Trends in global and regional hurricane intensity

(1)

Faculty of Science

Trends in global and regional hurricane

Intensity

Bachelor Thesis Future Planet Studies, Major Earth Sciences

Submitted by: Dhr. mr. Maarten Hugen

Date: 05-07-2018

(2)

(3)

Abstract

Global warming is increasingly linked to several major issues, such as the increase in frequency and intensity of extreme weather events such as hurricanes. Hurricanes are regularly responsible for damage and loss of life on a massive scale. This research aims to determine trends in high intensity hurricanes on both a global and regional level through the use linear regression and Two Sample T-tests and to which degree high intensity hurricanes are correlated to the sea surface temperature in hurricane formation zones through the use of Pearson’s bivariate correlation. This done through the use of hurricane data provided by National Oceanic and Atmospheric Administration (NOAA), the Joint Tycoon Warning Center and the Bureau of Meteorology Australia and sea surface temperature data provided by NOAA. The results show significant positive trends for both global and regional high intensity hurricanes with the exception of high intensity hurricanes in the Eastern Pacific Ocean, where the positive trend found was not statistically significant. The T-tests showed, albeit not significant in most cases, an increase in the mean number of high intensity hurricanes in the last 40 years. Moderately strong significant correlations were found between high intensity hurricanes and sea surface temperatures for global hurricanes, as well as hurricanes in the North Atlantic and the Eastern Pacific Oceans.

(4)

1. Introduction

Global warming has become a major problem that has been implicated in a myriad of issues worldwide. Global mean temperature has been steadily rising (Knutson et al., 2010; IPCC, 2014) and could have severe negative consequences (IPCC, 2014). This rise in global mean temperature has been linked most prominently to sea level rise (Vermeer & Rahmstorf, 2009; Nicholls & Cazenave, 2010; IPCC 2014) and the increase of extreme weather events (Easterling et al., 2000; IPCC, 2014), of which hurricanes are one of the more regularly occurring examples.

Hurricanes among the most destructive extreme weather events on the planet, both in regard to property damage as well as in regard to the loss of human lives. An example is hurricane Harvey, which caused roughly 125 billion US dollars of damage in the southern United States in 2017 (NOAA, 2017a) and was only one of the four hurricanes category 4 or stronger hurricanes on the Saffir–Simpson Hurricane Wind Scale (NOAA, 2009) to form in the North Atlantic basin that year (NOAA, 2018a). Hurricane Harvey however has not been the costliest hurricane on record. The costliest hurricane to date has been hurricane Katrina, which caused 160 billion US dollars’ worth of damage in the southern US in 2005 when taking inflation into account (NOAA, 2017a). Another example of the destructive potential of hurricanes is cyclone Nargis,category 4 on the Saffir–Simpson Hurricane Wind Scale, which was responsible for over 138.000 deaths in Myanmar in 2008 (Fritz et al., 2009). Furthermore, both property damage and human casualties are likely to rise in the future due the trend of people migrating to areas that are regularly hit by hurricanes (McGranahan et al., 2007).

Even though increases in hurricane intensity and frequency have been linked to global climate change, there are still uncertainties regarding this connection (Trenberth, 2005; Knutson et al., 2010). In this context, studies have been done to research hurricane frequency and intensity and their link to climate change. One such is the study, is the study done by Webster et al. (2005). There is a great quantity of hurricane data available dating as far back as 1901, however much of this data is seen as fairly unreliable. This is due to the fact that data before 1975 was not measured using satellites and is therefore much more prone to be distorted by lack of observations, poor observations or might not be available for all the relevant regions of the world (Webster et al., 2005). Previous studies have therefore been limited in the amount of available adequate data. Webster et al. (2005) for example were able to use only 30 years of data. Since then a lot more data has been collected.

(7)

can be explained by ongoing global climate change. This will be done through the use of the following main research question.

To what degree can trends be discerned regarding high intensity hurricanes in connection with rising average sea surface temperature?

The main research question will be answered by using the following three sub research questions

1. To what degree can a global trend be discerned in high intensity hurricanes?

2. To what degree can trends be discerned in high intensity hurricanes per major hurricane basin? 3. To what degree are these trends correlated to Sea Surface Temperature?

(8)

2. Theoretical Framework

2.1 Hurricanes

Hurricanes are defined as large scale rotary storms that form over the warm waters of the Tropics (Montgomery & Farrell, 1993). The formation of hurricanes requires a pre-existing low-pressure disturbance (Gray, 1998; Smith ,2013). If the pressure of the disturbance is low enough and four other requirements are met the disturbance can grow into a hurricane (Gray, 1998; Smith, 2013).

The most important factor in the formation of hurricanes is the sea surface temperature of the water above which a hurricane might form. The formation of hurricanes requires rising air that is warmer than the surrounding air up to a height of 12 kilometres, as well as a high atmospheric humidity up to the height of approximately 6 kilometres (Smith, 2013). The heating of the air and increase of humidity is the result of evaporation of warm sea water and the subsequent release of the latent heat when the water vapour condensates into clouds at higher altitudes (Gray, 1998; Smith 2013; Christopherson & Birkeland, 2013). For this dynamic to take place the sea surface temperature needs to be at least 26.5 °C (Palmen, 1948; Gray, 1998; Smith, 2013; Christopherson & Birkeland, 2013).

The second factor is vorticity. Hurricanes can only form above oceans between 5° and 12° degrees north or south of the equator (Emanuel, 2005). This is due to the fact that only these latitudes allow for the required spiral structure of hurricanes. It is extremely rare for hurricanes to form closer to the equator as there the Coriolis force is too low to allow for the formation of the spiral structure of hurricanes (Smith, 2013). Further away from the equator Coriolis force is sufficient but strong low-pressure areas are usually quickly filled up with air, and thus preventing formation (Gray, 1998; Smith, 2013).

The third factor is the lack of strong vertical wind shear. Hurricanes can only form when there is no or very little vertical wind shear. As wind shear can impede the formation of the vortex by disrupting the storms structure (Palmen, 1948; Gray 1998; Smith, 2013; Frank & Ritchie, 2001). High levels of wind shear are even able to disrupt hurricane development completely (Frank & Ritchie, 2001).

(9)

The fourth and rarest factor is the presence of an area above the forming hurricane of high relative pressure compared to the underlying low-pressure disturbance (Smith, 2013).

Figure 1 - vertical structure of a hurricane (Smith 2013)

When all these conditions are met a hurricane is able to form. Due to the combination of both the high and low-pressure areas rising warm humid air is drawn towards the low-pressure system (Smith, 2013; Christopherson & Birkeland, 2013). The rising air flows then out at great heights during this process more moisture and latent heat is released, increasing the strength of the hurricane (Smith, 2013; Christopherson & Birkeland, 2013). This process is shown in figure 1.

The previously mentioned requirements make that hurricane formation can only take place in specific areas. In figure 2 the average locations of hurricane formation and the corresponding tracks are shown. No hurricanes are shown for the Southern Atlantic Ocean as the water is generally not warm enough to allow hurricane formation (Walsh et al., 2016).

(10)

Figure 2 - World map showing the location and average annual frequency of hurricanes (Smith 2013)

Hurricanes are classified based on average sustained windspeed using the Saffir-Simpson Hurricane Windscale (NOAA, 2009). Disturbances that have sustained

windspeeds lower than 17 m/s are called tropical depressions and disturbances between 17 m/s and 33 m/s are called tropical storms. Hurricanes are classified as having sustained windspeeds over 33 m/s or higher. This study will focus only on hurricanes of the fourth and fifth categories, with average windspeeds between 58 m/s and 70 m/s and more than 70 m/s respectively (see table 1). Several different regional names exist for these kind of storms, in order to prevent confusion this paper will refer to all these storms as hurricanes.

2.2 Sea surface temperature

As previously mentioned sea surface temperature plays an important role in the formation of hurricanes. Sea surface temperatures of 26.5°C and higher are required for hurricanes to develop (Gray & Brody, 1967; Goldenberg et al., 2001) and are the major source of energy for hurricanes. It largely believed that higher sea surface temperatures will result in stronger hurricanes (Walsh et al.,

Saffir-Simpson Hurricane Wind Scale

Category Windspeed (m/s) One 33 - 42 Two 43 - 49 Three 50 - 58 Four 58 - 70 Five 70+

(11)

3. Data and Methods

3.1 Data

3.1.1 Hurricanes

This study will solely focus on hurricanes that have been tracked using satellite data in order to minimize bias between methods of measurement and to ensure data of adequate quality. Therefore, no data will be considered from before 1975. This study will consider hurricane data on a global level as well as on a regional level, considering the individual basins that see hurricane activity. Due to the difference in hurricane seasons between the northern and southern hemispheres (Webster et al., 2005)., hurricanes seasons in the southern hemisphere will be counted towards the year when the respective hurricane season commenced.

Hurricane data was gathered from the National Oceanic and Atmospheric Administration (NOAA, 2018b), the Joint Typhoon Warning Center (JTWC, 2018) and the Bureau of Meteorology Australia (BOM, 2018). A more elaborate overview of the data sources can be found in table 2.

Data Type Basins Time period Data provider Data source Hurricane North Atlantic,

Eastern Pacific 1975-2017 Unisys Weather NOAA Hurricane Western Pacific,

Northern Indian 1975-2017 Unisys Weather JTWC Hurricane Southern Indian 1975-2018 Unisys Weather JTWC Hurricane South Western

Pacific 2000-2018 Unisys Weather JTWC

Hurricane South Western

Pacific 1975-1999 Bureau of Meteorology Australia Bureau of Meteorology Australia Sea Surface

Temperature All basins 1979-2018 Climate Engine

CFS Reanalysis from NOAA Table 2 - Data providers and sources

(12)

3.1.2 Sea Surface Temperature

As is mentioned in chapter 2, sea surface temperature plays an important role in hurricane formation and intensity. However, there are few areas that can support hurricane development. In this research sea surface temperature data was only gathered from these areas. Six zones were determined to be relevant for the formation of hurricanes. These zones have been primarily defined based on longitude and latitude and can be found in table 3.

Basin Longitude Latitude

North Atlantic Ocean 80° to 20°W, 5° to 20°N Western North Pacific Ocean 120° to 180°E, 5° to 20°N Eastern North Pacific Ocean 90° to 120°W 5° to 20°N Southwestern Pacific Ocean 140° to 180°E 5° to 20°S North Indian Ocean 55° to 90°E 5° to 20°N South Indian Ocean 50° to 115°E 5° to 20°S Table 3 – longitudes and latitudes of hurricane formation zones

These zones were then further refined by excluding surface temperature data originating from land. Because land surface temperatures play no role in the formation of hurricanes and might distort the overall data as land temperature data shows much more fluctuation than sea surface temperature data. Six definitive zones were thus defined and can be found in figure 3.

Sea surface temperature data were only gathered during the hurricanes seasons of the respective hemispheres. The hurricane season for the Northern Hemisphere the interval from the month May up to and including October was used. For the Southern Hemisphere the interval from the month November up to and including April was used.

(13)

Figure 3 - Oceanic Basins. A: Northern Atlantic Ocean, B: Western Pacific Ocean, C: Eastern Pacific Ocean, D: South Western Pacific Ocean, E: Northern Indian Ocean, F: Southern Indian Ocean.

Sea surface temperature data was gathered through the use of the CFS Reanalysis from climateengine.org (NOAA, 2018c). The data consists of daily mean temperature of sea surface only for the earlier specified basins with spatial resolutions of 28.8 km grids for 1979 until 2011 and 19.2 km grids from 2011 until present, for the interval of 1979 until 2017 or April 2018 based on the specific oceanic basin.

3.2 Methods

3.2.1 Global and regional hurricane trends

In order to determine a trend in overall global hurricane intensity, hurricane data from all six basins were used. The hurricane tracking data were used to determine date and level of highest level of intensity that was reached on the Saffir-Simpson Hurricane Wind Scale. The hurricanes were consequently grouped by year and category on the Saffir-Simpson Hurricane Windscale Scale. Next,

(14)

the numbers of hurricanes of the fourth and fifth categories per year were determined as well as the number of hurricanes of category 4 and 5 combined per year. The results of which can be found in Appendix A. This was also done to determine trends in regional hurricane intensity except in those cases only hurricane data from the particular basin was used. The corresponding results can be found in Appendix B.

To determine a trend in intensity of the strongest hurricanes linear regression was performed between the time in years and the number of hurricanes of the fourth and fifth category combined (Reimann et al., 2008). After which F-tests were run to check whether the found trends were significant. Additionally, to check whether a significant difference between the number of high intensity of hurricanes is present over time Two-Sample T-tests were used between 2 groups of 20 years’ worth of hurricanes for global hurricanes (Burt et al., 2009) and Paired Sample T-tests for the individual basins (Burt et al., 2009).

3.2.2 Correlation with Sea Surface Temperature

Lastly, in order to check to which degree high intensity hurricanes are correlated to the Sea Surface Temperature Pearson’s bivariate correlation was used (Reimann et al., 2008). In this context the sea surface temperature which was gathered in the form of daily measurements in grids was averaged per grid. Consecutively the temperature data was averaged based on time. For the northern hemisphere the average temperature was calculated for the interval of May up to and including October and for the Southern Hemisphere the interval from the month November up to and including April was calculated.

Consequently, Pearson’s bivariate correlation was also used to calculate the correlation between the Sea Surface Temperature for the respective basins and the number of hurricanes of category 4 and 5 combined per year. In regard to the correlation between Sea Surface Temperature and high intensity hurricanes the averaged Sea Surface Temperature was further averaged over the basin to calculate the global yearly average Sea Surface Temperature of hurricanes formation zones.

This was then used in conjunction with the number of hurricanes of the fourth and fifth category combined to calculate Pearson’s bivariate correlation. Lastly, in order to check for significance F-tests were to check whether the found correlations were significant (Reimann et al., 2008).

(15)

4. Results

4.1 Global hurricane trends

Figure 4 - Number of global hurricanes of the fourth and fifth categories combined between from 1975 until 2017 including trend line.

Linear Regression

In the linear regression analysis, a significant positive trend in the number of high intensity, the combined hurricanes of the fourth and fifth category on the Saffir-Simpson Hurricane Wind Scale, hurricanes was found (see figure 4 and table 4). The complete regression statistics can be found in Appendix D.

Two Sample T-test

The Two Sample T-test resulted in a score of -1.449. However, the found difference between the two means did not reach the required significance threshold (see table 5).

Nr. of Observations

T-score

probability

Global

20 -1.449

7.7%

Table 5 - global hurricanes T-test results

offset

Trend line

slope

R

F-score

Nr. of years

probability

Global

10.14

0.238

0.50

13.767

43

0.06%

(16)

4.2 Hurricane trends per basin

Figure 5 - Number of regional hurricanes of the fourth and fifth categories combined between from 1975 until 2017 including trend line per Basin. Oceanic Basins. A: Northern Atlantic Ocean, B: Western Pacific Ocean, C: Eastern Pacific Ocean, D: South Western Pacific Ocean, E: Northern Indian Ocean, F: Southern Indian Ocean.

Linear Regression

The use of linear regression in order to determine whether linear trends could be determined between the amount of high intensity hurricanes over the period of 43 years resulted in positive trends in all six basins. However, the weak positive trend in the Eastern Pacific was found to be not significant. Trends in the five other basins were significant. Visualizations of the number of category 4 and 5 hurricanes from 1975 to 2017 and the corresponding trends for each basin can be found in figure 5. An overview of the results can be found in table 6 and the complete statistics of the linear regression statistics per basin can be found in Appendix E.

(17)

Paired T-test

The Paired T-tests resulted in negative T-scores in all basins except for the Eastern Pacific. Only one basin, the North Atlantic yielded a significant result. A complete overview can be found in table 7.

Basin

Nr. of Observations

T-score

Probability

North

Atlantic

20 -2.773

0.6%

Western

Pacific

20 -0.821

21.1%

Eastern

Pacific

20

0.865

19.8%

South

Western

Pacific

20 -1.652

5.7%

Northern

Indian

20 -1.677

5.5%

Southern

Indian

20 -0.365

35.9%

Table 7 - Per basin T-test results

Basin

offset

Trend

line slope

Multiple

R

F-score

Nr. of years

probability

North

Atlantic

0.791 0.033 0.30 4.249 43 4.5%

Western

Pacific

4.933 0.079 0.34 5.312 43 2.6%

Eastern

Pacific

2.495 0.014 0.01 0.280 43 59.9%

South

Western

Pacific

0.305 0.038 0.35 5.724 43 2.1%

Northern

Indian

0.013 0.019 0.37 6.403 43 1.5%

Southern

Indian

1.598 0.054 0.33 5.193 43 2.7%

(18)

4.3 Correlation with Sea Surface Temperature

4.3.1 Global correlation

Figure 6 – Average annual sea surface temperature for all hurricane formation zones averaged during hurricane season

Multiple R

F-score

Probability

Global hurricane

formation zones

0.36

5.532

2.4%

Table 8 – Global Pearson’s bivariate correlation results

The correlation between global high intensity hurricanes and the average Sea Surface Temperature of all the hurricane formation zones was found to be 0.36, with a significance of 0.024 (see table 8).

4.3.2 Correlation per basin

Figure 7 - Average yearly sea surface temperature per basin during hurricane season. NAO: Northern Atlantic Ocean, WPAC: Western Pacific Ocean, EPAC: Eastern Pacific Ocean, SPAC: South Western Pacific Ocean, NIO: Northern Indian Ocean, SIO: Southern Indian Ocean

(19)

Multiple R

F-score

Probability

North Atlantic

0.51

13.305

0.1%

Western Pacific

0.25

2.479

12.4%

Eastern Pacific

0.45

9.583

0.4%

South Western

Pacific

0.20

1.545

22.1%

Northern Indian

0.17

1.093

30.2%

Southern Indian

0.12

0.554

46.1%

Table 9 – Per basin Pearson’s bivariate correlation results

Only 2 basins resulted in significant correlations between high intensity hurricanes in the specific basin and the corresponding sea surface temperature in the hurricanes formation zones. The North Atlantic Ocean resulted in a moderately strong positive correlation with a multiple R of 0.51 with a high significance of 0.001. The Eastern Pacific likewise showed a moderately strong positive correlation with a multiple R of 0.45 and high significance of 0.004. The other basins showed weak positive correlations but were not statistically significant (see table 9).

(20)

5. Discussion

5.1 Global hurricanes

This study has found a significant positive trend in the number of high intensity hurricanes, category four and five on the Saffir-Simpson Hurricane Winds Scale. The found slope is quite steep and would suggest an increase of 2.4 high intensity hurricanes for every 10-year period.

The Two-Sample T-test resulted in a T-score of -1.449 (see table 5). This score confirms that there is an increase in the mean number of high intensity hurricanes between the two twenty-year periods. This would further substantiate the positive trend found. However, the found difference in means did not reach 95% confidence level. This likely due to the limited amount of data. If the trend continues and this test was repeated in the future with 10 or more years of data, the findings are likely to be significant.

This increase is not unexpected. Sea surface temperature is known to be an important factor in the cyclogenesis of hurricanes. It is widely understood that a higher average sea surface temperature should result in more latent heat for hurricanes and thus allow for higher intensity hurricanes as well (Holland 1997; Emanuel 2005; Knutson et al., 2010). As is shown in figure 6, the sea surface temperature averaged over all hurricane formation zones during hurricane season has been steadily rising over the last forty years. The increase in hurricane intensity of the last forty years is in line with current theory.

The increase in hurricane intensity is also corroborated by several hurricanes prediction models (Knutson & Tuleya 2004; Bengtsonn et al., 2007).

5.2 Regional hurricanes

As was mentioned in the previous paragraph, a significant positive trend was found for the number of global high intensity hurricanes. This trend is not uniformly distributed over the 6 basins that have been used in this research. Positive trends have been found for all 6 basins, however there is a disparity both in strength and significance of these trends.

High intensity hurricanes have been found to be increasing in all basins. The strongest increase was found in the Western Pacific Ocean (see table 6). The Western Pacific is currently the basin with the

(21)

The high sea surface temperature and high number of high intensity hurricanes is line with current theories about hurricane formation and intensity (Emanuel, 1991; Landsea, 2000).

The positive significant trend found in this research is line with a large number of previous studies on high intensity hurricanes in the Western Pacific (Webster et al., 2005; Stowasser et al., 2007; Chan 2008). However, there are several conflicting studies that have found negative trends for this basin as well (Wu et al., 2006; Chan & Liu, 2004; Camargo & Sobel, 2005). According to Wu et al. (2006) this likely due to the different methods used by the JWTC and Tokyo. The technique used by RSMC-Tokyo to determine average sustained wind speed causes the weaker tropical storms to be registered as stronger than would be the case with the traditional methods used by the JWTC as well as registering high intensity hurricanes as being weaker compared to the JWTC. Because this research has used best tracking data from the JWTC to determine the found trend, as well as for three other basins this research is compelled to endorse the traditional method used by the JWTC (Wu et al., 2006).

The most statistically significant trend was found in the Northern Indian Ocean (see table 6 and image 5E). The mean number of hurricanes for this basin is quite low compared to the other basins. This likely due to the presence of the landmass of the Indian subcontinent which effectively cuts through the formation zone, hindering the formation of hurricanes in this basin (Elsner et al., 2008). The trend found in this research is similar to the trend found by Singh et al. (2000). Singh et al. (2000) found a strong positive trend in hurricane intensity for the Northern Indian Ocean between 1877 and 1997 and note that high intensity hurricanes have doubled in the basin over that time period. Noted should be that a large part of this trend is based on historical data and may therefore not be as accurate. Another study by Krishna (2009) also found a significant positive trend but contributes it to a decrease in vertical wind shear.

In the case of the South Western Pacific this research found a significant positive trend (see table 6 and image 5D). The found trend is in concurrence with the study by Walsh et al. (2012), which also found a positive trend in the number of high intensity hurricanes in the South Western Pacific. However, this study suggests that even though there is a positive trend in high intensity hurricanes in the basin this may not result in an actual increase of high intensity hurricanes. According to Walsh et al. (2012), models predict a decrease in the total number of hurricanes in the South Western Pacific. It is not sure if this projected decrease in total hurricane frequency would outweigh the positive trend in high intensity hurricanes.

(22)

Positive trends were also found for the Eastern Pacific Ocean, the Southern Indian Ocean and the Northern Atlantic Ocean (see table 6 and image 5A and 5F). The trend found for the Eastern Pacific was fairly weak and the increase was far outside the 95% significance level (see table 6) and can therefore not be seen as a statistically significant. The positive trends for the Northern Atlantic Ocean and the Southern Indian Ocean were within the 95% significance level and are statistically significant. The positive trend found for the Indian ocean was the second strongest trend out of all six basins. This trend is in line with the earlier studies done by Kuleshov et al. (2010) and Webster et al. (2005). The trend found for the Northern Atlantic Ocean was fairly strong (see table 6 and image 5A). It is possible that this trend has been distorted due to a low in hurricane activity in the 1970s and 1980s (Trenberth & Shea, 2006). This temporary low in hurricane activity has been attributed to the influence of the Atlantic Multi-decadal Oscillation or AMO on sea surface temperatures (Ting et al., 2009; Ting et al., 2011; Trenberth & Shea, 2006). In the 1970s and 1980s the AMO was responsible for lowering sea surface temperatures and thus also hurricane activity. At the end of the 20th_{century the upswing} of the AMO might have skewed results further by increasing sea surface temperature and possibly hurricane activity as well (Ting et al., 2009). However due to length of the AMO the progression of the oscillation is not completely certain. If the current upswing of the AMO persist this might lead to more positive trend in the future, a downturn could likewise weaken the found positive trend (Ting et al., 2009).

Regardless of the influence of the AMO the found trend does agree with other studies into hurricane intensity in the Northern Atlantic Ocean. Both Balagru et al. (2018) who studied the intensity of hurricane intensity for the basin between 1986 and 2015 and Elsner et al. (2008) who studied the maximum hurricanes wind speed found positive trends in hurricane intensity.

Paired T-tests for five out of six basins returned negative T-scores further confirming that the mean number of high intensity hurricanes between the two twenty-year periods has increased (see table 7). Of these T-tests only one basin, the North Atlantic, met the 95% significance level. The eastern Pacific was the only basin with a positive score, however similar to the trend found for that basin the T-score was far outside the 95% significance level. It is likely that if these tests are run again in the future with more data that most basins would report significant results.

(23)

5.3 Correlation

The results have shown that a moderately strong positive correlation with a multiple R of 0.36 between the average sea surface temperature in hurricanes formation zones and the number high intensity global hurricanes can be assumed (see table 8). This is in line with Hoyos et al. (2006) who found a direct link between rising sea surface temperatures and high intensity hurricanes between 1970 and 2006. The level of correlation between global high intensity hurricanes and sea surface temperature suggests that sea surface temperature is just one of several factors that have an effect on this increase. Factors that could also be influencing the increase of high intensity hurricanes are vertical wind shear (Hoyos et al., 2006), the tropospheric temperature profile (Emanuel 2005; Trenberth et al., 2018) the El Niño Southern Oscillation (ENSO) (Emanuel 2005; Hoyos et al., 2006; Trenberth et al., 2018) and the proximity to land (Elsner et al., 2008). These factors are however out of the scope of this study.

The correlation between the number of global hurricanes of category 4 and 5 on the Saffir-Simpson Hurricane Wind Scale and sea surface temperature falls within the 95% significance level, however it might still be problematic to blindly assume this relationship on all levels. As this is not the case for the relationship between hurricane formation zones and the corresponding high intensity hurricanes in each individual basin.

High intensity hurricanes and average sea surface temperature were found to be moderately strongly correlated and within the 95% significance level for the Northern Atlantic and the Eastern Pacific Oceans (see table 9).

Northern Atlantic Ocean

The correlation was strongest for the North Atlantic Basin with a correlation with a multiple R of 0.51 and a significance of 0.001. Due to this the North Atlantic will be quite susceptible to changes in average sea surface temperature, whether due to global warming or fluctuations of the AMO (Trenberth and Shea, 2006).

Eastern Pacific Ocean

The average sea surface temperature in the Eastern Pacific Ocean was found to be moderately strong with a multiple R value of 0.45. It is especially of note that the correlation was this strong when it is taken into account that for this basin no statistically significant trend in hurricane intensity could be determined.

(24)

The correlation for the Western Pacific, South Western Pacific, Northern Indian Ocean and Southern Indian ocean were found to be weakly correlated, but we also no inside the 95% significance interval and are thus not statistically significant.

(25)

6. Conclusion

In conclusion this study has found that a significant positive trend exists in the number of global high intensity hurricanes, category 4 and 5 on the Saffir-Simpson Hurricane Wind Scale. For separate basins an increase in the mean number of high intensity global hurricanes in the last 40 years was also found, although not always significant this does suggest that high intensity hurricanes have been increasing. An increase of 2.4 high intensity hurricanes is expected globally every 10 years. This increase is not evenly distributed among the basins. However, the number of high intensity hurricanes is expected to increase in every basin. The increase ranges from 1.9 to 7.9 high intensity hurricanes per 100 years. No projection was made for the Eastern Pacific Ocean as the trend found in that basin was not statistically significant. Paired T-tests confirmed an increase of high intensity hurricanes in the last 40 years in the North Atlantic Ocean. This is also suggested for the other basins, excluding the Eastern Pacific, but these findings were not significant. A positive correlation with a multiple R of 0.36 was found for global high intensity hurricane increase and the average sea surface temperature in the designated hurricane formation zones. In regard to correlations per basin only significant correlations were found for the Northern Atlantic Ocean and the Eastern Pacific Ocean. The multiple R correlation coefficients were 0.51 and 0.45 respectively.

It is likely that up to a certain degree there is a connection between the increase of sea surface temperature and the increase and positive trends in high intensity hurricanes. This is at least the case for the basins where correlation was significant. This suggest a connection between the increase of global climate change and the increase of high intensity hurricanes as well. Further research should be considered regarding this possible indirect connection. As well as other factors that could play a role in the increasing strength of hurricanes.

(26)

7. References

Balaguru, K., Foltz, G. R., & Leung, L. R. (2018). Increasing magnitude of hurricane rapid intensification in the central and eastern tropical Atlantic. Geophysical Research Letters, 45(9), 4238-4247.

Bengtsson, L., Hodges, K. I., Esch, M., Keenlyside, N., Kornblueh, L., LUO, J. J., & Yamagata, T. (2007). How may tropical cyclones change in a warmer climate? Tellus a, 59(4), 539-561.

BOM (Bureau of Meteorology Australia), (2018). Database of past tropical cyclone tracks. Retrieved on April 24, 2018 from http://www.bom.gov.au/clim_data/IDCKMSTM0S.csv

Burt, J. E., Barber, G. M., & Rigby, D. L. (2009). Elementary statistics for geographers. Guilford Press. Camargo, S. J., & Sobel, A. H. (2005). Western North Pacific tropical cyclone intensity and ENSO.

Journal of Climate, 18(15), 2996-3006.

Chan, J. C., & Liu, K. S. (2004). Global warming and western North Pacific typhoon activity from an observational perspective. Journal of Climate, 17(23), 4590-4602.

Chan, J. C. (2008). Decadal variations of intense typhoon occurrence in the western North Pacific. In Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences (464(2089), 249-272.

Christopherson, R.W. & Birkeland, G. (2013). Elemental Geosystems. Glenview: Pearson Education, Inc.

Easterling, D. R., Meehl, G. A., Parmesan, C., Changnon, S. A., Karl, T. R., & Mearns, L. O. (2000). Climate extremes: observations, modeling, and impacts. Science, 289(5487), 2068-2074. Elsner, J. B., Kossin, J. P., & Jagger, T. H. (2008). The increasing intensity of the strongest tropical

cyclones. Nature, 455(7209), 92.

Emanuel, K. A. (1991). The theory of hurricanes. Annual Review of Fluid Mechanics, 23(1), 179-196. Emanuel, K. (2005). Increasing destructiveness of tropical cyclones over the past 30

years. Nature, 436(7051), 686.

Frank, W. M., & Ritchie, E. A. (2001). Effects of vertical wind shear on the intensity and structure of numerically simulated hurricanes. Monthly Weather Review, 129(9), 2249-2269.

Fritz, H. M., Blount, C. D., Thwin, S., Thu, M. K., & Chan, N. (2009). Cyclone Nargis storm surge in Myanmar. Nature Geoscience, 2(7), 448.

Goldenberg, S. B., Landsea, C. W., Mestas-Nuñez, A. M., & Gray, W. M. (2001). The recent increase in Atlantic hurricane activity: Causes and implications. Science, 293(5529), 474-479.

(27)

Gray, W. M. (1998). The formation of tropical cyclones. Meteorology and Atmospheric Physics, 67(1-4), 37-69.

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

Hoyos, C. D., Agudelo, P. A., Webster, P. J., & Curry, J. A. (2006). Deconvolution of the factors contributing to the increase in global hurricane intensity. Science, 312(5770), 94-97. IPCC (2014): Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to

the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp." JTWC (Joint Typhoon Warning Center) (2018). Best track archive. retrieved on April 21, 2018 from

http://www.metoc.navy.mil/jtwc/jtwc.html?best-tracks

Knutson, T. R. & Tuleya, R. E. (2004). Impact of CO2-induced warming on simulated hurricane intensity and precipitation: Sensitivity to the choice of climate model and convective parameterization. Journal of Climate. 17, 3477–3495.

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

Krishna, K. M. (2009). Intensifying tropical cyclones over the North Indian Ocean during summer monsoon—global warming. Global and Planetary Change, 65(1-2), 12-16.

Kuleshov, Y., Fawcett, R., Qi, L., Trewin, B., Jones, D., McBride, J., & Ramsay, H. (2010). Trends in tropical cyclones in the South Indian Ocean and the South Pacific Ocean. Journal of Geophysical Research: Atmospheres, 115(D1).

Landsea, C. W. (2000). Climate variability of tropical cyclones: past, present and future. Storms. Routledge, New York, 220-241.

McGranahan, G., Balk, D., & Anderson, B. (2007). The rising tide: assessing the risks of climate change and human settlements in low elevation coastal zones. Environment and urbanization, 19(1), 17-37

Montgomery, M. T., & Farrell, B. F. (1993). Tropical cyclone formation. Journal of the Atmospheric Sciences, 50(2), 285-310.

Nicholls, R. J., & Cazenave, A. (2010). Sea-level rise and its impact on coastal zones. Science, 328(5985), 1517-1520.

NOAA (National Oceanic and Atmospheric Administration) (2009) The Saffir-Simpson Hurricane wind scale (2009). Retrieved on May 3, 2018 from https://www.nhc.noaa.gov/aboutsshws.php NOAA (National Oceanic and Atmospheric Administration) (2017a) Costliest U.S. tropical cyclones

tables updated (2017) retrieved May 7, 2018 from https://www.nhc.noaa.gov/news/UpdatedCostliest.pdf

(28)

NOAA (National Oceanic and Atmospheric Administration) (2018a) 2017 Atlantic Hurricane Season (2017) retrieved May 7, 2018 from https://www.nhc.noaa.gov/data/tcr/?text

NOAA (National Oceanic and Atmospheric Administration) (2018b) NHC Data Archive. retrieved April 20, 2018 from https://www.nhc.noaa.gov/data/

NOAA (National Oceanic and Atmospheric Administration) (2018c). Climate Forecast System Reanalysis. Retrieved April 27, 2018 from http://cfs.ncep.noaa.gov/cfsr/

Palmen, E. (1948). On the formation and structure of tropical hurricanes. Geophysica, 3(1), 26-38. Reimann, C., Filzmoser, P., Garrett, R. G., & Dutter, R. (2008). Statistical data analysis explained:

applied environmental statistics with R. Chichester: Wiley.

Smith, K. (2013). Environmental Hazards: Assessing Risk and Reducing Disaster. Routledge.

Singh, O. P., Khan, T. A., & Rahman, M. S. (2000). Changes in the frequency of tropical cyclones over the North Indian Ocean. Meteorology and Atmospheric Physics, 75(1-2), 11-20.

Stowasser, M., Wang, Y., & Hamilton, K. (2007). Tropical cyclone changes in the western North Pacific in a global warming scenario. Journal of Climate, 20(11), 2378-2396.

Ting, M., Kushnir, Y., Seager, R., & Li, C. (2009). Forced and internal twentieth-century SST trends in the North Atlantic. Journal of Climate, 22(6), 1469-1481.

Ting, M., Kushnir, Y., Seager, R., & Li, C. (2011). Robust features of Atlantic multi-decadal variability and its climate impacts. Geophysical Research Letters, 38(17).

Trenberth, K. (2005). Uncertainty in hurricanes and global warming. Science, 308(5729), 1753-1754. Trenberth, K. E., & Shea, D. J. (2006). Atlantic hurricanes and natural variability in 2005. Geophysical

Research Letters, 33(12).

Trenberth, K. E., Cheng, L., Jacobs, P., Zhang, Y., & Fasullo, J. (2018). Hurricane Harvey links to Ocean Heat Content and Climate Change Adaptation. Earth's Future.

Vermeer, M., & Rahmstorf, S. (2009). Global sea level linked to global temperature. Proceedings of the National Academy of Sciences, 106(51), 21527-21532.

Walsh, K. J., McInnes, K. L., & McBride, J. L. (2012). Climate change impacts on tropical cyclones and extreme sea levels in the South Pacific—A regional assessment. Global and Planetary Change, 80, 149-164.

Walsh, K. J., McBride, J. L., Klotzbach, P. J., Balachandran, S., Camargo, S. J., Holland, G., ... & Sugi, M. (2016). Tropical cyclones and climate change. Wiley Interdisciplinary Reviews: Climate Change, 7(1), 65-89.

(29)

Wu, M. C., Yeung, K. H., & Chang, W. L. (2006). Trends in western North Pacific tropical cyclone intensity. Eos, Transactions American Geophysical Union, 87(48), 537-538

(30)

Appendices

Appendix A – Global Hurricanes

YEAR Total Hurricanes

Saffir-Simpson Hurricane Wind Scale

Category

5 Category 4 Category 3 Category 2 Category 1 High intensity (4 and 5)

1975 41 1 7 6 11 16 8 1976 35 1 12 4 4 14 13 1977 33 1 3 6 8 15 4 1978 49 1 9 8 12 19 10 1979 40 3 7 13 10 7 10 1980 55 2 4 20 13 16 6 1981 47 1 5 12 8 21 6 1982 51 2 8 12 9 20 10 1983 53 3 9 9 9 23 12 1984 60 1 13 10 10 26 14 1985 58 1 7 17 10 23 8 1986 51 3 6 9 14 19 9 1987 45 4 6 9 8 18 10 1988 43 3 13 6 6 15 16 1989 61 7 11 8 10 25 18 1990 62 4 11 7 16 24 15 1991 53 4 11 14 11 13 15 1992 65 5 22 8 11 19 27 1993 59 6 17 9 8 19 23 1994 57 5 18 10 12 12 23 1995 57 4 13 12 9 19 17 1996 62 5 14 10 12 21 19 1997 61 12 12 5 11 21 24 1998 60 5 11 9 14 21 16 1999 47 3 12 8 7 17 15 2000 40 3 9 4 9 15 12 2001 53 4 14 9 9 17 18 2002 45 8 14 7 6 10 22 2003 48 5 10 6 8 19 15 2004 50 8 18 5 8 11 26 2005 51 8 13 5 7 18 21 2006 45 5 13 9 5 13 18

(31)

2009 39 6 13 4 6 10 19 2010 36 2 11 5 6 12 13 2011 36 2 14 3 5 12 16 2012 49 3 8 12 10 16 11 2013 42 7 9 5 6 15 16 2014 49 8 13 5 8 15 21 2015 53 8 23 6 6 10 31 2016 42 4 12 6 6 14 16 2017 45 4 8 7 13 13 12

(32)

Appendix B – Hurricanes per Basin

North Atlantic Ocean

YEAR Total Hurricanes

Category

1975 6 0 1 2 2 1 1 1976 6 0 0 2 2 2 0 1977 5 1 0 0 0 4 1 1978 5 0 2 0 1 2 2 1979 5 1 1 0 1 2 2 1980 9 1 0 1 3 4 1 1981 7 0 1 2 1 3 1 1982 2 0 1 0 0 1 1 1983 3 0 0 1 0 2 0 1984 5 0 1 0 1 3 1 1985 7 0 1 2 0 4 1 1986 4 0 0 0 1 3 0 1987 3 0 0 1 0 2 0 1988 5 1 2 0 0 2 3 1989 7 1 1 0 2 3 2 1990 8 0 0 1 2 5 0 1991 4 0 1 1 1 1 1 1992 4 1 0 0 2 1 1 1993 3 0 0 1 1 1 0 1994 3 0 0 0 1 2 0 1995 11 0 3 2 3 3 3 1996 9 0 2 4 0 3 2 1997 3 0 0 1 0 2 0 1998 10 1 1 1 4 3 2 1999 8 0 5 0 3 0 5 2000 8 0 2 1 1 4 2 2001 9 0 2 2 1 4 2 2002 4 0 1 1 0 2 1 2003 7 1 1 1 1 3 2 2004 8 1 3 2 1 1 4 2005 15 3 2 2 1 7 5 2006 4 0 0 2 0 2 0

(33)

2010 11 0 4 1 3 3 4 2011 6 0 2 1 1 2 2 2012 9 0 0 1 3 5 0 2013 2 0 0 0 0 2 0 2014 6 0 1 1 1 3 1 2015 4 0 1 1 0 2 1 2016 7 1 1 1 1 3 2 2017 10 2 2 2 2 2 4

Western Pacific Ocean

Category

1975 14 1 3 1 3 6 4 1976 14 1 7 1 0 5 8 1977 11 0 3 1 3 4 3 1978 15 1 1 1 4 8 2 1979 14 2 2 4 4 2 4 1980 15 1 3 5 3 3 4 1981 16 1 3 2 5 5 4 1982 19 2 4 6 3 4 6 1983 12 3 3 0 1 5 6 1984 16 1 6 2 1 6 7 1985 17 1 0 5 6 5 1 1986 19 2 2 4 7 4 4 1987 18 4 4 4 3 3 8 1988 13 1 5 1 1 5 6 1989 21 5 3 1 3 9 8 1990 21 4 3 1 8 5 7 1991 20 3 6 2 4 5 9 1992 22 3 7 2 3 7 10 1993 20 1 5 3 2 9 6 1994 20 1 10 1 3 5 11 1995 15 3 3 1 3 5 6 1996 21 3 5 2 5 6 8 1997 21 10 1 1 4 5 11 1998 9 2 2 1 1 3 4 1999 12 1 1 3 0 7 2 2000 14 3 3 1 4 3 6 2001 21 2 5 5 5 4 7

(34)

2003 17 2 7 1 2 5 9 2004 21 4 8 1 4 4 12 2005 16 2 7 0 4 3 9 2006 15 4 6 1 2 2 10 2007 15 1 6 1 4 3 7 2008 12 1 4 3 3 1 5 2009 15 4 3 0 3 5 7 2010 8 1 1 2 1 3 2 2011 10 2 4 1 1 2 6 2012 16 3 4 3 1 5 7 2013 16 4 4 2 3 3 8 2014 11 4 3 0 1 3 7 2015 21 5 9 3 1 3 14 2016 14 3 5 2 2 2 8 2017 12 1 2 1 5 3 3

Eastern Pacific Ocean

YEAR Total Hurricanes

Category 5 Category 4 Category 3 Category 2 Category 1 High intensity (4 and 5) 1975 9 0 2 2 1 4 2 1976 9 0 4 1 1 3 4 1977 4 0 0 0 1 3 0 1978 14 0 5 2 2 5 5 1979 6 0 2 2 1 1 2 1980 7 0 1 2 2 2 1 1981 8 0 0 1 1 6 0 1982 12 0 1 4 2 5 1 1983 12 0 5 3 2 2 5 1984 13 0 4 3 3 3 4 1985 13 0 3 5 1 4 3 1986 9 0 3 0 1 5 3 1987 10 0 2 2 3 3 2 1988 7 0 2 1 2 2 2 1989 9 0 2 2 0 5 2 1990 16 0 4 2 2 8 4 1991 10 0 2 3 1 4 2 1992 16 0 7 3 2 4 7

(35)

1996 5 0 1 1 1 2 1 1997 9 2 3 1 1 2 5 1998 9 0 3 3 0 3 3 1999 6 0 1 1 2 2 1 2000 6 0 1 1 2 2 1 2001 8 0 2 0 2 4 2 2002 8 2 2 1 2 1 4 2003 7 0 0 0 4 3 0 2004 6 0 2 1 0 3 2 2005 7 0 1 1 2 3 1 2006 11 1 2 3 3 2 3 2007 4 0 1 0 0 3 1 2008 7 0 1 1 2 3 1 2009 8 1 2 2 1 2 3 2010 3 1 0 1 0 1 1 2011 10 0 5 1 0 4 5 2012 10 0 1 4 2 3 1 2013 8 0 0 1 1 6 0 2014 16 2 5 2 1 6 7 2015 16 1 9 1 2 3 10 2016 12 0 3 2 2 5 3 2017 9 0 2 2 1 4 2

South Western Pacific Ocean

Category

1974 3 0 1 1 0 1 1 1975 8 0 1 1 4 2 1 1976 5 0 0 0 1 4 0 1977 4 0 0 2 1 1 0 1978 5 0 0 2 2 1 0 1979 9 0 1 5 2 1 1 1980 10 0 0 5 2 3 0 1981 5 0 0 3 1 1 0 1982 6 0 0 1 1 4 0 1983 12 0 1 0 4 7 1 1984 10 0 0 3 2 5 0 1985 6 0 0 1 3 2 0

(36)

1987 3 0 0 1 1 1 0 1988 7 0 2 2 1 2 2 1989 5 0 1 0 1 3 1 1990 4 0 0 1 2 1 0 1991 9 0 0 7 1 1 0 1992 5 0 0 1 1 3 0 1993 9 5 1 0 1 2 6 1994 6 0 1 3 1 1 1 1995 9 1 1 3 1 3 2 1996 7 0 1 1 3 2 1 1997 4 0 0 1 2 1 0 1998 9 0 3 1 3 2 3 1999 5 0 0 1 0 4 0 2000 2 0 0 0 1 1 0 2001 3 1 0 0 1 1 1 2002 6 1 2 2 1 0 3 2003 3 1 0 1 0 1 1 2004 6 2 3 0 1 0 5 2005 5 1 0 1 0 3 1 2006 4 0 1 0 0 3 1 2007 4 0 0 3 0 1 0 2008 1 0 1 0 0 0 1 2009 6 1 2 1 1 1 3 2010 5 0 3 0 1 1 3 2011 1 0 1 0 0 0 1 2012 4 0 1 1 1 1 1 2013 3 0 2 0 0 1 2 2014 5 1 0 1 2 1 1 2015 5 1 1 0 3 0 2 2016 4 0 1 1 1 1 1 2017 4 0 1 0 3 0 1

Northern Indian Ocean

Category

1975 3 0 0 0 1 2 0

1976 0 0 0 0 0 0 0

(37)

1980 0 0 0 0 0 0 0 1981 2 0 0 0 0 2 0 1982 2 0 1 0 1 0 1 1983 0 0 0 0 0 0 0 1984 2 0 0 0 1 1 0 1985 0 0 0 0 0 0 0 1986 0 0 0 0 0 0 0 1987 0 0 0 0 0 0 0 1988 1 0 0 1 0 0 0 1989 0 0 0 0 0 0 0 1990 1 0 1 0 0 0 1 1991 1 1 0 0 0 0 1 1992 2 0 0 0 0 2 0 1993 2 0 0 0 0 2 0 1994 1 0 1 0 0 0 1 1995 2 0 0 1 0 1 0 1996 4 0 1 0 0 3 1 1997 2 0 1 0 0 1 1 1998 5 0 0 1 1 3 0 1999 3 1 1 1 0 0 2 2000 2 0 0 0 0 2 0 2001 1 0 1 0 0 0 1 2002 0 0 0 0 0 0 0 2003 2 0 0 0 1 1 0 2004 1 0 0 0 0 1 0 2005 1 0 0 0 0 1 0 2006 1 0 1 0 0 0 1 2007 3 1 1 0 0 1 2 2008 1 0 1 0 0 0 1 2009 1 0 0 0 0 1 0 2010 4 0 2 0 0 2 2 2011 1 0 0 0 0 1 0 2012 0 0 0 0 0 0 0 2013 3 1 0 0 0 2 1 2014 2 0 2 0 0 0 2 2015 2 0 1 1 0 0 1 2016 1 0 0 0 0 1 0 2017 1 0 0 0 0 1 0

(38)

Southern Indian Ocean

Category

1975 1 0 0 0 0 1 0 1976 1 0 1 0 0 0 1 1977 7 0 0 1 3 3 0 1978 8 0 1 3 2 2 1 1979 5 0 1 2 1 1 1 1980 14 0 0 7 3 4 0 1981 9 0 1 4 0 4 1 1982 10 0 1 1 2 6 1 1983 14 0 0 5 2 7 0 1984 14 0 2 2 2 8 2 1985 15 0 3 4 0 8 3 1986 15 0 1 4 4 6 1 1987 11 0 0 1 1 9 0 1988 10 1 2 1 2 4 3 1989 19 1 4 5 4 5 5 1990 12 0 3 2 2 5 3 1991 9 0 2 1 4 2 2 1992 16 1 8 2 3 2 9 1993 14 0 4 3 3 4 4 1994 17 1 4 6 4 2 5 1995 13 0 3 5 1 4 3 1996 16 2 4 2 3 5 6 1997 22 0 7 1 4 10 7 1998 18 2 2 2 5 7 4 1999 13 1 4 2 2 4 5 2000 8 0 3 1 1 3 3 2001 11 1 4 2 0 4 5 2002 10 2 1 2 1 4 3 2003 12 1 2 3 0 6 3 2004 8 1 2 1 2 2 3 2005 7 2 3 1 0 1 5 2006 10 0 3 3 0 4 3 2007 11 0 2 2 2 5 2 2008 6 0 1 3 0 2 1

(39)

2012 10 0 2 3 3 2 2 2013 10 2 3 2 2 1 5 2014 9 1 2 1 3 2 3 2015 5 1 2 0 0 2 3 2016 4 0 2 0 0 2 2 2017 9 1 1 2 2 3 2

(40)

Appendix C – Sea Surface Temperature

YEAR EPAC WPAC SPAC NAO NIO SIO Global average

1979 26.80 27.47 27.69 26.58 27.82 26.55 27.15 1980 26.83 28.24 27.87 26.73 27.85 26.51 27.34 1981 26.54 28.33 27.76 26.57 27.74 26.22 27.19 1982 27.05 27.78 27.49 26.06 27.84 26.27 27.08 1983 27.08 28.01 27.55 26.29 28.09 26.62 27.27 1984 26.51 28.13 27.75 25.93 27.39 26.30 27.00 1985 26.45 27.84 27.69 26.04 27.35 26.19 26.93 1986 26.76 27.98 27.58 25.92 27.56 26.24 27.01 1987 27.18 28.05 27.56 26.66 28.12 26.45 27.34 1988 26.28 28.27 27.84 26.40 27.97 26.67 27.24 1989 26.53 28.08 27.63 26.26 27.60 26.31 27.07 1990 26.99 28.11 27.69 26.48 27.75 26.52 27.26 1991 26.95 28.13 27.51 26.07 27.97 26.43 27.17 1992 26.92 27.97 27.36 26.08 27.68 26.26 27.05 1993 26.90 28.03 27.43 26.20 27.78 26.32 27.11 1994 26.99 28.30 27.61 25.98 27.64 26.55 27.18 1995 26.64 28.35 28.08 26.75 27.99 26.39 27.37 1996 26.75 28.48 28.04 26.35 27.66 26.65 27.32 1997 27.69 28.17 27.72 26.57 28.10 26.81 27.51 1998 27.03 28.34 28.08 26.84 28.25 26.61 27.53 1999 26.40 28.19 27.90 26.56 27.77 26.62 27.24 2000 26.74 28.26 28.05 26.31 27.89 26.44 27.28 2001 27.07 28.57 28.29 26.66 28.04 26.70 27.56 2002 27.32 28.51 28.05 26.47 28.28 26.73 27.56 2003 27.07 28.54 28.29 26.82 28.32 26.87 27.65 2004 27.19 28.46 28.18 26.84 27.97 26.85 27.58 2005 26.99 28.59 28.27 27.24 28.28 26.83 27.70 2006 27.32 28.46 27.82 26.91 28.21 26.82 27.59 2007 26.96 28.63 28.13 26.69 28.28 26.78 27.58 2008 27.04 28.41 28.24 26.86 28.12 26.63 27.55 2009 27.60 28.69 27.97 26.78 28.32 26.97 27.72 2010 26.67 28.45 28.10 27.22 28.37 26.85 27.61 2011 26.67 28.31 27.95 26.78 28.17 26.84 27.46 2012 27.08 28.36 27.93 26.68 28.10 26.89 27.51 2013 26.94 28.56 27.99 26.70 27.89 26.90 27.50 2014 27.46 28.73 28.11 26.47 28.24 27.01 27.67 2015 28.04 28.53 28.16 26.66 28.56 27.28 27.87

(41)

Appendix D – Global hurricane statistics

Regression Statistics Multiple R 0.501383969 R Square 0.251385885 Adjusted R Square 0.233127004 Standard Error 5.219492264 Observations 43 _df _SS _MS _F _{Significance F} Regression 1 375.0794 375.079 13.7678693 0.00061487 Residual 41 1116.967 27.2431 Total 42 1492.047

Coefficients Standard Error t Stat P-value

Intercept 459.6650559 128.0273 3.5904 0.00087447

X Variable 1 0.237994564 0.064141 3.71051 0.00061487

t-Test: Two-Sample Assuming Equal Variances Variable 1 Variable 2

Mean 14.6 17.2

Variance 38.7789474 25.6421053

Observations 20 20

Pooled Variance 32.2105263

Hypothesized Mean Difference 0

df 38 t Stat -1.4486866 P(T<=t) one-tail 0.07781365 t Critical one-tail 1.68595446 P(T<=t) two-tail 0.15562731 t Critical two-tail 2.02439416

(42)

Appendix E – Regional hurricane statistics

North Atlantic Ocean

Regression Statistics Multiple R 0.3064241 R Square 0.09389573 Adjusted R Square 0.07179563 Standard Error 1.33545012 Observations 43 df SS MS F Significance F Regression 1 7.57716702 7.57716702 4.24865553 0.04566132 Residual 41 73.1205074 1.78342701 Total ₄₂ _80.6976744

Coefficients Standard _Error t Stat P-value

Intercept -65.983087 32.756841 -2.01433 0.05056681

X Variable 1 0.03382664 0.01641093 2.0612267 0.04566132

t-Test: Paired Two Sample for Means

Variable 1 Variable 2

Mean 1.05 2.15

Variance 0.99736842 2.45

Pearson Correlation 0.09595783

Hypothesized Mean Difference 0

(43)

Western Pacific Ocean

Regression Statistics Multiple R 0.33870178 R Square 0.11471889 Adjusted R Square 0.09312667 Standard Error 2.79362362 Observations 43 df SS MS F Significance F Regression 1 41.4642102 41.4642102 5.31297301 0.02630404 Residual 41 319.97765 7.80433293 Total 42 361.44186 _Coefficients Standard

Error t Stat P-value

Intercept -151.26941 68.5239258 -2.2075414 0.03292988

X Variable 1 0.07913017 0.03432996 2.30498872 0.02630404

Mean 6.4 7.2

Variance 7.51578947 9.95789474

Pearson Correlation -0.0888236

Hypothesized Mean Difference 0

(44)

Eastern Pacific Ocean

Regression Statistics Multiple R 0.0824931 R Square 0.00680511 Adjusted R Square -0.0174192 Standard Error 2.22579064 Observations 43 df SS MS F Significance F Regression 1 1.39172455 1.39172455 0.2809213 0.59895684 Residual 41 203.119903 4.95414398 Total 42 204.511628 _Coefficients Standard

Error t Stat P-value

Intercept -26.12232 54.5957271 -0.4784682 0.63485868

X Variable 1 0.01449713 0.02735204 0.53002009 0.59895684

Mean 3.2 2.55

Variance 3.95789474 6.05

df 19 t Stat 0.86504937 P(T<=t) one-tail 0.19890132 t Critical one-tail 1.72913281 P(T<=t) two-tail 0.39780263 t Critical two-tail 2.09302405

(45)

South Western Pacific Ocean

Regression Statistics Multiple R 0.35002995 R Square 0.12252097 Adjusted R Square 0.10111904 Standard Error 1.27886882 Observations 43 df SS MS F Significance F Regression 1 9.3628813 9.3628813 5.72476314 0.0213922 Residual 41 67.0557233 1.63550545 Total 42 76.4186047

Intercept -73.899577 31.3532605 -2.3569982 0.02328024

X Variable 1 0.03760193 0.01571562 2.39264773 0.0213922

Mean 0.8 1.5

Variance 1.95789474 1.73684211

(46)

Northern Indian Ocean

Intercept -38.441256 15.3667807 -2.5015816 0.01644908

X Variable 1 0.01948052 0.00769864 2.5303848 0.01532871

Mean 0.3 0.65

Variance 0.22105263 0.66052632

(47)

Southern Indian Ocean

Intercept -105.41891 47.4847294 -2.2200591 0.03200297

X Variable 1 0.05421323 0.02378948 2.27887359 0.0279513

Mean 2.85 3.1

Variance 6.23947368 1.77894737

(48)

Appendix F – Correlation statistics

NAO EPAC WPAC SPAC NIO SIO Global formation _{zone average}

1979 26.60 26.38 27.47 26.82 27.28 25.58 26.69 1980 25.91 26.52 27.49 26.91 27.29 25.60 26.62 1981 25.86 26.29 27.67 26.93 27.16 25.49 26.57 1982 25.33 26.73 27.18 26.56 27.20 25.44 26.40 1983 25.66 26.63 27.24 26.79 27.22 25.74 26.55 1984 25.20 26.24 27.52 26.90 26.85 25.36 26.34 1985 25.17 26.12 27.30 26.86 26.86 25.32 26.27 1986 25.14 26.42 27.36 26.78 26.95 25.28 26.32 1987 25.83 26.77 27.38 26.63 27.38 25.69 26.61 1988 25.63 26.23 27.71 27.24 27.45 25.78 26.67 1989 25.38 26.12 27.51 26.95 27.05 25.35 26.39 1990 25.67 26.61 27.47 26.95 27.26 25.52 26.58 1991 25.33 26.66 27.37 26.79 27.33 25.64 26.52 1992 25.37 26.64 27.20 26.60 27.04 25.38 26.37 1993 25.39 26.69 27.29 26.43 27.18 25.35 26.39 1994 25.24 26.73 27.60 26.63 27.12 25.62 26.49 1995 25.85 26.51 27.59 27.13 27.29 25.56 26.65 1996 25.63 26.31 27.72 27.22 27.11 25.59 26.60 1997 25.81 26.98 27.54 26.73 27.38 25.63 26.68 1998 26.11 26.65 27.68 27.38 27.69 25.90 26.90 1999 25.66 26.11 27.79 27.17 27.30 25.60 26.61 2000 25.49 26.32 27.80 27.16 27.36 25.68 26.64 2001 25.86 26.59 27.99 27.32 27.47 25.82 26.84 2002 25.91 26.86 27.87 27.37 27.72 26.08 26.97 2003 26.05 26.80 27.89 27.30 27.77 25.99 26.97 2004 26.08 26.84 27.95 27.27 27.53 25.85 26.92 2005 26.37 26.64 27.96 27.45 27.61 25.94 26.99 2006 26.12 26.82 27.93 27.23 27.66 25.87 26.94 2007 26.02 26.65 28.04 27.49 27.66 25.99 26.98 2008 25.96 26.55 27.95 27.39 27.45 25.88 26.87 2009 25.90 27.05 28.04 27.26 27.76 26.02 27.00 2010 26.47 26.40 27.81 27.60 27.81 26.15 27.04 2011 26.00 26.21 27.81 27.24 27.50 25.88 26.78 2012 25.81 26.63 27.79 27.15 27.44 25.99 26.80 2013 25.94 26.60 27.94 27.21 27.42 25.95 26.84 2014 25.70 27.05 28.03 27.17 27.50 26.14 26.93 2015 25.89 27.57 27.83 27.16 27.85 26.31 27.10

Trends in global and regional hurricane intensity

Faculty of Science