Variability of the Gulf Stream

CHAPTER 1

INTRODUCTION TO THE RESEARCH

The Gulf Stream is an extremely important current located in the Atlantic Ocean that carries warm water from the coast of North America towards Europe. The Gulf Stream has a significant impact on the climates of both the southeastern portion of the United States and much of the area in northern Europe. Without the Gulf Stream, northern Europe would be very cold and their economy would be totally different. If the Gulf Stream would drastically change its path it would greatly affect life in northern Europe and cause significant problems for the world’s economy. As you can see, the Gulf Stream is an extremely important geophysical phenomenon and needs to be studied. This chapter will provide an introduction to this research project and focus on the basic goal of this research study.

The aim of this research project is to use sea-surface temperature data collected from the Advanced Very High Resolution Radiometer (AVHRR) which imaged the Atlantic Ocean over a period of two years to track the path of the north wall of the Gulf Stream. These data will be used to determine the variability of the path of the north wall of the Gulf Stream over time. The geographical positions of the Gulf Stream will be used to determine the variability of the path of the north wall of the Gulf Stream for an El Niño year compared to a normal year.

The null hypothesis is that the path of the north wall of the Gulf Stream does not significantly vary over time and holds the same path for an El Niño year as it does for a normal year.

• One limitation in this study is that we are only measuring the north wall of the Gulf Stream as opposed to measuring both the north and south walls to get more accurate results.
• Another limitation in this study is that we are measuring the coordinates of the north wall of the Gulf Stream by human interpretation rather than using a computer to get precise values. This makes the experimental results less repeatable.
• Another limitation used in this project is that we used only data from two civilian NOAA satellites instead of gathering data from military weather satellites as well.
• Another limitation of this study is that we only studied one current making it impossible to the see the variability of all currents and the total affect El Niño has on other bodies of water.

• One of the delimitation’s that affects this project is that we only have access to two years worth of sea-surface temperature data. If we had access to substantially more data this study could be much more thorough and thus provide us with more conclusive results on long term changes of climate.
• Another of the delimitation's that affects this study is the fact that many of the images of sea-surface temperature data were cloudy and hard to interpret. This complication led to the researchers slightly interpolating the path of the north wall of the Gulf Stream through cloudy areas, which could impact the results of the study. Also, the cloudiness led to the dismissal of many images giving the researchers less data to work with.

• One assumption made in this study is that the current of the Gulf Stream can be found by using sea-surface temperature data. Since the Gulf Stream is a current, defined by flowing water and not temperature, it is impossible to know if the exact position of the temperature matches the position of water flow of the Gulf Stream. We assume, however, that the north wall of the Gulf Stream is defined by the drastic change of temperature from cold water to warm water.
• Another assumption made in this study is that the effects of El Niño are consistent throughout the year and do not vary over the interval being studied.

1. Gulf Stream- warm water mid-Atlantic current flowing from southern Florida to northern Europe.
2. Advanced Very High Resolution Radiometer (AVHRR)- instrument used to measure emitted radiation for an area of about one square kilometer from a NOAA series satellite. This information can then be converted to sea-surface temperature.
3. National Oceanic and Atmospheric Administration (NOAA)- government organization responsible for the weather prediction that puts up the NOAA series of satellites.

In this project the variability of the path of the north wall of the Gulf Stream will be determined and the affect of El Niño on the path of the north wall of the Gulf Stream will be studied. From sea-surface temperature data the path of the north wall of the Gulf Stream will be mapped out for a period of two years. From this study we hope to find a correlation between the path of the north wall of the Gulf Stream and time and be able to draw significant conclusions from our data.

CHAPTER 2

BACKGROUND INFORMATION

The Gulf Stream has been a very important and well-studied current for many years. Due to the increased exploration of North America in the early 1500’s the Gulf Stream was discovered and became a very important part of trade and commerce between North America and Europe. In 1513, Ponce de Leon made the first reference to the Gulf Stream after having trouble crossing its path in the area just north of what is now Cape Canaveral (3,73). After Ponce de Leon’s discovery, the Gulf Stream became a well-known current to explorers from Europe and provided a fast route to take when returning from the America’s. In the late 1700’s the United States Postmaster-General, Benjamin Franklin, compiled one of the first accurate maps of the Gulf Stream to aid his ships in the delivery of mail between North America and London (3,74). Franklin, also, noticed the effect of wind on ocean circulation and came up with the idea of using sea-surface temperature to navigate and track the ocean (3,74). The Gulf Stream was at this time regarded as a fixed and stable way to return to Europe. Along with its importance in trans-Atlantic travel, the Gulf Stream controls much of the climate in Northern Europe. By bringing the waters of the Gulf of Mexico to Europe, the Gulf Stream creates warm winds that keep the climate very moderate as opposed to the harsh conditions that exist in corresponding latitudes throughout the world. From this effect on climate, the Gulf Stream therefore has a tremendous impact on the industry in Northern Europe by allowing people there to raise livestock and farm. The Gulf Stream has an enormous influence on climate and economic human life and determining and understanding its path and variability can be extremely useful.

In the 1960's, advances in oceanography led to a more thorough study of the Gulf Stream with more powerful measuring techniques. Multiple ship surveys organized by Fuglister (9,265) provided the first systematic description of the spatial structure between Cape Hatteras and the Grand Banks including the waters north and south of the Gulf Stream. Also, this period presented the first use of infrared radiation imaging to map the thermal patterns of ocean surfaces from orbiting satellites (14,4501). Through the new application of modern measuring techniques and abundant historical information, our knowledge of the Gulf Stream has expanded dramatically. This can be seen in Stommel’s (21) classic modern work on the physical descriptions of the Gulf Stream. The Gulf Stream has been found to be a wind-driven ribbon of high-velocity water forming the boundary between the warm waters of the Sargasso Sea and the cooler waters of the continental margin. The sharp gradient of temperature seen across the Gulf Steam is an expression of the balance between the horizontal pressure gradient force and the Coriolis force brought about by wind (3,79). With this understanding of the Gulf Stream it is possible to track its position through the use of sea-surface temperature data and in turn study its variability in greater depth. To use this method of tracking the Gulf Stream using sea-surface temperature, a way to easily determine these data accurately must be found.

One way to estimate the sea-surface temperature of a body of water is through the use of infrared radiation measurements taken from a satellite orbiting the earth. This technique for estimating sea-surface from measurements of radiance is based on the physics of blackbody radiation. By definition, a blackbody is a body that absorbs all incoming radiation. In 1893, Wien found the wavelength distribution of the radiation emitted from a blackbody, B(l ,T), to be:

where l is electromagnetic wavelength, T is the absolute temperature, and f(l T) is a function of the product of the wavelength and the temperature (6). This distribution function fit the observed data at the time and went to zero at the extremes of both short and long wavelengths. In 1901, Planck assumed that energy can only exist in discrete packets equal to hv where h is Planck’s constant and v is the electromagnetic frequency (6). From this and other thermodynamic considerations, Planck derived the blackbody radiance as:

where c is the speed of light and k is Boltzmann’s constant (6). With this equation the spectral radiance over a finite wavelength band can be determined. If N(l ₁,l ₂) represents the measured radiance in the region , then for a blackbody:

From the measurement of radiation emitted by a body, an associated temperature such that at that same temperature a blackbody would emit the same radiation can be found and this temperature is known as the brightness temperature. With the previous integral, the measurement of the radiance of a body within a specified wavelength window, the assumption that the ocean is very close to a blackbody and all other things being equal, a body’s true surface temperature can be thus determined (18,2).

Although this is an accurate way of estimating sea-surface temperature, there are some major limitations that must be overcome to ensure the data collected are truly accurate. First, clouds block the infrared radiation from the surface and, if cloudy regions cannot be identified, the cloud temperature can be incorrectly associated with sea-surface temperature. By identifying cloudy regions and ignoring the data in these areas, the clouds may slightly reduce the amount of data obtained, but will not affect the study as long as these data are ignored.

Another limitation of estimating sea-surface temperature is the problem of the atmospheric influences on the several ways the amount of radiation that reaches the radiometer. The intervening atmosphere can absorb some of the radiation emitted by the surface or it can emit its own radiation that can either go directly to the satellite or be reflected off the surface back to the satellite. Although this problem seems very significant, the effect of the atmosphere on radiation is wavelength dependent and can be easily corrected by the passive measurement of known radiation at different wavelengths to infer an atmospheric correction (18,2).

A third problem that must be overcome is the effect of solar radiation reflected off the surface on the measurement of the radiometer. This problem can, also, be corrected by the choice of wavelength, the identification of regions containing this problem, and by collecting the data at night.

For this study the optimum wavelengths used were AVHRR channels 4 and 5, which correspond to wavelengths of 11 and 12 micrometers respectively. These channels provided the most accurate results by not being significantly influenced by the effects of the atmosphere and solar radiation. The AVHRR channel 3, 4 micrometers, was also used at night because of its ability to make corrections for atmospheric interference, but not during the day because it is too sensitive to solar radiation. To obtain more accurate data in this study the data were compiled over a three-day span where the highest temperature at each spot on the ocean was used in the final data. The highest temperature was used because, if clouds were over the area, the cloud temperature could greatly affect the mean temperature and by using the highest temperature it would ensure that the temperature used was actually that of the ocean. By using three channels and compiling the data over a three-day period, the sea-surface temperature data can be assumed to be very accurate and account for all of the limitations presented by this method of estimation.

Global ocean circulation is controlled by the "thermohaline" conveyor belt. Wind forces and pressure gradients drive ocean surface motion. Pressure gradients are formed by water density, which is in turn determined by temperature (thermo) and salt concentration (haline);(2,1582). This "conveyor belt" describes the motion between ocean basins and vertically between the top and bottom of the ocean. Variability in ocean circulation can be caused by upsets in the balance of thermohaline circulation and, by observing variability in an ocean current, major geophysical changes might be able to be predicted.

Since circulation is a global phenomenon, changes in one area can effect the ocean and atmosphere thousands of kilometers away. Although the El Niño phenomenon has been observed for centuries, it has recently become clear that its effect is worldwide. El Niño events occur at several year intervals when atmospheric changes cause warm water to spread across the central Pacific. Such changes have been correlated with significant weather changes in North America and Europe. One aim in this research project is to determine if the position of the Gulf Stream changes between an El Niño year and a non-El Niño year. This change could help explain the weather changes already known to exist during an El Niño year. The problem will be distinguishing between typical seasonal variability and changes caused by El Niño.

Much previous work on Gulf Stream variability has been done which suggests a significant variability caused by several different factors. In 1975, Duing (5,129) used shipboard survey measurements to determine that Gulf Stream variability occurred over 4 to 10 day periods. Likewise, in 1978, Maul et al. (16,123) observed the Gulf Stream using the Geostationary Operational Environmental Satellite (GOES) to analyze the infrared data and obtain multiyear histories of Gulf Stream meanders. Maul (16,123) concluded that the meanders of the Gulf Stream were related to variability associated with 4 to 16 day and 40 to 100 day periods of flow. Although these studies provided researchers with the first background information in Gulf Stream variability, the measuring techniques were not advanced and were inaccurate. By using point measurements taken from ship surveys, Duing (5,129) did not have a great deal of data throughout the entire ocean and many errors could have occurred in collecting the data he did receive. Maul used the more advanced method of tracking the Gulf Stream from satellite, but this still presented a problem for his study. By using a geo-stationary satellite, the data received was low resolution compared to the more advanced orbiting NOAA satellites because of the greater distance from the earth and also the less sensitive radiometer it was equipped with. In addition to the low quality data, Maul (16,123) might have had a large amount of data for a single day, but did not conduct the research over a long period of time, thus limiting his study. These inaccuracies made the studies informative, yet inconclusive and the need for future research became apparent. In 1983, Watts (23,115) compiled much of the Gulf Stream data already collected from previous studies and found that variations occur over 4 to 14 day and 20 to 60 day periods. Watts (23,115) suggested that the Gulf Stream had seasonal and yearly changes in the meanders and began to determine reasons for these changes. In studying ocean circulation, Watts (23,115) attributed the variability of the Gulf Stream to changes in wind patterns, sea-mounts and the interaction with rings that have spun off. None of these causes of Gulf Stream variability are the sole cause and all of these factors along with other unknown factors are probably responsible for Gulf Stream variability. In this study, hopefully an accurate representation of the variability of the Gulf Stream can be found and through the comparison of data from a non-El Niño and an El Niño year the effect of Gulf Stream variability due to El Niño can be determined.

CHAPTER 3

RESEARCH METHODOLOGY

The Gulf Stream is an extremely important and well-known current located in the western Atlantic Ocean flowing between North America and Europe. The Gulf Stream has a significant impact on the climates of both the southeastern portion of the United States and much of the area of northern Europe. If the Gulf Stream were to drastically change its path it would greatly affect the climate and consequently the life and economy in northern Europe.

Global ocean circulation is dominated by the "thermohaline" conveyor belt that exchanges deep ocean water with surface water between ocean basins. The melting of ice in arctic regions dumping fresh water into the ocean or the heating of equatorial waters by El Niño may alter the thermohaline circulation and disturb surface currents. Thus, changes in near-surface ocean currents may not only affect global climate, but may be an early indication of such changes. This research project is an attempt to determine if it is possible to detect significant changes in the geographical location of the Gulf Stream between an El Niño and non-El Niño year. Another goal of this project is to estimate the day-to-day variability of the Gulf Stream. This measurement will determine how large year-to-year variations in the Gulf Stream must be to be considered statistically significant.

To understand and quantify the variation and position of the Gulf Stream, a way must be found to accurately and routinely measure the path taken by flowing water. The Gulf Stream is a warm water current and since the water of the Gulf Stream is significantly warmer than the surrounding water of the Atlantic Ocean, sea-surface temperature data can be used to estimate the path taken by the Gulf Stream. Other phenomenon, like the mixing of deep cold water with warmer surface water caused by storm mixing, can also effect sea-surface temperature. However, these other phenomenon, tend to be transient and unimportant in terms of the long-term variability of the Gulf Stream. From this important assumption, that sea-surface temperature is dominated by current flow, research can begin in the attempt to determine the variability of the Gulf Stream.

There are many ocean currents measurable by sea-surface temperature variation that are candidates for investigation. The Gulf Stream was chosen as the focus of this project because of its climate importance and the availability of locally received satellite data. In addition, the Gulf Stream was selected because it generally has a clearly defined northern edge that can easily be determined by the 5-10 degree temperature change across its border. The availability of the data and the importance of the Gulf Stream to climate made it an obvious selection.

Measuring Instruments

The instruments used to measure the sea-surface temperature data were the NOAA-12 and NOAA-14 Advanced Very High Resolution Radiometers (AVHRR). Recently the NOAA-15 was launched, but these data were collected before new satellite data were available. These AVHRR instruments measure infrared radiation emitted from the ocean. The intensity of the radiation is related to the sea surface temperature. The AVHRR instruments measure sea surface temperature at a 1 km resolution within the 2000 km wide swath of the instrument. Each NOAA satellite visits the region of the Gulf Stream six times a day, broadcasting the data as it is received. Taken together these satellites can image the geographical position of the Gulf Stream on a routine basis.

The Johns Hopkins University Applied Physics Laboratory has for the last several years routinely received AVHRR data from NOAA-12 and NOAA-14 and produced three-day composite images of sea surface temperature in the western North Atlantic. Three-day composites help remove the blocking effects of clouds. This imagery, which generally shows the Gulf Stream, is continually posted on the World Wide Web (http://fermi.jhuapl.edu/avhrr/).

1. Collect sea-surface temperature imagery in the form of GIF (Graphic Interchange Format) files for the northern portion of the Atlantic Ocean over a period from August 1996 to September 1998 using the NOAA-12 and NOAA-14 AVHRR. This imagery has geographic information embedded, so the relationship between position in the image and longitude and latitude is known.
2. Using IDL software that converts the click of the mouse button into longitude and latitude coordinates, estimate the path of the north wall of the Gulf Stream from the sea-surface temperature data by clicking on the path where the water temperature jumps from cold water to warm water. Latitude measurements at each degree of longitude from 85° W to 55° W are automatically stored in a text file.
3. Compare the resulting data examining the variability of the path of the north wall of the Gulf Stream with time while examining whether the mean Gulf Stream position between an El Niño and non-El Niño are statistically different.

• The current of the Gulf Stream can be found by using sea-surface temperature data. Since the Gulf Stream is a current, defined by flowing water and not temperature, it is impossible to know if the whole Gulf Stream is accounted for. We assume, however, that the north wall of the Gulf Stream is defined by the drastic change of temperature from cold water to warm.
• Atmospheric corrections applied to the raw AVHRR data by the Johns Hopkins University Applied Physics Laboratory properly adjust the temperature measurements for the effect of atmospheric water vapor.

• Measuring only the north wall of the Gulf Stream as opposed to measuring the north and south walls to get more accurate results.
• Measuring the coordinates of the north wall of the Gulf Stream by human interpretation rather than using a computer to get more repeatable values.
• Using only two years worth satellite’s data instead of gathering data from a group of different satellites as well.

From the sea-surface temperature data we will hopefully be able the get an accurate continuous representation of the path of the north wall of the Gulf Stream. By comparing the data taken over a period of two years we will look for a correlation between the path of the Gulf Stream and time and examine the effects El Niño has on the path of the north wall of the Gulf Stream. For each degree of longitude, we have a daily measurement of the latitude of the Gulf Stream. Thus, at each longitude, it is possible to compute the mean and standard deviation of latitude each year. Using standard tests for statistical significance, it is possible to determine whether the observed difference in the mean from year to year is a mere statistical fluctuation or a statistically significant difference.

CHAPTER 4

ANALYSIS OF THE DATA

The aim of this research paper is to use sea-surface temperature data collected from the AVHRR which imaged the Atlantic Ocean over a period of two years to track the path of the north wall of the Gulf Stream and to determine the variability of the path between an El Niño year and a normal year. From the satellite data the path of the Gulf Stream was extracted using a program which converts the click of a mouse into corresponding latitude and longitude measurements. By clicking on each longitude line at the point of the sharpest temperature change the path of the Gulf Stream can be found. Now that the satellite data has been processed, a way to interpret these data accurately and appropriately must be determined.

First, the descriptive statistics of mean and standard deviation must be determined to obtain a more accurate picture of the data and provide a basis for more in depth statistical evaluation. Also, the difference of the means from the two years will be evaluated to get a better picture of the variability in the path of the Gulf Stream. After examining the mean and standard deviation, the auto-correlation function for the measurements of the Gulf Stream is needed to determine the number of independent measurements collected in the data. This is necessary because the path of the Gulf Stream does not vary greatly day-to-day and since the measurements were taken daily, many measurements are not independent of one another. After the number of independent measurements is obtained, a two-sample t-test can then be used to determine if the difference in the two means is significant.

The null hypothesis is that the path of the north wall of the Gulf Stream does not vary over time and holds the same path for an El Niño year as it does for a normal year.

The raw data for this experiment was obtained by the AVHRR on the NOAA-12 and NOAA-14 satellites which compiled the data into GIF images, which displayed the sea-surface temperature of the northern portion of the Atlantic Ocean. The sea-surface temperature images were obtained for each day over the two-year period from May 1996 to June 1998. These images were placed on a compact disc and filled approximately 500 megabytes worth of data. Using the previously mentioned IDL program, the measurements of the position of the north wall of the Gulf Stream were interpreted from the sea-surface temperature images and compiled into text files. After all of the images had been processed the data was then compiled into one large 524 kilobyte text file which would serve as a basis for all future statistical calculations. This is an example of a small portion of the text file showing the longitude and latitude coordinates:

Figure 1 is an example of one the sea-surface temperature images:

(Figure 1)

Now that all of the raw data has been processed descriptive statistics can be used to better understand the data and get a clearer picture of the results of the experiment. From the data the mean and standard deviation for all of the data (Figures 2 & 5), the first year of data (Figures 3 & 6) and the second year of data (Figures 4 & 7) were calculated.

(Figure 2)

(Figure 3)

(Figure 4)

(Figure 5)

(Figure 6)

(Figure 7)

After examining the mean position of the north wall of the Gulf Stream the difference of the mean position between Year 1 and Year 2 can be calculated (Figure 8).

(Figure 8)

Now that the descriptive statistics have been calculated further analysis of this data is needed to determine if there is significant variability between the two years.

To further analyze the data using inferential statistics the two-sample t-test will be used to determine if the difference between the two means is significant. One problem that must be overcome before we use the two-sample t-test, is the determination of the number of independent measurements. Since the Gulf Stream does not move much from day to day, successive measurements are not independent and thus for 365 days of data there are not 365 independent measurements. The auto-correlation function allows us to determine what and really are by estimating how separated in time successive measurements must be in order to be considered independent.

An example will illustrate. Let be the latitude of the north wall of the Gulf Stream as a function of time index. For our case, represents the first day, represents the second day and so on. For convenience, we subtract off the mean latitude so that . The auto-correlation function is defined as:

where is the time lag index.

For example, when , the auto-correlation function becomes 1. As grows larger the product inside the summation becomes at times positive and at other times negative and the sum approaches zero. When the auto-correlation becomes zero, it is reasonable to assume independence. By using an auto-correlate function in the IDL program, the auto-correlation function was determined (Figure 9).

In this experiment, so every 25^th day is statistically independent, thus making and measurements.

With this information the two-sample t-test can be performed to make final conclusions about the data. The two-sample t-test is a test to determine to what degree, if any, the difference between two means is significant. This test is used to determine how large must be for two means to be statistically significant and not just a statistical fluctuation. The two-sample t-test is defined as:

(2)

where is the mean, is the number of significant measurements and is the standard deviation. After calculating the t-values a standardized table of values can be used to determine the significance of the measurements and either accept or reject the hypotheses, and . The following table (Table 1) provides t-test results for each longitude, the critical values, their corresponding levels of significance and the direction of the variation from Year 1 to Year 2:

(Table 1)

After obtaining all of the raw data and processing this data to make it useful, statistics were first used to get a better representation of the data and make it easier to interpret. The auto-correlation function was essential to the research because it provided the actual number of samples to make the findings more accurate. Also, the auto-correlation had yet to be determined for measurements taken of the Gulf Stream making its finding extremely useful. Through the use of the two-sample t-test, the significance of the difference between the two means can be found. The statistics used in this section can truly determine if there is significant variability of the path of the north wall of the Gulf Stream between an El Niño and a regular year.

CHAPTER 5

CONCLUSIONS

After the data has been collected and further analyzed, it is necessary to interpret the findings and draw statistically supported conclusions. In this study, we are trying to determine if El Niño affects the path of the north wall of the Gulf Stream. Two years of data, one that was an El Niño year and one that was not, were collected using sea-surface temperature data from an AVHRR. To determine if El Niño has an effect on the path of the Gulf Stream the difference between the means for the two years needs to tested for statistical significance. The null hypothesis stated that there would be no difference between the means of the two years. After determining the number of independent measurements using the auto-correlation function, the two-sample t-test was used to see if the data supports or rejects the null hypothesis.

The two-sample t-test is a test to determine to what degree, if any, the difference between two means is significant. This test is used to determine how large must be for two means to be statistically significant and not just a statistical fluctuation. The two-sample t-test is defined as:

where is the mean, is the number of significant measurements and is the standard deviation. In this study the auto-correlation function (Figure 8) found that at a period of twenty-five days the measurements of the path of the Gulf Stream were independent giving us fourteen measurements for each year and thirteen degrees of freedom. With this information the t-test can now be performed to determine the level of significance of the results found. After calculating the t-values, a standardized table of values can be used to determine the significance of the measurements. Table 1 displays the results of the two-sample t-test for the collected data in this experiment with the first column being the longitude measurement, the second column being the calculated t-value, the third column being the critical values with their corresponding levels of significance and the fourth column being the direction of the variation in the path of the Gulf Stream between year 1 and year 2.

(Table 1)

After obtaining all of the data, many steps were taken to analyze the data and draw meaningful conclusions. First, we calculated the mean and standard deviation for the position of the north wall of the Gulf Stream for all of the data as well as the first year and second year separately (Figures 1-6). From a time series of Gulf Stream positions the auto-correlation function was computed. The interpretation of the auto-correlation function plot (Figure 8) indicates that measurements approximately twenty-five days apart are statistically independent. As a result, we have fourteen independent measurements for each year. From the two-sample t-test (Figure 9) we found that at several locations across the path of the Gulf Stream there was a ninety percent certainty that the difference between the mean path was statistically significant. In addition, at various longitudes we could be ninety-two, ninety-five, and ninety-seven percent certain that the year-to-year difference was significant. The path of the Gulf Stream near the North American coast appeared to be relatively stable, but as the measurements were taken farther out into the ocean the path was more variable. Although the findings suggest that El Niño significantly effects the position of the Gulf Stream, the variability in the path could simply have been caused by normal year-to-year variations. Also, since the majority of the results were only ninety percent certain, further research to accumulate more statistics needs to be conducted to verify these preliminary results. In conclusion, we found that the auto-correlation function for the Gulf Stream indicates that Gulf Stream measurements are independent at approximately twenty-five day intervals and made a strong case that suggests that the mean paths of the Gulf Stream between an El Niño and a non-El Niño year are statistically different.

Recommendations for the replication or continuation of this project would include using a higher resolution radiometer with more channels to better make corrections for atmospheric interference. This would give the researcher clearer images to work with and insure more accurate results. Also, a computer program that would find the highest temperature gradient at each longitude and mark it as the path of the Gulf Stream could be written to process the raw satellite data. This would eliminate human error from the project and allow for more data to be analyzed by making data processing easier and faster.

Future research on the variability of the north wall of the Gulf Stream can and needs to be done. In this study, we were not able to distinguish if the variations were caused by El Niño or by unrelated and normal year-to-year variations. This research can be extended to determine the isolated effect El Niño has on the path of the north wall of the Gulf Stream by obtaining more data from El Niño years and non-El Niño years. In addition, as more years pass and more data is collected, the true variability of the Gulf Stream can be determined with greater precision. In the future, researchers studying the variability of the Gulf Stream can look for other causes of the observed variability and arrive at a more complete understanding of the causes of the variability of the Gulf Stream. Through the accumulation of more data and by studying the causes of the Gulf Stream, the relationship between Gulf Stream variability and long-term climate changes can be identified.

1. Bernstein, R.L. (1982). Sea surface temperature estimation using the NOAA-6 advanced very high resolution radiometer. Journal of Geophysical Research, 87, 9455-9465.
2. Broecker, W.S. (1997). Thermohaline Circulation, the Achilles Heel of Our Climate System: Will Man-Made CO2 Upset the Current Balance. Science, 278, 1582-1587.
3. Brown, J., Colling, A., Park, D., Phillips, J., Rothery, D., & Wright, J. (1989). Ocean Circulation. Oxford: Pergamon Press.
4. Dobson, E.B., Monaldo, F.M., Porter, D.L., Robinson, A.R., Kilgus, C.C., Goldhirsh, J., & Glenn, S.M. (1992). Radar Altimetry And Global Climate Change. Johns Hopkins APL Technical Digest,13(3), 431-444.
5. Duing, W., Mooers, C., and Lee, T. (1977) Low Frequency Variability in the Florida Current and Relationship to Atmospheric Forcing from 1972-1974. Journal of Marine Research,35, 129-161.
6. Eisberg, R.M. (1961). Fundamentals of Modern Physics. New York: John Wiley and Sons.
7. Fraedrich, K., Muller, K. and Kuglin, R. (1992). Northern Hemisphere Circulation Regimes During the Extremes of El Niño/Southern Oscillation. Tellus, Series A, Dynamic Meteorology and Oceanography, 44(1), 33.
8. Friedlander, A., and Tracey, K.L. (1989). Producing accurate maps of the Gulf Stream thermal front using objective analysis. Journal of Geophysical Research, 94, 8040-8052.
9. Fuglister, F.C. (1963). Gulf Stream 1960’s. Progress in Oceanography, 1, 265-373.
10. Gulev, S.K., Kadeyev, D.G., and Yashayev, I.M. (1988). Synoptic variability of watertemperature in the vicinity of the Gulf Stream and north atlantic current. Oceanography, 28, 562.
11. Hamilton, K. (1993). A General Circulation Model Simulation of El Niño Effects in the Extratropical Northern Hemisphere Stratosphere. Geophysical Research Letters, 20(17), 1803.
12. Jones, C.G., & Thorncroft, C.D. (1998). The Role Of El Niño in Tropical Cyclone Activity. Weather, 53(9) 324.
13. Kontoyiannis, H., and; Watts, D.R. (1994). Observations on the Variability of the Gulf Stream Path between 74 degree W and 70 degree W. Journal of Physical Oceanography, 24, 1999-2013.
14. Legeckis, R. (1978). A Survey of Worldwide Sea Surface Temperature Fronts Detected by Environmental Satellites. Journal of Physical Oceanography, 83, 4501-4522.
15. Malmgren, B.A., Winter, A. and Chan, D. (1998). El Niño--Southern Oscillation and North Atlantic Oscillation Control of Climate in Puerto Rico. Journal Of Climate, 11(10), 2713.
16. Maul, G.A., deWitt, P.B., Yanaway, A., and Baig, S.R. (1978). Geostationary Satellite Observations of Gulf Stream Meanders: Infrared Measurements and Time Series Analysis. Journal of Geophysical Research, 83, 6123-6135.
17. McClain, E.P., Pichel, W.G., and Walton, C.C. (1985). Comparative performance of AVHRR-based multi-channel sea surface temperature. Journal of Geophysical Research, 90, 11,587-11,601.
18. Monado, F.M. (1997). Primer on the Estimation of Sea Surface Temperature Using TeraScan Processing of NOAA AVHRR Satellite Data. Internal Publication for Johns Hopkins University Applied Physics Laboratory #S1R-96M-03, Laurel.
19. Porter, D.L., Glenn, S.M., Dobson, E.B., and Crowley, M.F. (1996). Extension and Validation of a Gulf Stream Geosat Synthetic Geoid. Journal Of Atmospheric And Oceanic Technology, 13(2), 514-531.
20. Stewart, R.H. (1985). Methods of Satellite Oceanography. University of California Press, Berkley.
21. Stommel, H. (1965). The Gulf Stream: A Physical and Dynamic Description . University of California Press, Berkley.
22. Vazquez, J. (1993). Observations on the Long-Period Variability of the Gulf Stream Downstream of Cape Hatteras. Journal of Geophysical Research, 98, 20133-20,147. Watts, D.R. (1983). Gulf Stream Variability. Eddies in Marine Science, 115-143.