a function of ENSO conditions.

2.3 Methods

2.3.1 Classification of ENSO events

The multivariate ENSO index (MEI) was used to identify ENSO cycles (Wolter and Timlin 1998). Compared to other ENSO indices, MEI is a more holistic approach that reflects the coupled nature of the ocean–atmosphere system through its incorporation of six predictors of sea level pressure, zonal and meridional surface winds, sea surface temperature, surface air temperature, and total cloud fraction. To overlap with the ENP hurricane season, El Niño (La Niña) events are defined by the 10 highest (lowest) years of sliding bimonthly MEI values, averaged from April–May to November–December, with the remaining 22 years defined as neutral years. Since this rank-based approach is

36 less influenced by environmental conditions of a particular season, leading to an asymmetrical number of El Niño and La Niña events, it is widely implemented in not only MEI (Klotzbach 2012), but also anomalies of other sea surface temperature–based

ENSO indices (Camargo and Sobel 2005).

2.3.2 TC activity and TC intensity

Historical TC track data, including the storm name and its 6-hourly geographic position by longitude and latitude, speed, and surface pressure, are obtained from the eastern North and Central Pacific hurricane database (HURDAT2) from the National

Hurricane Center. As in other works (e.g., Ralph and Gough 2009), TC data after 1971, when the routine use of satellite to monitor storm development, were included. Since the

ENP storm formation is spatially unequally distributed, data of TC activity is longitudinally divided at 112°W as the eastern (10-20° N and 85-112°W) and western

(10-20° N and 112-140°W) development regions (EDR and WDR, respectively; Ralph and Gough 2009). During occurrences when EDR storms migrate and establish their peak intensity in WDR, the tabulation of storm count and duration is allocated accordingly, based on the 6-hourly storm-track positions (Collins and Roache 2011).

The seasonal distribution of storm frequency, intensity, and duration during the

42-yr period was taken into account when deriving the two empirically based TC metrics.

Based on its surface wind speed, each TC is categorized as a tropical/named storm (18 m s-1), hurricane (33 m s-1) and intense (or major) hurricane (50 m s-1), which corresponds to categories 3, 4, or 5 on a Saffir–Simpson hurricane intensity scale (Simpson 1974). An overall measurement of the net tropical cyclone (NTC; Gray et al. 1994) activity index is

37 defined as

NTC = (%NS + %H + %MH + %NSD + %HD + %MHD) / 6, (Equation 2-1)

where each season’s percentage value is weighted against the entire period mean (1981–

2010) and is used for the six measures of seasonal activity: named storms (NS), named storm days (NSD), hurricanes (H), hurricane days (HD), intense hurricanes (IH), and intense hurricane days (IHD). In addition, TC intensity is measured through the power dissipation index (PDI; Emanuel 2005), defined as:

τ dt, (Equation 2-2) �� = �

where the maximum surface wind speed (Vmax) is summed over each storm’s lifetime

(τ) every 6 h for all storms of at least tropical storm strength. To make PDI more manageable, seasonal values are accumulated as annual aggregates and displayed at a scale of 1.0 × 107m3 s-3.

2.3.3 Statistical analyses

Seasonal NTC and PDI values for the MDR are stratified into the EDR and WDR

(Figure 2-1). To account for seasons with no WDR storms, time series for NTC (PDI) are normalized through the square root (natural log) transformation within MDR and its subdivisions. The Mann–Kendall test (Sprent 1989) is utilized to determine if trends of the time series for both storm indices are statistically significant (p < 0.05). The

38 magnitude of the trend is calculated through the Theil–Sen slope (Helsel and Hirsch

1991).

Seasonal NTC and PDI values are statistically compared to reveal any spatial and temporal effects. An analysis of variance (ANOVA) is applied to assess whether there is any statistically significant difference among different groups of MDR storm measurements (NTC and PDI). Seasonal storm measures are compared and grouped according to two conditions: 1) temporal classification of annual ENSO event (El Niño,

La Niña, or neutral years) and 2) spatial division of MDR into EDR and WDR. With each storm measurement stratified into groups of ENSO conditions and MDR subdivisions, a two-way ANOVA is applied to test for both individual effects, as well as any interaction from the two factors. An interactive effect from the two factors is examined if relative

39 differences due to ENSO influences are consistent at both EDR and WDR. Since there are three pairings (El Niño–La Niña, El Niño–neutral, and La Niña–neutral) of ENSO phases, a post hoc test using Tukey’s method of multiple comparisons is applied to determine which ENSO pairing(s) is accounted for the different group means (p < 0.05).

A correlation analysis is conducted to measure the degree of relationship for both measures of TC activity and TC intensity with the ENSO index of MEI. A quantile– quantile (Q–Q) plot and the Shapiro–Wilk test are applied to assess the normality of both storm indices prior to the decision of selecting a parametric or nonparametric testing for the correlation analysis. A Q–Q plot demonstrates the fitted distribution of seasonal storm measurements and displays the spread of data deviations from the respective line of normal distribution. In total, forty years of storm metrics were regressed against standardized values (as theoretical quantiles) of a normal distribution. Data normality was improved through the best parameter (lambda) estimated from the Box–Cox transformation and quantitatively proven through the Shapiro–Wilk test. A least squares regression is used to correlate MEI values with the two TC measures. A Pearson coefficient of correlation r provides a degree of association between NTC and PDI with

MEI. The coefficient of determination r2 is applied to explain the amount of variance that is captured by the linear regression. The correlation analysis is extended when the entire region is subsequently divided into EDR and WDR. Residuals of observed and predicted values from linear regression are assessed for data independence. A Ljung–

Box test is applied to examine the null hypothesis of residual independence in a correlogram (autocorrelation plot).

To distinguish the relative contribution of storm frequency, intensity, and duration

40 to PDI, we derive three additional PDI-derived indices. Essentially, the PDI time series is modified so that each of its surrogates only varies with storm frequency, intensity, or duration only. The average intensity (Ii) of a storm during its lifetime (Li) can be represented as Ii = PDIi / Li, where i indices each individual storm. Hence, PDI1, PDI2 and PDI3 can be calculated when the storm intensity, duration, and count are varied, respectively, according to each hurricane season, while the other two terms are averaged across all seasons of the entire 42-yr period (Camargo and Sobel 2005). Correlations calculated with MEI and PDI1, PDI2 and PDI3 can quantitatively determine and separate the contributions of storm intensity, duration, and frequency to PDI.

2.4 Results

2.4.1 Classification of ENSO events

Table 2-1 highlights the 10 highest (El Niño) and lowest (La Niña) MEI values of bimonthly averages from April–May to November–December, with an additional 22 years classified as neutral years during the 1971–2012 period. Our annual partitioning of

ENSO phases is in general agreement with results of Lupo et al. (2008) and Irwin and

Davis (1999). In particular, we show the strongest El Niño year (indicated by the highest

MEI average) to occur in 1997, when climatological impacts and extreme hydrometeorological events devastated the southwestern United States (Changnon 1999).

Nevertheless, the choice of ENSO classification schemes and the length of analysis period could have led to some disagreement in the identification of ENSO events. For instance, using the sea surface temperature anomaly to classify ENSO events, Lupo et al.

41 Table 2-1: Years from 1971-2012 are classified by ENSO phases based on the averages of the eight sliding bimonthly MEI values* from April-May to November-December.

El Niño Neutral La Niña 1972 (+ 1.50) 1976 (+ 0.52) 1971 (- 1.33) 1982 (+ 1.69) 1977 (+ 0.76) 1973 (- 1.28) 1983 (+ 0.96) 1978 (- 0.17) 1974 (- 0.86) 1987 (+ 1.72) 1979 (+ 0.61) 1975 (- 1.58) 1991 (+ 1.02) 1980 (+ 0.46) 1988 (- 1.11) 1992 (+ 0.97) 1981 (+ 0.00) 1999 (- 0.83) 1993 (+ 1.12) 1984 (- 0.17) 2007 (- 0.70) 1994 (+ 0.95) 1985 (- 0.30) 2008 (- 0.40) 1997 (+ 2.42) 1986 (+ 0.76) 2010 (- 1.26) 2002 (+ 0.88) 1989 (- 0.28) 2011 (- 0.60) 1990 (+ 0.31) 1995 (- 0.14) 1996 (- 0.27) 1998 (- 0.07) 2000 (- 0.29) 2001 (+ 0.02) 2003 (+ 0.27) 2004 (+ 0.52) 2005 (+ 0.14) 2006 (+ 0.73) 2009 (+ 0.88) 2012 (+ 0.48) *Values are provided in parentheses

(2008) deems 1992–94 as neutral years. However, our analysis shows 1992 as an El

Niño year, associated with the greatest NTC and PDI values reflected by the highest number (24) of storm genesis with the longest duration spent at each (tropical storm, hurricane, and major hurricane) stage. In addition, the extension of our data from that of

Lupo et al. (2008) leads to the identification of three La Niña years (2008, 2010, and

2011).

42 2.4.2 Time series of storm activity and intensity

TC measures of frequency, intensity, and duration are summarized as metrics of seasonal storm activity (NTC) and intensity (PDI). For comparison purposes, the summary statistics of storm count and lifetime (in days) are binned into storm strengths of tropical/named storm, hurricane, and major hurricane in EDR (Figure 2-2) and WDR

(Figure 2-3). Both EDR and WDR experienced heightened storm activity in terms of the number of tropical storms developed from the early 1980s to 1992 and 1993, respectively, when unusually high proportions of tropical storms strengthened into major hurricanes. However, while it is followed by 1994 and 1999 when no EDR hurricanes grew into major hurricanes, no WDR storms of even NS strength were found in 1996.

Overall, the entire region experiences an annual TC count of 15 storms. When

MDR is subdivided, while similar seasonal frequencies of NS, H, and MH (Table 2-2) are shown between EDR and WDR, the origin of a few WDR storms was pertained within

WDR boundary. Instead, an average of four WDR storms were initiated in EDR and migrated into WDR, while reaching their maximum intensities. Since these EDR- originated storms establish their peak intensity in WDR, more than half of them strengthened into hurricanes or major hurricanes, thereby greatly extending the length of storm duration during all three storm stages.

The time series of the seasonal overall TC activity (NTC; Figure 2-4) and TC intensity (PDI; Figure 2-5) are displayed from 1971 to 2012. Although the Mann–

Kendall test confirms the Theil–Sen slope values of MDR storm measurements to be statistically insignificant, negative trends of NTC and PDI imply that there are decreasing number of storms with long duration and fewer intense storms of long duration,

Figure 2-2 Time series for the seasonal a) frequency (TS, H and MH) and b) duration (TSD, HD and MHD) at EDR from 1971-2012

Figure 2-3 Time series for the seasonal a) frequency (TS, H and MH) and b) duration (TSD, HD and MHD) at WDR from 1971-2012.

45 Table 2-2: TC activity in the eastern North Pacific basin in the main (eastern and western) development regions from 1971-2012; while seasonal averages from 1981-2010 are incorporated into the derivation of NTC.

NTC 42-year (1971-2012) Average 30-year (1981-2010) Average Measures MDR EDR WDR MDR EDR WDR NS 15 8.0 7 15.4 8.0 7.5 NSD 60.4 25.2 35.3 62.1 25.6 36.5 H 8.5 4.3 4.2 8.4 4.1 4.3 HD 24.8 9.3 15.5 25.5 9.5 16.0 MH 3.9 2.0 2 3.9 1.9 2.1 MHD 7.0 2.4 4.5 7.5 2.5 5.0

Figure 2-5 Time series charts for PDI (1.0 × 107.m3 s-3) at a) MDR with horizontal dashed lines showing values at the 25th and 75th percentiles and when TC development region is subdivided into b) EDR/WDR during 1971-2012. Dashed vertical lines are colored to distinguish El Niño (red), La Niña (blue) and neutral (grey) years, in consistent with a). statistically insignificant, negative trends of NTC and PDI imply that there are decreasing number of storms with long duration and fewer intense storms of long duration, respectively (Figures. 2-4a and 2-5a). Such a recent decrease in the ENP storm development is consistent with findings from Schultz (2007), Lupo et al. (2008), and Wu et al. (2008), but opposite from that experienced in adjacent regions of the North Atlantic

(Webster et al. 2005; Collins 2010) basin, marked by consecutive (2004 and 2005) peaks of North Atlantic hurricane seasons. Overall, this continuous increase of TC activity

(Steenhof and Gough 2008) and TC intensity (Wu et al. 2008) in the North Atlantic basin over a long time record is widely known to be enhanced by local environmental

48 conditions of thermodynamic and dynamic factors (Emanuel 2005; Goldenberg and

Shapiro 1996; Goldenberg et al. 2001).

When MDR is subdivided at 112°W (Ralph and Gough 2009), while negative trends of both NTC (Figure 2-4b) and PDI (Figure 2-5b) values are still maintained in

WDR, NTC (Figure 2-4a) and PDI (Figure 2-5a) measures for EDR storms have decreased at a slower pace for NTC, with PDI experiencing virtually no change through the 42-yr period. Though trends of the regional contrast are determined as statistically insignificant, there is growing evidence that WDR storm activity is regulated by environmental controls that are different from that in EDR (Collins and Mason 2000;

Ralph and Gough 2009).

One such large-scale ocean–atmosphere climate phenomenon that can affect the interannual variability of both NTC (Figure 2-4) and PDI (Figure 2-5) is the influence of

ENSO events. In general, El Niño years are identified with greater values of storm indices derived from seasonal TC activity. On the other hand, La Niña years are associated with some of the least active TC seasons. From Table 2-3, when seasonal mean values of six NTC measures are binned to the appropriate ENSO events, both the frequency and duration measures are consistently highest during El Niño years, follow by neutral and La Niña years. However, as there is one year (1996) without any WDR storm formation, direct spatial comparison between the six NTC measures under ENSO influences would violate fundamental assumptions (e.g., data normality) of parametric statistics. Presumably, an ENSO-induced difference is greater in WDR where drastic meteorological shifts resonate more with ENSO phase changes (Collins 2007). Although such an assessment of ENSO influences on ENP storms is only qualitative, statistical

49 approaches attributing ENSO events to the regional TC activity and TC intensity using metrics that include storm frequency, maximum sustained wind speed, and duration will follow in the next subsection.

Table 2-3: Mean seasonal (1971-2012) values of six NTC measures subdivided into EDR and WDR during three ENSO conditions. Region NTC measures El Niño La Niña Neutral NS 8.1 7.9 8.1 NSD 27.5 25.4 24.0 H 4.4 4.2 4.3 HD 10.8 8.9 8.8 MH 2.8 1.6 1.8 EDR MHD 3.2 2.0 2.3 NS 8.7 5.1 7.1 NSD 46.7 27.6 33.6 H 5.3 3.2 4.2 HD 22.7 11.4 14.0 MH 2.8 1.5 1.9 WDR MHD 7.5 2.7 4.0

2.4.3 Statistical analyses

2.4.3.a Two-way analysis of variance

The mean differences of the NTC and PDI are compared using a two-way

ANOVA under two factors: temporal classification of annual ENSO conditions and spatial division of MDR (Tables 2-4 and 2-5). The F test confirms that temporal differences of NTC and PDI can be attributed to ENSO variation though we do not know under which ENSO conditions are both indices different. Since the F test alone cannot detect exactly which ENSO pairing(s) is attributed to a significant difference between both storm measures, Tukey’s method of multiple comparisons identifies NTC and PDI

50 Table 2-4: A two-way analysis of variance is tested for effects of ENSO and regional division on the NTC activity index (%) a) Response: NTC (%) Df Sum Sq Mean Sq F value Pr(>F) ENSO 2 33418 16709.0 9.7441 <0.0001 Region 2 647 323.7 0.1888 0.828 ENSO:Region 4 3751 937.7 0.5469 0.702 Residuals 117 200629 1714.8

Tukey multiple comparisons of means ENSO Differences p-va lues

El Niño- La Niña 44.889 <0.0001 El Niño-Neutral 32.524 0.002 La Niña-Neutral -12.365 0.367

Region

Differences p-values

EDR-MDR 3.347 0.927 EDR-WDR 5.510 0.815 MDR-WDR 2.163 0.969

ENSO:Region Differences p-values La Niña:EDR-El Niño:EDR -24.668 0.920 Neutral:EDR-El Niño:EDR -21.081 0.919 Neutral:EDR-La Niña:EDR 3.586 1.000 La Niña:MDR-El Niño:MDR -46.899 0.228 Neutral:MDR-El Niño:MDR -34.297 0.431 Neutral:MDR-La Niña:MDR 12.602 0.997 La Niña:WDR-El Niño:WDR -63.100 0.024 Neutral:WDR-El Niño:WDR -42.193 0.170 Neutral:WDR-La Niña:WDR 20.907 0.922

51 Table 2-5: A two-way analysis of variance is tested for effects of ENSO and regional division on the PDI (scale of 1.0 × 107 m3 s-3) b) Response: PDI (107 m3 s-3) Df Sum Sq Mean Sq F value Pr(>F) ENSO 2 1.846 0.923 7.962 0.001 Region 1 1.277 1.277 11.015 0.001 ENSO:Region 2 0.641 0.320 2.762 0.069 Residuals 78 9.045 0.116

Tukey multiple comparisons of means ENSO Differences p-values El Niño- La Niña 0.403 0.001 El Niño-Neutral 0.304 0.004 La Niña-Neutral -0.099 0.528

Region Differences p-values EDR-WDR 6.871 0.001

ENSO:Region Differences p-values La Niña:EDR-El Niño:EDR -0.165 0.887 Neutral:EDR-El Niño:EDR -0.127 0.925 Neutral:EDR-La Niña:EDR 0.038 1.000 La Niña:WDR-El Niño:WDR -0.642 0.001 Neutral:WDR-El Niño:WDR -0.482 0.005 Neutral:WDR-La Niña:WDR 0.161 0.818

to be only statistically different for pairings of El Niño–La Niña and El Niño–neutral comparisons at p < 0.05 (Table 2-4). In previous work, though more El Niño storms became intense hurricanes, the analysis of monthly TC statistics does not yield a significantly greater TC activity during El Niño years (Lupo et al. 2008). In contrast, our results demonstrate that both storm measurements are significantly different under both

ENSO pairings, which are also acquainted to influence the mean genesis location of ENP

52 storms (Irwin and Davis, 1999).

The second main effect of regional difference demonstrates that only seasonal

PDI values are statistically different (p < 0.05) between EDR and WDR. A follow-up regional comparison of TC measures using Tukey’s method is shown only as a confirmatory analysis since there are only two levels (EDR and WDR) of MDR subdivisions. Regional variations of TC intensity (PDI) due to ENSO events further confirm that MDR should be viewed as two distinct development regions (Collins and

Mason 2000; Ralph and Gough 2009). While the average WDR storm intensity (PDI) is noticeably higher during El Niño years, the mean difference of the overall TC activity

(NTC) is not detected (Table 2-4), possibly because of the derivation of seasonal NTC values. The derivation of the NTC index is sensitive to the seasonal variability of relative measures of all TC components of the index to the respective 1981–2010 climatological means (Table 2-2). Because the seasonal (1981–2010) averages of storm duration measures are distinctly higher in WDR, its seasonal NTC values during any given year can be comparably similar, statistically indifferent, to EDR values. For instance, 2009 saw the WDR storm duration (95 days) close to be 3 times more than that of EDR (36 days), with a comparable difference for the total storm count. However, owing to the fact that many EDR-originated storms spend the majority of lifetimes in WDR, because long- term averages of all three NTC duration measures in WDR are substantially greater, the regional difference of NTC values in 2009 is virtually the same. As such, even if there is a substantial spatial difference of any individual NTC component in any given year, NTC is collectively bounded by all six NTC parameters’ 30-yr averages (Table 2-2).

Since there is a combination of effects considered for statistical comparison, an

53 interaction effect of the regional and ENSO-induced temporal (ENSO region; Tables 2-4 and 2-5) factors on both NTC and PDI values is assessed. An interactive outcome occurs when rankings of relative NTC and PDI values based on ENSO conditions shift in ranks between regions. Not only do our results show that the average measures of both storm indices are greatest during El Niño years, followed by neutral and La Niña years (Figures.

2-6a and 2-6b), but also, both MDR subdivisions maintained such relative ranks of storm metrics due to ENSO modification. This consistency of the result in the F test (p < 0.05) demonstrates the absence of an interaction effect of a combined ENSO-region factor across NTC (Table 2-4) and PDI (Table 2-5). However, Tukey’s multiple comparisons of the interaction effect demonstrate significant differences between NTC and PDI resulting from different ENSO comparisons that are not equally distributed within MDR subdivisions. In particular, the differences of NTC and PDI values within the ENSO pairing of El Niño–La Niña are maintained only in WDR.

2.4.3.b Correlation analysis

Relatively greater NTC and PDI values during El Niño years, followed by neutral and La Niña years suggest that MEI values are positively linked to both indices.

However, the use of the parametric correlation analysis is deemed appropriate only when both storm metrics have been normalized. When TC data are binned into EDR and

WDR, the data distribution of each storm measure varies. Visual inspections of Q–Q plots show that an absence of WDR storm activity may have caused non-normal

57 distributions of both NTC (Figure 2-7a) and PDI (Figure 2-7b) indices. In particular,

WDR storm development during 1996 went completely dormant with no storms generated west of 112°W. After both datasets were transformed, distributions of both

NTC (Figure 2-7c) and PDI (Figure 2-7d) are closer to the line of normality. Under the null hypothesis that the sampled data come from normal distributions, a subsequent

Shapiro–Wilk normality test was applied to test against the assumption of data normality.

The results of the Shapiro–Wilk test calculate a W statistic that is used to calculate the correlation of the sampled data after they have been ordered and standardized and what the samples would have been if they were drawn from a normal distribution and ordered.

Significance of the Shapiro–Wilk test subjectively determines both WDR storm metrics were normalized after data transformations.

A parametric testing of correlation analysis is used to highlight the extent of NTC

(Figure 2-8) and PDI (Figure 2-9) relationships with MEI. Time series of residuals

(observed -correlated) are plotted with their correlograms for both WDR storm metrics from 1971 to 2012 (Figure 2-10). To test for the presence of serial independence of the residuals, subsequent Ljung–Box test statistics (chi-squared values) show that there is no statistical evidence (p < 0.05) for nonzero autocorrelations. Positive associations for both

MDR storm measurements with MEI are determined statistically significant for PDI, not

NTC (Table 2-6). However, when MDR is subdivided, such a positive linear relationship is significant for both TC indices for WDR storms only (p < 0.05). The presence of such statistical significance confirms that the ENSO influence on TC activity is most pronounced in WDR, even though more EDR storms are generated during the local hurricane season. A relatively higher r2 in WDR indicates a greater NTC and PDI

Figure 2-8 Linear regressions of MEI values with data transformed for NTC at a) MDR, b) EDR and c) WDR under three ENSO conditions during 1971-2012. Pearson coefficient of correlation (r) and correlation of variation (r2) are provided in Table 2-4.

NTC PDI MDR EDR WDR MDR EDR WDR r 0.28 0.12 0.35 0.32 0.17 0.35 r2 0.08 0.01 0.12 0.10 0.03 0.12

Figure 2-10 Time series of residuals (observed-predicted) NTC values and correlograms (autocorrelation plots) are evaluated with Ljung-Box test statistic (chi-squared value), with its p-value, based on the first 20 lags at WDR for a) NTC and b) PDI (scale of 1.0 × 107 m3 s-3) during 1971-2012.

variability that is explained by MEI values. Residuals (observed - predicted) for the

WDR storm metrics are plotted in Figure 2-10; while, subsequent Ljung–Box test statistics (chi-squared values) show that there is no statistical evidence (p < 0.05) for nonzero autocorrelations.

2.5 Discussion

The influence of ENSO on different parameters of MDR storm activity at the

ENP is statistically addressed in this study. Previously, a lack of such statistical evidence in the basin could be attributed to inconsistent classification schemes in distinguishing

ENSO phases of El Niño, La Niña, and neutral conditions, in addition to the absence of

63 reliable TC data on a longer time scale (Collins 2007; Schultz 2007; Lupo et al. 2008).

Though maximum NTC and PDI values in 1992 could be related to a greater midlevel moist static energy during the peak (July–September) hurricane season (Wu and Chu

2007), the recent 2007, 2008, and 2010 La Niña years are marked as the three most meager TC years of the entire data analysis (1971–2012). NTC and PDI values for these three years were well below seasonal averages of 97.56% and 1.20× 107 m3 s-3, respectively, over 10 La Niña years. In terms of the same storm measures of TC activity and intensity, when MDR is subdivided, 2011 is found to exhibit the second least-active

TC season of all 10 La Niña years in WDR.

Alternative considerations of MDR divisions from Irwin and Davis (1999) and

Collins and Mason (2000) also have attributed to the detection of the difference of regional TC sensitivity to ENSO oscillation. The MDR division at 112°W depicts a substantially greater TC intensity in WDR (Ralph and Gough 2009), where many

EDR storms migrate and achieve their peak intensities prior to dissipation. While there are more storms originated from EDR, its overall TC activity and TC intensity are less influenced by environmental changes on a seasonal scale; rather, it is WDR storms that are more responsive to ENSO-induced environmental influences (Collins and Mason

2000). Such a regional difference of storm sensitivity to changes of the external environment is also consistent with that of the original longitudinal division of MDR storm frequency of at least tropical storm strength into eastern and western geographical boundaries (Collins and Mason 2000).

Another important aspect is revealed when focusing on the correlation strengths of MEI with NTC and PDI in MDR and within its subdivision. Overall, PDI is more

64 sensitive to ENSO influences than NTC (Table 2-6). Among factors of storm intensity

(PDI1), duration (PDI2) and count (PDI3attributed to PDI variability (Figure 2-11), only storm intensity (PDI1) demonstrates a significant (p < 0.05) effect on the correlation of

PDI with ENSO signal (Table 2-7). Such a result is contrary to northwestern Pacific storms where Camargo and Sobel (2005) found that only the duration variability attributes significantly to the ENSO signal on PDI. However, when MDR is subdivided, seasonal variations of all three storm factors are significantly sensitive to the ENSO signal on WDR storm intensity, while their influences are little to minimal in EDR. As the derivation of PDI is directly related to the total storm count, such associations between (PDI1 versus MEI) show comparable strength with the relationship of PDI and

MEI (Table 2-6). Overall, the subdued r2 of ENSO with PDI, and also NTC, suggests other local environmental forcings are involved at explaining the annual variation of seasonal TC measures.

Figure 2-11 Boxplot summaries for data distributions of a) PDI1, b) PDI2 and c) PDI3 (scale of 1.0 × 107 m3 s-3) at MDR, EDR and WDR. Grey diamonds represent the regional averages

66 Table 2-7: Correlations of PDI1, PDI2 and PDI3 with bimonthly (May-November) MEI values in MDR, EDR and WDR from 1971-2012. Bolded values are statistically significant at p<0.05.

PDI1 PDI2 PDI3 MDR 0.33 0.19 0.27 EDR 0.18 -0.06 -0.02 WDR 0.34 0.30 0.37

Changes to sea surface temperature (SST) due to ENSO modification can be linked to the atmospheric alteration of convective mechanisms for fueling storm development (Collins 2007; Klotzbach and Blake 2013). One of the crucial indicators in determining ENSO signal (i.e., SST) has been well known for its local effect on the ENP storms (Ralph and Gough 2009). Since WDR ocean temperatures are closer to the

26.5°C threshold for TC formation (Gray 1968), a warm anomaly arising from ENSO changes may trigger the genesis of more storms and vice versa when a cold SST anomaly occurs. In particular, fluctuating SST could reinforce changes to the availability of atmospheric moisture and dictate the seasonal variability of WDR storm activity (Collins and Roache 2011). While the seasonal sea surface temperature in EDR is already well above the critical threshold for storm formation (Ralph and Gough 2009), fluctuation of the sea surface temperature–dependent MEI may be less crucial to the region’s overall

TC activity and storm intensification. These thermodynamic controls offer explanations to why dramatic differences of the two indices appear only when MDR is subdivided.

The fact that the ENP storm activity and intensity in WDR are more sensitive to changes of ENSO conditions is geographically consistent with neighboring storm development regions. In the Southern Hemisphere, the strength of the correlation

67 between southwestern Pacific storm frequency and the Southern Oscillation index gradually shifts in direction westward from the 170°E longitude of boundary division

(Basher and Zheng 1995). This means that in contrast to the ENP basin, more South

Pacific storms develop during the La Niña phase. Similarly, at the adjacent northwestern

Pacific basin, there are more La Niña storms that reached the strength of a typhoon, equivalent to a hurricane, at its easternmost boundary near the International Date Line

(Chan 1985; Wang and Chan 2002). Such a close proximity to the neighboring WDR could result into a greater storm development, which is directly observed through the eastward extension of the monsoon trough crossing the Central Pacific boundary during extreme El Niño years (Clark and Chu 2002).

Apart from its direct influences on moist static stability (Malkus and Riehl 1960),

ENSO is also indirectly acquainted with changes in the wind direction of the tropospheric profile. As demonstrated in the North Atlantic basin, ENSO is known to control wind shear in other TC basins where such a dynamical factor is primarily responsible for TC intensity (Gray 1984a; Jones and Thorncroft 1998; Landsea 2000; Goldenberg et al.

2001) and is noted to be highly correlated with the primary ENP storm genesis region

(Table 2-6). Because low wind shear is observed over much of the MDR in 1992 during such a strong El Niño event, it could also play an important role at reinforcing WDR convection through the reduction of wind shear by enhancing the upper-level westerly wind and the lower-level easterly wind (Jones and Thorncroft 1998). When wind shear and other dynamical factors are included in the seasonal genesis parameter (Gray 1977), a greater number of El Niño storms is observed to coexist with higher dynamical potential values than during La Niña conditions (Clark and Chu 2002).

68 2.6 Conclusion

The longitudinal division of MDR has greatly facilitated our understanding of environmental influences at the ENP basin. We have statistically evaluated and compared the impact of ENSO on TC activity and intensity between the neighboring subdivisions of EDR and WDR. In combination with the classification of years of El

Niño, La Niña, and neutral conditions, the spatial and temporal influences on the ENP storm activity and intensity are quantitatively resolved over a 42-yr period. Although the physical forcings accompanying ENSO phase changes are not directly addressed, previous studies have identified a combination of local thermodynamic and dynamic factors in speculating El Niño–induced shifts of more intense storms with longer lifetimes.

Measures of the overall seasonal TC activity and intensity are expressed empirically as indices of NTC and PDI, respectively. Accounting for the nonnormality of storm data in WDR where fewer storms are observed seasonally, both TC indices have been transformed and tested for statistical differences between every pairwise combination of three ENSO phases. Overall, both indices are only proven statistically different between El Niño and La Niña years. When MDR is subdivided, such a contrast is only maintained in WDR. This difference of the regional sensitivity to ENSO is supported by results of the correlation analysis, which demonstrates the strength of correlations in WDR to be stronger for the seasonal TC activity than TC intensity between 1971 and 2012. Statistical comparisons of regional TC indices at different

ENSO conditions validate previous findings that the effects of environmental influences over ENP storms would have been overlooked if MDR was not subdivided (Collins and

69 Mason 2000; Ralph and Gough 2009).

Between storm parameters of frequency, intensity, and duration, only the seasonal variability of TC intensity contributes significantly to the ENSO signal on PDI.

However, this independent analysis on the influence of PDI due to ENSO-induced storm measurements is also subject to regional variation. When comparing between MDR subdivisions, PDI is found to be significantly sensitive to seasonal fluctuations for all

WDR storm parameters, with storm count being relatively more important.

The findings generally underscore the societal importance of ENSO on TC activity and intensity. Although most of the ENP storm development initiated in EDR, many storms did not reach maximum intensity until entering WDR. During El Niño years, there are more WDR storms with higher intensity and longer lifetimes that develop than during La Niña years. A direct consequence is that as the storms are transitioned from EDR, ENSO influences on seasonal WDR environmental conditions could fuel further storm development and prolong storm tracks affecting islands in the Central

Pacific basin. Hence, the determination and incorporation of ENSO as a predictor for short-term, seasonal storm forecasts could prove to be a practical tool that better anticipates the seasonal outlook of WDR storm activity and intensity and the potential

TC-inflicted damages upon landfalls. Since North Atlantic storm damage and the relationship with the oscillation of ENSO phases have already been well documented

(Hebert et al. 1997; Pielke and Landsea 1999), future work on such a data archive should include additional input on the relationship of ENSO events with the societal cost associated with ENP storm landfalls on the Pacific islands and the continental U.S..

70 2.7 References

Basher, R. E. and X. Zheng, 1995: Tropical cyclones in the Southwest Pacific: Spatial patterns and relationships to Southern Oscillation and sea surface temperature. Journal of Climate, 8, 1249-1260.

Bove, M. C., J. B. Elsner, C. W. Landsea, X. Niu, and J. J. O’Brien, 1998: Effect of El Niño on U.S. landfalling hurricanes, revisited. Bulletin of the American Meteorological Society, 79, 2477–2482.

Camargo, S. J., and A. H. Sobel, 2005: Western north Pacific tropical cyclone intensity and ENSO. Journal of Climate, 18, 2996–3006.

Chan, J. C. L., 1985: Tropical cyclone activity in the northwest Pacific in relation to the el Nino/Southern oscillation phenomenon. Monthly Weather Review, 113(4), 599-606.

Changnon, S. A., 1999: Impacts of 1997—98 El Niño Generated Weather in the United States. Bulletin of the American Meteorological Society, 80, 1819–1827.

Chu, P.-S., 2004: ENSO and tropical cyclone activity. Hurricanes and Typhoons: Past, Present, and Potential, R. J. Murnane and K. B. Liu, Eds., Columbia University Press, 297–332.

Chu, P. -S., and J. Wang, 1997: Tropical cyclone occurrences in the vicinity of Hawaii: Are the differences between El Niño and non-El Niño years significant? Journal of Climate, 10(10), 2683-2689.

Clark, J. D. and Chu, P.-S. 2002: Interannual variation of tropical cyclone activity over the central North Pacific. J. Meteor. Soc. Japan, 80, 403–418.

Collins, J., 2010: Contrasting High North-East Pacific Tropical Cyclone Activity. Southeastern Geographer, 50, 83-98.

Collins, J. M., and I. M. Mason, 2000: Local environmental conditions related to seasonal tropical cyclone activity in the Northeast Pacific basin. Geophysical Research Letters, 27, 3881-3884.

71 Collins, J. M., and D. R. Roache, 2011: The 2009 Hurricane Season in the Eastern North Pacific Basin: An Analysis of Environmental Conditions. Monthly Weather Review, 139(6), 1673-1681.

Emanuel, K. A. (2005). Increasing Destructiveness of Tropical Cyclones over the Past 30 Years. Nature , 436, 686-688.

Frank, W.M., and G. S. Young, 2007: The Interannual Variability of Tropical Cyclones, Monthly Weather Review, 135, 3587–3598.

Goldenberg, S. B. and L. J. Shapiro, 1996: Physical mechanisms for the association of El Niño and West African rainfall with Atlantic major hurricane activity. Journal of Climate, 9, 1169–1187.

Goldenberg, S. B., C. W. Landsea, A. M. Mestas-Nuñez, and W. M. Gray, 2001: The recent increase in Atlantic hurricane activity: Causes and implications. Science, 293, 474–479.

Gray, W. M., 1968: The Origin of Tropical Disturbances and Storms. Monthly Weather Review., 96, 669-700.

Gray, W. M., 1977: Tropical cyclone genesis in the western North Pacific. Journal of the Meteorological Society of Japan, 55, 465–482.

Gray, W. M., 1984a: Atlantic seasonal hurricane frequency: Part I. El Nino and 30 mb quasi-biennial oscillation influences. Monthly Weather Review, 112, 1649-1668.

Gray, W. M., 1984b:. Atlantic seasonal hurricane frequency. Part II: Forecasting its variability. Monthly Weather Review, 112, 1669–1683.

Gray, W. M., C. W. Landsea, P. W. Mielke Jr., and K. J. Berry, 1994: Predicting Atlantic basin seasonal tropical cyclone activity by 1 June. Weather and Forecasting, 9, 103-115.

Hebert, P. J., J. D. Jarrell, and M. Mayfield, 1997: The deadliest, costliest, and most intense United States hurricanes of this century (and other frequently requested hurricane facts). NOAA Tech. Memo. NWS TPC-1, Miami, FL, 30 pp. [Available from National Hurricane Center, 11691 SW17th Street, Miami, FL 33165.]

Helsel, D. R., and R. M. Hirsch, 1991: Statistical methods in water resources. Techniques of Water Resources Investigations of the USGS, 266 pp.

72 Jones, C. G. and C. D. Thorncroft, 1998: The role of El Niño in Atlantic tropical cyclone activity. Weather, 53(9), 324-336.

Klotzbach, P. J., 2011: El Niño-Southern Oscillation's Impact on Atlantic Basin Hurricanes and U.S. Landfalls. Journal of Climate, 24, 1252-1263.

Klotzbach, P. J., 2012: El Niño-Southern Oscillation, the Madden-Julian Oscillation and Atlantic basin tropical cyclone rapid intensification. Geophysical Research Letters, 117, D14104.

Klotzbach, P. J., and S. E. Blake, 2013: North-Central Pacific Tropical Cyclones: Impacts of El Niño–Southern Oscillation and the Madden–Julian Oscillation. Journal of Climate, 26, 7720-7733.

Landsea, C.W., 2000: El Niño/Southern Oscillation and the Seasonal Predictability of Tropical Cyclones. In El Niño and the Southern Oscillation: Multiscale Variability and Global and Regional Impacts, H.F. Diaz and V. Markgraf, Cambridge University Press, 149-181.

Lupo, A. R., T. K. Latham, T. H. Magill, J. V. Clark, C. J. Melick, and P. S. Market, 2008: Interannual Variability of Hurricane Activity in the Atlantic and East Pacific Regions. National Weather Digest, 32, 119-133.

Malkus, J. S. and H. Riehl, 1960: On the dynamics and energy transformations in steady- state hurricanes. Tellus, 12, 1-20.

Molinari, J., D. Vollaro, S. Skubis, and M. Dickinson, 2000: Origins and mechanisms of eastern Pacific tropical cyclogenesis: A case study. Monthly Weather Review, 128, 125– 139. Nicholls, N., 1979: A possible method for predicting seasonal tropical cyclone activity in the Australian region. Monthly Weather Review, 107(9), 1221-1224.

Nicholls, N., 1992: Recent performance of a method for forecasting Australian seasonal tropical cyclone activity. Australian Meteorological Magazine, 40, 105-110.

Pielke Jr., R. A., and C. N. Landsea, 1999: La Niña, El Niño, and Atlantic hurricane damages in the united states. Bulletin of the American Meteorological Society, 80(10), 2027-2033.

Ralph, T. U., and W. A. Gough, 2009: The influence of sea-surface temperatures on Eastern North Pacific tropical cyclone activity. Theoretical and Applied Climatology, 95, 257-264.

Rasmusson, E. M. and T. H. Carpenter, 1982: Variations in tropical sea surface temperature and surface wind fields associated with the Southern Oscillation/El Niño, Monthly Weather Review, 110, 354–384.

73 Ropelewski, C. F. and P. D. Jones, 1987: An extension of the Tahiti-Danvin Southern Oscillation index. Monthly Weather Review, 115, 2161-2165.

Schultz, L., 2007: Some Climatological Aspects of Tropical Cyclones in the Eastern North Pacific. National Weather Digest, 32, 45-54.

Simpson, R. H., 1974: The hurricane disaster potential scale. Weatherwise, 27, 169-186.

Sprent, P., 1989: Applied nonparametric statistical methods. Chapman and Hall, London, 259 pp.

Steenhof, P. A. and W. A. Gough, 2008: The impact of tropical sea surface temperatures on various measures of Atlantic tropical cyclone activity. Theoretical and Applied Climatology, 92, 249-255.

Trenberth, K. E., 1997: The definition of El Niño. Bulletin of the American Meteorological Society, 78 (12), 2771-2777.

Wang, B., and J. C. L. Chan, 2002: How strong ENSO events affect tropical storm activity over the western north Pacific. Journal of Climate, 15(13), 1643-1658.

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.

Wolter, K., and M. S. Timlin, 1998: Measuring the strength of ENSO events - how does 1997/98 rank? Weather, 53, 315-324.

Wu, L., B. Wang, and S. A. Braun, 2008: Implications of tropical cyclone power dissipation index. International Journal of Climatology, 28, 727-731.

74 3 Chapter 3: The Impact of El Niño-Southern Oscillation on Eastern North Pacific Tropical Cyclone Tracks

3.1 Abstract

The impact of El Niño-Southern Oscillation (ENSO) on the genesis and trajectory of eastern North Pacific tropical cyclones from 1971-2013 is investigated. During the local hurricane season (May 15 – November 30), storms are grouped into El Niño, La

Niña and neutral phases based on the Multivariate ENSO Index. When compared to the neutral and La Niña phases, it was found that although El Niño storms typically form closer to the equator, their downgradation points are located further poleward. In addition to a significantly greater latitudinal movement, the longitudinal/westward transport of ENP storms is also greater during El Niño years.

When storms are longitudinally binned into eastern (10–20°N and 85–112°W) and western (10–20°N and 112–140°W) development regions (EDR, WDR), genesis and trajectory of WDR storms are found to be more sensitive to ENSO influences. Even though zonal distances between locations of storm genesis and maximum storm intensity tend to be strongly modulated by ENSO influences, the latitudinal extension of the WDR storm track is significantly greater during El Niño years. Since WDR storms tend to experience a longer northward trajectory, it was found that they are also more likely to dissipate outside of the northernmost boundary. Though most WDR storms establish maximum lifetime intensities further westward in the main development region (MDR), when they landfall, they are observed to have a greater influence in the southwest U.S. than EDR storms.

75 3.2 Introduction

El Niño-Southern Oscillation (ENSO) has a profound impact over the global climate through its long-range influence over local environmental conditions (Quinn et al., 1981; Changnon, 1999; Higgins et al., 1999; Sobel and Camargo, 2005). Such

ENSO-generated environmental variation is critically important to the regional tropical cyclone (TC) development and its track movement (Camargo et al., 2007; Klotzbach,

2012; Jin et al., 2014). For instance, ENSO-related rainfall anomalies in parts of Asia are largely regulated by TC activity, which in turn is affected by transient changes in the intensities of low-level wind flows and tropical convection in the equatorial Pacific

(Ashok et al., 2007; Pradhan et al., 2011). In combination with other large-scale circulation anomalies, these ENSO-induced variations contribute greatly to the seasonal

TC variability and its storm impacts (Chen and Tam, 2010; Chen, 2011; Pradhan et al.,

2011). While differences of ENSO-induced climatic patterns largely depend on spatially distinctive sea surface temperature anomalies along the equatorial Pacific (Rasmusson and Carpenter, 1982; Ashok et al., 2007), ENSO fluctuation is commonly linked to variability of TC landfalls through modifications of TC formation and downgradation locations (Evans and Allen, 1992; Irwin and Davis, 1999; Saunders et al., 2000; Camargo et al., 2008; Kim et al., 2009; Pradhan et al., 2011).

Though ENSO conditions are generated in the equatorial Pacific Ocean, they are capable of generating strong local and remote impacts in multiple TC basins (Alexander et al., 2002). In the North Atlantic basin, TC reductions are remotely affected by El Niño condition through enhancement of local wind shear (Jones and Thorncroft, 1998;

Goldenberg and Shapiro, 1996; Aiyyer and Thorncroft, 2006) and relative humidity in the

76 mid-troposphere (Camargo et al., 2007; Finlayson, 2015). As a result, less (more) U.S. hurricane landfalls are observed (Klotzbach, 2011) in reducing (elevating) monetary damages experienced during El Niño (La Niña) years (Hebert et al., 1997; Pielke and

Landsea, 1999). The lower hurricane damage during El Niño years can be attributed to fewer landfall hurricanes threatening the U.S. East Coast due to stronger steering winds which direct storms towards the North Atlantic Ocean (Colbert and Soden, 2012).

During this warm (El Niño) phase of ENSO, the weakening of Trade Winds due to the abnormally warm eastern Pacific Ocean is accompanied by enhanced tropical convection

(Jin et al., 2014) and reduced wind shear (Wu and Chu, 2007; Zhang and Wang, 2015).

These El Niño-induced effects also likely trigger an eastward displacement of the monsoon trough, causing an above-normal TC season in the Central Pacific basin (Wu and Lau, 1992; Chu and Wang, 1997). At the eastern tropics of the North Pacific basin, a similar effect is extended into its western boundary (Jien et al., 2015), possibly in conjunction with elevations of thermodynamic forcing (Ralph and Gough, 2009) and relative humidity (Collins and Mason, 2000; Collins and Roache, 2011).

TC development in other basins has been directly linked to ENSO fluctuation, specifically through shifts in their primary areas of storm genesis. Though Chan (1985) and Wang and Chan (2002) concluded the number of typhoons in the western Pacific is modulated by ENSO, Lander (1994) did not find that the total number of storms changed.

However, it was generally agreed that the main genesis region of western Pacific storms typically shifts southeast (northwest) during El Niño (La Niña) events (Chan, 1985;

Lander, 1994; Wang and Chan, 2002). This movement in the mean TC genesis location is mainly guided by the ENSO-induced shift in the local monsoon trough (Lander, 1994;

77 Chen et al., 1998; Chia and Ropelewski, 2002). In the southern hemisphere, the genesis location of southwest Pacific storms also experiences a similar eastward shift during El

Niño years (Hastings, 1990; Basher and Zheng, 1995). Overall, there is strong evidence that during El Niño years, areas of TC genesis at the western Pacific Ocean are displaced eastward, generally following areas of positive SST anomaly. While in the south Indian

Ocean, its main cyclogenesis area often drifts closer to the western Australian coastline during La Niña periods (Kuleshov et al., 2008). Due to the proximity of the main TC genesis areas, landfalls are more frequent during El Niño (La Niña) conditions along the western perimeter of Pacific Ocean (Australia).

In the eastern North Pacific (ENP) basin, while more ENP storms with greater longevity and intensity are more likely to form during El Niño years (Landsea and Gray,

1989; Jien et al., 2015), there remains a regional contrast in the distribution of TC genesis locations. In previous work, when main development region (MDR) of TC activity was divided into an eastern and western development region (EDR and WDR, respectively), the relationship between seasonal environmental factors and TC activity is only significant in the WDR (Collins and Mason, 2000; Ralph and Gough, 2009). While many WDR storms were triggered at EDR (Jien et al., 2015), such a relationship was not found in the EDR where a higher density of storm tracks is observed. As such, ENSO influence on ENP storm genesis and downgradation locations and track movement should consider subdividing MDR as two separate regions.

The current understanding of the geographic positioning of ENP storm tracks has not considered the spatial stratification of ENP storms into EDR and WDR. While a westward expansion of ENP storm tracks was initially detected during the El Niño phase

78 (Schroeder and Yu, 1995; Kimberlain, 1999), recent analysis of the current TC climatology showed a more active WDR storm season, highlighted by storms with greater intensity (Jien et al., 2015). When ENSO signatures were defined using the

Southern Oscillation Index, the mean longitude of TC genesis (downgradation) points is

5.7° (8.5°) westward during the El Niño phase relative to the La Niña phase (Irwin and

Davis, 1999). However, less is known of ENSO influence on ENP storms’ genesis location and track movement as a function of the MDR sub-regions. Although neither the latitudinal nor longitudinal length of TC track is preferentially extended by either

ENSO phase, storm development outside the ENP primary genesis area is strongly modulated by ENSO phases (Irwin and Davis, 1999). This ENSO-modulated lengthening of ENP storm tracks could be revealed if data were to be extended from the end of the 1997 hurricane season used in Irwin and Davis (1999).

Although previous work shows that shifts in ENP storm tracks are not significantly modified by the interannual variability of ENSO (Irwin and Davis, 1999), the objective of this study is to detect if recent changes in ENP storm origins and downgradation locations due to ENSO fluctuation have led to shifts in the poleward and westward movement of storm tracks. In addition, we hypothesize the stratification of historical TC climatology into EDR and WDR based on where maximum lifetime storm intensities are achieved will better differentiate possible regional differences of ENSO influence. Since a poleward shift in maximum TC winds would significantly alter the geographical distribution of storm-related risks (Kossin et al., 2014), it is also of interest to investigate and regionally compare how locations of the highest sustained winds respond to ENSO.

79 3.3 Data

To extend the analysis of Irwin and Davis (1999), TC events up to 2013 are incorporated into the current analysis. We only include TC data during the satellite observation era since 1971, thus 43 years of historical TC track data are considered.

They were accessed from the National Hurricane Center and modified as the best-track data (HURDAT2; Landsea and Franklin, 2013), which includes each storm’s geographic position and wind speed every six hours. Our region of TC record reflects the National

Hurricane Center’s area of ENP responsibility, though ENP storm tracks beyond 140°W are also included. TC data are also restricted to the local hurricane season from May 15 to November 30.

With respect to the classification of ENSO events, the multivariate ENSO index

(MEI) was used (Wolter and Timlin, 1998). MEI uses sea level pressure, zonal and meridional surface winds at the 850mb level, sea surface temperature, surface air temperature and total cloud fraction and is considered as a more holistic approach than measures of sea surface temperature anomalies in the Oceanic Niño Index or the

Southern Oscillation Index, which only takes into account of standardized differences in sea level pressure between Tahiti and Darwin, Australia.

3.4 Methods

MEI is well correlated with other ENSO indices that solely depend on either sea surface temperature anomalies or sea surface pressure differences (Trenberth, 1997;

Wolter and Timlin, 1998). To reflect the ENP hurricane season, bimonthly MEI values from April-May to November-December were averaged and ranked (Jien et al., 2015).

80 Years with the ten highest (lowest) MEI values were defined as El Niño (La Niña) phases, with the remaining 23 years in the neutral phase (Figure 3-1).

Figure 3-1 Bimonthly MEI averages from April-May to November-December with the ten highest (lowest) marked as El Niño (La Niña) years

To determine if ENSO-induced differences in storm genesis and downgradation points and track movement can be compared at the regional level, TCs are binned into

EDR and WDR categories, according to where a storm’s maximum lifetime intensity is achieved, not according to location of genesis (Collins and Roache, 2011; Jien et al.,

2015). Since there are three ENSO phases and two regions, a two-way ANOVA was applied to test if geographic coordinates of ENP storms are sensitive to ENSO and regional factors. However, a two-way ANOVA does not assess the combined effect of both factors nor does it directly evaluate individual comparisons of storm track characteristics among all three ENSO pairings. A post-hoc Tukey method of multiple comparisons was employed to distinguish which, if any, ENSO pairings show

81 significantly different genesis and downgradation points, leading to latitudinal

(northward) and longitudinal (westward) track movements at the 5% significance level.

This statistical evaluation would also identify if any ENSO-induced difference is maintained when ENP storms are regionally stratified as EDR and WDR storms, based on the longitudinal location where the maximum lifetime intensity is achieved.

Complementing the work of Irwin and Davis (1999) who found ENSO had an important impact on storms that developed outside of the TC primary genesis region, it is of interest to examine if ENSO also significantly affects storms that tracked outside of the area. To determine the presence of ENSO influence on storms that tracked outside the primary genesis region, the area, different from what was recommended in Irwin and

Davis (1999), was replaced by MDR and subdivided according to the geographical boundaries for EDR (10–20°N and 85–112°W) and WDR (10–20°N and 112–140°W;

Figure 3-2), as was done in Ralph and Gough (2009) and Jien et al. (2015). Since the dominant TC track heads in a west to northwest direction, the number of storms that traverses the northernmost and westernmost boundaries for both EDR and WDR is likely dependent on ENSO variability. Storms are tallied for those that exceed 20°N for both

EDR and WDR, and west of 112°W and 140°W for EDR and WDR respectively. A Chi- square (χ2) test is employed to assess if the percentage of storm count existing outside the two subdivisions is significantly different among the three ENSO groups.

To pinpoint the location where the maximum TC winds are achieved, the impact of translational speed is accounted for. Relative velocity at each position of storm track is calculated by subtracting the sustained wind speed from the translational speed. As such, the latitudinal and longitudinal coordinates where the maximum relative winds

Figure 3-2 Map of EDR and WDR boundary divisions of the eastern North Pacific basin with surrounding North American major cities.

were established were compared among the different ENSO pairings when storms were regionally stratified. Latitudinal and longitudinal differences that show the distance in which an average storm travels to gain peak intensity are also determined for the three

ENSO phases. Though there is an absence of temporal trend (Jien et al., 2015), maximum lifetime TC winds among the ENSO groups were also compared between ENP subdivisions.

3.5 Results

Including storms occurring in the Central Pacific basin, the entire HURDAT2

Eastern Pacific storm archive has 713 TCs that reached a maximum intensity of at least tropical storm strength (≥ 18 m s-1) during the period of 1971 to 2013 inclusive. If we refine this to storms of ENP origin during the local hurricane season (May 15 to

83 November 30), 650 storms (~91%) remain. Of the 650 storms, when ENP basin is subdivided, 288 of them reached maximum intensities in EDR; while, there were 362 such storms in the WDR. When storms are partitioned to each ENSO group, since there is a greater number of neutral years, the total storm count during the neutral phase is more than that combined for the El Niño and La Niña years.

3.5.1 Impacts of El Niño-Southern Oscillation and regional divisions

3.5.1.a El Niño-Southern Oscillation differences

In the absence of storm stratification into MDR subdivisions (EDR, WDR), the results of ENSO impact on ENP storm tracks are generally in agreement with Irwin and

Davis (1999), with minor differences mostly attributed to a different ENSO classification system. While Irwin and Davis (1999) used a threshold-based method in the application of SOI, a ranked-based method, MEI, is adopted in this work to determine annual ENSO classification. In addition, the analysis period is updated to reflect the most current TC climatology (1971-2013). During this 43-year timespan, an annual average of 16.9 (13) storms per year formed during El Niño (La Niña) years. Of all three ENSO phases, neutral years are associated with, on average, the lowest number (10.6) of annual storm formation.

When all ENP storms are considered, ENSO conditions have a clear impact on their latitudinal and longitudinal positions. Differences of mean latitudes and longitudes of TC initial genesis and downgradation locations between ENSO categories are shown in Figures 3 and 4. Although every ENSO comparison exhibits changes in TC genesis location, no single ENSO pairing demonstrates differences for both the average

Figure 3-3 Distributions of a) latitudes and b) longitudes of genesis locations for all ENP storms. Green diamonds represent averages within each ENSO phase

Figure 3-4 Distributions of a) latitudes and b) longitudes of downgradation locations for all ENP storms. Green diamonds represent averages within each ENSO phase

86 latitudinal (Table 3-1) and longitudinal coordinates (Table 3-2). The average genesis position for El Niño storms tends to form closer to the equator (Table 3-1) while La Niña storms are displaced further eastward (Table 3-2). On the other hand, the mean latitude and longitude of TC downgradation point both display (p < 0.05) differences between El

Niño and La Niña conditions (Tables 3 and 4). While the mean dissipation point for El

Niño storms is significantly displaced poleward (Table 3-3). When compared with the other two ENSO groups, La Niña storms are significantly displaced more eastward than storms that formed during the El Niño and neutral phases (Table 3-4).

Even though Irwin and Davis (1999) did not find significant ENSO-induced differences in the latitudinal and longitudinal movements, our analysis shows both measurements demonstrate significant sensitivities to ENSO conditions. In particular, the poleward movement is more responsive to variations in ENSO conditions (Figure 3-5).

Though El Niño and La Niña storms display the greatest inequality in their latitudinal drift, almost all three ENSO pairings show significant differences at p < 0.05 (Table 3-5).

Such differences in the northward/poleward trajectory are shown to be statistically different at p < 0.05 between two ENSO pairings: El Niño and La Niña and El Niño and neutral storms. With respect to the latitudinal drift, the El Niño storms travel 2.6° west and 1.4° north on average further than La Niña storms and neutral storms respectively.

On the other hand, only between El Niño storms and neutral storms is the mean longitudinal extension (Table 3-6) and storm days (Table 3-7) statistically distinct at p <

0.05. The average storm lifetime during El Niño years is sustained the longest (7.2 days;

Table 3-7) and travelled the furthest (about 4025 km), with an average translational speed of 6.5 m s-1.

87 Table 3-1: A two-way analysis of variance is tested for effects of ENSO influence and regional division on latitudes of storm genesis locations

Sum Mean F Probability DF Squared Squared value (>F) EN SO 2 138.9 69.43 12.03 <0.001 Region 1 13.0 12.96 2.25 0.13 ENSO:Region 2 17.2 8.61 1.49 0.23 Residuals 644 3716.7 5.77

ENSO Diff p values La Niña-El Niño 0.57 0.10 Neutral-El Niño 1.09 <0.001 Neutral-La Niña 0.52 0.09

Region Diff p values WDR-EDR -0.28 0.14

ENSO :Region Diff p values La Niña:EDR-El Niño:EDR 1.06 0.12 Neutral:EDR-El Niño:EDR 1.38 <0.001 Neutral:EDR-La Niña:EDR 0.32 0.94 La Niña:WDR-El Niño:WDR 0.07 1.00 Neutral:WDR-El Niño:WDR 0.89 0.02 Neutral:WDR-La Niña:WDR 0.82 0.20

88 Table 3-2: A two-way analysis of variance is tested for effects of ENSO and regional divisions on longitudes of storm genesis locations

Sum Mean Probability Longitude (°W) DF Squared Squared F value (>F) ENSO 2 1014 506.90 7.64 <0.001 Region 1 17553 17553.00 264.65 < 2.20E-16 ENSO:Region 2 17 8.60 0.13 0.88 Residuals 644 42713 66.30

ENSO Diff p values La Niña-El Niño -3.16 <0.001 Neutral-El Niño -0.06 1.00 Neutral-La Niña 3.10 <0.001

Region Diff p values WDR-EDR 10.39 <0.001

ENSO :Region Diff p values La Niña:EDR-El Niño:EDR -1.00 0.98 Neutral:EDR-El Niño:EDR 1.33 0.89 Neutral:EDR-La Niña:EDR 2.33 0.35 La Niña:WDR-El Niño:WDR -1.30 0.92 Neutral:WDR-El Niño:WDR 0.56 0.99 Neutral:WDR-La Niña:WDR 1.85 0.64

89 Table 3-3: A two-way analysis of variance is tested for effects of ENSO and regional divisions on latitudes of storm downgradation locations

Sum Mean F Probability Latitude (°N) DF Squared Squared value (>F) ENSO 2 345.5 172.74 6.39 <0.001 Region 1 118.0 117.95 4.36 0.04 ENSO:Region 2 167.8 83.89 3.10 0.05 Residuals 644 17421.0 27.05

ENSO Diff p values La Niña-El Niño -2.00 <0.001 Neutral-El Niño -0.31 0.80 Neutral-La Niña 1.69 <0.001

ENSO Diff p values WDR-EDR -0.85 0.04

ENSO Diff p values La Niña:EDR-El Niño:EDR -0.47 1.00 Neutral:EDR-El Niño:EDR 0.64 0.97 Neutral:EDR-La Niña:EDR 1.11 0.68 La Niña:WDR-El Niño:WDR -3.54 <0.001 Neutral:WDR-El Niño:WDR -0.96 0.63 Neutral:WDR-La Niña:WDR 2.58 0.01

90 Table 3-4: A two-way analysis of variance is tested for effects of ENSO and regional divisions on longitudes of storm downgradation locations

Sum Mean F Probability Longitude (°W) DF Squared Squared value (>F) ENSO 2 3158 1579.00 7.82 <0.001 Region 1 70565 70565.00 349.27 < 2.20E-16 ENSO:Region 2 239 119.00 0.59 0.55 Residuals 644 130112 202.00

ENSO Diff p values La Niña-El Niño -6.27 <0.001 Neutral-El Niño -4.03 0.01 Neutral-La Niña 2.24 0.28

Region Diff p values WDR-EDR 20.83 <0.001

ENSO :Region Diff p values La Niña:EDR-El Niño:EDR -1.86 0.98 Neutral:EDR-El Niño:EDR -0.56 1.00 Neutral:EDR-La Niña:EDR 1.30 0.99 La Niña:WDR-El Niño:WDR -2.37 0.91 Neutral:WDR-El Niño:WDR -3.24 0.40 Neutral:WDR-La Niña:WDR -0.87 1.00

Figure 3-5 Distributions of a) northward and b) westward storm track movements all ENP storms. Green diamonds represent averages within each ENSO phase

92 Table 3-5: A two-way analysis of variance is tested for effects of ENSO and regional division on latitudinal movement of storm tracks

Sum Mean F Probabili Latitudinal Shift DF Squared Squared value ty (>F) ENSO 2 499.5 249.74 9.22 <0.001 Region 1 52.7 52.71 1.95 0.16 ENSO:Region 2 77.5 38.77 1.43 0.24 Residuals 644 17437.2 27.08

ENSO Diff p values La Niña-El Niño -2.57 <0.001 Neutral-El Niño -1.40 0.01 Neutral-La Niña 1.17 0.07

Region Diff p values WDR-EDR -0.57 0.17

ENSO :Region Diff p values La Niña:EDR-El Niño:EDR -1.53 0.55 Neutral:EDR-El Niño:EDR -0.74 0.94 Neutral:EDR-La Niña:EDR 0.79 0.90 La Niña:WDR-El Niño:WDR -3.61 <0.001 Neutral:WDR-El Niño:WDR -1.85 0.04 Neutral:WDR-La Niña:WDR 1.76 0.20

93 Table 3-6: A two-way analysis of variance is tested for effects of ENSO and regional divisions on the longitudinal movement of storm tracks

Sum Mean F Probability Longitudinal Shift DF Squared Squared value (>F) ENSO 2 1814 907.10 4.52 0.01 Region 1 17730 17729.70 88.27 < 2.00E-16 ENSO:Region 2 129 64.60 0.32 0.73 Residuals 644 129348 200.90

ENSO Diff p values La Niña-El Niño -3.10 0.15 Neutral-El Niño -3.97 0.01 Neutral-La Niña -0.86 0.82

Region Diff p values WDR-EDR 10.72 <0.001

ENSO :Region Diff p values La Niña:EDR-El Niño:EDR -0.86 1.00 Neutral:EDR-El Niño:EDR -1.89 0.95 Neutral:EDR-La Niña:EDR -1.03 1.00 La Niña:WDR-El Niño:WDR -1.08 1.00 Neutral:WDR-El Niño:WDR -3.80 0.22 Neutral:WDR-La Niña:WDR -2.72 0.78

94 Table 3-7: A two-way analysis of variance is tested for effects of ENSO and regional divisions on total storm duration (days)

Sum Mean F Probability Duration DF Squared Squared value (>F) ENSO 2 78.7 39.33 3.38 0.03 Region 1 379.4 379.37 32.60 <0.001 ENSO:Region 2 20.3 10.14 0.87 0.42 Residuals 644 7493.7 11.64

ENSO Diff p values La Niña-El Niño -0.66 0.22 Neutral-El Niño -0.82 0.03 Neutral-La Niña -0.16 0.89

Region Diff p values WDR-EDR 1.53 <0.001

ENSO :Region Diff p values La Niña:EDR-El Niño:EDR -0.70 0.85 Neutral:EDR-El Niño:EDR -0.56 0.89 Neutral:EDR-La Niña:EDR 0.14 1.00 La Niña:WDR-El Niño:WDR -0.01 1.00 Neutral:WDR-El Niño:WDR -0.80 0.37 Neutral:WDR-La Niña:WDR -0.79 0.62

95 With respect to the storms’ maximum wind speed, the average maximum lifetime intensity of all 650 storms is 33.7 m s-1, expressed as the relative winds (maximum sustained winds – translational speed). The average maximum relative intensity

(36.5 m s-1) during El Niño years is statistically stronger than those that formed during the neutral phase (Table 3-8). Between the same two ENSO phases, the geographic location (Table 3-9 and 10) where the average maximum relative winds occurred are also statistically different. In fact, the statistical significance of longitudinal differences extends across all three ENSO pairings (Table 3-10). Because stronger storms require a longer time to establish maximum intensity, it is likely that El Niño storms took longer to develop and reach maximum intensity than storms that had occurred during the neutral phase (Table 3-11).

3.5.1.b Regional differences

When only the presence of regional division is considered when comparing geographic coordinates and track movement, the latitudinal difference between regions only becomes evident as storms are downgraded (Figure 3-6). Table 3-1 shows although the average latitudinal coordinates of storm genesis location between EDR and WDR are not dramatically different (Figure 3-7), a significant (p < 0.05) regional difference is observed for their average dissipation locations (Tables 3 and 4). With respect to the longitudinal coordinate, WDR storms have a pronounced westward extension (Table 3-6) compared to EDR storms. Since many WDR storms began as tropical depressions within the EDR before transitioning and establishing peak intensities upon crossing the MDR longitudinal division at 112°W, it is generally expected that the longitudinal shift measured from storm genesis to downgradation for WDR storms is longer than EDR

96 Table 3-8: A two-way analysis of variance is tested for effects of ENSO and regional divisions on maximum storm intensity

Sum Mean F Probability Highest.Relative DF Squared Squared value (>F) ENSO 2 1879 939.27 4.12 0.02 Max_Region 1 7 7.15 0.03 0.86 ENSO:Max_Region 2 305 152.40 0.67 0.51 Residuals 644 146990 228.25

ENSO Diff p values La-El -2.96 0.21 Neutral-El -4.05 0.01 Neutral-La -1.10 0.76

Region Diff p values WDR-EDR 0.21 0.86

ENSO :Max_Region Diff p values La:EDR-El:EDR -5.20 0.37 Neutral:EDR-El:EDR -5.35 0.18 Neutral:EDR-La:EDR -0.15 1.00 La:WDR-El:WDR -1.04 1.00 Neutral:WDR-El:WDR -3.30 0.45 Neutral:WDR-La:WDR -2.25 0.91

97 Table 3-9: A two-way analysis of variance is tested for effects of ENSO and regional divisions on latitudes of the locations of maximum storm strength

Sum Mean F Probability Max_Lat DF Squared Squared value (>F) ENSO 2 107.6 53.78 6.36 <0.001 Max_Region 1 43.3 43.34 5.13 0.02 ENSO:Max_Region 2 18.2 9.08 1.07 0.34 Residuals 644 5443.5 8.45

ENSO Diff p values La-El 0.27 0.70 Neutral-El 0.91 <0.001 Neutral-La 0.64 0.08

Region Diff p values WDR-EDR -0.52 0.02

ENSO :Max_Region Diff p values La:EDR-El:EDR 0.72 0.72 Neutral:EDR-El:EDR 1.13 0.11 Neutral:EDR-La:EDR 0.41 0.93 La:WDR-El:WDR -0.29 0.99 Neutral:WDR-El:WDR 0.73 0.29 Neutral:WDR-La:WDR 1.02 0.17

Table 3-10: A two-way analysis of variance is tested for effects of ENSO and regional divisions on longitudes of the locations of maximum storm strength

Sum Mean Probability Max_Lon DF Squared Squared F value (>F) ENSO 2 2321 1160 16.70 <0.001 Max_Region 1 38815 38815 558.56 < 2.20E-16 ENSO:Max_Region 2 359 179 2.58 0.08 Residuals 644 44752 69

ENSO Diff p values La-El -5.57 <0.001 Neutral-El -2.92 <0.001 Neutral-La 2.65 0.01

Region Diff p values WDR-EDR 15.45 <0.001

ENSO :Max_Region Diff p values La:EDR-El:EDR -0.61 1.00 Neutral:EDR-El:EDR 0.63 1.00 Neutral:EDR-La:EDR 1.24 0.91 La:WDR-El:WDR -4.07 0.03 Neutral:WDR-El:WDR -2.89 0.04 Neutral:WDR-La:WDR 1.18 0.93

99 Table 3-11: A two-way analysis of variance is tested for effects of ENSO and regional divisions on longitudes of the locations of maximum storm strength Sum Mean F Probabilit Max_Day DF Squared Squared value y (>F) ENSO 2 49.75 24.88 7.71 <0.001 Max_Region 1 94.69 94.69 29.34 <0.001 ENSO:Max_Region 2 2.2 1.10 0.34 0.71 Residuals 644 2078.18 3.23

ENSO Diff p values La-El -0.64 0.01 Neutral-El -0.63 <0.001 Neutral-La 0.02 1.00

Region Diff p values WDR-EDR 0.76 <0.001

ENSO :Max_Region Diff p values La:EDR-El:EDR -0.35 0.88 Neutral:EDR-El:EDR -0.38 0.72 Neutral:EDR-La:EDR -0.04 1.00 La:WDR-El:WDR -0.60 0.30 Neutral:WDR-El:WDR -0.67 0.02 Neutral:WDR-La:WDR -0.07 1.00

100

Figure 3-6 Distributions of latitudes and longitudes of ENP storm downgradation locations based on regional division. Hollow diamonds represent the regional averages

Figure 3-7 Distributions of latitudes and longitudes of ENP storm genesis locations based on regional division. Hollow diamonds represent the regional averages

101

storms. Figure 3-8 shows the westward drift is strikingly higher for WDR storms. Even when these EDR-derived WDR storms are excluded (not shown), the remaining WDR storms that were initiated within WDR still had a statistically significant difference in the longitudinal distance travelled. In addition, the removal of these transitioning storms also leads to a regional difference in the latitudinal shift.

Storms that transitioned from EDR to WDR also play a critical role when comparing TC characteristics and locations where storm intensities peaked. Except for the mean latitudinal position, significant differences of TC characteristics and locations where maximum lifetime storm intensities were attained were not found when comparing

EDR and WDR storms. Because these transitioning storms are sustained over a longer lifespan, their average intensity is usually stronger (Wang and Chan, 2002; Chan, 2005).

Figure 3-8 Distributions of latitudinal and longitudinal movements of ENP storm genesis locations based on regional division. Hollow diamonds represent the regional averages

102

Once these transitioning storms are removed, EDR storms were found to be much stronger than WDR storms that were not derived from EDR. When EDR-originated

WDR storms were excluded, WDR (EDR) storms still experience a greater westward

(northward) progression between locations of TC genesis and its maximum lifetime intensity.

3.5.1.c Combined El Niño-Southern Oscillation and regional differences

When storms are analyzed on an annual basis and linked to ENSO conditions, the

ENSO influence is regionally distinct. Although El Niño years are slightly favoured for

TC formation, EDR storm frequency is more evenly distributed among the three ENSO phases. In WDR, however, the distribution of storm frequency within the three ENSO phases is more distinct. In particular, the mean storm count during the ten El Niño years

(11.9 storms per El Niño year) is almost twice as many as those that formed during La

Niña years (6 storms per La Niña year), with the number of storms that formed during the neutral phases falling in between the two extreme ENSO phases (Figure 3-9).

When ENP storms are regionally stratified as EDR and WDR storms, ENSO- induced variations in the genesis (Figures 3-10a and 3-10b) and downgradation (Figures.

3-11a and 3-11b) locations only exist for their mean latitudinal coordinates. For the genesis position, EDR and WDR storms are statistically different when comparing El

Niño and neutral conditions (Table 3-1). EDR and WDR storms during El Niño years tend to form southward of storms that developed during the neutral phase. Based on results from the Tukey analysis for multi-factor comparison, there is no statistical

103

Figure 3-9 Comparison of the total and annual/seasonal average storm frequency in MDR subdivisions during all three ENSO phases

104

Figure 3-10 Distributions of latitudes and longitudes for EDR and WDR storm genesis locations within each ENSO phase. Green diamonds represent individual averages

105

Figure 3-11 Distributions of latitudes and longitudes for EDR and WDR storm downgradation locations within each ENSO phase. Green diamonds represent individual averages

difference in the longitudinal position between any ENSO pairing for either EDR or

WDR storms (Table 3-2). Even though the initiation points of EDR storms are as equally sensitive to ENSO fluctuation as WDR storms, the downgradation positions are regionally dependent on ENSO influences. For WDR storms, the average latitudinal coordinate during El Niño is not statistically different from storms that occurred during the neutral phase (Table 3-3). However, where El Niño storms typically form closer to the equator, they are downgraded at a more poleward position than storms during the other two ENSO phases (Table 3-4). Although the longitudinal displacement from its genesis and dissipation points is significantly shifted only when the regional stratification is considered (EDR and WDR), this displacement is found to be statistically similar when

EDR and WDR storms are categorized by ENSO events (Table 3-6). In fact, detection of strong ENSO influences occurs only for WDR storms when poleward shifts between the

106 regions are compared. A statistically different northward shift is found when comparing

El Niño and La Niña and El Niño and neutral conditions (Table 3-5). Overall, WDR storm tracks experience the greatest latitudinal shift during the El Niño phase (Figure 3-

12a). In contrast the average longitudinal shift is not statistically different for any ENSO pairing (Figure 3-12b).

The average longitudinal position of the maximum relative winds is more sensitive to ENSO phases (Table 3-10). The magnitude of the largest latitudinal change between any ENSO pairing for either EDR or WDR storms is just under 0.4°. However, such an ENSO-induced effect on the longitudinal distance in storm development is only apparent for WDR storms (Table 3-12). Generally, the average peak intensity of WDR storms are established more westward during the El Niño phase than the other two ENSO conditions though the resulting longitudinal shift from where they were first detected is only significantly different between El Niño and neutral conditions.

3.5.2 Primary Genesis Region

Since it is generally observed the predominantly westward moving TCs tends to be deflected poleward, the number of storms that surpass the westward and northward ranges of their primary genesis region is investigated and regionally compared. In contrast to Irwin and Davis (1999), a northern boundary of 20°N is shared by both EDR and WDR. Meanwhile, our western boundaries for EDR and WDR are at 112°W and

140°W respectively. Though a larger (smaller) number of EDR (WDR) storms had drifted past its poleward boundary, approximately similar proportions of EDR (76%) and

WDR (72%) storms have extended beyond either northward or westward of their prime

107

Figure 3-12 Distributions of latitudinal and longitudinal movements for EDR and WDR storm downgradation locations within each ENSO phase. Green diamonds represent individual averages

108 Table 3-12: A two-way analysis of variance is tested for effects of ENSO and regional divisions on the longitudinal shift between locations of storm genesis and maximum storm strength Sum Mean F Probability DF Squared Squared value (>F) ENSO 2 960 480.20 9.55 <0.001 Max_Region 1 4164 4163.70 82.82 < 2.20E-16 ENSO:Max_Region 2 243 121.70 2.42 0.09 Residuals 644 32377 50.30

ENSO Diff p values La-El -2.41 0.01 Neutral-El -2.87 <0.001 Neutral-La -0.46 0.81

Regio n Diff p values WDR-EDR 5.06 <0.001

ENSO :Max_Region Diff p values La:EDR-El:EDR 0.39 1.00 Neutral:EDR-El:EDR -0.70 0.99 Neutral:EDR-La:EDR -1.09 0.89 La:WDR-El:WDR -2.78 0.15 Neutral:WDR-El:WDR -3.45 <0.001 Neutral:WDR-La:WDR -0.67 0.99

109 genesis regions. However, a greater regional contrast is observed for storms that had crossed both longitudinal and latitudinal divisions. Specifically, during El Niño years, there are three times more WDR storms that had advanced past their poleward boundary than during La Niña years (Table 3-13).

Based on the earlier observation that the mean latitudinal position where WDR storms are downgraded is more sensitive to ENSO fluctuations than EDR storms, the

ENSO effect on the number of storms that had reached outside their prime genesis areas is explored. Because the WDR storm subset that tracked poleward of its northward boundary embodies the greatest number of storms, it is associated with the greatest

ENSO contrast. Table 3-13 shows group percentages of total storms within each ENSO phase that extended outside of the northern and western boundaries before dissipating. A higher χ2 than the criterion value (χ2 = 5.99) at p < 0.05 means the observed percentages of total storms that reached the boundaries at different ENSO groups is statistically significant. The analyses have shown only the WDR storm group that advanced outside their northernmost boundary is statistically different among ENSO phases. Figure 3-13a1 and (Figure 3-13b) shows a greater (smaller) proportion of WDR (EDR) storms that had reached at least 20°N is observed to landfall in the U.S..

3.6 Discussion

While the location of storm origins has been inconsistently different between all three pairings of ENSO groups, variations between El Niño and La Niña phases are stronger for the downgradation point and track length. In particular, this ENSO-induced

1 Though Hurricane Iniki (1992) gained its maximum intensity in the Central Pacific Ocean, it is recorded as an ENP storm in HURDAT2.

110 Table 3-13: Comparison of the group percentage of EDR and WDR storms in crossing the northern-most boundary. To reject the null hypothesis of similarity between three ENSO phases at p < 0.05, the χ2 value must exceeds 5.99.

ENSO Proportion Percentage χ2 EDR storms crossed 20°N El Niño 32/60 53.3 La Niña 38/70 54.3 2.47 Neutral 98/155 63.2 WDR storms crossed 20°N El Niño 63/109 57.8 La Niña 19/60 31.7 16.76 Neutral 110/196 56.1

(a)

111 (b)

Figure 3-13 a) EDR and b) WDR storm tracks that had crossed 20°N and made landfalls

difference in storm characteristics is more significant for the latitudinal component of the storm tracks, leading to a greater poleward extension for El Niño storms. Relative to La

Niña phase, storms that occurred during El Niño years are located further westward when formed and more poleward when dissipated. This variation in the lengthening of a northward drift for WDR storms could be the result of a large scale, environmental pattern induced by ENSO fluctuation. Similar convergence of TC genesis and storm activity towards the International Date Line is also noted with Northwest Pacific (Clark and Chu, 2002; Wang and Chu, 2002) and Southwest Pacific (Basher and Zheng, 1995)

TCs, which generally shift eastward, towards the equator, during El Niño years.

112 Based on the results presented here, WDR storms are significantly more sensitive to ENSO fluctuation. During El Niño years, the average genesis position for WDR storms hovers around the equator, sharing a similar ENSO response in the adjacent

Northwest Pacific and Central Pacific storms (Chen et al., 1998; Wang and Chan, 2002;

Chan, 2005). In the Northwest Pacific basin, this is often associated with both the eastward extension and equatorial movement of the monsoon trough where the effect of

Walker Circulation has been linked with local tropical convection (Chen et al., 1998).

Such an eastward migration of tropical convergence zone (Chen et al., 1998; Chan, 2000) might have contributed to the ENSO-induced differences in WDR. During El Niño years, this shift in the main storm genesis area is accompanied by the propagation of cyclonic vorticity into the eastern Pacific Ocean, corresponding with anomalous equatorial westerlies and subtropical easterly flows (Chu and Wang, 1997; Chan, 2005).

Since TC genesis locations during El Niño years generally shift to the southeast quadrant of the Northwest Pacific basin, it would greatly extend the storm development stage and provide an environment where stronger storms are favoured (Wang and Chan,

2002; Chan, 2005). Similarly, in the ENP basin, an El Niño-induced transport of storm genesis location may have contributed to an overall higher WDR storm intensity. At the

Northwest Pacific basin, storm intensification is highly dictated by the relative vorticity associated with monsoon trough (Chen et al., 1998; Wang and Chan, 2002). Likewise, an

El Niño-induced migration of cyclonic flow near the eastern periphery of the Northwest

(Chen et al., 1998; Chan, 2000) and Central (Chu and Wang, 1997) Pacific basins could have also shifted locations of maximum lifetime winds of WDR storms further westward when compared with storms that formed during La Niña years (Figure 3-14). This

113 westward displacement of maximum storm intensity would support the fact that there is a greater TC influence in the Central Pacific basin (between 140°W and 180°) during El

Niño years (Chu and Wang, 1997).

Figure 3-14 Locations where maximum relative winds are achieved for EDR and WDR storms under three ENSO phases

A greater northward movement of El Niño storm tracks in other TC basins is likely linked to dynamical mechanisms that extend storm duration and propel its dissipation stage further poleward. For instance, the poleward movement of Southwest

Pacific TCs is particularly reinforced by the steering wind flow at the mid-level (700-

500-mb) troposphere during El Niño years (Camargo et al., 2007; Chand and Walsh,

2009). Responding to El Niño conditions, the importance of meridional component of this mid-tropospheric wind flow (Camargo et al., 2007; Chand and Walsh, 2009) and the

114 associated meridional heat exchange (Wang and Chan, 2002) to TCs at other regions could also be vital for recurving WDR storm tracks.

3.6 Storm Impact

The exposure of coastal populations to ENP storm impact can be mapped and visualized as they are stratified into the three ENSO phases. While storm genesis location tends to remain attached to the North American continent, the typical track path of ENP storms usually aims directly at the Hawaiian Islands (Camargo et al., 2008).

During La Niña years, a significant eastward extension of storm origin greatly exposes

Mexico and countries of Central America (Figure 3-15a) to hurricane risk. On the other hand, the average genesis point for El Niño storms is displaced further westward into the open water where the least societal impact is expected. When storm downgradation locations are taken into consideration (Figure 3-15b), El Niño storms continue to display a stronger westward shift than storms that occurred during the other two ENSO phases.

Because El Niño storms are also detected further poleward when dissipated, their area of storm track influence is substantially expanded.

While ENSO has an important role in determining the extent of a poleward shift, when storms are regionally stratified, the latitudinal difference in the downgradation point is mostly detecTable 3-with WDR storms. In contrast, the mean latitudinal coordinate of storm formation position is not significantly affected by the regional stratification of storms (Figure 3-16a). Although more EDR storms are observed to landfall, most are downgraded well short of striking the continental U.S.. Whereas, when

WDR storms landfall, the proportion that reached continental U.S. is substantially greater

115 (a)

(b)

Figure 3-15 Storm locations at the a) genesis, and b) downgradation stages under three ENSO phases

116 (a)

(b)

Figure 3-16 Locations of a) genesis, and b) downgradation points for EDR and WDR storms

117 than U.S. landfalling storms from EDR (Figure 3-16b). Since WDR storms are generally displaced further westward, Hawaiian Islands are also more vulnerable to their influence

(Figure 3-16b).

While differences of storm genesis location between El Niño and neutral phases are equally sustained in both regions (Figure 3-17a), WDR storms have a strong ENSO connection in the mean downgradation location (Figure 3-17b) and latitudinal lengthening (Figure 3-12a) of storm tracks. In addition, potential risks associated WDR storms are not equally distributed among all ENSO phases. There has been virtually no

WDR storm landfall recorded during all of the ten La Niña years in North America. In particular, WDR storms are steered away from the southwest U.S. during La Niña years.

In contrast, the southwest U.S. is particularly vulnerable to WDR storm influence during non-La Niña years. Part of the reason could be storms that developed during the La Niña phase are on average weaker and are less likely to be sustained over a long distance.

3.7 Summary and Conclusion

While a global shift of TC’s maximum intensity has been ascribed to climate change (Kossin et al., 2014), this study shows that the interannual variability of ENSO events is actively involved in the modification of ENP storm genesis, downgradation and track locations. In agreement with Irwin and Davis (1999), differences of storm genesis and downgradation locations are detected. In addition, differences of latitudinal and longitudinal storm movements due to ENSO modification are also found when the most current TC climatology from 1971-2013 is used. While differences of latitudinal and longitudinal shifts were mostly found to be distinctly different between El Niño and La

118 (a)

(b)

Figure 3-17 Locations of EDR and WDR storms at the a) genesis, and b) downgradation under three ENSO phases

119 Niña storms, they do not occur between every ENSO pairing. When MDR is subdivided, regional sensitivity to variations in ENSO phase also emerges. Overall, EDR storms are not responsive to ENSO phase change; while, the positions of storm formation and downgradation, as well as track movement of WDR storms are more significantly affected by ENSO phases. This had led to a greater northward transport of WDR storms during El Niño years than any other ENSO phases.

While the latitudinal component of WDR storms is more sensitive to changes in

ENSO phase, the proportion of WDR storms that reached the northernmost (20°N) boundary was detected to vary significantly among ENSO phases. El Niño storms have a greater proportion of storms that tracked poleward of the WDR boundary than La Niña and neutral phases combined. As such, this ENSO-induced variability of poleward transport plays a key role in limiting the number of storm that recurve and, ultimately, caps the number of storm impact in the U.S.. During El Niño years, a greater northward extension of WDR storm tracks might have contributed to the expansion of storm influence into southwest U.S..

In response to ENSO fluctuation in the Northwest Pacific domain (Wang and

Chan, 2002), it is speculated the eastward extension of monsoon trough due to El Niño forcing greatly facilitates a longitudinal difference in ENP storm tracks. From the onset of storm genesis till achieving its maximum lifetime wind, the strongest westward drift is found to occur during El Niño years, when the highest average storm intensity is also found. In response to the migration of monsoon trough, it is likely that dynamical features such as low-level vorticity and mid-level steering flow are enhanced during El

Niño years for strengthening WDR storms and shifting their eventual track movement.

120 While it was previously identified that thermodynamic factors such as the mid-level relative humidity (Collins, 2010) and sea surface temperature (Ralph and Gough, 2009) are more critical to the seasonal activity of WDR storms, dynamic features due to ENSO modification likely play a greater role in reinforcing the latitudinal displacement of the genesis and downgradation locations of WDR storms, thereby, raising the variability of latitudinal movements.

121 3.8 References:

Aiyyer, A. R., and C. Thorncroft (2006), Climatology of vertical wind shear over the tropical Atlantic, Journal of Climate, 19, 2969–2983, doi:10.1175/JCLI3685.1.

Alexander, M. A., I. Bladé, M. Newman, J. R. Lanzante, N-C. Lau, and J. D. Scott, 2002: The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans. Journal of Climate, 15, 2205–2231.

Ashok, K., S. K. Behera, S. A. Rao, H. Weng, and T. Yamagata, 2007: El Niño Modoki and its possible teleconnection. Journal of Geophysical Research., 112, C11007, doi:10.1029/2006JC003798.

Basher, R. E. and X. Zheng, 1995: Tropical cyclones in the Southwest Pacific: Spatial patterns and relationships to Southern Oscillation and sea surface temperature. Journal of Climate, 8, 1249-1260.

Camargo, S. J., A. W. Robertson, S. J. Gaffney, P. Smyth and M. Ghil, 2007: Cluster analysis of typhoon tracks. part II: Large-scale circulation and ENSO. Journal of Climate, 20(14), 3654-3676.

Camargo, S. J., A. W. Robertson, A. G. Barnston, and M. Ghil, 2008: Clustering of eastern north Pacific tropical cyclone tracks: ENSO and MJO effects. Geochemistry, Geophysics, Geosystems, 9(6) Q06V05. Chan, J. C. L., 1985: Tropical cyclone activity in the northwest Pacific in relation to the el Nino/Southern oscillation phenomenon. Monthly Weather Review, 113(4), 599-606.

Chan, J. C. L., 2005: Interannual and interdecadal variations of tropical cyclone activity over the western north Pacific. Meteorology and Atmospheric Physics, 89(1-4), 143-152.

Chand, S. S., and K. J. E., Walsh, 2009: Tropical cyclone activity in the Fiji region: Spatial patterns and relationship to large-scale circulation. Journal of Climate, 22(14), 3877-3893.

Changnon, S. A., 1999: Impacts of 1997—98 El Niño Generated Weather in the United States. Bulletin of the Meteorological Society, 80, 1819–1827.

Chen, G., 2011: How does shifting Pacific Ocean warming modulate on tropical cyclone frequency over the South China Sea? Journal of Climate, 24(17), 4695-4700.

122 Chen, T.-C., S.-P. Weng, N. Yamazaki, and S. Kiehne, 1998: Interannual variation in the tropical cyclone formation over the western North Pacific. Monthly Weather Review, 126, 1080–1090.

Chen, G., and C. Y. Tam, 2010: Different impacts of two kinds of Pacific Ocean warming on tropical cyclone frequency over the western North Pacific. Geophysical Research Letters, 37, L01803, doi:10.1029/2009GL041708.

Chia, H. H., and C. F. Ropelewski, 2002: The interannual variability in the genesis location of tropical cyclones in the northwest Pacific. Journal of Climate, 15(20), 2934- 2944.

Clark, J. D., and P.-. Chu, 2002: Interannual variation of tropical cyclone activity over the central north Pacific. Journal of the Meteorological Society of Japan, 80(3), 403-418.

Chu, P. -., and J. Wang, 1997: Tropical cyclone occurrences in the vicinity of Hawaii: Are the differences between El Niño and non-El Niño years significant? Journal of Climate, 10(10), 2683-2689.

Colbert, A. J., and B. J. Soden, 2012: Climatological variations in north Atlantic tropical cyclone tracks. Journal of Climate, 25(2), 657-673.

Collins, J., 2010: Contrasting High North-East Pacific Tropical Cyclone Activity. Southeastern Geographer, 50, 83-98.

Collins, J. M., and I. M. Mason, 2000: Local environmental conditions related to seasonal tropical cyclone activity in the Northeast Pacific basin. Geophysical Research Letters, 27, 3881-3884.

Collins, J. M., and D. R. Roache, 2011: The 2009 Hurricane Season in the Eastern North Pacific Basin: An Analysis of Environmental Conditions. Monthly Weather Review, 139(6), 1673-1681.

Evans, J. L. and R. J. Allan, 1992: El Niño/Southern Oscillation modification to the structure of the monsoon and tropical cyclone activity in the Australasian region, International Journal of Climatology, 12, 611–623.

Finlayson, M., 2015: Influence of ENSO on Tropical Cyclones in North Atlantic in a high-resolution climate model (Bachelor thesis). Retrieved from http://wesscholar.wesleyan.edu/cgi/viewcontent.cgi?article=2470&context=etd_hon_thes es

Goldenberg, S. B. and L. J. Shapiro, 1996: Physical mechanisms for the association of El Niño and West African rainfall with Atlantic major hurricane activity. Journal of Climate, 9, 1169–1187.

123 Hastings, P. A., 1990: Southern Oscillation influences on tropical cyclone activity in the Australian/South-west Pacific region, International Journal of Climatology, 10, 291–298.

Higgins, R. W., Y. Chen and A. V. Douglas, 1999: Interannual variability of the North American warm season precipitation regime. Journal of Climate, 12(2-3), 653-680.

Jien, J. Y., W. A. Gough, and K. Butler, 2015: The Influence of El Niño–Southern Oscillation on Tropical Cyclone Activity in the Eastern North Pacific Basin. Journal of Climate, 28, 2459–2474.

Jin, F.-F., J. Boucharel, and I.-I. Lin, 2014: Eastern Pacific tropical cyclones intensified by El Niño delivery of subsurface ocean heat, Nature, 516, 82–85, doi:10.1038/nature13958.

Jones, C. G. and C. D. Thorncroft, 1998: The role of El Niño in Atlantic tropical cyclone activity. Weather, 53(9), 324-336.

Kim, H. M., P. J. Webster, and J. A. Curry, 2009: Impact of shifting patterns of Pacific Ocean warming on North Atlantic tropical cyclones, Science, 325, 77–80.

Kimberlain, T. B. 1999. The effects of ENSO on North Pacific and North Atlantic tropical cyclone activity. Proc. 23rd Conference on Hurricanes and Tropical Meteorology, Dallas, Texas, Amer. Meteor. Soc., Boston, 250–253.

Klotzbach, P. J., 2011: El Niño-Southern Oscillation's Impact on Atlantic Basin Hurricanes and U.S. Landfalls. Journal of Climate, 24, 1252-1263.

Klotzbach, P. J., 2012: El Niño-Southern Oscillation, the Madden-Julian Oscillation and Atlantic basin tropical cyclone rapid intensification. Geophysical Research Letters, 117, D14104.

Kuleshov, Y., L. Qi, R. Fawcett, and D. Jones, 2008: On tropical cyclone activity in the Southern Hemisphere: Trends and the ENSO connection. Geophysical Research Letters, 35, L14S08, doi:10.1029/2007GL032983.

Kossin, J. P., K. A. Emanuel, and G. A. Vecchi, 2014: The poleward migration of the location of tropical cyclone maximum intensity. Nature, 509(7500), 349-352.

124 Lander, M. A., 1994: An exploratory analysis of the relationship between tropical storm formation in the western north Pacific and ENSO. Monthly Weather Review, 122(4), 636- 651.

Landsea, C. W., and J. L. Franklin, 2013: Atlantic Hurricane Database Uncertainty and Presentation of a New Database Format. Monthly Weather Review, 141 (10), 3576–3592, doi:10.1175/ MWR-D-12-00254.1.

Landsea, C. W. and W. M. Gray, W.M., 1989: Eastern North Pacific tropical cyclone climatology—low frequency variations. Report of WG 6.2 (Northeast Pacific Ocean) of IWTC–II, WMO, WMO/TD, no. 319, Geneva.

Pielke Jr., R. A., and C. N. Landsea, 1999: La Niña, El Niño, and Atlantic hurricane damages in the United States. Bulletin of the Meteorological Society, 80(10), 2027-2033.

Pradhan, P. K., B. Preethi, K. Ashok, R. Krishnan, and A. K. Sahai, 2011: Modoki, Indian Ocean Dipole, and western North Pacific typhoons: Possible implications for extreme events. Journal of Geophysical Research., 116, D18108.

Quinn, W. H., D. O. Zopf, K. S. Short, and R. T. W. Kuo Yang, 1981: Historical trends and statistics of the southern oscillation, El Nino, and Indonesian droughts (Peru). Fishery Bulletin, 76(3), 663-678.

Ralph, T. U., and W. A. Gough, 2009: The influence of sea-surface temperatures on Eastern North Pacific tropical cyclone activity. Theoretical and Applied Climatology, 95, 257-264.

Schroeder, T. A. and Z. Yu, 1995: Interannual variability of Central Pacific tropical cyclones. Preprints of the 21st Conference on Hurricanes and Tropical Meteorology, Amer. Meteor. Soc., Miami, Florida, 437-439.

125 Sobel, A. H., and S. J. Camargo, S. J, 2005: Influence of western north Pacific tropical cyclones on their large-scale environment. Journal of the Atmospheric Sciences, 62(9), 3396-3407.

Trenberth, K. E., 1997: The definition of El Niño. Bulletin of the Meteorological Society, 78 (12), 2771-2777.

Wang, B., and J. C. L. Chan, 2002: How strong ENSO events affect tropical storm activity over the western north Pacific. Journal of Climate, 15(13), 1643-1658.

Wolter, K., and M. S. Timlin, 1998: Measuring the strength of ENSO events - how does 1997/98 rank? Weather, 53, 315-324.

Wu, G. and N. C. Lau, 1992: A GCM simulation of the relationship between tropical- storm formation and ENSO. Monthly Weather Review, 120(6), 958-977.

Zhang, G., & Wang, Z. (2015). Interannual variability of tropical cyclone activity and regional Hadley circulation over the northeastern Pacific. Geophysical Research Letters, 42(7), 2473-2481.

126 4 Chapter 4: Near-time Sea Surface Temperature and Tropical Cyclone Intensity in the Eastern North Pacific basin

4.1 Abstract

Although a significant relationship between near-time sea surface temperature

(SST) and tropical cyclone (TC) intensity has been found for many major TC basins, this topic has not been explored in the eastern North Pacific basin. When the main development region of the eastern North Pacific Ocean is subdivided into eastern (EDR) and western (WDR) development regions, SSTs show a weak, yet significant, positive relationship with intensities of the six-hourly TC observations and storms’ maximum strengths in WDR. This SST-storm intensity relationship is most apparent for the maximum lifetime TC intensity of WDR major hurricanes. Though storm-induced SST reduction is detected as early as a week prior to storm passages in both MDR subdivisions, stronger EDR storms tend to inhibit further intensification by generating a more pronounced sea surface cooling than WDR storms. Contrary to the North Atlantic basin, the SST achievements of the major hurricane status in the ENP basin are bounded by lower thresholds in EDR (27°C) and WDR (25°C). Such a difference in the minimum

SST requirement to achieve hurricane status is found to further differ between EDR and

WDR storms.

When intensity observations are binned into SST intervals, the upper bound value of TC intensity is found to increase with SST. Compared to the previous TC climatological analysis (Whitney and Hobgood, 1997), the maximum relative wind speed

127 has increased for SST bins of 27°C (>26.5°C and < 27.5°C) or higher. While a linear function was determined previously as the best empirical fit for ENP maximum potential intensity (ENPMPI) for each SST bin (Whitney and Hobgood, 1997), other means of curve fittings such as the exponential decay (increase form) function also shown skill at representing the SST-dependent ENPMPI in the WDR. When storm observations are regionally stratified, the rate of increasing maximum potential intensity with SST flattens out towards the highest SST category.

4.2 Introduction

Sea surface temperatures (SSTs) have been long thought to be one of the major determining factors in limiting the structure and development of tropical cyclones (TCs;

Miller, 1958; Gray, 1969; Merrill, 1988; Evans, 1993; DeMaria and Kaplan, 1994;

Whitney and Hobgood, 1997; Saunders and Lea, 2008). However, other more recent research indicates that the development of TCs depends on more than just an absolute threshold of SST value. Vecchi and Soden (2007) for example noted that remote changes of the zonal mean, and all-tropics mean, SSTs and their differences from the local SST provide a more effective metric to categorize TC intensification. Despite this remote control of local TC intensity, near-time SSTs remain one of the primary indicators in regulating the upper boundary of storm intensity and its lifetime maximum wind speed.

Miller (1958) proposed a direct relationship between SST and the minimum sea level pressure in the surrounding air below the storm eye. Within the context of future climate change scenarios, Emanuel (1987) pointed out that the maximum reduction in sea level pressure experienced through SST warming pertains only to the reduction of the

128 lowest sustainable pressure achieved by the most intense TC. While a range of environmental conditions could influence the lifetime storm intensity (Gray, 1969), SSTs have been acknowledged as a limit to TC’s maximum potential intensity (MPI).

However, this direct relationship appears to be less important at explaining TC intensity above a certain SST value (Evans, 1993; Michaels et al., 2006). For example, because dynamic influences such as vertical wind shear are found to be key in determining the intensities of North Atlantic storms (Gray, 1984; Saunders and Lea, 2008), it appears TC intensity could be more sensitive to other factors such as atmospheric dynamical influences rather than a single thermodynamic limit alone (Michaels et al., 2006).

SSTs have been found to act as a cap to MPI for TCs at a global scale (Merrill,

1987). Using the most reliable storm observations for each TC development basin, intensity data were binned according to the corresponding SST groups to derive empirical relationships of SSTs and their maximum sustained TC winds for the North Atlantic

Ocean (DeMaria and Kaplan, 1994), Northwestern Pacific Ocean (Zeng et al., 2007) and eastern North Pacific Ocean (ENP; Whitney and Hobgood, 1997) and, most recently, Bay of Bengal (Kotal et al., 2009), part of North Indian basin. While the degree in which TC intensity’s bahaviour with SST influence varies among basins, a positive correlation between the maximum TC intensity and SST is generally agreed for all TC areas.

However, the fact that most TCs do not achieve their SST-bounded MPI suggests other mechanisms are at play.

Instead of assigning each TC observation according to the SST group that it was detected, maximum sustained winds of all TC tracks could be directly linked to SST values (Evans, 1993; Michaels et al., 2006). Results of correlating SST values with TC

129 intensities of multiple basins also caution the overreliance of a single SST predictor at explaining the sustained maximum TC winds (Evans, 1993; Henderson-Sellers et al.,

1998; Goldenberg et al., 2001). Such an attempt to determine the nature of the SST-TC relationship is further complicated by SST cooling along storm tracks (Mei and Pasquero,

2013). Using a relatively higher resolution SST dataset, Michaels et al. (2006) shared a similar concern in addressing TC intensification by using SSTs alone. Although a SST threshold of 28.25°C had been determined for North Atlantic TCs to reach maximum intensities at the major hurricane strength (≥ 50 m s-1), further SST warming does not significantly contribute to increasing maximum TC wind speeds (Michaels et al., 2006).

Observations of TC intensity have been linked to climatological SSTs in multiple basins (Evans, 1993) and updated for the North Atlantic TCs (Michaels et al., 2006).

However, such research for the ENP basin has been only presented in Whitney and

Hobgood (1997) but has not yet been evaluated statistically. Though it is expected that

ENP storm intensities respond to SST change as found in other basins (Evans, 1993;

Michaels et al., 2006), there may be differences in which ENP storm intensity shows its

SST dependence. Direct attribution of near-time SST values to TC winds would also be an improvement over previous basin-wide studies (Emanuel, 2005; Webster et al., 2005) and reveal how contemporaneous SST would facilitate the ultimate TC intensity at different levels of storm strength.

The limiting effect of SST on maximum ENP storm intensity was better understood by attributing SST as the upper bound of MPI. Based on the exponential function in fitting the maximum TC intensity using SST in the North Atlantic basin

(DeMaria and Kaplan, 1994), a linear function has been extended for ENP storms

130 (Whitney and Hobgood, 1997). After all storms’ translational speeds have been removed from every six-hourly observations of sustained winds, the eastern Pacific MPI (EPMPI; m s-1) is developed as such:

EPMPI = Co + C1 (SST), (Equation 4-1)

-1 whereas SST (°C) is associated with a slope (parameter estimate) of C1 = 79.17262 m s

-1 -1 and a y-intercept (constant) of C0 = 5.361814 m s °C . However, it has been nearly two decades since this empirical relationship on the limiting effect of SST on maximum ENP storm intensity was first documented (Whitney and Hobgood, 1997). Due to the SST warming observed worldwide (Xie et al., 2010) and locally (Ralph and Gough, 2009), we hypothesize that the upper bound of maximum ENP storm intensity should have also shifted. If this direct relationship holds, then EPMPI pertaining to each SST group is expected to increase as well. As such, the SST-dependent theory of maximum potential intensity can be refined using the most recent (1982-2013) SST climatological dataset which has a higher temporal-spatial resolution.

The objective of this study is to quantify and update the relationship between TC intensity and its underlying SSTs for ENP storms. Though there is more than one convention (Collins and Mason, 2000; Ralph and Gough, 2009) to derive the longitudinal boundary as to how the main development region of ENP basin should be sub-divided, to better understand its relationship with its environment, there is a general agreement that the local storm activity should be longitudinally stratified into eastern (EDR; 10-20°N and 85-112°W) and western (WDR; 10-20°N and 112-140°W) development regions.

Hence, regional differences of how TC intensity responds to SST fluctuation will be investigated. Given the previous understanding of the regional sensitivity to

131 environmental influences (Collins and Mason, 2000; Ralph and Gough, 2009; Jien et al.,

2015), it is expected that intensities of all WDR storm observations and their lifetime maxima would be more sensitive to SST changes.

Section 2 describes details of the TC and SST datasets and methods used to analyze them. Section 3 provides one rationale to divide ENP basin longitudinally when examining possible environmental linkages. In section 4, the spatial misalignment of the warmest SST and maximum storm intensity is examined. Section 5 explores the impact of SST on the maximum lifetime intensity. Section 6 updates the empirical functions by incorporating SST as the dependent variable using the most recent TC climatology, followed by conclusion in section 7.

4.3 Data and Methods

Storm track data from the Central Pacific and eastern North Pacific basins are retrieved from the best track HURDAT2 (Landsea et al., 2013) of National Hurricane

Center. It records TC characteristics such as a storm’s one-minute maximum sustained wind speed and its geographic positions at 6-hourly intervals. Though non-developing tropical depressions are better integrated in the current storm dataset, the underestimation of storm intensity using aerial reconnaissance led to our exclusion of storm data during the pre-satellite era. Due to uncertainty in wind speed estimation below the tropical storm designation, all non-developing tropical depressions are excluded in our analysis.

Modifications from the previous HURDAT (Davis et al., 1984) version include the maximum radii distances of 34kt (18 m s-1), 50kt (26 m s-1) and 60kt (31m s-1) at all four quadrants of cardinal directions. In addition, reports of storm track positions other

132 than the synoptic (00, 06, 12, and 18Z) timeframes, when locations of intensity maxima and landfalls are detected, were recently added to the storm dataset after 2012. Because the goal of the study is to refine the analysis of the relationship of SST and TC intensity from Whitney and Hobgood (1997), the consideration of TC climatology in the study coincides with the data availability of the daily SST dataset, covering the period of 1982-

2013.

Since only storm data within ENP boundary are of interest for this work, a total of

1445 six-hourly TC observations that lie west of 140°W were eliminated. During the 32- year span, a total of 496 ENP storms were associated with 12657 six-hourly TC observations, approximately 25.5 six-hourly observations per TC. In comparison, an average of 24.1 six-hourly observations are associated with each storm during the 1963-

1993 period from Whitney and Hobgood (1997), slightly a shorter storm duration per TC than the analysis period of the study. Of the 1445 ENP storm observations eliminated beyond 140°W, 84 of them had continued to track further westward, past the

International Date Line.

Daily mean SST values are retrieved from Optimum Interpolation SST dataset from the National Oceanic and Atmospheric Administration / National Climatic Data

Center (NOAA/NCDC) with a spatial resolution of 0.25° by 0.25°2. This refined spatial- temporal resolution of SST record had been applied to investigate the TC-induced SST response (Hart et al., 2007; Dare and McBride, 2011a) but has not been applied to update the relationship of ENP TC intensity with SST. Of the 12657 six-hourly observations,

188 were observed over land. Hence, SST values over 12469 TC observations from

2 NOAA High Resolution SST data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their Web site at http://www.esrl.noaa.gov/psd/

133 1982-2013 are matched and extracted based on the linear interpolation of nearby SST grids. Because TCs had been shown to reduce local SST, averages of SST values one week prior to and after each storm passage at each TC position are also included to monitor the progression of SST-dependent TC intensity. In total, a time series of 15 days of SSTs are obtained in association with each TC position.

The SST control of the maximum TC intensity has been established by a clear statistical linkage when SST is treated as a continuous variable. However, because there remain uncertainties in the measurement of storm wind speed at the tropical depression stage, records with wind speed below 18 m s-1 are often considered unreliable (Collins,

2010). While the SST effect at various TC stages is of inherent interest, only maximum wind speeds for named storms (NS; sustained wind speed between 18 to 32 m s-1), hurricanes (H; sustained wind speed between 33 to 50 m s-1) and major hurricanes (MH; sustained wind speed greater than 50 m s-1) are quantitatively related to the underlying seawater temperature. Any TC observation over land is of necessity associated with missing SST data. Consequently, a total of 188 observations and 2 storms that had their maximum lifetime relative velocities detected onshore are removed from the correlation analysis.

To update the previous Eastern Pacific Maximum Potential Intensity developed over the 1963-1993 period (Whitney and Hobgood, 1997), all TC tracks are binned into a total of 13 SST groups from 19°C to 31°C, each with a 1°C range. As an alternative way to demonstrate the impact of SST on TC intensity, daily SSTs seven days prior to storm arrivals were averaged and rounded to the nearest whole number. As evidence of local

SST warming, storm intensities of 30°C and 31°C SST groups were extended from a

134 range of 19-29°C groups of the previous TC climatological record (Whitney and

Hobgood, 1997). The maximum intensity values at each SST group are extracted to develop an empirical function that best fit these data points. Results were repeated for each of the MDR subdivisions. Due to a lack of data for SST groups below 19°C (23°C),

27 (16) observations were removed from WDR (EDR).

To further understand the longitudinal variation of ENP storm activity and how

TC intensity would respond to SST differences, TC track record is longitudinally binned into subdivisions of MDR: EDR and WDR. Analyses are repeated when storm observations are spatially partitioned within EDR and WDR boundaries according to their six-hourly locations, while the regional affiliation of each storm is linked to the region where the maximum intensity is achieved (Collins and Roache, 2011). To compare with previous analysis performed by Whitney and Hobgood (1997), the relative velocity of storm track is complied and calculated by subtracting its translational speed from the maximum wind speed. Translational speed is calculated by dividing the total distance travelled 6 hours before and after the storm’s current position, except at the initiation and downgradation points where 6-hourly movements between the first two points and last two points are computed respectively (DeMaria and Kaplan, 1994;

Whitney and Hobgood, 1997; Mei et al., 2012). The average translational speed for

12657 tracks is about 4.4 m s-1, comparable to the value of 4.7 m s-1, noted in Whitney and Hobgood (1997).

135 4.4 Longitudinal Division of ENP basin

Of the 496 ENP storms identified, many did not remain within the region where they first formed. Table 4-1 shows if storms were regionally separated based on locations where they were originally detected, more than three quarters of seasonal storm count were derived from EDR. However, many had achieved their maximum winds and spent the majority of their lifespan upon entering WDR. During this type of transition, others have suggested that these EDR-originated systems should be designated as WDR storms (Collins and Mason, 2000; Collins and Roache, 2011). As such, there tends to be

Table 4-1: Annual Number of storm count (NS, H and MH) stratified between EDR and WDR averaged from 1982-2013 based on the location where the peak intensity is established. Total storm count is also sorted according to where storms were initially detected

EDR WDR Initial 11.8 3.8 NS 6.7 8.8 H 3.6 5.0 MH 3.0 5.6

a greater number of WDR storms produced during any storm season within a given annual cycle. Of the total 496 ENP storms between 1982-2013, EDR to WDR storm ratio is approximately 3:4. In other words, a net movement of EDR-originated storms into WDR occurs every season. Although it is much more common for WDR storms to originate from EDR, Rosa (1994) is the only exception where its region of translocation is reversed. Rosa (1994) originated in WDR and unconventionally migrated to EDR,

136 where it peaked as a Category 2 (wind speed greater than 42.7 m s-1) storm prior to dissipation over Mexico.

When the six-hourly TC observations are regionally assigned to where each of the

496 ENP storms attained its maximum lifetime intensity, TCs that tracked beyond 140°W are also included. Of all six-hourly track records associated with WDR storms, only 19 out of 8679 WDR storm tracks are recorded to have reached land. In comparison, a greater proportion (170 out of 5425) of EDR storm track points made landfall. While

WDR storms are generally expected to travel westward, sometimes passing over Hawaii, only one had crossed Hawaii. This is due to a combination of the fact Hawaii has a smaller land surface and that most of WDR storm landfalls on the North American continent require westerly winds to induce strong TC recurvature. Interestingly, when comparing the average translational speed between EDR and WDR landfalling storms, significant differences were detected. The average translational speed (15.3 m s-1) associated with these EDR landfalling storm records is almost twice as large as the average (7.4 m s-1) of all landfalling WDR storm tracks. Part of the difference is due to an overall more rapid recurvature of EDR storms when making landfalls.

During the period of 1982-2013, the annual average number of ENP storms count has been decreasing. This is mostly attributed to a below-normal number of WDR storms.

On the other hand, the number of EDR storms stays relatively the same. Using the Theil-

Sen slope estimator, a reduction of 0.17 WDR storms per year (5.3 storms over 32 years) is observed to be statistically significant at the 5% confidence level (Figure 4-1). This value is large considering the seasonal average of storm count is 8.8 storms in the region.

Since WDR storm activity represents a significant proportion of ENP storm activity, this

137 reduction of WDR storm count is a major contributor to a decreasing trend (significant at p < 0.05, not shown) of total ENP storms during the same time period.

During an average hurricane season, more than half of the total ENP storm count (15.5) average per season became WDR storms. However, the most recent (1981-2013) TC climatology indicates the number of EDR-originated WDR storms has diminished. Trend analysis shows the annual proportion of these transitioning WDR storms to the total ENP storm count has decreased significantly (p < 0.05) averaging to a reduction of 5.1 WDR storms over the time period (Figure 4-2). In addition, the reduced WDR storms matches well with the diminishing number of EDR-to-WDR storms. As such, this downward trend implies seasonal WDR storm frequency is becoming less reliant on EDR for storm initiations, leading to further storm growth and development in the WDR.

Figure 4-1 Annual proportions of WDR storms to all ENP storms from 1982-2013

138

Figure 4-2 Annual proportions of EDR-derived WDR storms to the total WDR storm count from 1982-2013

While a reduced number of WDR storms may seem that WDR storm seasons are progressively less active, it should be interpreted more carefully. Most of the downward trend of total WDR storm (Figure 4-1) is explained by the lack of transfer of storms from the EDR to WDR (Figure 4-2). At the same time, the number of WDR storms that originated from the WDR boundary has remained stable. In fact, the numbers of WDR hurricanes and major hurricanes have both increased significantly (Collins and Mason,

2000).

139 4.5 Displacements of maximum TC intensity and maximum SST and initial genesis point

Under the influence of the tropical easterly flow, ENP storms predominantly track westward and gradually divert poleward, over cooler tropical water prior to dissipation.

Many of them can be traced as far upstream as African easterly waves that only intensify to TC strength after crossing the North Atlantic Ocean, the Caribbean Sea and the Gulf of

Mexico. While most storms form close to the North American coast, these storms typically undergo intensification and obtain their maximum lifetime intensity a considerable distance from the coast. Figures 4-3a and 4-3b show the northwest tendency for storms before gaining peak intensity. On average, it requires a storm to shift 3.7° northward and 9.6° westward of its original identification point to achieve its maximum lifetime wind speed.

When storms are separated into EDR and WDR, regional variations of latitudinal

(Figure 4-4a) and longitudinal (Figure 4-4b) differences are observed despite the average locations of their maximum lifetime intensities are found northwest of their initial detections. While the mean latitudinal difference between the initial storm detection and its maximum intensity of EDR is only 0.4° northward of WDR storms, WDR storms’ mean longitudinal difference is 5.3° westward of EDR storms. Although a greater longitudinal shift is detected for WDR storms, it is mostly attributed to the fact that 58% of WDR storms were initiated in the EDR and experienced longer periods of intensification than those WDR storms that were originated within WDR. However, even if WDR storms that originated from EDR are removed, WDR’s mean longitudinal difference is still maintained at 2.4° greater, significant at the 5% confidence interval,

140 than that of EDR storm’s average longitudinal shift. The ability for WDR storms to track over a greater distance before achieving maximum intensity is linked to a longer longitudinal shift. This is supported by a clear right-tailed distribution for longitudinal differences, mainly attributed to WDR storms (Figure 4-4b).

Although ocean temperature tends to be warmer where storms initially formed, their genesis locations are not necessarily the warmest SST experienced during a storm’s lifecycle. The warmest SST tends to provide the most optimum condition for TC development. However, its lifetime maximum intensity is rarely achieved where the warmest SST is encountered. On average, the maximum intensity is located northwest of where the highest weekly SST is found a week prior to the storm passage. However, the latitudinal (Figure 4-5a) and longitudinal (Figure 4-5b) differences are more evenly distributed compared to displacements between maximum intensity and initial storm genesis location (Figures 4-3a and 4-3b). Such a comparison also demonstrates the highest SSTs are experienced closer to maximum lifetime intensities than initial genesis locations of ENP storms. When storms are detached from a warm sea surface, storm intensity is maintained possibly in response to a delayed convective feedback. An average delay of 1.7 days is observed for a storm to establish peak intensity after experiencing its warmest seawater versus approximately 3 days after a storm was first generated. This coincidence where the highest storm intensity is matched with the highest SST was found for only 18 storms, representing less than 4% of the total storm count.

The spatial misalignment of maximum SST and maximum intensity is generally displayed by North Atlantic storms though the extent of displacement differs compared to

141 (a)

(b)

142 (a)

(b)

143 (a)

(b)

144

ENP storms. While two-thirds of North Atlantic storms met its maximum SST within 5° latitude of maximum lifetime intensity (Michaels et al., 2006), 83% of ENP storms had done so. Overall, this latitudinal displacement for ENP storms’ maximum intensity is

2.2° northward of where maximum SST is achieved, half of North Atlantic storms’

(Michaels et al., 2006). These comparisons relative to North Atlantic storms signify ENP storms’ stronger linkage to areas with the warmest SST in acquiring maximum intensity.

For storms that took an eastward recurvature or landfall, the highest SSTs are often found near shore when their strengths are diminishing, possibly in conjunction with extratropical transitioning. Interestingly, peak intensities for these storms took less time to establish than those that experience warmest SST at their earliest stage of development. Overall, the northward distance between warmest SST and maximum TC intensity is greater for TC observations in WDR (Figure 4-6a). As contrasting environmental conditions near shore and in the open water could dictate a regional difference of TC development, WDR reveals a strong longitudinal displacement (Figure

4-6b) between locations of the warmest SST and maximum lifetime intensity. Although this could also be attributed to WDR storms that have encountered the maximum SST while they were still in the EDR; 137 (or 40%) of WDR storms match this criterion. This is lower than those storms (58%) that formed in EDR but achieved maximum lifetime intensities in WDR. Even when these storms are excluded, storms that attained maximum intensity and maximum SST in the WDR still maintain a 2° westward lead over EDR storms.

145 (a)

(b)

Figure 4-6 a) Distribution of latitudinal differences between locations of maximum SST and maximum lifetime intensity for EDR and WDR storms from 1982-2013 and b) same as a), but for longitudinal differences

146 4.6 Correlation between SST and TC intensity

4.6.1 All Observations

Since the onset of a storm induces a cold wake (due to turbulence induced entrainment of below surface waters) and thus obscures the SST-TC intensity relationship, a seven-day SST average prior to the storm arrival was used to correlate with the intensity observations. Possibly due to a large sample size (n = 12469), Figure

4-7 shows the linear regression fit which is significant (p < 2.2 × 10-16) with a slope of

0.889 and r2 value of 0.020 for storm track positions observed prior to crossing the westernmost boundary division at 140°W. Though the explained variance is small

(though significant), it is higher than a similar correlation analysis performed in the North

Atlantic basin (Michaels et al., 2006). If storm tracks that traversed west of ENP boundary are also included (not shown), the linear regression is still significant. Such a significant linkage between SST and TC intensity is consistent with that of Ralph and

Gough (2009), which considered the correlation of TC activity of all storm strengths and basin-wide SST at a monthly scale. Although our analysis shows the explained variance is small, this association of near-time SSTs and TC wind speeds is consistent with

Michaels et al. (2006). Though SST certainly provides a certain degree of influence on the upper bound of TC intensity, SST is clearly far from being the sole control of TC intensity. Even when elevated SSTs are observed, other atmospheric factors such as wind shear diminish TC strength (Landsea et al., 1998; Maue, 2009). The fact that the r2 in the study is higher than a similar analysis conducted for North Atlantic storms

(Michaels et al., 2006) underscores greater SST influences on ENP storm intensities.

147

Figure 4-7 Scatterplot and regression fit for all observations of TC intensity against the weekly SSTs averaged over seven days prior to storm passage

Upon investigating how TC intensity is related to SST at a daily scale, the temperature measurement at the ocean surface demonstrates a two-way, SST-TC relationship. Though it is shown in Figure 4-7 that SST directly contributes to TC intensification, storm passage is capable of cooling the ocean surface at least a week prior to its arrival (Cione and Uhlhorn, 2003; Mei and Pasquero, 2013). Figure 4-8 shows the

SST averages seven days before and after (total of 15 days) for locations of all 12469 storm tracks. Generally, this cooling remained for at least one week since the storm has departed, with the greatest SST reduction experienced during the period between one day before onset until the day of storm arrival. Such a sharp drop in the ocean surface temperature is likely the result of a greater vertical mixing due to a close proximity to the radius of maximum winds and storm movement, which entrains cooler waters beneath the

148

Figure 4-8 Daily SST response (with standard error) to the influence of storm passage over a 15-day time series

sea surface. Such a cooling effect lasts longer and is more pronounced for locations that experience storms with lower translational speed and higher storm intensity (Dare and

McBride, 2011a). Compared to the North Atlantic TCs, since the average translational speed is higher, SST reduction associated with a slower storm movement is greater in the

ENP basin.

When all 12469 observations of storm intensity are correlated with their same- day SSTs (i.e. Day 0), a weak relationship is found (slope = 0.63078 and r2 = 0.01094).

Though the result of a linear least-squares fit maintains its statistical significance (p < 2.2

× 10-16), the fact that the greatest SST decrease occurs during the day prior to storm arrival weakens the near-time SST-TC intensity relationship. The effect of SST cooling

149 on the SST-TC intensity relationship is likely to be most important when storms are undergoing rapid intensification, particular during September, the peak month of ENP hurricane season (Kaplan et al., 2010). Once the TC-induced rate of SST cooling diminishes, storms are found to intensify at a faster rate (Cione and Uhlhorn, 2003).

While the choice of daily SST at least a week before storm arrival does not dramatically change the explained variance for the linear fit with TC intensity (Table 4-2), the use of weekly SST dataset (Michaels et al., 2006) may not be as closely aligned to the initial storm detection as daily level. For instance, when a storm has just formed, the SSTs averaged during the past week may be more critical at determining storm genesis than TC intensification. Figure 4-9 compares a 15-day time series of SST averages centred at the genesis stage and the time when a storm’s peak intensity is observed. From the day before until the day after the detection of maximum lifetime intensity, the most dramatic

SST decline can be discerned. This is mostly likely due to a combination of a slower translational speed and a strong TC wind speed in churning up colder water from below (Mei and Pasquero, 2013). Since weak TC winds are associated with fast-moving, decaying storms, virtually little to no SST recovery is observed past the storm genesis stage. While, the average SST recovery at locations that just experienced peak storm intensities is 0.06°C warmer than the average of all TC observations.

4.6.2 MDR Subdivisions

When TC observations are subdivided and binned into EDR and WDR, the sensitivity of TC intensity to SST is regionally dependent. With the addition of a linear regression of least-squares fit, Figure 4-10a (b) shows a scatterplot of EDR (WDR) storm

150 Table 4-2: Regression analysis of storm intensity of all EDR and WDR observations with daily SST at seven days prior to storm passage with Day 0 being the arrival day of storm passage

EDR WDR Region / Days Before slope r2 slope r2 7 -0.79 0.01 1.55 0.06 6 -0.81 0.01 1.56 0.06 5 -0.72 0.00 1.60 0.06 4 -0.67 0.00 1.61 0.06 3 -0.69 0.00 1.60 0.06 2 -0.67 0.00 1.62 0.06 1 -0.63 0.00 1.59 0.06 0 -1.43 0.02 1.32 0.04 7-day Average -0.79 0.00 1.64 0.06

Figure 4-9 Same as in Figure 4-8, but the 15-day SST is centred around the storm genesis and maximum lifetime TC intensity. Dashed lines represent the daily range of the highest rate of SST reduction

151 (a)

(b)

Figure 4-10 a) Relative wind speeds for all EDR storm observations and b) same as a) but for all WDR storm observations

152 intensity and SST averaged over a 7-day period prior to TC detection. Compared to

EDR, WDR storm intensities are more responsive to SST differences. In fact, the slope is nearly twice as large for MDR storms during any day of the week prior to the TC arrival.

From Table 4-2, the worst fit is found at Day 0, when the negative (positive) slope is strongest (weakest) at EDR (WDR). In the EDR, this corresponds to the time when the greatest rate of SST reduction due to storm influences is observed (Figure 4-11). This

SST reduction in EDR continues until a local minimum is observed two days after the storm appearance. While the temperature of the ocean surface is noted to reduce at a slower pace a week prior until the day before EDR storms arrive, SST is observed to start cooling only four days prior to TC arrival until the local minimum appears the day after the WDR storm passage. An overall greater SST reduction during storm passages is supported by stronger TCs in EDR where less cold/warm water stratification is induced during the local hurricane season so that cold water is more accessible below a shallower

EDR thermocline.

Subdividing ENP storm observations into EDR and WDR sectors is also critically important to the understanding of the regional difference in which TC wind speed responds to SSTs at different storm development stages. In particular, the relationship of

WDR storm intensity with the underlying SST shows a greater significance when the maximum lifetime intensity is examined. Such a finding of SST control over the maximum lifetime TC intensity would not have been possible if the ENP basin had not been subdivided. Similar to the above result, little association with SST is detected on the day when maximum relative wind is detected. On average, the greatest correlations are observed two days before TC arrivals when the impact of TCs on ocean temperature

153

Figure 4-11 Same as Figure 4-8, but separated for storm passages of EDR and WDR observations. Dashed lines represent the reference SSTs of local minima over for EDR and WDR

is less pronounced. This observation has been previously hypothesized but not proven until the use of the daily SST dataset. In terms of ocean surface cooling, EDR storm observations generate greater SST reductions at least a week prior to storm geneses and the attainment of maximum intensities; however, TC passages at both regions display little evidence of SST recoveries (Figures 4-12a and 4-12b). However, dramatic differences of SST responses between EDR and WDR are noticed for locations that experience passages of maximum lifetime intensities (Figure 4-12b). EDR storm records tend to be accompanied by greater SST reduction than WDR’s.

Regional sensitivity for the maximum lifetime intensity to SSTs is found to exist between EDR and WDR storms. Figure 4-13a (b) depicts the scatterplots of maximum

154 (a)

(b)

Figure 4-12 a) Same as Figure 4-11, but for the average SST conditions during storm genesis and b) same as a) but during the maximum lifetime intensity

155

(a)

(b)

156

lifetime intensity of relative winds with average weekly SSTs, seven days prior to TC arrival for all 283 (213) WDR (EDR) storms. The highest slope for the linear regression fit is found to occur four days prior to TC occurrence. Coincidently, this is also when the local SST maximum is observed during when the time series of 15 daily averages of

WDR lifetime maximum intensity is found (Figure 4-12b). Overall, the results underscore the importance of SSTs only for WDR storm intensification. Similar to

Michaels et al. (2006), the importance of SST threshold to reach certain levels of storm intensity in the North Atlantic basin is found. This study found the importance of an SST threshold of 25°C (24°C) below which no TC of a major hurricane (hurricane) status is observed in WDR. Significantly, this minimum SST requirement is well below the

28.25°C SST requirement for North Atlantic major hurricanes. Compared to EDR, due to its seasonally higher SSTs, higher thresholds are required to achieve both hurricane

(25.25°C) and major hurricane (27°C) strengths. This highlights the absolute SST requirement to sustain major hurricanes is not uniform across and within TC development basins.

Although a SST threshold of 28.25°C is identified for North Atlantic storms to achieve major hurricane status, Michaels et al. (2006) shows North Atlantic TC winds that achieved maximum intensities past this threshold has little to no relationship with corresponding SSTs. Even if only the intensities of major hurricanes were investigated,

SST does not seem to be the overriding factor at determining the maximum TC winds.

However, at ENP basin, the maximum intensity of WDR major hurricanes demonstrate significant relationship with daily SSTs. Figure 4-14 shows the SST and TC intensity for

157 WDR major hurricanes three days before relative winds are achieved. A linear regression analysis on this day has the highest slope and explained variance. To compare with the previous result shown in the North Atlantic basin (Michaels et al., 2006), this correlation test was also replicated using maximum winds (relative velocity + translational speed); and, the result also shown similar statistical significance at p < 0.01. Though maximum intensity of WDR major hurricanes are critically dependent of SST conditions, many

WDR storms remain weakly developed despite encountering favourable SST conditions that are well above the minimum requirement (25°C) to attain major hurricane strength.

In contrast, while maximum lifetime TC intensity for all EDR storms is not found to positively correlate with ocean surface temperature, a weak, but positive, relationship was found for maximum winds of EDR major hurricanes and their underlying SSTs.

Figure 4-14 Scatterplot and linear regression of maximum lifetime intensity of relative velocity for WDR storms that reached the strength as major hurricanes with SST at the locations where storm passages occur three days ago

158 4.7 Upper bound of TC Intensity by SST Groups

In addition to statistically linking SSTs to observations of TC winds, the impact of increasing SST on the maximum sustained winds can also be explored by stratifying records of TC wind speed into SST groups based on the work of DeMaria and Kaplan

(1994). By rounding average SST values a week prior to TC arrival to the nearest integer, TC winds have generally maintained the same relationship with increasing SSTs as found by Whitney and Hobgood (1997). Overall, warmer SSTs are able to sustain a higher maximum TC intensity. Compared with Whitney and Hobgood (1997), an update to the current TC record shows the maximum intensities at some of the highest SST categories are increasing.

Table 4-3 summarizes the 1982-2013 storm intensities in each SST bin, updated from the1963-1993 climatology (Hobgood and Whitney, 1997). The creation of two extra

(30°C and 31°C) SST bins are now required, though it has been noted that there were few storm observations from the 1963-1993 climatology that encounter SSTs greater than

29.5°C. Similar proportions (77%) of the total TC intensity observations were found in the 26°C or higher SST bins, though the highest average TC intensity has shifted from

26°C (Whitney and Hobgood, 1997) to the 27°C bin (Table 4-3). A gradual decrease in the average TC intensity with increasing SST is partly due to degrading storms recurving over shallow and warm water prior to dissipation at landfall. It could be also attributed to a tendency for storms to form and develop over the highest SSTs and slowly decay while passing over cooler water (Whitney and Hobgood, 1997).

Though the total number of analysis years is one year more than that of Whitney and Hobgood (1997), there are an additional 1595 SST observations associated with the

159 TC record. This value is large considering the average number of the SST-TC record, between 1963-1993, per year is 356.8, roughly averaging 36 more observations/year during the 1982-2013 period. Although it is difficult to attribute the recent spurt of TC activity to climate change due to an inconsistent monitoring of TC tracks when compared with data prior to the 1970s, the recent SST warming has shifted the corresponding surge of TC observations towards the distribution with higher SST values. Figure 4-15 shows that most of the surge in TC detection associated with SST warming occurs preferentially in the higher SST bins. Likewise, less of an increase in TC observations are observed to be distributed at lower SST values. Thus, this improvement of thermodynamic condition is able to allow storms to extract energy to reach higher maximum TC intensities.

Table 4-3: Summary of TC intensity records stratified into SST bins

Sample Average Maximum SST (°C) Size Intensity (m s-1) Intensity (m s-1) 19.0 34 7.4 12.2 20.0 74 8.5 16.2 21.0 148 10.2 28.3 22.0 280 12.2 38.8 23.0 475 13.8 40.1 24.0 714 16.4 48.7 25.0 952 21.6 55.7 26.0 1386 24.4 60.0 27.0 2367 24.4 66.3 28.0 3304 21.4 68.5 29.0 1966 20.4 74.9 30.0 708 20.1 78.2 31.0 34 18.6 50.4

160

Figure 4-15 Anomaly of TC observations in each SST bin during 1982-2013 compared to 1963-1993

Consequently, storms (in Table 4-4) that are responsible for achieving the highest maximum intensity in each SST bin differ from those found in Table 2 of Whitney and

Hobgood (1997). TCs Linda (1997), Kenna (2002) and Rick (2009) had all exceeded the maximum strength of Trudy (1990), which was found to be the most intense storm from an earlier 1963-1993 period (Whitney and Hobgood, 1997).

Since it is of interest to understand the limiting effect of SSTs on storm intensity, only the maximum, 99th, 95th, 90th and 50th percentiles of TC winds are plotted in Figure

4-16. Initial linear increases are noted for 90th percentiles or above though the rate of increase flattens, or even decreases, as bins of higher SSTs are approached. Only the maximum intensity values maintained a linear growth rate with increasing SSTs.

Because Figure 4-17 shows that the maximum intensity dips dramatically at the 31°C bin, its value is not included in the empirical derivation of MPI, possibly due to a lack of

161 observation in the highest SST category. Using a locally weighted scatterplots smoothing

known as LOWESS (not shown), the fit is best described as a linear relationship. The

updated (1982-2013) relationship between empirical MPI and SST at ENP basin has an

equation of the form,

ENPMPI = A (SST) + B, (Equation 4-2)

where ENPMPI (m s-1) represents the revised the ENP maximum potential intensity, with

a constant of B = -91.4863 m s-1 and a slope of A = 5.7975 m s-1 °C-1. Overall, 96% of

total data variance is explained by ENPMPI. Compared to Whitney and Hobgood

(1997), this updated ENPMPI sees an increase in slope showing the linear fit is stronger

than the past TC climatology (Figure 4-17).

Table 4-4: Details on storms that attained the highest relative winds at each SST bin

Maximum Relative SST Latitude Longitude Winds Winds (°C) Year Name (°N) (°W) (m s-1 ) Strength (m s-1) Region 19.0 2001 NARDA 16.3 139.6 15.4 TD 12.2 WDR 20.0 2005 KENNETH 16.5 139.4 23.1 NS 16.2 WDR 21.0 2013 GIL 13.7 138.2 33.4 H 28.3 WDR 22.0 2004 ISIS 16.3 135.7 43.7 H 38.8 WDR 23.0 2008 CRISTINA 14.1 133.7 46.3 H 40.1 WDR 24.0 1995 ADOLPH 17.8 108.8 51.4 MH 48.7 EDR 25.0 2006 EMILIA 30.8 125.2 59.2 MH 55.7 WDR 26.0 1988 CARLOTTA 21.7 123.7 64.3 MH 60.0 WDR 27.0 1984 NORBERT 19.4 116.3 72.0 MH 66.3 WDR 28.0 1998 KENNA 18.3 108.3 74.6 MH 68.5 EDR 29.0 1997 LINDA 17.7 110.3 79.7 MH 74.9 EDR 30.0 1997 LINDA 17.1 109.6 82.3 MH 78.2 EDR 31.0 2009 JIMENA 15.7 105.5 54.0 MH 50.4 EDR

162

Figure 4-16 Storm relative winds at the maximum intensity and 99th, 95th, 90th and 50th intensity percentiles at all SST bins after translational speeds have been accounted

Figure 4-17 Scatterplot and regression line drawn for the maximum storm intensity attained by each SST bin with and without the maximum intensity at the 31°C bin

163 When MDR is subdivided, it is evident that there are regional differences in the distribution and magnitude of TC intensities among the SST bins. A warmer (colder)

SST condition is experienced at EDR (WDR), where 16 (slightly over a thousand) TC track records are associated with SST bins lower than the 23°C category (Tables 4-5 and

4-6). Towards the upper limit of TC intensity distribution, over 95% of EDR storm observations (Table 4-5) are in the 26°C category or greater, while only 64% for the

WDR (Table 4-6). When the maximum intensity is regionally compared for each SST bin, maximum intensities observed at the WDR are higher at SST categories of 25°C and

27°C (Table 4-6), while the highest attained TC intensities at other SST bins are higher at

EDR (Table 4-5). Although maximum TC intensity is not unanimously favoured within a particular MDR subdivision, the average storm intensity at each SST bin is consistently stronger at EDR (Table 4-5).

When maximum, 99th, 95th, 90th and 50th percentiles of TC intensity are plotted for EDR (Figure 4-18a) and WDR (Figure 4-18b), regionally specific positive relationships are observed for TC intensities. In the EDR, a more smoothly increasing maximum TC intensity is observed while WDR storm intensity tends to experience a steeper increase. At TC intensities of lower percentiles, this increase is slower towards the higher SST categories for both regions. When the average intensity within each SST bin is compared between regions, observations from both regions show less sensitivity to warmer SST conditions than the maximum intensity. Differences between the two regions are dramatically different towards TC winds at the higher percentiles. WDR storm intensities greater than the 90th percentile of each SST bin are more dependent of

SSTs than EDR’s.

164 Table 4-5: Summary of EDR storm intensities stratified into SST bins

SST (°C) Sample Average Maximum Bin Size Intensity (m s-1) Intensity (m s-1) <23.5 16 20.9 45.3 24 33 23.6 52.2 25 48 29.1 48.7 26 178 30.2 60.3 27 597 27.8 62.7 28 1728 22.3 73.3 29 1714 18.1 78.2 30 685 16.7 62.2 31 27 13.3 33.1

Table 4-6: Summary of WDR storm intensities stratified into SST bins

SST (°C) Sample Average Maximum Bin Size Intensity (m s-1) Intensity (m s-1) <18.5 27 6.9 11.8 19 34 7.4 12.2 20 74 8.5 16.2 21 148 10.2 28.3 22 278 12.1 38.8 23 461 13.8 40.1 24 681 16.1 45.2 25 904 21.4 55.7 26 1208 23.2 60.0 27 1770 24.6 66.5 28 1576 24.3 66.2 29 252 21.2 68.7 30 23 12.0 21.9 31 7 10.7 13.7

165 (a)

(b)

Figure 4-18 a) Same as Figure 4-16., but only for EDR storm observations and b) Same as Figure 4-16., but only for WDR storm observations

166 Since the ENP basin exhibits longitudinally different SST conditions for TC development, ENPMPI derived above might not best represent the upper limit of maximum intensity as a function of SST category when local storms are regionally divided. Based on the above findings, the MPI in the EDR (ENPMPIe) and the WDR

(ENPMPIw) may require different empirical models than that was performed for the entire basin. Due to a limited record of maximum TC intensity associated with each SST bin (Table 4-5), only data from 24°C to 30°C bins are used for ENPMPIe. By examining the LOWESS curve (Figure 4-19), it appears a linear function is still the most suitable function in describing the data distribution (up to 29°C) for ENPMPIe, with an equation

ENPMPIe = Ae (SSTe) + Be, (Equation 4-3)

-1 -1 -1 -1 where ENPMPIe (m s ) has a constant of Be = -35.756 m s and Ae = 3.64 m s °C .

However, compared with ENPMPI, ENPMPIe’s residual standard error has grown to

7.637 m s-1. An ANOVA was attempted to compare the difference in residual error between a linear and a curve (polynomial) fit. Though a curve fit leads to a smaller error

(3.73 m s-1), it is not statistically different from the error associated with a linear function fit (Equation 4-3) at p > 0.05.

For WDR storms, while the data distribution of maximum TC intensity is observed to shift towards warmer SST bins, the LOWESS curve indicates the growth rate of maximum TC intensity has started to slow at the 26°C SST bin. Interestingly, this is also generally observed as the minimum SST requirement for TC genesis and development (Palmén, 1948; Dare and McBride, 2011b). This flattening of growing maximum TC intensity appears to initiate at even cooler SST bins for the lower

167

Figure 4-19 Scatterplot for maximum storm intensity attained by each SST bin of all EDR observations. A locally weighted scatterplot smoothing (LOWESS) is applied

percentiles curves (Figure 4-18b). As such, even with a higher SST, other environmental influences may be more important in determining the ultimate maximum TC intensity.

When linear and quadratic (curve) functions are compared for data fitting, both functions had parameter estimates that are found to be highly statistically significant. Overall, the

ANOVA results indicate that the reduced predicted error of a quadratic fit is significantly lower than that of a linear function (p < 0.01). Thus a linear function is deemed inadequate to represent ENPMPIw. Instead, an exponential decay (increase form) function is found to be more suitable, with the following equation

-Aw (SST-To) ENPMPIw = Cw + Bw e , (Equation 4-4) where To (°C) is a specified as the reference temperature and Aw, Bw and Cw are constants of the parameter estimates. With To =29 as the highest SST bin, using a non-linear least-

168 squares fit, Aw = 100.40171, Bw = -29.90966 and Cw = 0.11039. Figure 4-20 shows the

-1 data and the fitted function for ENPMPIw, with a residual standard error of 2.764 m s .

Comparatively, a linear regression fit generated an error of 4.192 m s-1. The choice of the exponential decay function in the form of (Equation 4-4) is indicative of the flattening of the fitted curve, starting at the 26°C SST bin. Though it was not taken into consideration for the functioning fit, the reduced slope is also evident for North Atlantic storms at SST bin above 28°C (DeMaria and Kaplan, 1994). In our case, due to data limitations, there remains some uncertainty for the reduced growth to continue beyond the 29°C, the reference SST bin below which ENPMPIw applies.

4.8 Conclusion

For TC development basins worldwide, SSTs have been recognized as the major thermodynamic limit of maximum TC intensity. As the long-term conditions of the ocean surface shift, TC intensities are expected to strengthen with SST increase.

Although the spatial-temporal resolution of SST data has remained a challenge in pinpointing the SST condition underlying storm tracks, near-time (daily) SST with a spatial resolution of 0.25° by 0.25° grids are employed to document its extent of impact on ENP storm winds. Due to a difference of spatial sensitivity to environmental conditions, while SST warming tends in general to elevate storm intensity, its limiting effect on ENP storm intensity is shown to vary regionally. Because the inhibiting effect of SST cooling rate is stronger (weaker) during EDR (WDR) storm passage, SST impact on TC intensity is greater for WDR storm observations and their maximum lifetime intensities.

Two ways of exploring the SST-TC intensity relationship are presented in the

169

Figure 4-20 Non-linear least-squares (NLS) fitting of maximum storm intensities bounded by SST bins at WDR, with the LOWESS curve added

study: 1) Correlating SST values with TC intensity and 2) Stratifying TC observations into SST bins of 1°C interval and then relating TC intensities of all track record with SST category. Using linear least-squares regression, results of the study indicates a statistically significant relationship exists between SST and TC intensity of all tracks.

Although SST may not be the only factor in contributing to TC intensity, the association presents a stronger linkage than North Atlantic storms. Further investigation on the maximum lifetime intensity of major hurricanes shows the SST influence on TC intensity is greatest for the most intense storms. Stratification of TC winds according to near-time

SST bins supports the previous finding of the SST-dependent MPI in the ENP basin, indicating a stronger relationship when the most recent TC climatology is compared to the past.

170 Under a regional difference in SST conditions, ENP storms are longitudinally divided to understand regional sensitivity of storm intensity to a SST limit. Compared to

EDR, maximum lifetime intensities for WDR storms are more responsive to SST fluctuations. When the maximum winds of WDR storms are categorized into different levels of TC intensity, major hurricanes display a stronger association with contemporaneous SST values. In contrast to the North Atlantic region where an absence of such relationship was found, subdividing MDR has facilitated a regional variation in which TC intensity is correlated with SST. An examination of minimum SST requirements to achieve certain TC strengths also reveals regionally distinct SST thresholds to attain maximum intensities for hurricanes and major hurricanes. Compared to the SST limit (28.25°C) for attaining major hurricanes in the North Atlantic basin, lower SST thresholds are required to establish EDR (27°C) and WDR (25°C) major hurricanes in the ENP basin.

In contrast to Whitney and Hobgood (1997), an updated ENP storm climatology revealed that the relationship between the maximum TC intensities and near-time SSTs is non-linear. While the linear association between maximum storm intensity with SST has increased with a changing climate, when MDR is subdivided, the extent of such relationship is stronger for WDR storm observations until a certain SST limit is approached. Part of this may be attributed to a less potent SST reduction by WDR storm passages. The decay in the growth rate of ENP maximum storm intensities signals an upper limit of maximum TC winds as the highest SSTs are approached.

Despite a downward trend experienced in the transport of EDR-originated WDR storms, SST warming supports the strengthening of the strongest storms. However,

171 WDR storms indicate such a response becomes non-linear with a slower rate of increase as warmer SSTs are approached. In addition, the association of TC intensity to SST is further challenged by observations of relatively weak intensity over warm SSTs before progressing and intensifying over cooler waters. Although other atmospheric factors could also have an important role in modifying TC intensity, the application of near-time

SST greatly contributes to the understanding that, in addition to the regional difference of

SST requirement for genesis and intensification of major hurricanes, the eventual intensity of major hurricanes in the WDR are greatly limited by SST conditions encountered prior to the peak of its lifetime intensity.

With the realization of higher maximum lifetime intensity and MPI under a continuous SST warming, future increase of ocean temperatures may raise the intensity of damaging storm winds. Since stronger storms tend to be sustained over longer distances, SST warming may also lengthen TC influence for a greater spatial extent.

Upon landfall, storm track may extend further inland, posing a direct threat to coastal safety. From a forecasting point of view, the possibility of an extended TC influence upon landfalling tropical cyclones may result in the broadening of uncertainties associated with near-time projection of storm path.

172 4.9 References

Cione, J. J., and E. W. Uhlhorn, 2003: Sea surface temperature variability in hurricanes: Implications with respect to intensity change. Monthly Weather Review, 131, 1783–1796.

Collins, J. M., and I. M. Mason, 2000: Local environmental conditions related to seasonal tropical cyclone activity in the Northeast Pacific basin. Geophysical Research Letters, 27, 3881-3884.

Collins, J. M., and D. R. Roache, 2011: The 2009 Hurricane Season in the Eastern North Pacific Basin: An Analysis of Environmental Conditions. Monthly Weather Review, 139(6), 1673-1681.

Dare, R. A., and J. L. Mcbride, 2011a: Sea surface temperature response to tropical cyclones. Monthly Weather Review, 139(12), 3798-3808.

Dare, R. A., and J. L. Mcbride, 2011b: The threshold sea surface temperature condition for tropical cyclogenesis. Journal of Climate, 24(17), 4570-4576.

Davis, M. A. S., G. M Brown, and P. Leftwich, 1984: A Tropical Cyclone Data Tape for the Eastern and Central North Pacific Basins, 1949-1983: Contents, Limitations, and Uses. NOAA Technical Memorandum NWS NHC 25, http://www.nhc.noaa.gov/pdf/NWS- NHC-1984-25.pdf

DeMaria, M., and J. Kaplan, 1994: Sea surface temperature and the maximum intensity of Atlantic tropical cyclones. Journal of Climate, 7, 1325–1334.

Emanuel, K. A., 1987: The dependence of hurricane intensity on climate. Nature, 326(6112), 483-485

Emanuel, K. A., 2005: Increasing Destructiveness of Tropical Cyclones over the Past 30 Years. Nature , 436, 686-688.

Evans, J. E., 1993: Sensitivity of tropical cyclone intensity to sea surface temperature. Journal of Climate, 6, 1133– 1140.

Goldenberg, S. B., C. W. Landsea, A. M. Mestas-Nuñez, and W. M. Gray, 2001: The recent increase in Atlantic hurricane activity: Causes and implications. Science, 293, 474–479.

Gray, W. M., 1984: Atlantic seasonal hurricane frequency: Part I. El Nino and 30 mb quasi-biennial oscillation influences. Monthly Weather Review, 112, 1649-1668.

173 Hart, R. E., R. N. Maue, and M. C. Watson, 2007: Estimating local memory of tropical cyclones through MPI anomaly evolution. Monthly Weather Review, 135(12), 3990-4005. Henderson-Sellers, A., et al., 1998: Tropical cyclones and global climate change: A post- IPCC assessment, Bulletin of the American Meteorological Society, 79, 19–38.

Jien, J., W. A. Gough and K. Butler, 2015: The Influence of El Niño-Southern Oscillation on Tropical Cyclone Activity in the Eastern North Pacific Basin. Journal of Climate, 28(16), 2459-2474.

Kaplan, J., M. DeMaria, and J. A. Knaff, 2010: A revised tropical cyclone rapid intensification index for the Atlantic and eastern north Pacific basins. Weather and Forecasting, 25(1), 220-241.

Kotal, S. D., P. K. Kundu, and S. K. R. Bhowmik, 2009: An analysis of sea surface temperature and maximum potential intensity of tropical cyclones over the Bay of Bengal between 1981 and 2000. Meteorological Applications, 16(2), 169-177.

Landsea, C. W., G. D. Bell, W. M. Gray, and S. B. Goldenberg, 1998: The extremely active 1995 Atlantic hurricane season: Environmental conditions and verification of seasonal forecasts. Monthly Weather Review, 126(5), 1174-1193.

Landsea, C. W., and J. L. Franklin, 2013: Atlantic hurricane database uncertainty and presentation of a new database format. Monthly Weather Review, 141(10), 3576-3592.

Maue, R. N., 2009: Northern hemisphere tropical cyclone activity. Geophysical Research Letters, 36(5), L05805, doi:10.1029/2008GL035946

Mei, W., and C. Pasquero, 2013: Spatial and temporal characterization of sea surface temperature response to tropical cyclones. Journal of Climate, 26(11), 3745-3765.

Merrill, R. T., 1988: Environmental influences on hurricane intensification, Journal of the Atmospheric Sciences, 45, 1678– 1687.

Miller, B. I., 1958: On the maximum intensity of hurricanes. Journal of Meteorology, 15, 184–195.

Palmén, E., 1948: On the formation and structure of tropical hurricanes. Geophysica, 3, 26–38.

Ralph, T. U., and W. A. Gough, 2009: The influence of sea-surface temperatures on Eastern North Pacific tropical cyclone activity. Theoretical and Applied Climatology, 95, 257-264.

174

Saunders, M. A., and A. S. Lea, 2008: Large contribution of sea surface warming to recent increase in Atlantic hurricane activity. Nature, 451(7178), 557-560.

Vecchi, G. A., and B. J. Soden, 2007: Effect of remote sea surface temperature change on tropical cyclone potential intensity. Nature, 450(7172), 1066-1070.

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, L., L. Tao, and Q. Ding, 2010: Influence of sea surface warming on environmental factors affecting long-term changes of Atlantic tropical cyclone formation. Journal of Climate, 23(22), 5978-5989.

Xie, S. -., C. Deser, G. A. Vecchi, J. Ma, H. Teng, and A. T. Wittenberg, 2010: Global warming pattern formation: Sea surface temperature and rainfall. Journal of Climate, 23(4), 966-986.

Zeng, Z., Y. Wang and C. -. Wu, 2007: Environmental dynamical control of tropical cyclone intensity - an observational study. Monthly Weather Review, 135(1), 38-59.

175 5 Chapter 5: Conclusion

5.1 Summary and Discussion:

The devastating consequences of TCs have far-reaching implications for many areas of the world. One of the most immediate challenges is to understand how various

TC characteristics respond to natural variability and climate change. This thesis has contributed to closing the knowledge gap of the relationship between ENP TCs and

ENSO fluctuation and recent SST changes and variability. The results presented in

Chapters 2 to 4 have demonstrated that a longitudinal division in the distribution of ENP storm activity substantially contributes to the detection of more responsive WDR storm characteristics to environmental influences. The six main research questions were explored and these are addressed below for each chapter.

In Chapter 2, the first two research questions were addressed:

1. How do storm metrics that measure seasonal storm activity and intensity vary as

a function of ENSO conditions in the main development region (MDR)?

Since previous attempts at attributing ENP storm variability to ENSO fluctuation based on metrics of Ocean Niño Indices and the Southern Oscillation Index were inconclusive, alternative ENSO indices that depend on variations of multiple environmental fields may be better suited for ENSO detection. Using the MEI (Wolter and Timlin, 2011) to determine ENSO events, ENSO control on seasonal variations of

176 TC intensity (PDI) is statistically established using the most recent storm climatology.

While only the degree of association (r = 0.32 and r2 = 0.10) between PDI and MEI is considered significant, it is quite similar to the bivariate correlation (r = 0.28 and r2 =

0.08) of NTC and MEI.

2. How do storm metrics that measure seasonal storm activity and intensity vary as

a function of ENSO conditions compare between the MDR subdivisions?

When ENP storms are regionally divided into the EDR and the WDR, only metrics of

WDR storm activity (NTC; Gray et al., 1994) and intensity (PDI; Emanuel, 2005) show significant associations with the MEI-based ENSO index. Both NTC and PDI exhibit the same degree of association (r = 0.35 and r2 = 0.12) with MEI. In agreement with Collins and Mason (2000) and Ralph and Gough (2009), our findings further suggest that the regional categorization of TC events is necessary to fully understand environmental influences on ENP storm activity. Emphasis of ENSO on seasonal TC predictability at other basins (Nichols, 1985; Landsea, 2000; Camargo and Sobel, 2005; Camargo et al.

2007) highlights the importance of such connection in the seasonal forecast of WDR storms.

In Chapter 3 two more research questions were addressed:

177 3. How do locations of storm genesis, maximum lifetime intensity and

downgradation and track movement vary as a function of ENSO conditions in the

MDR?

While favourable environmental conditions during El Niño years often led to a more heighted hurricane season (Chapter 2), the results in Chapter 3 highlight an ENSO impact on TC track positions and movements. Although ENSO-induced differences at the genesis and dissipation locations were previously detected in Irwin and Davis (1999), we also demonstrate the effect of ENSO on storm track movement. In agreement with Irwin and Davis (1999), comparisons of latitudinal and longitudinal locations show strong

ENSO modulation among the different ENSO groups. These ENSO-induced differences are also seen in the variations in the latitudinal and longitudinal extensions of storm tracks.

4. How do locations of storm genesis, maximum lifetime intensity and

downgradation and track movement vary as a function of ENSO conditions

compare between MDR subdivisions?

Important distinctions in the regional variation of storm track locations and movements due to ENSO are also detected when ENP storms are binned into the appropriate MDR subdivisions. WDR storms tend to dissipate at positions more northward during El Niño years and their storm track distance is further lengthened and statistically different than those that occurred during La Niña and neutral years. Such a

178 poleward extension leads to greater storm longevity and is partly attributed to an overall stronger storm intensity during El Niño years. ENSO is shown to have a significant impact on those storms that traverse outside of WDR’s poleward boundary at 20°N. Of those WDR storm tracks that had extended outside of their poleward boundary, there are more recurving storms that landfalled during El Niño years than the other two ENSO phases.

Although WDR storms reach maximum intensities far from ashore, they are more capable of landfalling in areas that are generally considered less vulnerable to TC influences posing an enhanced risk. Because the southwest U.S., for example, is rarely devastated by recurving ENP storms, local residents are usually less prepared when confronting storm-related hazards. Since it has been one of the National Hurricane

Center’s priorities to minimize forecasting errors associated with the path of storm recurvature, disseminating relevant information to the general public has been challenging. Because it has been shown that recurving storms tend to travel at a greater speed, forecasts of storm track characteristics after TC recurvatures must be as accurate and as quick as possible. A fast and reliable forecast in the estimated storm arrival time and its expected strength is critical for the public and decision makers to arrange the necessary resources to prepare for potential storm impacts.

In Chapter 4 the final two research questions are addressed:

5. How do near-time SSTs link to the observations of TC intensities in the MDR

subdivisions?

179

Although SSTs are often linked to provision of the required energy for storm intensification, ENP storm intensity is only partly explained by SSTs. As observed in other basins, it has been shown that while SST values are critically important for the establishment of stronger ENP storms, SSTs are not the overriding factor at determining storm intensity. Instead, a combination of other environmental factors could have a greater role when a certain SST condition is met. A lack of support for this direct relationship between SST and TC intensity was partly attributed to a delayed TC response in which the maximum lifetime intensity is often achieved only after the highest

SST is encountered. In addition, due to the ocean surface cooling generated by TC winds themselves, the maximum lifetime intensity is often attained near the local SST minimum of storm passage.

In terms of the SST limitation to storm intensity, distinctly different responses between WDR and EDR storms are also apparent in these results. While WDR storm intensity corresponds positively to SST conditions, it is opposite for EDR storm intensity.

When the maximum lifetime storm intensities are considered, SSTs appear to act as a limiting factor on the regulation of occurrence of hurricanes and major hurricanes. One particularly noteworthy finding is the minimum SST threshold for ENP major hurricanes to attain maximum intensities is not uniformly observed when compared with other basins and noticeably different when the distribution of ENP storms is longitudinally divided. For major hurricanes in the EDR and WDR, their respective SST thresholds are

27°C and 25°C, both of which are substantially lower than the minimum SST, 28.25°C, in the North Atlantic basin.

180 Compared with EDR storm intensity, SSTs are considered to be more relevant in contributing to some of the highest intensities attained by WDR storms. Specifically, major hurricanes (with sustained winds exceeding 50 m s-1) in WDR have a greater response to SST fluctuations than storms of hurricane or tropical storm strengths. Such an association is important considering that the most powerful storms are often associated with longest storm tracks and responsible for the majority of TC-related damage than weaker TCs. As a greater proportion of WDR storms that landfall tend to occur in the southwest U.S., the ability of near-time SSTs to regulate WDR storm intensity could potentially imply further storm-inflicted consequences if SSTs warm.

6. How has the storm’s maximum potential intensity been limited by SSTs using the

most recent SST record?

Although the recent SST warming coincides when a declining trend in EDR- originated WDR storms, it was demonstrated that near-time SSTs influence ENP maximum storm intensity (ENPMPI). As a result of a shift towards a warmer basin-wide

SSTs, the maximum attainable storm intensities under some of the highest SST categories have increased. It was shown previously that a linear function best characterizes the relationship between ENPMPI and near-time SSTs (Whitney and Hobgood, 1997).

Through an update of the recent climatological period, contemporaneous SSTs are found to have a greater influence over ENPMPI.

When ENP storms are regionally separated, ENPMPI at WDR displays a stronger relationship than EDR storms’ MPIs. However, for some of the highest SST bins,

181 corresponding MPIs at both regions gradually stabilize. Rather than using a linear regression to illustrate the empirical response from ENPMPI, an alternative function of exponential decay (increasing form) is proposed to limit ENPMPI at WDR. This leads to the conclusion that the regional rise of ENPMPI in the WDR within the warmer SST bins does not occur linearly. Under a continuous future SST warming, a shift towards the strengthening of storm intensity may likely extend TC influences further poleward, thereby rendering more regions vulnerable to storm risks. As such, future research should focus on the anticipated societal impact in correspondence to changes in storm characteristics due to climate variability and change.

5.2 Future Research

The results from this thesis demonstrate the significance of climate variability, especially the ENSO variation, on storm characteristics. Further research should focus on changes in storm characteristics due to projected climate change and their impacts on society. Chapters 2 and 3 examined the impact of storm activity due to ENSO at an interannual scale. It would of great interest to investigate how TCs respond to ENSO as intense El Niño years are projected to occur at a greater frequency and perhaps greater magnitude. A better understanding of the Pacific Decadal Oscillation (PDO; Mantua et al., 1997) and its link to seasonal ENP storm activity and intensity would greatly benefit the forecasting community at explaining seasonal variations of the local hurricane season.

Although PDO signatures (Hare, 1996; Zhang, 1996) are more dominant poleward of the

MDR’s northernmost boundary at 20°N, the anticipated pattern of a poleward shift for future TC tracks may lead to important implications for ENP TCs. Additionally, the

182 remote control of the Atlantic Multi-decadal Oscillation in the North Atlantic Ocean could have a role at driving seasonal variation of storm activity in the adjacent ENP basin

(Collins, 2010). AMO is demonstrated to affect seasonal North Atlantic TC development through local controls of wind shear and SSTs (Goldenberg et al., 1999; Wang et al.,

2008). However, other than SST (from chapter 4) and mid-level relative humidity

(Collins and Roache, 2011), dynamic factors such as vertical wind shear and relative vorticity have been less rigorously reviewed for their connections with TC variability in the ENP over a long-term, climatological period. Though thermodynamic environment may play a greater role at portraying the regional difference of storm development during a specific season (Collins and Roache, 2011), further analyses are required to understand the long-term importance of dynamic factors to metrics of TC activity and intensity, relative to thermodynamic controls.

183 5.3 References

Collins, J. M., and I. M. Mason, 2000: Local environmental conditions related to seasonal tropical cyclone activity in the Northeast Pacific basin. Geophysical Research Letters, 27, 3881-3884.

Collins, J. M., and D. R. Roache, 2011: The 2009 Hurricane Season in the Eastern North Pacific Basin: An Analysis of Environmental Conditions. Monthly Weather Review, 139(6), 1673-1681.

Emanuel, K. A., 2005: Increasing Destructiveness of Tropical Cyclones over the Past 30 years. Nature , 436, 686-688.

Gray, W. M., C. W. Landsea, P. W. Mielke Jr, and K. J. Berry, 1994: Predicting Atlantic basin seasonal tropical cyclone activity by 1 june. Weather and Forecasting, 9(1), 103- 115.

Hare, S. R., 1996: Low-frequency climate variability and salmon production. Ph.D. thesis, University of Washington, Seattle, 306 pp. [Available from University Microfilms, 1490 Eisenhower Place, P.O. Box 975 Ann Arbor, MI 48106.]

Landsea, C. W., 2000: El Niño-Southern Oscillation and the seasonal predictability of tropical cyclones. In press in El Niño : Impacts of Multiscale Variability on Natural Ecosystems and Society, edited by H. F. Diaz and V. Markgraf.

Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bulletin of the American Meteorological Society, 78(6), 1069-1079.

Nicholls, N., 1985: Predictability of interannual variations of Australian seasonal tropical cyclone activity. Monthly Weather Review, 113(7), 1144-1149.

Ralph, T. U., and W. A. Gough, 2009: The influence of sea-surface temperatures on Eastern North Pacific tropical cyclone activity. Theoretical and Applied Climatology, 95, 257-264.

Rappaport, E. N., J. L. Franklin, L. A. Avila, S. R. Baig, J. L. Beven II, E. S. Blake, . . . A. N. Tribble, 2009: Advances and challenges at the national hurricane center. Weather and Forecasting, 24(2), 395-419.

Zhang, Y., 1996: An observational study of atmosphere-ocean interactions in the Northern Oceans on interannual and interdecadal time-scales. Ph.D. thesis, University of

184 Washington. [Available from University Microfilms, 1490 Eisenhower Place, P.O. Box 975 Ann Arbor, MI 4

185 Appendix I – Conditions for tropical cyclone development

The number and type of TCs, along with their duration, are highly regulated by seasonal changes in climatic conditions. Conditions of oceanographic and atmospheric factors that limit TC development have been identified (Gray, 1968; Gray, 1979) and divided into favourable thermodynamic and dynamic conditions as follows:

1. Warm ocean water at least 26.5°C at the sea surface overlying a deep mixed layer in providing surface air-ocean enthalpy exchange. 2. A steep drop in atmospheric temperature (or increase in lapse rate) so air can be lifted and cools, forming convective clouds 3. An abundance of moisture (or high relative humidity) at the mid-level troposphere to maintain the development of convective system 4. Pre-existing near surface disturbance with a high convergence of surface wind and divergence of wind aloft. It is generally a low pressure system that has sufficient vorticity. 5. A distance usually 5° latitude poleward of equator so Coriolis force can maintain tropical disturbance through gradient wind balance. 6. Low vertical wind shear between the lower and upper troposphere. A large magnitude of vertical wind shear can reduce TC formation or inhibit further TC development by disrupting the organization of convective clouds. References:

Gray, W. M., 1968: The Origin of Tropical Disturbances and Storms. Monthly Monthly Weather Review, 96, 669-700.

Gray, W. M., 1979: Observational inferences concerning the occurrence, structure and dynamics of tropical cyclones. Australian Meteorological Magazine, 27(4), 197-211.

186 Appendix II – Saffir-Simpson classification of tropical cyclone intensity

TC activity can be divided into measures of TC frequency and duration based on various categorization schemes to classify TC strength. TC is a general term that includes hurricanes (with wind speed greater than 33m/s) in the North Atlantic and eastern North Pacific (ENP) basins, typhoons in the Northwest Pacific basin and severe tropical cyclones in the southwest Pacific and Indian basins (Landsea et al., 1996). In North America, once a TC reaches a one-minute maximum surface wind of at least 33m/s, it is locally categorized according to the Saffir-Simpson scale (Table A-1). With an increase in its wind strength, a hurricane’s destructive power increases.

Table A-1: Classification of TC strength for storms that originated in the eastern North Pacific and Atlantic basins. TCs greater than 33m/s are categorized as hurricanes on the Saffir-Simpson Scale.

Saffir-Simpson Scale - Category *TD **TS 1 2 3 4 5 Maximum Sustained -1 Wind Speed (ms ) <18 18-33 33-42 43-49 50-58 59-69 9-6

* TD represents tropical depression ** TS represents tropical storm

Regionally dependent factor such as ENSO has shown importance in its significant correlations with TC frequency (Nicholls, 1984), as well as its genesis region and eventual track. Consequently, any projected regional differences of TC characteristics critically hinge on the accuracy of ENSO prediction. However, the short- term forecast of ENSO development is usually strongest during the winter, which may

187 hamper its effect on northern hemispheric storms where storm seasons peak from the late summer to early fall (Camargo et al., 2007). In addition, the challenge of ENSO forecasting would also impair any definitive conclusion that can be drawn from the likely

TC impact due to climate change (Henderson-Sellers et al., 1998). Alternative definition of extreme El Niño events using rainfall anomalies have called for the increased likelihood of extreme El Niño conditions in the two centuries. However, any concluding remark on the expected changes of TC characteristics associated with an increased frequency of extreme El Niño conditions would follow a historical review of such relationships.

References:

Nicholls, N., 1984: The southern oscillation, sea-surface-temperature, and interannual fluctuations in Australian tropical cyclone activity. Journal of Climatology, 4(6), 661- 670.

188