Monitoring coral bleaching by satellite thermal products a case study in the a case study in the southern east sea

MONITORING CORAL BLEACHING BY SATELLITE THERMAL PRODUCTS: A CASE STUDY IN THE SOUTHERN EAST SEA, VIETNAM A Thesis presented to the Faculty of the Graduate School at the Dr.. The unde

MONITORING CORAL BLEACHING

BY SATELLITE THERMAL PRODUCTS:

A CASE STUDY IN THE SOUTHERN EAST SEA, VIETNAM

A Thesis presented to the Faculty of the Graduate School at the

Dr Cuizhen Wang, Thesis Supervisor

MAY 2012

The undersigned, appointed by Dean of the Graduate School, have examined the thesis entitled

MONITORING CORAL BLEACHING

BY SATELLITE THERMAL PRODUCTS:

A CASE STUDY IN THE SOUTHERN EAST SEA, VIETNAM

Presented by Ngan Thuy Le

A Candidate for the degree of Master of Arts in Geography

And hereby certify that in their opinion it is worthy of acceptance

Professor Cuizhen Wang

Professor Michael A Urban

Professor Robert Sites

Dr Wang working hard on the IDL programming to help me approach the best results Without her contributions, my enthusiasm in studying coral reefs could not be possible

I also would like to thank Dr Mike Urban and Dr Robert Sites who were willing

to attend my Thesis Committee and gave me an adequate time to finish my thesis as satisfactorily as I expected There is no word can express my appreciation to their patience

It is an honor for me to meet and learn from Vietnamese specialists working at the Oceanography Institute in summer 2011 I am heartily thankful to Mr Phan Kim Hoang and Mr.Tong Phuoc Hoang Son, who shared with me their plenty of precious knowledge about the coral reefs in Vietnam Their sharing inspired me and helped me clarify my research questions Particularly, Mr Son did not hesitate to spend his valuable time teaching me working with the remote sensing data I deeply admire his devotion and her passion to share his experiences with me

I should also send my thanks to Mr Le Chi Lam, my mentor in Vietnam and Mr Nguyen Thanh Minh at the Department of Natural Resources and Environment in Khanh Hoa Province for their technical supports

Besides, I offer my regards and blessings to all faculty members and staff in the Department of Geography as well as my friends who supported me in any respect during the completion of my Master program I really appreciate Dr Matt Foulkes who questioned me: “Why did you do that?” to force me think deeply about the purpose of my

study In addition, special thanks to Dr Joe Hobbs, Chair of the Department and Director

of the Vietnam Institute, who gave me opportunity to study at MU and always, cares for

me with his kindness heart

Moreover, I would not overcome many challenges and depressions without the encouragements from my greatest friends who are working or studying in Vietnam, Europe and at MU I wish I could express completely my love to my “intimate sisters”, Nga Bui and Ha Phan, who are always beside me, understand my obstacles and support me as much as they can

Last but not least, I would like to send my honest gratitude to my family for their countless love and confidence on me This thesis is a special gift for my beloved parents

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ii

LIST OF FIGURES vii

LIST OF TABLES .x

LIST OF ABBRIVIATIONS AND EQUATIONS xi

ABSTRACT xii

CHAPTERS 1 INTRODUCTION 1

1.1 Coral Reefs 2

1.2 Threats to Reefs .5

1.3 Coral Bleaching .6

1.4 Predicting coral bleaching by sea surface temperature .9

1.5 Research Objectives .12

2 CASE STUDY 13

2.1 Status of Coral Reefs in Vietnam .13

2.2 Status of Coral Reefs at Con Dao and Phu Quoc Archipelagos .15

2.3.Coral bleaching at Con Dao Archipelago .19

2.4 Coral bleaching at Phu Quoc Archipelago 21

2.5 Global Thermal Anomalies in 2010 .22

3 MATERIALS AND METHODS .25

3.1 Data Description .25

3.1.2.MODIS Aqua SST Products .27

3.1.3 NOAA Coral Reef Watch Datasets 29

3.1.4.Observation Sites and Mainland Boundary .30

3.2 Data Pre-processing .31

3.2.1 Extracting SST Value From AVHRR and MODIS Products .32

3.2.2 Excluding of Invalid Pixels .33

3.3 Data Processing .37

3.3.1 Calculating SST Climatology and Maximum Monthly Mean (MMM) .37

3.3.2 Calculating Bleaching HotSpot 39

3.3.3 Calculating Degree Heating Week (DHW) .39

3.3.4 Generating of Bleaching Alert .40

3.3.5 Statistic analysis of MODIS- and AVHRR-derived SST Value .41

4 RESULTS AND DISCUSSIONS .43

4.1 Comparison of MODIS- and AVHRR-derived SST Value .43

4.1.1 Results of Correlation Analyses .43

4.1.2.Results of Paired Comparisons .46

4.2 Comparison of SST in the Southwestern and Southeastern Areas of the East Sea49 4.3 Thermal Stress in Bleaching Years 54

4.3.1 Thermal stress in 1998 .56

4.3.2 Thermal Stress in 2005 .61

4.3.3 Thermal Stress in 2010 .65

4.4 Comparison of the 50 km and 4 km Bleaching Prediction Products .68

4.4.1 Bleaching Threshold and Weekly SST Products .69

4.4.2 Mapping HotSpots and Bleaching Prediction .70

5 CONCLUSIONS 74

5.1 Major Findings .74

5.2 Limitations .75

5.3 Future Envisions .77

APPENDIX A Outputs from Linear Regression of the 2005 Weekly SST Data 79

B Outputs from Paired Comparison of the 2005 Weekly SST Data 86

BIBLIOGRAPHY 93

LIST OF FIGURES

FIGURES PAGE

Figure 1.1: Anatomy of a coral polyp 3

Figure 1.2: Major coral reef sites are red dots on the world map .4

Figure 1.3: Coral bleaching process .8

Figure 1.4: Trends in coral bleaching, 1980 – 2010 .9

Figure 2.1: Reef distribution in the coastal waters of Vietnam .14

Figure 2.2: Location of observed sites at Con Dao in 1994-2004 .17

Figure 2.3: Location of observed sites at Phu Quoc in 1994-2007 .17

Figure 2.4: The percentage of hard coral cover observed at Con Dao in 1994-2004 .18

Figure 2.5: The percentage of hard coral cover observed at Phu Quoc in 1994-2007 18

Figure 2.6: Coral bleaching was observed in the southwestern of Con Dao archipelago in October, 2005 .20

Figure 2.7: Bleaching alert graph (based on the 50 km spatial resolution data set) at Con Dao in 2010 .21

Figure 2.8: Coral bleaching was observed at Phu Quoc archipelago in May (a) and August (b), 2010 .22

Figure 2.9: Global temperature anomalies since 1880 .23

Figure 2.10: Coral bleaching alert areas in Western Pacific in summer 2010 .24

Figure 3.1: Locations of the three Virtual Stations in the Southern areas of East Sea 30

Figure 3.2: Location of the case study in East Sea .32

Figure 3.3: Example of Mean Sea Surface Temperature derived from AVHRR products34 Figure 3.4: The weekly SST in 2010 before S-Golay Smoothing .35

Figure 3.5: The weekly SST in 2010 after S-Golay Smoothing .36

Figure 3.6: The one-year SST climatology curve of one pixel .38

Figure 3.7: Distribution of the 2005 weekly SST samples .42

Figure 4.1: Scatter plot of the 2005 SST (degree Celsius) data overlaid with the regression line, and 95% confidence and prediction limits 44

Figure 4.2: Concordance of SST values derived from NASA MODIS and NOAA AVHRR products for Heron Island in 2005-2006 44

Figure 4.3: Scatter plot of the SST data in winter months (January to April and November to December of 2005) overlaid with the regression line, and 95% confidence and prediction limits .45

Figure 4.4: Scatter plot of the data in summer months (May to October of 2005) overlaid with the regression line, and 95% confidence and prediction limits .46

Figure 4.5: Comparison of weekly SST values derived from AVHRR PF and MODIS products in January and May 2005 48

Figure 4.6: Sea surface temperature climatology from January to June 50

Figure 4.7: Sea surface temperature climatology from July to December .51

Figure 4.8: 12 months sea surface climatology at Phu Quoc .53

Figure 4.9: 12 months sea surface climatology at Con Dao .53

Figure 4.10: Location of Observed Reefs and relative SST pixels at Phu Quoc 55

Figure 4.11: Location of Observed Reefs and relative SST pixels at Con Dao 55

Figure 4.12: Thermal stress at Phu Quoc in 1998 .57

Figure 4.13: Thermal stress at Con Dao in 1998 .58

Figure 4.14: Degree Heating Week from April to July 1998 .60

Figure 4.15: Degree Heating Week from April to October 1998 .60

Figure 4.16: Degree Heating Week in October 1998 .61

Figure 4.17: Thermal stress at Phu Quoc in 2005 .62

Figure 4.18: Thermal stress at Con Dao in 2005 .63

Figure 4.19: Degree Heating Week from May to October 2005 .65

Figure 4.20: Thermal stress at Phu Quoc in 2010 .66

Figure 4.21: Thermal stress at Con Dao in 2010 .67

Figure 4.22: Degree Heating Week from April to October 2010 68

Figure 4.23: Comparison of SST and bleaching threshold between the 50km (a) and the 4km (b) product .69

Figure 4.24: The 50km spatial resolution HotSpot map at East Sea in June and July 2010 71 Figure 4.25: The 4km spatial resolution HotSpot map in June 2010 71

Figure 4.26: The 4km spatial resolution HotSpot map in July 2010 72

Figure 4.27: The 50km spatial resolution Bleaching Alert map in July 2010 .72

Figure 4.28: The 4km spatial resolution Bleaching Alert map in July 2010 .73

LIST OF TABLES

Table 2.1: Cover (%) of coral and other benthos at 4 permanent monitoring sites in

2007-2008 16

Table 3.1: Location of the permanent observed reefs at Con Dao and Phu Quoc in 1994 – 2007 31

Table 3.2: The minimum monthly value of SST at two virtual stations, Con Dao and Tao Island 35

Table 3.3: ID and location of pixels chosen to build the SST time series data at Con Dao and Phu Quoc 37

Table 3.4: The flexible global tool for monitoring of coral bleaching 41

Table 4.1: 12 months sea surface climatology and Bleaching Threshold at Phu Quoc and Con Dao .52

Table 4.2: Degree Heating Week (0C weeks) at Phu Quoc in 1998 .57

Table 4.3: Degree Heating Week (0C weeks) at Con Dao in 1998 .59

Table 4.4: The cover of corals (%) and percentage of bleaching at Con Dao in October 1998 62

Table 4.5: Degree Heating Week (0C weeks) at Phu Quoc in 2005 .62

Table 4.6: Degree Heating Week (0C weeks) at Con Dao in 2005 .64

Table 4.7: Degree Heating Week (0C weeks) at Phu Quoc in 2010 .66

Table 4.8: Degree Heating Week (0C weeks) at Con Dao in 2010 .67

LIST OF ABBRIVIATIONS AND EQUATIONS

ABBRIVIATIONS

CRW Coral Reef Wacth

MODIS Moderate Resolution Imaging Spectroradiometer

NOAA National Oceanic and Atmospheric Administration

Equation 3.1: Convert AVHRR SST value to degrees C 33 Equation 3.2: Convert MODIS SST value to degrees C 33 Equation 3.3: Calculate HotSpot value 39 Equation 4.1: Linear relationship between the 2005 MODIS-derived weekly SST and the

2005 AVHRR-derived weekly SST .44 Equation 4.2: Linear relationship between the summer 2005 MODIS-derived weekly SST and the summer 2005 AVHRR-derived weekly SST .45 Equation 4.3: Linear relationship between the winter 2005 MODIS-derived weekly SST and the winter 2005 AVHRR-derived weekly SST .45

ABSTRACT

Thermal stress is marked as an essential cause of coral bleaching Corals begin losing the symbiotic algae and their color when sea temperature exceeds one to two degree Celsius

above the summer maxima Thus, monitoring thermal anomalies of seawater has become an imperative need The U.S National Oceanic and Atmospheric Administration (NOAA) has developed methods to predict bleaching based on sea surface temperature (SST) achieved from satellite images However, the predictions usually underestimate at local and regional scales due to the low spatial resolution (50 km) As a contribution to improve prediction of coral bleaching in the East Sea, Vietnam, my research examined SST derived from higher spatial resolution products (4 km), including the Advanced Very High Resolution Radiometer (AVHRR) and the Moderate Resolution Imaging Spectroradiometer (MODIS) Two parameters, HotSpot and Degree Heating Week (DHW) were applied to describe thermal stress in historical bleaching years at two archipelagos in Vietnam, Phu Quoc and Con Dao The major findings include: (1) The difference of SST in the southwest and southeast areas of the East Sea was revealed through SST climatology data That difference explained the phenomenon of fewer bleaching events at Phu Quoc because coral living in different temperature conditions can have different bleaching thresholds (2) Coral reefs at my study area suffered a severe thermal stress in 2010 with longer duration and higher DHW value than in 1998 and

2005 (3) The application of 4 km spatial resolution data prevented underestimating thermal anomaly and provided bleaching maps with more details than with 50 km resolution data

Keywords: Coral Bleaching, Sea Surface Temperature, Con Dao, Phu Quoc, East Sea, Vietnam, Remote Sensing

1987, Lalli and Parsons 1995)

As one of the most diverse ecosystems on Earth, coral reefs play an important role not only in the biodiversity of nature but also in the survival of many coastal communities Reefs provide habitat for a variety of marine species, protect shorelines from severe storms, prevent coastal erosion and filtrate pollution water from mainland However, the worldwide coral reefs are under serious pressures caused by human activities and climate change (Burke et al 2011) Among those stressors, coral bleaching

is marked as an essential contributor to global decline of reefs during the last decade (Wilkinson 2008)

Previous studies in the laboratory and field highlighted that thermal stress is the primary cause of mass bleaching, including the elevated sea temperature and the prolonged suffering thermal anomaly, (Glynn and D'Croz 1990, Glynn 1996, Brown 1997a, Lesser 2011) Thus, monitoring thermal anomalies of seawater has become an imperative need

Since remote sensing is the most effective means of acquiring global and continuous surface temperature, the U.S National Oceanic and Atmospheric Administration (NOAA) has developed a series of products to monitor the near-real-time thermal stress within the Coral Reef Watch (CRW) program Those products are based on studies about sea surface temperature (SST) anomaly, bleaching HotSpot and Degree Heating Week (DHW) to determine the potential bleaching level of most reefs around the world (Goreau and Hayes 1994, Montgomery and Strong 1995, Liu et al 2005)

However, the CRW alert products sometimes underestimate bleaching at regional and local scales due to the low spatial resolution (50 km) (Haines et al 2007, Weeks et

al 2008) As a contribution to improve prediction of coral bleaching in the southern area

of the East Sea, my research examined SST derived from higher spatial resolution satellite images (4 km), such as those acquired from the Advanced Very High Resolution Radiometer (AVHRR) and the Moderate Resolution Imaging Spectroradiometer (MODIS)

1.1 Coral Reefs

Coral reefs are underwater structures formed by colonies of thousands tiny stony corals called polyps, living in a symbiotic relationship with microscopic algae known as zooxanthellae Those algae shelter inside the gastrodermal tissues of polyps They utilize carbon dioxide produced by coral respiration and other compounds such as nitrates, and phosphates from the metabolic processes of corals for their photosynthesis (Figure 1.1)

In return, zooxanthellae provide corals oxygen and a various source of nutrients, including glucose, glycerol, and amino acids They also help the host corals in removing

wastes and building calcium carbonate skeletons (Barnes 1987, Lalli and Parsons 1995, Levinton 1995, Barnes and Hughes 1999)

Figure 1.1: Anatomy of a coral polyp Source: Encyclopedia Britannica http://www.britannica.com/EBchecked/topic/137037/coral

Since the symbiotic between corals and their zooxanthellae depends mostly on photosynthesis with high requirement of light intensity, reef sites primarily are distributed

in warm and clear shallow waters of the tropical and sub-tropical oceans, ranged between

300 North and 300South latitude (Figure 1.2) Although some coral reefs also are observed at higher latitudes or cold water, their diversity, such as the amount of species,

is lower than reefs in warmer regions (Barnes 1987)

Additionally, observations at 1000 reef sites indicated that coral requires discrete conditions of temperature, salinity, nutrient, aragonite saturation, turbidity and depth For example, average temperature at reefs ranges from 210C to 29.50C and the salinity is limited from 23.3 to 41.8 parts per thousand (ppt) Although corals can occur from 7 m to

91 m under sea level, the most prolific reefs occupy depths of 18 – 27 m Any anomalies exceed those limits may break the symbiotic relationship and therefore, cause distress or death of corals (Kleypas et al 1999, Hoegh-Guldberg 2011)

Figure 1.2: Major coral reef sites are red dots on the world map

Source: NOAA's National Ocean Service, Education Division http://oceanservice.noaa.gov/education

Coral reefs play a significant role on marine ecosystem Indeed, shallow water reefs provide habitats for more than 25% of all known marine species, including over 4,000 species of fish and 800 types of coral while they only occupy less than 0.015% area

of the world oceans (Spalding et al 2001, Burke et al 2011)

Moreover, recent studies conducted by the World Resource Institute also

countries and territories from the ravages of storms, and provide a source of food and income for millions of coastal people (Burke et al 2011)

1.2 Threats to Reefs

Although population residing within 30 km of reefs and less than 10 km from the coast has reached more than 275 million (Burke et al 2011), few people know their

“neighbor” ecosystem formed about 200 million years ago, surviving through plenty of

natural disturbance events such as the ice ages or climate change (Brown 1997b, Wilkinson 1998, Souter et al 2000)

Upon the most up-to-date report from the World Resources Institute (Burke et al 2011), coral reefs of the world are distributed in six regions: Middle East, Atlantic, Indian Ocean, Australia, Pacific and Southeast Asia More than 55% of these reefs are in the Southeast Asia and Pacific, while the best-reserved reefs are in Australia with 16.4% reef cover

The report also emphasized that coral reefs are in deep danger due to both human and natural disturbances Since human populations and coastal pressures increased dramatically in the 20th century, more than 60% of reefs are under immediate and direct threats from one or more local sources caused by human activities, including overfishing and destructive fishing, coastal development, watershed-based pollution and marine-based pollution

Human exploitation disturbs the ecosystem, break the calcium carbonate structure and decrease the number of species rapidly As a result, coral reefs become too weak to withstand natural events although healthy corals generally are resilient and eventually recover Other disturbances rised from human activities such as water pollution or

greenhouse gases are intensifying the frequency, magnitude, duration, and spatial distribution of natural hazards (Nystrom et al 2001) For example, plankton blooming and growth of alga organisms due to eutrophication can stifle corals or replace their habitat (Jones and Endean 1976) Disease outbreaks and predations also cause serious decline of reefs (Barnes and Hughes 1999)

Indeed, the percentage of worldwide reefs rated as threatened would increase up

to approximately 75% when local threats are combined with global threats, such as the increase of average temperature and acidification in oceans (Burke et al 2011)

Several disturbances are counted as natural threats to reefs They include hazards induced by climate events such as tropical typhoons, hurricanes and cyclones Particularly, the El Niño with its consequences such as the increased sea surface temperatures, decreased sea level, and altered salinity due to excessive rainfall can put coral under an extremely stressful condition (Forrester 1997) When corals suffer stress for a long time, they will expel their symbiotic algae and cannot uptake enough nutrients

to survive Elevated temperature and thermal anomalies in recent years are reported as being responsible for a mass bleaching of coral while rising acidity reduces coral growth rates and their ability to maintain physical structure (Burke et al 2011)

1.3 Coral Bleaching

Healthy reefs are seen as yellow or brownish color due to the photosynthetic pigments of their symbiotic algae If corals lose their zooxanthellae, the concentration of photosynthetic pigments declines, the living coral tissues become pale and the white of the skeleton underneath becomes visible This phenomenon is usually described as coral bleaching (Glynn 1996, Brown 1997a) (Figure 1.3)

Although corals are able to catch prey and feed themselves, most of them will be starved and weaken without their zooxanthellae Depending on the duration and level of stresses, reefs can either die or survive If the bleaching is not too severe, and the stressful conditions decrease in a short time, the affected colonies will regain their symbiotic algae within several weeks to months However, frequent or severe bleaching can induce reduction in skeletal growth, reproductive activity, and disease resistance Bleached corals often die if the stress persists and reefs suffering severe mortality following bleaching can take many years or decades to recover (Glynn 1996, Burke et al 2011)

Bleaching does not equally occur over single coral colonies within coral communities or across reef zones Under the same stress conditions, some species are more susceptible to bleaching than others In some cases, only the upper surface or lower surface of the colony is affected In others, bleached tissue appears as a circular patch or

in the shape of a ring or wedge (Glynn 1996, Burke et al 2011)

Many studies have been conducted to find causes of bleaching At a local scale, bleaching occurs under the effect of environmental stressors such as increased ultraviolet radiation, elevated sea temperature, fluctuations in light intensity, reduced salinity, bacterial infection, or a combination of those At regional and global scales, scientists emphasized high sea temperature as the primary cause of mass bleaching events (Glynn

1996, Brown 1997a, Lesser 2011) Observations and experiments also asserted that corals begin to bleach when sea temperature exceeded one to two degrees Celsius above the highest summer temperature (Berkelmans and Willis 1999, Reaser et al 2000, Donner et

al 2005)

Figure 1.3: Coral bleaching process Source: Great Barrier Reef Marine Park Authority http://www.gbrmpa.gov.au/about-us

The bleaching across the Pacific in 1982 was the first described bleaching event Since then, a growing number of “mass bleaching” events have been reported via the worldwide reef observations (Brown 1997a) Approximately 370 observations of coral bleaching were reported globally in the period of 1980-1997 while the amount of reports

in 1998-2010 was more than 3700 (Burke et al 2011)

The trend of coral bleaching is associated with global climate change and El Niño/ La Niña events For instance, the most notable bleaching event occurred in 1998 coincided with the extremely high temperature, responding to an unusually strong El

Niño sequence in 1997 Subsequently, extensive bleaching events were observed on the Great Barrier Reef of Australia in 2002 and parts of the Caribbean in 2005 The number

of severe bleaching reported in 1998, 2002 and 2005 was higher than in other years (Figure 1.4) Although the report is still in process, the World Resource Institute strongly predicted a worldwide bleaching event at multiple regions in 2010 because that year was one of the warmest years on record (Burke et al 2011, Hoegh-Guldberg 2011)

Figure 1.4: Trends in coral bleaching, 1980 – 2010 Source: Reefs at Risk Revisited (Burke et al 2011)

1.4 Predicting Coral Bleaching by Sea Surface Temperature (SST)

The need for improved understanding, monitoring, and predicting coral bleaching became imperative after the beaching event in 1998 Causes of bleaching and coral responses were examined in the laboratory and the field However, there are several

limitations from both approaches The laboratory approach is limited by the sensitivities

of most coral species to laboratory handling and the difficulties in imitating field conditions Meanwhile, the finding of field studies remain elusive because the shallow reef is an extremely complex and heterogeneous environment in both temporal and spatial dimensions (Richmond and Wolanski 2011)

Since thermal stress is highlighted as the primary cause of bleaching, monitoring sea temperature is an effective tool to assess coral health (Weeks et al 2008, Castillo et

al 2010, CRW 2011) Sea temperature data generally are acquired in two ways The

direct measurement, so-called in situ method, offers an accurate ambient temperature

even over short temporal and spatial scales but it is costly and time consuming, and

therefore, in situ data records generally are limited The second approach to thermal data

on reefs is collecting the sea surface temperature derived from satellite observations

Although some researchers doubt that satellite thermal products can describe accurately temperatures in coastal or subsurface environments (Brown 1997a, Castillo and Lima 2010), remote sensing is still at the forefront in detecting thermal anomalies on reefs because it is the most efficient way of acquiring data (Donner et al 2005, Haines et

al 2007, Weeks et al 2008)

Digital remote sensing of coral reefs began in the mid-1970s with the first Landsat mission (Smith et al 1975, Hochberg 2011) Early studies utilized moderate spatial resolution (20 – 80 m), broadband (60-100 nm bandwidth) and multispectral (two

or three wavebands in visible-infrared spectrum) satellite images to detect reefs and delineate reef geomorphology (Bina et al 1978, Loubersac et al 1988, Hochberg 2011) More recently, technical advances led to a development of high quality remote sensing

data such as the commercial high-spatial-resolution images and hyperspectral satellite products, which have been applied in a variety of coral reef studies such as habitat mapping, physical environment monitoring and bleaching prediction (Hochberg 2011, GEFCoral 2010)

In early 1997, the NOAA‟s National Environmental Satellite, Data, and

Information Service (NESDIS) began producing web-accessible, satellite-derived, global near-real-time nighttime SST products to monitor thermal conditions associate to coral bleaching and to assess the intensity of bleaching stress around the globe This activity evolved into a crucial part of NOAA's Coral Reef Watch (CRW) program in 2000 The CRW utilizes the near-real-time sea surface temperature derived from the Advanced Very High Resolution Radiometer (AVHRR) at 0.5 degree (approximately 50 km) spatial resolution to monitor thermal stress and predict coral bleaching The resultant composite SST products are calibrated to 1 m depth and produced twice weekly (Goreau and Hayes

1994, Montgomery and Strong 1995, Liu et al 2005, Strong et al 2006)

In commonly recognized products, the potential bleaching is identified by Degree Heating Week (DHW) and HotSpot value (Strong et al 2006) DHW is the accumulation

of thermal stress that coral reefs have experienced over the past 12 weeks up to and including the most current product update HotSpot is an anomaly product based on value

of the Maximum of the Monthly Mean (MMM) SST HotSpot value of 10C is a threshold for thermal stress leading to coral bleaching

The CRW products were criticized as less successful in detecting severe bleaching conditions at local scales because of their use of a constant thermal threshold (summer maximum monthly mean) and low spatial resolution (50 km) (Weeks et al

2008) To improve the prediction of coral bleaching, Weeks et al (2008) combined the 4

km spatial resolution AVHRR Pathfinder product and the 1 km resolution MODIS SST product to develop a seasonal and spatial thermal threshold for coral bleaching

In that research, the 50 km prediction with 10C thresholds from the NOAA CRW was compared with the 4 km prediction with the seasonal baseline thresholds at Heron Island and Keppel Island in the Great Barrier Reefs, Australia The 4 km prediction combined SST data derived from the AVHRR climatology (1985-2001) and the weekly SST data derived from MODIS to examine the thermal stress during bleaching event in austral spring and summer months (November- March) of 2005–2006 and 2001–2002 at the two islands Results showed at the near shore site where beaching was undetected in the 50 km satellite product, the seasonally adjusted thermal threshold produced a greatly improved consistency between accumulated heating and bleaching severity

1.5 Research Objectives

As an effort to improve the prediction coral bleaching at local scale, my research applied and modified the method introduced by Weeks et al (2008) to detect and predict bleaching of coral reefs in the southern area of the East Sea The research could meet the

purpose above through following objectives:

- Develop a thermal stress approach to examining historical coral bleaching with monthly SST from AVHRR satellite nighttime images in a span of 1985-2009;

- Apply the approach to develop a bleaching alert map in 2010 with weekly sea surface temperature from AVHRR and MODIS satellite nighttime images;

- Compare bleaching detection between the 50 km products released by CRW and the 4 km products in this research

Chapter Two

Case study

2.1 Status of Coral Reefs in Vietnam

Vietnam is a coastal country in the Southeast Asia The north and west of the country share borders with China, Laos and Cambodia while the south and east border the East Sea and Gulf of Thailand Vietnam has more than 3200 km shoreline, stretching through more than 15 degrees of latitude (8030‟North to 230

North) There are more than

3000 inshore and offshore islands that are favorable for the growth of coral reefs The exeptions are locations strongly affected by river inputs with low salinity and high turbidity such as the Red River Delta or the Mekong Delta (Wilkinson 1998, Vo et al 2006) (Figure 2.1)

According to the report, Status of Coral Reefs of the World: 1998 (Wilkinson

1998), reefs in Vietnam are distributed in five regions:

- The Gulf of Tokin, northern Vietnam

- The central coast

- The Gulf of Thailand, southwestern Vietnam

- The offshore islands and atolls in parts of the Spratlys claimed by Vietnam

More than 400 hard coral species belonging to 79 genera and 18 families occur in these areas For this reason, Vietnam is considered as one of the countries with high generic richness of coral fauna (Vo et al 2006) Observations from 1998 to 2008 showed that reefs in the southern region are more diverse than in the northern region The number

of coal species observed in southern reefs was 277 while in northern reefs it was 165 species (Vo 1998, Vo et al 2006)

Figure 2.1: Reef distribution in the coastal waters of Vietnam The two archipelagos in this research are

identified by the red rings (Source: Vo et al 2006 Modified by Ngan Le 2012)

Most threats to coral reefs in Vietnam come from human activities The Reef at Risk model conducted in 2002 assessed that human disturbances placed 96% of reefs in danger, with nearly 75% at high or very high threat A more detailed result indicated that destructive fishing, including trawling or using dynamite and cyanide, and overfishing were the major stressors of coral reefs Besides, sediment from upland sources and costal development also threatened the ecosystem (Burke et al 2002)

Natural threats to reefs in Vietnam include physical damage from typhoon and

bleaching caused by thermal stress For instance, in 1997, an unusual typhoon Lynda

destroyed most reefs at Con Dao archipelago in the southeastern East Sea After that, an extensive coral bleaching was observed from June to October 1998 The percentage of degraded coral at Con Dao was up to 70% within depth of 15 m Moreover, about 70–80% of coral at shallow water (1-2 m) was dead due to the severe bleaching Bleaching also affected many reefs at other sub-regions but reefs at Con Dao archipelago suffered the most severe devastation (Wilkinson 1998, Vo 2000)

2.2 Status of Coral Reefs at Con Dao and Phu Quoc Archipelagos

In 2007-2008, the Department of Exploitation and Conservation of Fishery Resources, Vietnam carried out a survey at four Marine Protected Areas (MPAs), including:

(1) Bach Long Vi Island in the Gulf of Tokin, northern Vietnam

(2) Con Co Island in the central coast

(3) Con Dao Archipelago in the southeastern Vietnam

(4) Phu Quoc Archipelago in the Gulf of Thailand, southwestern Vietnam

The survey indicated that reefs at Phu Quoc and Con Dao had the highest species diversity and reef percent cover (Table 2.1) These two archipelagos also shared 65% of common species because both are located in the southern East Sea and are very close geographically In 2008, the cover of live coral at Phu Quoc reached 36.44% and 29.71%

at Con Dao Meanwhile, the proportion of dead coral at two archipelagos was higher than the other sites, with 10.19% at Phu Quoc and 15.24% at Con Dao

Table 2.1: Cover (%) of coral and other benthos at 4 permanent monitoring sites in 2007-2008

Source: www.decafirep.gov.vn

Study site Dead

coral

Fleshy seaweed

Hard Coral Other

Rubble

of dead coral

Rock

Recent Dead Coral

Soft coral Sand Sponge

Bach

Long Vi 6.47 23.24 1.74 2.06 1.32 38.82 0.15 0.44 11.32 0.15 Con Co 8.09 8.94 27.62 0.85 6.44 23.43 0.95 9.59 14.09 0.00 Con Dao 15.24 5.16 29.71 1.45 6.92 28.14 0.00 2.7 10.65 0.02 Phu Quoc 10.19 0.58 36.44 2.4 3.08 19.33 0.00 0.19 27.79 0.00 Mean 12.05 7.05 28.88 1.48 5.82 26.72 0.25 3.91 13.8 0.03 From 1994 to 2004, the Oceanography Institute of Vietnam observed frequently the reef cover and number of marine species at eight sites of Con Dao (Figure 2.2) and six sites in the southern area of Phu Quoc (Figure 2.3) The observations at Con Dao were conducted in July 1994, May 1998, July 1999, June 2000, September 2001, June 2002 and April 2004 (Figure 2.4) The reefs at Phu Quoc were checked in March 1994, December 2000, April 2002, April 2003, June 2005 and December 2007 (Figure 2.5) (Vo

et al 2008)

Figure 2.2: Location of observed sites at Con Dao in 1994-2004 (Vo et al 2008)

Figure 2.3: Location of observed sites at Phu Quoc in 1994-2007 (Vo et al 2008)

Figure 2.4: The percentage of hard coral cover observed at Con Dao in 1994-2004 (Vo et al 2008)

Figure 2.5: The percentage of hard coral cover observed at Phu Quoc in 1994-2007 (Vo et al 2008)

During the observing period, the mean proportion of hard coral at Phu Quoc decreased less than that at Con Dao, with around 10% decline from 1994 to 2007 whereas the decrease at Con Dao was more than 15% from 1994 to 2004 (Fig 2.4 and 2.6) Although bleaching was not the only cause of the degradation of hard coral at the two archipelagos, it was considered as the major factor because the notable decline was observed after the bleaching event in 1998 After 1998, hard coral began to recover slightly

2.3 Coral bleaching at Con Dao Archipelago

The Con Dao archipelago, located in the southeastern waters of Vietnam at 80

37‟-80 48‟ North and 1060

32‟-1060 45‟ East, is characterized with diverse terrestrial and marine ecosystems The archipelago consists of 14 islands and approximate 1000 ha coral reefs with more than 320 hard corals species recorded Coral reefs here occur widely from littoral to the depth of 17 – 20 m Their structure is mainly composed of typical and non-typical fringing reefs (Vo 2000)

The observation carried out in 1994-1995 indicated that coral reefs at Con Dao was the best protected reefs in Vietnam with the average reef cover higher than 50%

However, affected by typhoon Linda in November 1997, the bleaching event in summer

1998 destroyed a large area of coral reefs at Con Dao, reduced the percentage of reef cover to 25% (Vo et al 2005) After that, coral reefs recovered very slowly

The Reef Check observation at Con Dao in April 1998 reported that exposed reefs

showed major hard coral cover lost (up to 100%) due to the typhoon Linda, while

repeated surveys at the sites which were not hit by the typhoon revealed that high coral cover was still present The observations in October 1998, which found that the percent area of coral colonies suffering bleaching ranged from 0 to 74.2% at eleven observed sites with an average bleaching rate of 37% (Vo 2000)

In October 2005, bleaching was observed in the northeastern waters of the archipelago, with mortality of corals ranging from 20% to nearly 100% The mass mortality is more serious in reef flats at depth of 2-3 m than in reef slopes at depth of 5-6

m (Figure 2.6) Most non-movable benthic invertebrates died and the density of reef fish

decreased seriously on impacted reefs However, there was no bleaching observed in the southeastern area of the archipelago (Hoang et al 2008)

The analysis of water samples and the data recorded by the Station of Meteorology and Hydrology at Con Dao showed that water temperature in Con Dao was higher than 300C in many days during the October of 2005, especially more than 310C on

11th and 12th of that month Meanwhile, salinity also reduced with value less than 25‰ in seven days The combined impact of those two factors was considered as the reason for mass mortality A hypothesis that fresh water from river mouths of the Mekong River system was dispersed to the archipelago during this period was recommended for further studies (Hoang et al 2008)

Figure 2.6: Coral bleaching was observed in the southwestern of Con Dao archipelago in October, 2005

(Source: Hoang et al 2008)

In 2010, thermal observation for Con Dao, published by the NOAA CRW, predicted that a severe bleaching would occur in July 2010 due to the prolonged thermal stress in this year (Figure 2.7) However, the Oceanography Institute in Vietnam did not conducted a reef check to assess the level of impact at Con Dao Thus, there are no ground data regarding the bleaching in 2010 at this archipelago

Figure 2.7: Bleaching alert graph (based on the 50 km spatial resolution data set) at Con Dao in 2010 Source: http://coralreefwatch.noaa.gov/satellite/vs/southeastasia.html#ConDao_Vietnam

2.4 Coral bleaching at Phu Quoc Archipelago

Located completely in the Gulf of Thailand, Phu Quoc, Nam Du and Tho Chu are three main archipelagos in the southwestern waters of Vietnam Those archipelagos are recognized as significant biodiversity and fisheries areas of Vietnam

In Phu Quoc, reefs occupy in the shallow waters of the south (An Thoi) and the north-west (Ghenh Dau, Cua Can) (Vo et al 2005) Status of reefs at those sites ranges from good to severe impact by destructive and over fishing The observations in 2002-

2003 found that marine living resources on coral reefs declined, especially the commercial species (Nguyen 2003) A sign of negative impacts to marine environment

by sedimentation and pollution from mainland agricultural was recorded

Particularly, Phu Quoc is a famous tourism attraction of Vietnam Along with impacts of fishing activities, the increase of frequent tourists also affects the coral reefs directly or indirectly

The newest observation (unpublished) conducted by the Oceanography Institute of Vietnam in May and August, 2010 reported that coral bleaching was observed at most reefs of Phu Quoc due to the thermal anomalies in 2010 (Figure 2.8)

Figure 2.8: Coral bleaching was observed at Phu Quoc archipelago in May (a) and August (b), 2010

(Source: the Oceanography Institute at Nha Trang, Vietnam) (unpublished)

2.5 Global Thermal Anomalies in 2010

2010 was the warmest year in the 131-year record of global surface temperatures (1880 to 2010) (NASA's Goddard Institute for Space Studies (GISS) 2011) The analysis

of global surface temperature found that 2010 was approximately 2.030C warmer than the average temperature from 1951 to 1980 (Hansen et al 2010) The GISS analysis is based

on weather data of more than 1000 meteorological stations around the world, satellite observations of sea surface temperature and Antarctic research station measurements

Since the weather stations are not distributed evenly enough across the globe to provide meaningful measurements, a relative measure called a “temperature anomaly” was calculated to track whether global temperatures are changing instead of absolute global average surface temperatures (Figure 2.9) In order to generate temperature

long-term average called “base period” The base period serves as a point of reference against which climate change can be tracked The GISS data indicated that the temperature trend showed a warming climate by approximately 0.650 C per decade since the late 1970s, and the last decade has been the warmest on record (Hansen et al 2010)

Figure 2.9: Global temperature anomalies since 1880 The thin line represents annual data, while the thick

line is a five-year running mean (Source: NASA Earth Observatory/Robert Simmon)

In addition, the 2010 Global Climate Highlights released by NOAA indicated that the global ocean surface temperatures in 2010 was the third warmest on record, at 0.490C above the 20th century average of sea temperature Consequently, the NOAA Coral Reefs Watch‟s Thermal stress Outlook announced that the 2009-2010 El Niño caused an intensive thermal stress in most of the northern Indian Ocean and Southeast Asia regions

Although the rainy season was expected to relieve the high thermal stress in these regions and to promote the recovery of bleached corals, reefs in the eastern East Sea experienced a significant bleaching (Figure 2.10)

Figure 2.10: Coral bleaching alert areas in Western Pacific in summer 2010 Source:

http://coralreefwatch.noaa.gov/satellite/bleachingoutlook/outlook_messages/bleachingoutlook_20100608_f

or_2010junsep.html

In brief, the study area of this research focused on the reefs at Con Dao and Phu Quoc With rich observations of satellite imagery at higher resolutions in past 30 years, the research examined the differences of historical bleaching events in these two archipelagos, and predicted their bleaching in 2010 with the apparent thermal anomaly in this year These high-resolution findings will filled in the gaps of the 50 km bleaching

Chapter Three

Materials and Methods

3.1 Data Description

3.1.1 AVHRR Pathfinder (PF) Version 5 SST Products

NOAA has been measuring sea surface temperatures with satellites since 1972 One of the primary sources of infrared data for SST measure is the Advanced Very High Resolution Radiometer (AVHRR) carried on NOAA's Polar Operational Environmental Satellites (POES) series, beginning in 1978 The POES satellite system offers the advantage of daily global coverage, by making near-polar orbits roughly 14 times daily (NOAA Coral Reef Watch 2011)

AVHRR is a broad-band scanner to acquire images, with five or six wide spectral bands (depending on the model) in visible, near-infrared, and thermal infrared portions of the electromagnetic spectrum The first two are centered at red (0.6 micrometer) and near-infrared (0.9 micrometer) regions, the third one at around 3.5 micrometers, and the last two bands are the emitted thermal radiation at around 11 and 12 micrometers, respectively Only the last two bands at 11 and 12 micrometers are used to derive SST (Kilpatrick et al 2001)

The NOAA/ NASA AVHRR Oceans Pathfinder Program developed global SST

at 9.28 km resolution beginning in early 1990s The AVHRR Pathfinder Version 5 SST Project (Pathfinder V5) is a reanalysis of the AVHRR data stream developed by the Rosenstiel School of Marine and Atmospheric Science (RSMAS) and the NOAA National Oceanographic Data Center (NODC) The reanalysis reprocessing used an

improved version of the Pathfinder algorithm and processing steps to produce daily global SST and related parameters back to 1981, at a resolution of approximately 4

twice-km This SST time series represents the longest continuous global ocean physical measurement from space

Distribution of AVHRR Pathfinder products is a collaborative effort between the NASA Physical Oceanography Distributed Active Archive Center (PO.DAAC) and NODC They also offer temporal averages for 5-day, 7-day, 8-day, monthly, and yearly periods The file format for all data is Hierarchy Data File (HDF) format (Casey et al 2010) More detailed information could be accessed at the NODC Website: http://www.nodc.noaa.gov/SatelliteData/pathfinder4km

For coral watch data, the nighttime SST are preferred to eliminate daily warming caused by solar heating at the sea surface (primarily at the "skin" interface, 10-20 µm) during the day and to avoid contamination from solar glare Compared with daytime SST and day-night blended SST, nighttime SST provides a more conservative and stable estimate of thermal stress conducive to coral bleaching Nighttime SST also compare

favorably with in situ SST at one meter depth (Montgomery and Strong 1995)

In this research, the “climatology” SST, showing the major characteristics of

seasonal variation of sea temperature at Con Dao and Phu Quoc, were produced from the AVHRR Pathfinder Level 3 Monthly Nighttime SST Version 5 data This data set covers

300 images (12 months x 25 years) to provide monthly SST from 1985 to 2009 at 0.044 degrees (Latitude) x 0.044 degrees (Longitude) spatial resolution

In other to examine the thermal stress occurring in 1998 and 2005, the AVHRR Pathfinder Level 3 8-Day Nighttime SST Version 5 products were used Each year includes 45 weekly SST images at the same spatial resolution with the monthly product

Data are available at the following links:

ftp://podaac-ftp.jpl.nasa.gov/allData/avhrr/L3/pathfinder_v5/monthly/night/04km/ ftp://podaac-ftp.jpl.nasa.gov/allData/avhrr/L3/pathfinder_v5/8day/night/

3.1.2 MODIS Aqua SST Products

The MODIS sensor was first on-board the Terra polar-orbiting spacecraft launched on 18 December 1999 The second MODIS sensor was on-board the Aqua launched on 24 May 2002 As both Terra and Aqua satellites have a sun synchronous, near-polar orbit, a specific area on the Earth‟s surface is over flown by each satellite approximately every 12 hours and therefore, the sensors aboard provide both daytime and nighttime images Over the study area, Terra over flights occurred approximately between 10:00 and 11:00 (local time) during the day and between 22:00 and 23:00 during the night; Aqua passes occurred between 13:00 and 14:00 during the day and between 01:00 and 02:00 at night MODIS captures data in 36 spectral bands between 0.405 μm and 14.385 μm, with spatial resolutions at nadir of 250 m (2 bands), 500 m (5 bands) and

1000 m (29 bands) (Shutler et al 2005)

SST is derived from the 11 and 12 μm thermal infrared (IR) bands and the 3.959

μm and 4.050 μm mid-IR bands Since the short thermal infrared bands near 4 μm are

affected by bright reflective sources such as sun glint, the short wavelength SST is only available for nighttime while the long wavelength SST product is available for both daytime and nighttime Different algorithms are used to calculate the SST for short