|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
EVALUATION OF THE MODIFIED ECFL LURCS RIP CURRENT
FORECASTING SCALE AND CONDITIONS OF SELECTED
RIP CURRENT EVENTS IN FLORIDA
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
This document is dedicated to my parents and brothers.
I would like to thank my family for their support and advice during my past two-
and-a-half years at UF. I always loved to go home.
I thank Dr. Robert Thieke for allowing me to take part in the rip current program
and for finding the funds to pay me along the way. I also thank him for his advice,
assistance, and cheerful optimism (even in the case of a $15,000 current profiler being
buried under three feet of sand and water).
My thanks go to Dr. Andrew Kennedy and Dr. Robert Dean, my supervisory
committee, for their open doors and for taking their time to provide assistance.
Students (past and present) whom I would like to thank are Jason Engle and Jamie
MacMahan for their enormous assistance with field work and other issues that arose
during the past year. Great appreciation goes out to Oleg Mouraenko for his MatLab
expertise. I would also like to thank my good friend, Enrique Gutierrez, for his help and
advice throughout this study.
TABLE OF CONTENTS
ACKNOWLEDGMENT S .............. .................... iv
LI ST OF T ABLE S .........__.. ..... .___ .............._ vii..
LIST OF FIGURES ........._.___..... .__. ..............viii...
AB STRAC T ................ .............. xi
1 INTRODUCTION ................. ...............1.......... ......
2 RIP CURRENT S .............. ...............4.....
3 IMPORTANCE OF PREDICTING RIP CURRENTS .............. ....................7
4 STUDY SITES AND DATA ................. ...............11........... ...
Volusia County ................. ...............11.......... .....
Site Description ................ ...............11.................
Rip Current Rescue Data ................. ...............12................
W ave Data .............. ...............12....
Tidal Data ................ ...............16.................
Photographic Data ................. ...............17.......... .....
M eteorological Data ................ ...............17.................
Beach Population Data ................ ...............17........... ....
Panhandle Counties .............. ...............18....
Site Description ................ ...............18.................
Rip Current Death Data ................. ...............18................
W ave Data .............. ...............19....
5 STATISTICAL ANALYSIS .............. ...............21....
Volusia County ......................... ...............2
Rip Current Rescue Statistics .............. ...............21....
W ave and Tide Statistics .............. ...............24....
Panhandle Counties ...................... ...............30
Rip Current Drowning Statistics .............. ...............30....
W ave and Tide Statistics ................ ...............31................
Value of W ind Statistics ................. ...............35........... ...
6 METEOROLOGICAL ANALYSIS ................. ...............37........... ....
Wind, Pressure, and Frontal Systems .............. ...............37....
Volusia County 1996 ................. ...............38................
Volusia County 2002 ............. ...... ._ ...............44..
Florida Panhandle 2003 .............. ...............52....
Summary ............. ...... ._ ...............59...
7 MODIFIED ECFL LURCS SCALE .............. ...............61....
A analysis .............. ...............61....
8 SUMMARY AND CONCLUSIONS .....__.....___ ..........__ ...........71
LIST OF REFERENCES ............. ...... .__ ...............744..
BIOGRAPHICAL SKETCH .............. ...............766....
LIST OF TABLES
5.1: Rip current related deaths in Florida Panhandle counties during the summer of
2003 ................ ............... ......... ........ ......... ........ ...._..31
6.1: Conditions for high rescue days in Volusia County during June/July 1996............38
6.2: Wave conditions for five rip current outbreaks in Volusia County, 1996. ........._....41
6.3: Estimated conditions for May 7th May 20th, 2002 on Volusia County beaches. ..45
7.1: Parameter values used by the Modified ECFL LURCS for May 7th-20th, 2002,
Volusia County, FL. ............. ...............64.....
LIST OF FIGURES
4.1: Map showing study site in Volusia County, Florida and the location of the Sontek
wave gage and camera used for still and time-lapse photos ................. ................11
4.2: Output comparison between WDS and DIWASP. ............ ......................1
4.2. Continued ................. ...............15.......... .....
4.3: Study area and location of NDBC buoys #42040, located 64 nm south of Dauphin
Island, AL at water depth 237m, and #42039, located 115 nautical miles east-
southeast of Pensacola, FL at water depth 284m. ..........._.....__ ...............19
5.1: Daily beach population compared to daily rescues and daily normalized rescues for
May 7th through 20th, 2002. Normalized rescues are daily rescues divided by the
daily population multiplied by 10,000. ............. ...............23.....
5.2: Frequency distributions of A) deep water significant wave height, B) wave period,
C) deep water dominant direction D) directional spreading, and E) tidal stage. May
7th through 20th, 2002, 10:00am through 5:00pm. Rip rescue data is normalized by
beach population. ............. ...............26.....
5.3: Entire record of deep water wave heights for Volusia County, FL and 10:00am
through 5:00pm rescues. May 7 May 20, 2003 ................ ................. ....._28
5.4: Entire record of deep water wave directions for Volusia County, FL and 10:00am
through 5:00pm rescues. May 7 May 20, 2003 ................ ............... ...._...29
5.5: Entire record of tidal stage for Volusia County, FL and 10:00am through 5:00pm
rescues. May 7 May 20, 2003. ............. ...............30.....
5.6: Wave conditions at Walton County, estimated from NDBC buoy #42039. Eight
drownings occurred on June 8, 2003, most likely between 10:00am and 6:00pm
(the first "probable hours of drowning" section). One drowning occurred on June
9th, most likely during the second "probable hours of drowning" section. ..............33
5.7: Wave conditions at Bay County, estimated from NDBC buoy #42039. Two
drownings occurred on July 2, 2003. .........._...._ ......_. ......_. ...............34
5.8: Wave conditions at Pensacola estimated from NDBC buoy #42040. Four
drownings occurred on August 31 (Sunday, Labor Day weekend), 2003. ...............35
6.1: Surface weather maps for A) June 27th, 1996 at 7:00am EST through J) July 1st,
1996 at 7:00pm EST. The pressure given is the last three digits of the actual
pressure (987=998.7mb, 024=1002.4mb). ............. ...............39.....
6.1. Continued .............. ...............40....
6.2: Surface weather maps for A) June 18, 1996 at 7:00pm EST through C) June 19,
1996 at 7:00pm EST. ................. ...............42....... ....
6.3: Surface weather maps for A) August 9, 1996 at 7:00am EST through D) August 10,
1996, 7:00pm EST. ........... _............ ...............43..
6.4: Surface weather map for June 5, 1996 at A) 7:00am EST and B) 7:00pm EST......43
6.5: Surface weather maps for July 13, 1996 at A) 7:00am EST and B) 7:00pm EST...44
6.6: Surface weather maps for July 16, 1996 at A) 7:00am EST and B) 7:00pm EST...44
6.7: Surface weather maps for A) May 9, 2002 at 7:00am EST through J) May 13, 2002,
7: 00pm EST ............... ...............45..
6.7. Continued .............. ...............46....
6.7. Continued .............. ...............47....
6.8: Time-lapse photographs of Ormond Beach, Volusia Co. on May 11Ith, 2002 at A)
10:31am and B) 11:01am and May 12th at C) 11:33am and D) 1:33pm. .................47
6.8. Continued .............. ...............48....
6.9: Surface weather maps for A) May 18, 2002 at 7:00am EST through B) May 18,
2002, 7:00pm EST ................. ...............48........... ....
6. 10: Snapshots of Ormond Beach, Volusia Co. on May 18th, 2002 at A) 12:01pm and B)
2:01pm. One rescue occurred around each of these times. ................ ................. 49
6.11: Surface weather maps for A) May 7, 2002 at 7:00am EST through D) May 8, 2002,
7: 00pm EST ............... ...............49..
6.11. Continued .............. ...............50....
6.12: Surface weather maps for A) May 14, 2002 at 7:00am EST through D) May 15,
2002, 7:00pm EST. ...._.._................. ........._.._.......5
6.12. Continued .............. ...............51....
6. 13: Time-lapse photos of Ormond Beach, Volusia Co. on A) May 14th, 2002 at 2:31Ipm
and B) May 15th at 3:09pm. One rescue occurred around each of these times. ......5 1
6.14: Surface weather maps for A) May 19, 2002 at 7:00pm EST and B) May 20, 2002 at
7:00am EST. ................. ...............52.._._._ ....
6.15: Wave conditions estimated from NDBC buoy #42040. One drowning occurred on
each day from May 9th May 11Ith, 2003 ....__.. ............_... .........._ ..........53
6.16: Surface weather maps for A) May9th, 2003 at 7:00am EST through F) May 11Ith,
2003 at 7:00pm EST ................. ...............54...............
6.18: Surface weather maps for A) June 7th at 7:00am EST through F) June 9th, 2003 at
7:00pm EST............... ...............56..
6.19: Surface weather maps for July 2nd at A) 7:00am EST and B) 7:00pm EST. ...........57
6.20: Wave conditions estimated from NDBC buoy #42039. Two drownings occurred
on July 2nd............ ...............58....
6.21: Surface weather maps for the period A) August 30th, 7:00am through D) August
31st, 7:00pm. Eastern standard time. ............. ...............58.....
6.21. Continued .............. ...............59....
6.22: Wave conditions estimated from NDBC buoy #42040. Four drownings occurred
on August 3 1st............. ...............59...
7.1: Example computation of the Modified ECFL LURC S checkli st. ................... .........62
7.2: Modified LURCS rip current predictive scale results for May 7th-20th, 2002,
Volusia County, FL. ............. ...............63.....
7.3: Modified LURCS rip current predictive scale results, using normalized rescues, for
May 7th-20th, 2002, Volusia County, FL. ............. ...............65.....
7.4: Spectra depicting A.) wave direction at 4m depth and B.) wave frequency for May
11th, 2002 in Volusia County, FL ................. ...............66........... ..
7.4: Spectra at 4m water depth depicting A.) May 14th, 2002, 11:00am and B.) May
16th, 2002, 12:00pm. Volusia County, FL ................. ...............67........... ..
7.5: Time-lapse photo of Ormond Beach, Volusia Co. May 7th, 2002 at 1:01pm. One
rescue occurred between 12:00pm and 1:00pm. ............. ...............68.....
7.6: Entire record of directional spreading for Volusia County, FL and 10:00am through
5:00pm rescues. May 7 May 20, 2003 .......... ................ ........... ..........70
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
EVALUATION OF THE MODIFIED ECFL LURCS RIP CURRENT FORECASTING
SCALE AND CONDITIONS OF SELECTED RIP CURRENT EVENTS INT FLORIDA
Chair: Robert Thieke
Major Department: Civil and Coastal Engineering
Sea state parameters and meteorological conditions associated with rip current
events were evaluated for sites in Volusia County, Florida, and counties in the Florida
Panhandle. Data from a two-week period in Volusia County were used to make an
unbiased evaluation of the Modified ECFL LURCS rip current forecasting scale. Rip
current rescues performed by county lifeguards during this time were used as markers for
rip current occurrences. The Modified ECFL LURCS scale was used to hindcast the rip
current threat at these times. An evaluation of the scale was made by comparing the
calculated threat of occurrence to actual rip rescue occurrences. The scale predicted rip
currents for the maj ority of days containing rescues and did not predict rips on days
without rescues. A further study of rip current occurrences along the Florida Panhandle
revealed some similarities in wave characteristics and meteorological events that may
create sea conditions which force rip currents. Some conditions, however, are dissimilar
and suggest that specific locations have their own range of key parameters, and a specific
range of parameters cannot be used to accurately predict rip current threat in different
locales. Meteorological events at the time of rip occurrences were evaluated in an effort
to study the effect of passing weather systems, which induce varying wave and wind
directions and strongly alter the sea state.
As part of an ongoing study at the University of Florida, meteorological maps,
photographs, wave and tide data were used in an attempt to characterize dangerous rip
current events in various sites throughout Florida. The Modified ECFL LURCS, a rip
current predictive index developed by Engle (2003), was used in the evaluation. This
served as a blindfolded test of the index.
Rip currents, on average, result in more deaths in Florida than hurricanes, tropical
storms, lightning and tornadoes combined (Lascody 1998). Beaches in Volusia County,
on Florida' s east coast, and in the counties of Florida' s Panhandle, which border the Gulf
of Mexico, have high numbers of rip current rescues and deaths compared with other
Florida beaches. Volusia County averages more rip current rescues each year than all
other Florida counties combined (Lascody 1998). Panhandle counties, including
Escambia, Santa Rosa, Walton, and Bay County experienced a high number of rip current
drownings and rescues during the summer of 2003.
In an effort to predict the formation of rip currents, Lushine (1991) developed the
Lushine Rip Current Scale (LURCS), an empirical forecasting technique that utilizes
wind direction and velocity, swell height, and the time of low tide to forecast rip current
danger in South Florida (Engle 2003). The LURCS scale was modified for use in east
central Florida (ECFL LURCS) by changing the tidal factor and including swell period
The ECFL LURCS was modified by Engle (2003) in order to reduce the amount of
false alarms predicted by the scale. Two wind factors were removed from the scale and
three parameters were added: 1) an improved tide factor, 2) a wave direction factor, and
3) a directional spreading factor. These modifications were based on directional wave
data obtained for Volusia County for 1996 and improved the accuracy of the ECFL
LURCS scale for that specific site.
In this study, with the intention of testing the Modified ECFL LURCS scale, a two-
week data set from the Daytona Beach area was evaluated. This period corresponded to a
field experiment which included the deployment of a directional wave gage. This study
also had the added parameter of daily beach population which was used to normalize the
number of daily rip current rescues.
It was thought that wave and sea characteristics that create rip currents in Volusia
County may also create rips in other parts of Florida. To investigate this theory, wave
characteristics and tidal stage at sites along the Florida Panhandle were evaluated on days
marked by rip current drownings during the summer of 2003. An evaluation of the
parameters favorable for rip current formation in Volusia also showed a positive
correlation to rip formation in the Panhandle, although values and ranges of the
Large-scale meteorological events such as pressure systems may play a part in the
creation of rip current producing conditions. Meteorological maps, time-lapse and still
photos taken on days with high numbers of rip current rescues were evaluated to examine
possible connections between rip currents and weather patterns. On days with high
numbers of rescues and deaths, all available data, including wave and sea state
parameters, meteorological maps, photographs and beach population are presented in an
attempt to describe as fully as possible all factors that may lead to events of rip current
Rip currents are relatively narrow currents, generated in the near shore, which
move offshore through the surf zone. The word "rip" may be derived from the idea that
the current rips through the sand and offshore bar, creating channels through which the
current flows. The offshore velocity in rip currents can exceed 2 m/s (4.5 mi/hr), and
they contribute to the death toll at beaches by carrying unwary swimmers directly
offshore into deep water (Dean and Dalrymple 2002). They can form along beaches with
varying topography, or flat beaches with sandbars, as well as near structures such as piers
Tidal stage, wave height, period, wave direction, and directional spreading have
been correlated to rip current formation. Shepard et al. (1941) noted the relation between
larger waves and stronger rip currents as well as a connection with tidal stage. In 1958,
McKenzie correlated wave direction with the orientation of rip currents, stating that rips
commonly turn diagonally across the surf zone into the direction of approaching waves.
Engle's (2003) research in Volusia County, Florida indicated an especially strong
correlation of rip current rescues with both wave direction and directional spreading.
Sonu (1972) and Lushine (1991) also found positive correlations between shore-normal
wave angles and rip current occurrence. Lushine specifically referred to winds of
sustained onshore direction, which would imply that the wave direction was most likely
also onshore. According to anecdotal evidence from the Panhandle, rip current accidents
are more likely to occur after a period of sustained onshore wind.
Surf-zone topography has also been found to have a strong influence on rip current
formation and strength. Rip currents are associated with longshore variations of bottom
contours in the nearshore, where rip currents occur in the deeper regions and shoreward
transport occurs over the shallower regions. This topography includes beach cuspate
features as well as deeper channels occurring periodically in the offshore bar. Beaches
along Volusia County are straight, and longshore-alternating bars and rip channels are
reasonably common. The beaches are fairly two-dimensional, but when rip currents are
present, low relief cuspate features and feeder channels can be seen in the swash zone.
Sonu (1972) related rip currents to surf-zone undulations. His study area was
Seagrove Beach, in the Florida Panhandle, in the vicinity of data collected for this study.
Beaches here tend to be more three-dimensional than those of Volusia County. Beach
cuspate features, and the nearshore circulation they induce (where waves tend to break on
the cusp horns, then run out of the embayments), may form more permanent channels for
rip current action. It is possible that tidal stage is not as important here as in Volusia
County, where the more two-dimensional beach profile may require certain sea
conditions for rip channels to open, and lower water levels for the rip cells to strongly
develop. However, at Seagrove Beach, Sonu (1972) positively correlated increasing rip
current velocities to lower tide levels. It was thought that low tide caused stronger
breaking on the bar, increasing set-up and radiation stress which would strengthen rip
current intensity. For a more in depth discussion of the general characteristics and causes
of rip currents, the reader is referred to Engle (2003).
The difficulty with using rescue data as a marker for rip current occurrence is that
people must be in the water for rescues to occur. Rip currents are probably occurring
during conditions that include long period, large onshore waves (and probably onshore
wind). However, due to the fact that no swimmers are in the water when conditions are
rough, a lack of rescue data does not verify this.
Lushine observed that rip currents would sometimes continue to occur once the
wind turned clockwise after a several day period of moderate to strong onshore wind. He
also noted that rip currents generated by swells, especially when these swells are
decreasing, can be particularly hazardous, because local winds may be light, and bathers
may be deceived into thinking surf conditions are safe (Lushine 1991). It appears that rip
rescues begin to occur with a drop in wave height, decreasing (or more offshore) winds,
and a general improvement in the weather. An increase in rescues during such
"improving" conditions was noted by both Lushine (1991) and Engle (2003). Often, this
type of improving weather and sea state is a characteristic of a passing frontal system in
IMPORTANCE OF PREDICTING RIP CURRENTS
Rip currents are responsible for about 150 deaths every year in the United States
(Hauserman, 2003). Florida has a high number of rip current rescues and drownings each
Figure 3.1: Number of rip current drownings throughout Florida for the period 1989-
1996, totaling 180 rip current drownings.
Both the Panhandle of Florida and Volusia County, containing Daytona Beach,
have a high number of beach going tourists in the summer months. Volusia county
lifeguards rescued 2,399 people from rip currents in 2001, accounting for 68% of all
rescues performed (Volusia County 2003). During the summer of 2003, counties along
the Panhandle recorded an unprecedented amount of rescues and deaths. There have
been, "upwards of 40 drowning deaths since Jan. 1, 2000," in the Panhandle counties of
Gulf, Bay, Walton, Okaloosa, Santa Rosa and Escambia (Hauserman, 2003).
In 1991 Lushine noted that little had been published about the number of
drownings caused by rip currents or about attempts to operationally forecast their
occurrence. He then developed the LURCS scale to predict rip current threat.
The scale works by assigning values to specific ranges of parameters. Those
ranges that correlate to the highest incidence of rip rescues are assigned the highest
values. The sum of all parameter values is the "scale" of the rip current threat, and
warnings are issued to the public depending on this scale exceeding a certain threshold
Subsequent modifications to Lushine's scale by Lascody (1998) and Engle (2003)
seem to indicate that parameters (and their ranges) affecting rip current development are
site specific. The parameters included in Lushine's scale were wind direction, wind
speed, swell height, time of low tide, and a persistence factor relating to previous days'
conditions. Lascody's modifications included the addition of a wave period parameter
and a miscellaneous factor to account for higher astronomical tides. Engle removed the
wind parameters and introduced a directional spreading parameter and a more detailed
tide parameter, among other modifications. Each researcher' s modifications to an
existing scale lead to the notion of parameters being site specific. For instance, the effect
of tidal stage, may be more important at one site than another. Similarly, the range of
tide that correlates to the strongest rips may also differ between sites.
It is because of this that quite possibly one rip current forecasting scale cannot
adequately predict rip currents for a large expanse of coastline, such as east central
Florida. It is possible that a scale needs to be developed for each region, based on certain
parameters and ranges.
For instance, on June 8, 2002, Walton County beaches in the Panhandle recorded
eight rip current deaths. Data describing the primary forcing parameters are comparable
to days in Volusia County with large numbers of rescues: relatively shore normal wave
direction, and decreasing wave height. However, tidal fluctuation on June 8th, in Walton
County, was the smallest of the month. The Modified ECFL LURCS, developed for east
central Florida, could not be directly applied to beaches in Walton County even though
some of the forcing signatures are similar. A new scale would have to be developed
including parameter ranges most influential to that region.
An improvement that might be made to prediction efforts, and could very likely be
applicable to sites throughout, and outside of, Florida, is accounting for large-scale
meteorological events. This study found a correlation between the occurrence of rip
currents and the presence of pressure and frontal systems. Storms usually accompany
these systems, bringing strong winds and larger waves with long periods.
It is the purpose of this study to predict rip currents, not rip current rescues. Rip
current rescue data, however, are used to mark the existence of rip currents in the belief
that where there are rescues, there are currents. Rescue logs from Volusia County are
very detailed and clearly indicate what type of rescue took place. Rip current rescues are
always noted with the word "RIP" to distinguish them from other types of rescues such as
"swimmer in distress," or "overturned jet ski." Logistical problems limit the opportunity
to measure rip currents directly since they are sporadic and not always located in the
same place. Therefore rip rescue data, despite their shortcomings, represent the most
available and credible evidence available.
STUDY SITES AND DATA
The coastline of Volusia County, including Daytona Beach, was the study site on
Florida' s east coast.
:F~agl r__________ ~~~L tionr of: Soutek wav~e
Sotkwvegg n caerue for stllad timne-lapehos
Beaces lon ths aea re elaiveystoraih an adadpridclysae
rip chanels occr fairl frequetly inteofhr ar h vrg eahsoefo h
upper_~ bec aet h et f lsr s14,adte ensdmn imtri
0.2mm ~ ~ ~ ~ ~ at th hrln Cale ta.19) he contietlshlis7kofhrean
the bottom slopes mildly to this point. The beach and beach face are relatively two-
dimensional, and offshore contours are relatively shore-parallel. Semidiurnal tides have
an approximate maximum range of 2 meters. During the summer months (the time of
this and Engle's (2003) study) wave directions are normally from the southeast.
Northeasterly wave directions occur during most of the winter and are characteristic of
offshore summer storms. The average deep water wave height for 1996, according to
Engle's study (2002), was 0.7 meters. The average wave height during this study was
Rip Current Rescue Data
Rescue data were made available by the Volusia County Beach Patrol. Rescue logs
are kept for each of 103 lifeguard towers along Volusia's beaches. The logs list: type of
rescue, time of rescue, tower number, and number of victims. Rip rescue data from all
towers in Volusia County were used in this study. However, only rip rescues from the
hours of 10 a.m. to 5 p.m. were considered since these are the hours during which the
beach is most populated. This procedure hopefully reduces the likelihood of a rip current
event not being "recognized" due to a lack of swimmers
Directional wave data were collected by a Sontek wave gage. The gage was
deployed roughly '/ miles from shore in 4 meters of water off Ormond Beach in Volusia
County (see Figure 4.1i). This site was chosen for its proximity to where bathymetric data
was being collected in the hopes of mapping rip channels. The gage uses one pressure
sensor and three velocity beams, which measure water velocities in the x (east- positive),
y (north- positive), and z (up- positive) directions. An internal compass records
velocities relative to magnetic north. The package samples at a rate of 2 Hz, and
recorded data in hourly bursts. Each of 322 bursts for the period contained 4200 samples.
Spikes in the time series were replaced by linearly interpolated values.
Available data covered a two-week period from May 7th through May 20th, 2002.
The pressure/velocity data were analyzed by a suite of Matlab programs, titled DIWASP
(Directional Wave Spectra Toolbox for Matlab). DIWASP used the IMLM (Iterative
Maximum Likelihood Method) of spectral estimation in order to compute wave height,
period, and direction for the two-week period. The directional spectra were then created
from this output. Frequency resolution for the directional spectra was 0.01 Hz, and
directional resolution was 1 degree. Wave direction used in this study corresponds to the
dominant direction (Dp) output from DIWASP. This is the direction with the highest
energy integrated over all frequencies. Directional spreading was calculated using
Matlab code written and used by Engle (2003) for his study.
Statistics were derived from the directional energy density spectra, S(f,6), as
follows. The kth moment of the spectral density function, denoted mk, is defined as:
mk kXS~f,0 Od
The kth angular moment of the spectral density function, denoted dmk, is defined as:
dmk k~Slf,0 fd6
Engle (2003) computed directional spreading as:
However, he did not remove the mean direction from all directions when calculating the
second angular moment to input into the calculation for directional spreading. If the peak
of the directional spectrum was around zero (in which case the mean would be close to
zero), failure to remove the mean would not create a significant difference in the
calculation. However, if the peak is not centered close to zero, removing the mean
becomes important. The following additions were made to his code:
mean direction (D) -
One =0 D
dm2 new2,,;Slf, 0,,,fd6
In order to verify the DIWASP output, the same pressure/velocity data was run
through another suite of Matlab programs created by Nortek, titled 'WDS' (Wave
Directional Spectrum). Output from this suite includes wave height, period, direction,
and directional spreading.
A. Significant Wave Height (m)
7 8 9 10 11 12 13 14 15 16 17 18 19 20
Figure 4.2: Output comparison between WDS and DIWASP. A) significant wave height,
B) peak period, C) peak wave direction, and D) directional spreading. Zero
degrees corresponds to shore-normal. Counter-clockwise from shore-normal
is the positive direction.
7 8 9 10 11 12 13 14 15 16 17 18 19 20
C. Peak Wave Direction (deg)
7 8 9 10 11 12 13 14 15 16 17 18 19 20
B. Peak Period (s)
D. Directional Spreading (deg)
Figure 4.2. Continued
Figure 4.2 shows that wave heights and periods between the two suites agreed well.
Wave directions given by WDS were consistently more shore normal than DIWASP,
however the general directions (northeasterly or southeasterly) agreed. Engle used
DIWASP in his study, and the DIWASP data were used in this study for a few reasons.
The first reason is that the DIWASP wave directions showed better agreement with
photographic evidence showing a pronounced shift in wave direction from southeast to
northeast on May 14th and 19th. WDS does not show such a marked change in direction.
It is also important that DIWASP computes a dominant wave direction (Dp) as well as a
peak wave direction (WDS only calculates peak wave direction; therefore, that is the only
period comparison shown in Figure 4.1). Engle (2003) computes directional spreading
related to the dominant direction (Dp), which is the direction with the highest energy
integrated over all frequencies. WDS computes directional spreading for the peak
direction integrated over the peak frequency. This difference is the main reason why
directional spreading results from WDS were smaller than those from DIWASP. In order
to stay consistent with Engle (2003), the dominant wave period from DIWASP and
Engle's program for directional spreading were used in this study.
Tidal data were retrieved from a web-based tide predictor
"http://tbone.biol. sc.edu/tide/index.html" Comparison with tidal data from 1996 used by
Engle indicated a constant difference of +0.53m due to the use of a different tidal datum
(the predictor uses mean lower low water). Therefore, this factor was subtracted from all
2002 predicted tidal data in an effort to keep parameters used in this study and Engle's as
similar as possible.
Photographs of the beach and ocean near the field site in Ormond Beach, Volusia
County, Florida were taken using an automated camera situated on the roof of a beach-
front condominium approximately 1 mile south of the site (see Figure 4.1).
The camera location was chosen because it is 1) in Volusia County, just north of
Daytona Beach, 2) shoreward of where the Sontek wave gage was deployed, and 3) the
site of bathymetric surveys.
For the two-week study period in 2002, a snapshot was taken at the start of every
hour, and then a 3-minute time-lapse photo was taken. Unfortunately, not many time-
lapse photos exist. Due to technical difficulties, the time-lapse photos were often only
taken as snapshots.
Weather maps were retrieved from the website
"http:.//weather.unisys. com/archive/sfc~map/"'. Composite surface maps contain the
following analyses: radar summary, surface data plot, frontal locations and pressure
contours. Surface data is reported hourly from places like airports and automated
observing platforms. These data are updated hourly at around 30 minutes past the hour.
Frontal data are only available every 3 hours, so fronts may not exactly match the
Beach Population Data
Beach attendance was estimated from entrance ramps along Volusia County
beaches. These ramps are controlled by Republic Parking, and records are kept of the
number of cars entering the beach each day. This gives a rough estimate of the number
of people on the beach when rip rescues occur but not necessarily the number of people
in the water. However, it is a reasonable conclusion that on days when there are many
cars at the beach, there will be more people in the water compared to days with few cars.
During the summer of 2003 rip current deaths were recorded in four different
counties including Bay, Escambia, Santa Rosa, and Walton along the Florida Panhandle
(see study area in Figure 4.3).
A basic analysis of wave data, tide conditions and meteorological conditions was
conducted for this time period. It was the goal of this study to make a preliminary
investigation into rip current forcing parameters in the Panhandle and to evaluate if they
were comparable to those found in Volusia County.
Beaches in the Panhandle study area are sandy with very common cuspate features
on the beach face and surf zone undulations. Sonu (1972) noted that the surf zone is
relatively shallow with 1 meter or less depth over the inner bar. Outside the inner bar, the
bottom drops steeply to about 5 meters, and then rises to about 4 meters at the outer bar,
approximately 200 meters offshore. The offshore topography is smooth. Semidiurnal
tides in this area have an approximate maximum range of 0.7 meters. The average deep
water wave height during the days studied in 2003 was approximately 1.2 meters at the
Rip Current Death Data
The beaches along Florida' s Panhandle are not monitored as densely as those of
Volusia County. For this reason, and due to a lack of record keeping, there is not much
data available regarding rip current rescues. Rip current deaths, however, were well
documented by periodicals during the summer of 2003. Through this source, the day and
area where drownings occurred are known. However, the approximate time that the
victim was caught in the rip current is not known.
Wave height, period, and wave directional data were taken from NOAA National
Data Buoy Center buoys: #40239 for Bay and Walton Counties, and #42040 for
Figure 4.3: Study area and location of NDBC buoys #42040, located 64 nm south of
Dauphin Island, AL at water depth 237m, and #42039, located 115 nautical
miles east-southeast of Pensacola, FL at water depth 284m.
Wave heights and directions were then shoaled and refracted, using linear wave
theory, to 10m water depth. Waves with periods of 7 through 9 seconds would travel
from buoy #42039 to the coast in approximately 4 through 5.5 hours and from buoy
#42040 to the coast in approximately 2.5 through 3 hours. Dropouts occur in the
shoaled/refracted data where wave directions at the 10m mark were calculated to be
coming from a direction greater than 90 degrees (0 degrees corresponds to south) or less
than -90 degrees. These wave directions would correspond to waves not coming to
shore. Tidal data was retrieved from the same tide predictor as noted above for Volusia
County data, and the same web-based weather map provider used for Volusia County was
used for the Panhandle.
Rip Current Rescue Statistics
It is the purpose of this study to predict the occurrence of rip currents most
dangerous to beach goers and not rip current rescues or drownings. Like lightning
injuries due to a thunderstorm, rip current rescues are influenced by both meteorological
and human factors. Weather forecasters do not predict how many people will be struck
by lightning during a certain storm because it is unpredictable how many people will go
out in bad weather, or will go to a place where a lightning strike is likely to occur. It is
the same case with rip currents. There are far too many anthropogenic factors involved,
such as whether people will be in the water, or whether they will go near rip channels
when rip currents are occurring. Like severe weather prediction, it is more practical to
warn people of dangerous conditions, then depend on them to heed the warning.
Due to factors such as cold temperatures or very large waves (both of which tend to
keep people out of the water), rip currents may be occurring on many days, but there will
likely be no rescues to "mark" them. This is a shortcoming with the use of rescue data in
place of actual rip current measurements. However, Lushine (1991) Lascody (1998) and
Engle (2003) used rescue statistics and found that the benefits of rescues as a long term,
dense record of rip current events outweighed the drawbacks, and thus considered them
as a valid source of data for creating their predictive scales.
The logistics of placing and maintaining instruments in the field to measure rip
currents are complex and can present many obstacles. This is due to the constantly
shifting nature of beaches (like those in Volusia Co. and the Panhandle) with a sandy bed
as well as the migratory nature of rip channels. Instances of instrument burial were fairly
common when in situ measurements were attempted. Some structures, such as rock
jetties, have permanent rip channels nearby where measurements would be far easier, but
these are not the naturally occurring sandbar rip channels that lead to most rescues and
An analysis of rescues and wave data is limited by the length of the data set. This
study's analysis from Volusia County constituted a period of only two weeks, which
severely limited the number of rescues that were analyzed. This had a large effect on the
accuracy of the Modified ECFL LURC S due to the fact that the number of rescue days
was small. Therefore, if one is missed, it has a greater negative effect on the scale's
accuracy than if there were a longer time series with a larger amount of predicted rescue
Beach population data bolsters the reliability of rescue data in marking days of high
rip current occurrence. Total daily rescues can be normalized by daily beach population.
Hence, days with many rescues and few people on the beach would represent a higher rip
threat than days with many rescues and many people on the beach. Normalization of rip
rescues per hour was accomplished by dividing by the total number of cars that entered
the beach that day (beach ramp data was only available on a daily, not hourly, basis).
This number was multiplied by 10,000 for ease of use. This can be visualized by
comparing May 10th and 12th in Figure 5.1. The 10th had fewer rescues than the 12th.
However, since population was also low on the 10th, the relative risk, reflected by
normalized rescues, is actually higher than that of the 12th. As an extreme case, consider
the days of May 7th and 20th, which both had one rescue. Beach attendance on the 7th WaS
relatively normal, so the normalized rescues are slightly higher than the raw rescues.
However, on the 20th, there was a very low population which gives a great deal of weight
to the one rescue and results in a spike in the normalized rescues. However, with only
one rescue to use as a marker, the extrapolation to a dramatically greater risk is somewhat
tenuous. At any rate, the comparison demonstrates the value of beach population data.
Raw and Normalized Rescues vs. Beach Population
7 8 9 10 11 12 13 14 15 16 17 18 19 20
Figure 5.1: Daily beach population compared to daily rescues and daily normalized
rescues for May 7th through 20th, 2002. Normalized rescues are daily rescues
divided by the daily population multiplied by 10,000.
By observation, the times of greatest attendance at the beach are the hours between
10 a.m. and 5 p.m. In order to evaluate parameters during the time of highest attendance,
and highest number of rescues, only rescues between these hours were considered. The
time of the study (May 7th through May 20th, 2002) is also a time of high beach
attendance and includes a major beach holiday. Because of this, the study was not
adversely affected by a seasonal low beach attendance.
Wave and Tide Statistics
Four wave parameters including deep water wave height, wave period, deep water
wave direction, and directional spreading along with tidal stage are compared to
normalized rip current rescues in this section. Each parameter is individually related to
normalized rescues using a double bar histogram. In each histogram, light colored bars
represent parameter frequencies for the entire period (May 7th through 20th, from 10:00am
to 5:00pm each day). The number of observations in each bin range was normalized by
the total number of observations. The dark colored bars represent the frequency of
normalized rip current rescues within specified parameter ranges. For instance, the deep
water wave height histogram (Figure 5.2(A)) shows how often wave heights of 0.5m, Im,
1.5m, etc. occurred during the entire period, compared to how often rip rescues occurred
when waves were at those heights. Therefore, if wave heights of 0.9m had a small
frequency, while rip rescues at that height had a high frequency, there would be a strong
positive correlation between that wave height and rip current occurrence. The relative
magnitude of the normalized rip rescue probability compared to the overall probability
within each bin of the double histogram is effectively a measure of the risk.
The parameters and rip rescues were recorded on an hourly basis, therefore the
values of tidal stage, wave height, period, direction, and spreading are fairly precise for
each rescue. If four rescues occurred between 12 p.m. and 1 p.m. during a day, then the
parameters for that hour were recorded four times.
The sum of the squared difference (SSD) between the light and dark bars is given
on each plot. A higher SSD means a high correlation (positive or negative) between rip
occurrence and that parameter. For example, there might be a high positive correlation
between 0.9m wave height and rips and a high negative correlation between 1.6m wave
height and rips. These combined would give the same SSD as having a high positive
correlation for both. A higher SSD simply says there is a correlation, not whether the
correlation is positive or negative.
Wave height and direction were shoaled and refracted to deep water (consistent
with the approach of Engle (2003)) for use in the Modified ECFL LURCS scale.
Figure 5.2(A) shows that 82% of rescues occur at wave heights between 0.7m and
1.3m. The highest correlation is at 0.9m with 34% of rescues happening while wave
heights only reach this height 14% of the time. These values are higher than those found
by Engle (2003). He found that 63% of all rescues occurred with wave heights between
0.45 and 0.85 meters. A possible outlier in this study occurs at 2.4m. This high wave
height only occurred 3% of the time, but rescues at this height occurred 12% of the time.
It must be remembered that only two weeks of data were used for this study which
included a relatively small amount of rescues (37 rescues). The event in question
occurred on May 20th. There was one rescue on that day but a low beach population of
754 cars. Since the rescues were normalized by population, a great amount of weight
was given to this one rescue.
Wave periods between 8 and 9 seconds have the strongest positive correlation with
rescues. This agrees well with Engle (2002), Lushine (1991) and Lascody (1998) who
found strong correlations between rip current rescues/drownings and longer period
2 ~Deep Water H113 (meters) Overall Probability
I I I Normalized Rip Rescue Probability
O 0.5 1 1 .5 2 2.5 3
2 4 6 8 10 12 14
2~0.5 Deep Water Dominant Direction degreese)
c 0 degrees = shore-normal. Counter-clockwise is positive direction.
-80 -60 -40 -20 0 20 40 60 80
D.1 irectional Spreading (degrees)
ssd =0.033 h n,
25 30 35 40 45 50 55
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5.2: Frequency distributions of A) deep water significant wave height, B) wave
period, C) deep water dominant direction D) directional spreading, and E)
tidal stage. May 7th through 20th, 2002, 10:00am through 5:00pm. Rip rescue
data is normalized by beach population.
Rescues had the best correlation with deep water wave directions of -20 degrees to
-39 degrees (Figure 5.2(C)). Engle found the best correlation between 20 to -35 degrees.
During the summer, most waves come from the southeast (negative angles). It must be
remembered that data for this study was from a relatively short time period, compared to
Engle's, and few days had waves from positive angles. The high correlation that Engle
found between directly onshore (0 degrees) waves and high rip rescue occurrence was not
apparent in this data set.
Directional spreading showed a weaker correlation than in Engle's study. This
study found directional spreading of less than 39 degrees occurred 75% of the time and
accounted for 91% of rescues. Engle (2003) found that directional spreading less than 35
degrees accounted for 75% of rescues while waves occurred in that range just 37% of the
time. However, in this study, the SSD was the lowest of all parameters.
Tidal stage (Figure 5.2(E)) had the strongest positive correlation with rip rescues in
this study. This is reflected in an SSD of 0. 16, the highest of all parameters. The greatest
positive correlation is with tides in the -0.5 to -0.4 meter range. This range accounts for
68% of rescues while occurring only 25% of the time. Engle found the range from
-0.75m to -0.45m accounted for 62% of rescues while occurring 42% of the time.
During this study, the tide never dropped below -0.5m (between 10:00am and 5:00pm) as
it did in Engle's study. It is thought that more rescues occur at lower tide levels because
there is stronger breaking on the bar which increases set up. Lower water levels also
force water flow through the more hydrodynamically efficient rip channels instead of
over the shallow bar.
Comparison plots of the entire two-week period were useful in further evaluation of
parameters with a high SSD. These plots include parameter values from all hours during
the two-week period but only the rescues (not normalized) that occurred between
10:00am and 5:00pm.
Figure 5.3 demonstrates that the maj ority of rescues occur during decreasing wave
height. This relation was also noted by Engle (2003), Lushine (1991), and Lascody
(1998). The clusters of rescues occurring during wave heights under 1.25m also lead to
the conclusion that lower wave heights can pose more of a threat to beach goers; not
because there are no rips with higher waves, but because bathers do not perceive the
danger when waves are smaller.
Deep Water Wave Height (m) and Rescues
-Deep Water Wave Height (m)
o 1 rescues
+ 2 rescues
3~ -a 3 rescues
CI 4 rescues
05/08 05/09 05/10 05/11 05/12 05/13 05/14 05/15 05/16 05/17 05/18 05/19 05/20
Figure 5.3: Entire record of deep water wave heights for Volusia County, FL and
10:00am through 5:00pm rescues. May 7 May 20, 2003.
Figure 5.4 demonstrates that most rescues were grouped in the first week when
wave directions were more consistent. During the second week, wave direction
fluctuates widely, possibly due to local wind waves moving at different directions from a
longer period swell. Only 5 rescues occur during this confused sea state. This supports
the conclusion that consistent, normally directed wave angles are more conducive to rip
Deep Water Wave Direction (d g.) and Rescues
Deep Water Wave Direction (deg.)
o 1 rescues
60 + 2 rescues
a 3 rescues
a 4 rescues
05/08 05/09 05/10 05/11 05/12 05/13 05/14 05/15 05/16 05/17 05/18 05/19 05/20
Figure 5.4: Entire record of deep water wave directions for Volusia County, FL and
10:00am through 5:00pm rescues. May 7 May 20, 2003.
The relation between lower tide levels and rescues is very evident in Figure 5.5.
During the first week, there is at least one rescue at every other low tide (this would be
the daylight low tide when the beach was populated). Only three rescues occurred near
high tide during the entire period. Also of note is the amount of rescues occurring on a
falling tide. Six out of seven rescues that occurred in the mid-tide range (between -0.2
and 0.2 meters) of the first week occurred during a falling tide.
Engle (2003) referred to anecdotal evidence from beach patrol staff suggesting that
rip current related rescues may occur more frequently during outgoing tide. However, he
found no correlation between ebbing tide and rescues. Sonu (1972) noted that rip current
intensity increased with a falling tide due to increased breaker activity on the bar.
Tidal stage (m) and Rescues
05/08 05/09 05/10 05/11 05/12 05/13 05/14 05/15 05/16 05/17 05/18 05/19 05/20
Figure 5.5: Entire record of tidal stage for Volusia County, FL and 10:00am through
5:00pm rescues. May 7 May 20, 2003.
Rip Current Drowning Statistics
The same anthropogenic factors that arise when relating rip current rescues to rip
current occurrence also arise with rip current drowning statistics. A further complication
is that drownings (thankfully) are much less commonplace than rescues. Also, only the
day (and not the time) of drowning is known. So, conditions at the exact time a victim
was caught in a rip are unknown. Since rescue data is unavailable from the Panhandle
counties, only those few days with records of rip current drownings were evaluated in
this study. A summary of the days follows.
Table 5.1: Rip current related deaths in Florida Panhandle counties during the summer of
Date (2003) Deaths Site County
18-Mar 1 Panama City Bay
27-Mar 1 Pensacola Escambia
7-Ap 1 Pensacola Escambia
9-May 1 Perdido Ke Escambia
10-May 1 Navarre Beach IEscambia
11-May 1 Pensacola Escambia
8-Jun 8 various Walton
9-Jun 1 Pensacola Escambia
2-Jul 2 various Bay
13-Jul 1 Panama City Bay
31-Au 4 Pensacola Escambia
Wave and Tide Statistics
It is possible that there are certain signatures of rip current outbreaks in Florida.
These signatures include relatively long period swells, small directional wave angles to
shore, low directional spreading, and declining wave energy. These are characteristics
noted by Engle (2003) and others, including this study, for the east coast of Florida. In an
attempt to find similarities with these signatures in the Panhandle, wave and tide statistics
were analyzed on days with rip related drownings. As can be seen in Table 5.1, June 8th,
July 2nd, and August 31st were the days with the greatest number of documented rip
current drownings. These three days will be the focus of this section.
Figure 5.6 shows wave conditions in Walton County on the biggest rip current
drowning day of the summer. Hour zero is 12:00am on June 7th. During the probable
hours the drownings took place (from around 10:00am through 6:00pm), there is a
noticeable decline in wave height while the wave period stays fairly long. These
conditions could fool beachgoers into thinking that conditions are safe since the waves
have calmed down. However, the energy within the waves is still reasonably high.
Lascody (1998) refers to a sequence of events where moderate/strong onshore winds
generate choppy surf and strong rip currents, but people stay out of the water due to the
waves. Then, as winds subside and conditions improve, people go into the surf, and long
period swells result in the formation of rip currents and large numbers of rescues.
Lushine (1991) also states that rip currents generated by decreasing swells can be
particularly dangerous since local winds may be light, and bathers may be deceived into
thinking surf conditions are safe.
Also notable in Figure 5.6 is the crossing of wave directions through zero, showing
that wave directions were more shore normal before the rip current outbreak. This
change in direction may be the result of a passing frontal system. The wave direction
before and during the probable hours of drowning were fairly shore normal at 20 degrees.
Tidal stage has been found to be an important factor for rip current threat in
Volusia County, as well as in southeast Florida. Sonu (1972) also found a correlation
between higher rip current velocities and a falling tide at Seagrove, in the Florida
Panhandle. However, on June 8th, tide data from Panama City Beach, Florida indicates
that the tidal difference was very small during the probable hours of drowning. A low
tide at 9:26am was 0.26m and a high tide at 2:06pm was 0.29m. The next tide was at
11:09pm and was 0. 12m. These tides are relatively average for this area. The average
sea level for the months of June, July and August, 2003 was 0.25m. The lowest and
highest tides of June took place a week later and were -0. 16m and 0.62m, respectively.
The small tidal difference during this rip current outbreak leads to the notion that tidal
stage may not be as important at this site as it is in Volusia County. This may be
attributed to the existence of beach cusps prevalent at Panhandle sites which act as
3n~ h ,~
0 0 20 30 40 50 60 70 80
Fiue5.6 WaecndtosatWlo Cutesiae fro NDCbo409
high pero (for the Panhatindl) and wavenCoy dietiont crsin hroug zero before therip
currenigt drownings occurred. The tialdffrne was 20.56mos ateybew 10:00am o00m atd
5:0p.Op a fr "roal or donn scin O donn
catalysts for rip channel formation. A lower tide may be necessary in the relatively 2-D
bathymetry of Volusia County in order for rip channels to form and function.
Estimated W2ave Conditioni for Wa~lton County: June 7 June 9, 2003
0 10 20 30 4 01 50 60 70 80
0 10 20 30 4C1 50 60 70 80
0 10 20 30 40 50 60 70 80
Figure 5.7: Wave conditions at Bay County, estimated from NDBC buoy #42039. Two
drownings occurred on July 2, 2003.
Conditions for the period August 30th through September 1st are shown in Figure
5.8. There is a relatively constant wave height combined with wave direction crossing
through zero before the drownings. Wave period is high before the drownings and then
relatively constant at around 6 seconds. Tidal stage at 10:00am was 0.296m and was
0.303m at 5:00pm. Again there was a very small tidal difference during the hours of
probable drowning and no tides lower than average.
Es~timnated Wrve Conditions for Bay County: July 1 July 3, 2003
Prob.ble hours of drowrt
Estimated Wave Conditions for Pensacola: August 30 Septermber 1, 2003
0 10 20 30 4 50 60 70
O 10 il U) 4~3 : our ~
Figur 5.8:Wv odtosa enaoaetmtdfo NB uy#24.Fu
this sudy 5.were to further evlat eEngcl e's (2003) Ending inD Vousi ount and0 to
compare these Eindings with rip current outbreaks in the Panhandle. After statistical
research on Volusia County sites, Engle opted to remove wind direction and wind speed
from the Modified ECFL LURCS scale. In keeping with Engle's research, neither of
these parameters were used during this study to evaluate rip currents in either Volusia
County or the Panhandle. However, Engle's Modified ECFL LURCS scale was created
with data specifically from Volusia County. If a rip current predictive index is created
for sites in the Panhandle, wind parameters should not necessarily be rej ected on the basis
of Engle' s or this study.
Wind speed and direction are parameters in both the LURCS scale and the ECFL
LURCS scale. The LURCS scale was created with data specifically from southeastern
Florida, and the ECFL LURCS scale was created with data from east-central Florida.
The ECFL LURCS scale was then modified by Engle (2003) based on directional wave
data from Volusia County. This brings out the point that parameters will have varying
importance depending on the site. Therefore one scale can not necessarily be used for an
entire region, east-central Florida for example, but may need to be modified to fit
conditions for a smaller area.
Wind, Pressure, and Frontal Systems
Since waves and wind have been found to have a great effect on rip currents, it is
evident that the large-scale weather patterns which create such forces may be a key in the
prediction of rip occurrence. Lascody (1998) pointed out that although rip currents are an
oceanographic phenomena, meteorological factors influence their development. He
found that some rip current outbreaks followed passing frontal systems and made note of
the effect of pressure differences affecting wind. He stated that the analysis of
wind/wave reports showed that specific synoptic weather patterns were usually
identifiable on days with large numbers of rip current rescues or drownings. Lushine
(1991) also refers to long-term weather patterns being associated with a 10-year cycle in
rip current drownings.
Cold fronts often move over the Florida peninsula from the northwest, pressure
differences affect wind flow, and low pressure systems create strong surf. A combination
of these weather patterns seems to be strongly related to rip rescues and drownings. Cold
fronts, for example, can be characterized by 1) sharp temperature changes over a
relatively short distance, 2) shifts in wind direction, 3) pressure changes, 4) clouds and
precipitation patterns, and 5) changes in the air's moisture content (Ahrens 1994).
Shifts in wind direction may be of particular importance to rip current occurrence.
Often, winds are onshore as a front approaches, building swell with higher period waves.
Then, as the front passes, the wind rotates offshore and the weather improves. Offshore
wind, smoothing the sea surface, combined with decreasing swell height and clearing
skies make surf conditions more attractive. Swimmers may enter the water although
wave energy is still high and rip currents are present.
Volusia County 1996
Volusia County beaches are affected by northwest cold fronts as well as low
pressure systems in the Atlantic. In the open ocean, areas of low pressure create large,
high period swell that impacts the coast. Often, lows move from west to east over the
Florida peninsula. As the low moves offshore and the fetch increases, wave heights can
often decrease as period increases, creating a signature of rip current outbreaks.
Precipitation and clouds are also associated with the low which keep bathers off the
beach. Once the low has passed however, weather improves and bathers enter the water
although dangerous rips still exist.
A weather pattern such as this may have influenced a rip current outbreak during
the dates researched by Engle (2003). On Friday, June 28th, 1996 there were 23 rip
current rescues, around 5 rescues on Saturday, 32 rescues on Sunday, and 30 rescues on
Monday, July 1st. Conditions for the high rescue days are summarized below.
Table 6.1: Conditions for high rescue (>15) days during a rip current outbreak in Volusia
County during. June/July 1996.
Date Day # of Rip Wave Ht. Wave Per. Wave Dir. Spreading Time of Level of
Rescues (m) (s) dg dg Low Tide Low Tide (m)
6/28 Fri 23 0.83 6.9 20 37 11am -0.6
6/30 Sun 32 0.86 6.3 -5 29 1m -0.8
7/1 Mon 30 0.64 8.5 -3 31 2pm -1.0
A low pressure system was in the Atlantic while another low crossed the state from
the Gulf of Mexico on the 27th and 28th. Figure 6. 1 shows weather surface maps at two
times (7:00am and 7:00pm) per day in sequential order.
-- -- -- l
I G4 5
A. 67 E;5
61 ~r 62Y
6 28/96 -7:00pm
F. 61 7
Figure 6.1: Surface weather maps for A) June 27 ', 1996 at 7:00am EST through J) July
1"t, 1996 at 7:00pm EST. The pressure given is the last three digits of the
actual pressure (987=998.7mb, 024=1002.4mb).
Figure 6.1. Continued
387 3; 7
6 30 96 -7:00am
L.. E 8112 :,10
7 i9 i
On June 27th, two low pressure areas can be seen around the Florida peninsula. The
offshore low would have created storm swell which impacted Volusia County.
The front became stationary during the morning of the 29th, and there were
relatively few rescues although it was a Saturday. Possibly, this is due to afternoon rain
showers or the north wind creating a longshore current. Wave height increased by the
30th. Wind and wave directions crossed through zero and began arriving just south of
By Monday the period had increased while wave heights dropped. Skies were
clear. Morning and afternoon winds were offshore, flowing from high to low pressure
areas across the state. These conditions resulted in 30 rescues.
Engle (2003) referred to five other rip current outbreaks in Volusia Co. during
1996. The wave conditions for these days are listed in Table 6.2.
Table 6.2: Wave conditions for five rip current outbreaks in Volusia County, 1996.
Date Day # of Rip Wave Ht. Wve Per. Wave Dir. Spedingl Time of Level of
Rescues (m) (s) (deg) (deg) Low Tide Low Tide
6/5/96 Wed 31 0.56 8.8 -9.00 32 5pm -0.7
6/1 9196 Wed 30 0.62 7.9 -8.00 30 5p -0.6
7/1 3/96 Sat 45 0.74 8.2 -23.00 n.a. 1pm -0.6
7/1 6/96 Tue 22 0.74 5.7 -32.00 n.a. 3pm -0.7
8/1 0/96 Sat 23 0.48 8.7 -18.00 n.a. 12p3m -0.6
Of these days, only June 19th and August 10th Seem to be strongly associated with
local weather patterns. As can be seen in Figure 6.2, on the 19th a high pressure area was
positioned in the Gulf of Mexico. Easterly winds on the Atlantic coast of Florida during
the night of the 18th became calm by morning. When rescues occurred on the afternoon
of the 19th, winds were likely from the southeast and pressure had increased (Figure 6.2
115 73 74 4
.. .11 7 70"
7 140 73
6/18/96- 7:00pm 6/19/96 -7:00am
Figure 6.2: Surface weather maps for A) June 18, 1996 at 7:00pm EST through C) June
19, 1996 at 7:00pm EST.
The outbreak on August 10th also had an associated low as can be seen in Figure
6.3. Offshore low pressure generating east winds on the 9th probably caused the long
period swell seen on August 10th. The low moves across Florida during the early
morning of the 10th. Morning winds were calm.
Weather conditions for the June 5th outbreak (Figure 6.4) were composed of high
pressure in the Gulf of Mexico and offshore winds in the morning. Winds were onshore
by early evening, two hours after low tide (5:00pm).
The July 13t'" outbreak had clear skies, and a high pressure area over Florida
causing a south wind to flow toward lower pressure (Figure 6.5).
7E 158 *
6/5 96 -7:00am 6/5./96 -7:00pm
Figure 6.4: Surface weather map for June 5, 1996 at A) 7:00am EST and B) 7:00pm EST.
72 167 *
84 143 d. 10z
D. 6o 7 r
8/10/96 -7:00am 8/10/9 -7:00pm
Figure 6.3: Surface weather maps for A) August 9, 1996 at 7:00am EST through D)
August 10, 1996, 7:00pm EST.
7/w%/9 -7:ooam 7/13/96 -7:00pm
Figure 6.5: Surface weather maps for July 13, 1996 at A) 7:00am EST and B) 7:00pm
On the morning of the 16th (Figure 6.6) rain was present, winds were calm, and a
stationary front was present in the northern southeast U.S. High pressure over north
Florida caused a southeasterly wind by early evening.
90 115 83 134 149
LA 86 15,8
I I i~
b198 82~ 212
8 1 91i
7/16/96 -7:00am 7/16/96 -7:00pm
Figure 6.6: Surface weather maps for July 16, 1996 at A) 7:00am EST and B) 7:00pm
Volusia County 2002
During the 2002 study, the period with the most rescues was from May 9th through
May 13th, 2002. Conditions for these days are in Table 6.3.
Table 6.3: DIWASP's estimated conditions (from Sontek wave gage) from 10:00am
through 5:00pm for May 7th May 20th, 2002 on Volusia County beaches.
Wave height and direction are deep water values, shoaled and refracted from
the Sontek wave gage. The tidal elevation given is the lowest tide during
daylight hours (not necessarily between 10:00am and 5:0p)
Day I# of Rip Beach Wave Ht. IWave Dir. Wave Dspr Time of Low Tide
Rescues IPopulationl (m) Deg. IPeriod (s)l Deg. ILow Tide Level
07 Tues. 1 2836 1.01 -14.50 9.88 34.96 11:39 -0.40
08Wed. 1 2938 0.91 -16.66 8.82 35.41 12:17 -0.43
09Thurs. 2 2996 0.83 -26.73 9.15 33.79 12:53 -0.47
10 Fri. 5 3842 0.76 -26.67 10.15 33.17 13:28 -0.5
11 Sat. 12 6980 0.93 -37.50 8.31 36.89 14:02 -0.51
12 Sun. 9 8197 1.12 -17.65 5.69 38.04 14:36 -0.53
13 Mon. 2 2576 0.81 -29.09 7.97 36.54 15:11 -0.53
14 Tues. 1 1488 1.00 39.00 5.73 40.87 15:49 -0.53
15 Wed. 1 2625 1.58 35.16 6.42 39.39 16:30 -0.52
16 Thurs.l O 2602 1.45 -20.00 5.69 41.07 17:16 -0.50
17 Fri. O 3500 1.05 -35.97 7.32 36.61 18:09 -0.48
18 Sat. 2 6216 0.74 -39.34 7.05 34.49 19:10 -0.46
19 Sun. O 654 1.89 63.00 7.67 37.16 8:12 -0.46
20Mon. 1 754 2.48 -13.00 10.22 37.96 9:11 -0.51
Weather maps from the 9th through the 13th (Figure 6.7) show a cold front
approaching Florida from the north, which stalls and never pushes through the state.
2002,Y 7:0p ET
Figure 6.7. Continued
5/13/02 -7:00am 5/13/02 -7:00pm
Figure 6.7. Continued
Winds feed into the front throughout the period. They switch from calm, or from
the south, during the mornings to a more onshore direction by early evening. These
winds created a southeasterly wave field as can be seen in Table 6.3. Skies were partly
cloudy with calm morning winds turning onshore by the early evenings of the highest
rescues days (the 11th and 12th.
Figure 6.8 shows time-lapse photos of the 11Ith and 12th. On both days, some rip
channel formation can be seen along the outer bar, especially to the north (top) of the
photo. Channels seem to be more defined on the 11Ith, which had more rescues and a
lower population (see Table 6.3).
Figure 6.8: Time-lapse photographs of Ormond Beach, Volusla Co. on May 11 I', 2002 at
A) 10:31am and B) 11:01am and May 12th at C) 11:33am and D) 1:33pm.
Figure 6.8. Continued
Two rescues on the 18th of May seem to have occurred during weather conditions
similar to those of the 9th through the 13th
5/18/02 s~~~~e -70a 5/80 -7:00pmst~Bl~~ g
Figue 69: Srfae wethe map fo A) ay 8, 202 t 7:0amEST hrogh B Ma
18,~ 2002, 7:00pm EST.
A7 1 4 notw senfotst boeFoia(iue69, but dosnt oesotw r
thoghtestt.Thscussasotestrywidad 3 ege av iecin
calme6.. Poulfatio warelatively hih nrmtesapshorA a 1,20 ts7a in T Figre g 6.10 aferoo
skiesg were claear ands se a condtions apeared mid.ad-9dge aedrcin
Figure 6. 10: Snapshots of Ormond Beach, Volusia Co. on May 18th, 2002 at A) 12:01pm
and B) 2:01pm. One rescue occurred around each of these times.
Every other day with rescues during the 2002 study is associated with a cold front
moving across the state. Rescues on May 7th and 8th occur in the wake of a cold front
moving through during the early morning of the 7th. JUSt the tail of the front can be seen
in the lower right of Figure 6. 11(A).
After the front, winds on the morning of the 7th were calm. Wind speeds at the
offshore buoy were calm while onshore winds picked up during the afternoons of both
days. These days had wave directions under -20 degrees (Table 6.3).
5//0 -70a 5/7/0 -7:00pm
8,,~~: 2002, 7:0mET
5/8/0 -7:0a 5// 7:0p
Rescues onMy1t,15h n 0h a eascitdwt hepsaeo
notwetcodfrns asin l h wa thrug Foida iue61 hw ete
70 79 70~ 7J44
5/14/02 -7:00am 5/14/2 -7:00pm
Figure 6.12 Surace weather1" maps for A)May 14, 2002 ate 7:00a tES through D)Ma
on eah 15, 2002 7:00pm EST.23p o h 4had n rud300mo h 5
5/15/02 -7:00am 5/15/2 -7:00pm
Figure 6.12. Continued
This northeast flow continues and wave heights build to 1.58 meters on the 15th
Time-lapse photographs from around the times of the rescues can be seen in Figure 6. 13.
Figure 6. 13: Time-lapse photos of Ormond Beach, Volusia Co. on A) May 14th, 2002 at
2:31pm and B) May 15th at 3:09pm. One rescue occurred around each of
The time-lapse photos show breaking on the bar, but not any well defined rip
channels through the bar. However, some areas show signs of channel development. It
must be remembered that the rescues happened in one spot along the entire coastline of
Volusia County and not necessarily exactly at this site in Ormond Beach where this
camera is located.
A fast-moving cold front swept through Florida on the 20th of May, 2002 (Figure
6.14). Beach population was very low and one rip rescue occurred.
5/19/02 -7:00pm~ 5/0/2 -7:00a
Figure 6.4 Surac wete asfrA My1,20 t70pmETadB a0
By the morin ofte2twnsa h fsoebo ee1-2kosfo h
havue a.4 calm or weakhe locl in feld) asocated9 with athm Onshore winds are better
like to coldfronts that pushthroughs a the stat (Ma 14t -o 15th) 8p2 curnots app tea
lnkedto lowhc pressure systemoswhic ros ath Vlstat fromnt west oes. Daereaingho swell
ha er h ight and inceain period reut as thesytems moh ve farther into the Atlantic.
Low prelsisur eand e frontalsytems alosemt hv get nleneo rip current ubek Vls Cnt
occurroence in the Panhanle. Tis reae yEgion of03) Florida e is not imatdbyln erio
ground swell like the east coast due to the relatie small4t siz and) shalo drepth of tea
Gulf of Mexico. Instead, onshore wave direction caused by onshore winds may be the
most significant signature of rip currents in this area.
Every day evaluated in this study, where rip current outbreaks occurred in the
Panhandle, had some form of low pressure system associated with it. Most of these
systems were close in proximity to the Panhandle. However, on July 13th (1 drowning)
and August 31st (4 drownings), rain and wind over the Panhandle seem to be associated
with distant lows across the upper southeastern states.
The outbreak in Escambia County, from May 9th through 11Ith (3 deaths), seems to
be associated with a strong low in the mid-eastern U.S. which created a strong northerly
wind flow across the Gulf, leading to onshore wind (and waves) along the Panhandle.
Figure 6.15 shows estimated wave conditions for this period.
Estimated Wanve Conditions f~or Escambia County: May 8 May 12, 2003
0 20 40 60 80 100 120
Figure 6.15: Wave conditions estimated from NDBC buoy #42040. One drowning
occurred on each day from May 9th May 11t 2003.
10:00am through 6:00pm each day.
5 9 03 -7:00am 74
67 81 757
5/10/03 -7:00am 7la
ve .T 72
5/11/o -7:oam 0"
rr~i~ 1 P
I B .--
:-;- 8~0~1 86~20 111
o.~ ~-!: 'B 8~ II
i 7Pi I11 i: 14
`~~/9/03 :p:qpp, I ? 1-~ ~R~S~PSLI
Wave data were shoaled and refracted, using linear wave theory, from deep water
offshore buoys to 10m water depth. Probable hours of drowning are from around
.~ ~ II'r .I
16% 1 82
Figure 6.16: Surface weather maps for A) May9th, 2003 at 7:00am EST through F) May
11Ith, 2003 at 7:00pm EST.
On May 8th, wind direction was from the south at 8-12 knots. Figure 6.16 shows
weather conditions during the following three days when one drowning occurred on each
day. Wind velocity at an offshore buoy reaches 13-17 knots at 7:00am on May 9th
(Figure 6. 16(A)). The onshore wind flow continues feeding into the approaching low
pressure system at around 8-12 knots during the rest of the outbreak. This continuous
onshore wind created a shore normal wave field as can be seen in Figure 6.15. Wave
directions reach zero before each drowning and cross zero before the drowning on May
9th. The drownings begin after a decline in fairly high, long period waves. Each
drowning probably occurred during decreasing wave height.
A strong northerly wind flow during June 7th fed into an approaching cold front.
Eight drownings occurred the next day in Walton County, and one drowning occurred on
the 9th in Pensacola. Figure 6.17 shows estimated wave conditions during this outbreak.
3Estun~ated Wave Conditsons for Walton County: June! 7 June 9, 2003
0~~ 10~ 20~ 30 40 50 0 7 8
0n 10- 20 0 4 50 60 0 6
0 10 2 0 4 0 6 0 6
Fiur 6.7 aecniiosetmtdfomNB uy#239 ihrwig
ocurdon Jun 8t ndoe cure o un
r -76 155
I r qr154
7 03 7*00
r OUY' 0i~ 6
6//0 -7:0a 8
6/8 03 -7-00am 7
82 142. +4
6 0 :00am 76 7
The onshore wind flow continued on the 8th and 9th as can be seen in Figure 6.18.
I 3-:0p ;
j ,Ey I,~ 84
72 II 7 5
Figure 6.18: Surface weather maps for A) June 7th at 7:00am EST through F) June 9th,
2003 at 7:00pm EST.
Two cold fronts approach the Panhandle on the 7th. South winds at the offshore
buoy are 18-22 knots and decrease to 8-12 knots on the 8th. Local winds are lighter.
Offshore buoy winds clock to the west by the morning of the 9th and are calm by early
Two rip current drownings on July 211d in Bay County may have been associated
with the passage of a cold front that skims across the upper Gulf of Mexico as it moves in
an easterly-northeasterly direction. South winds feed into the passing system as seen in
7 ~~ ~ ~ a oa -?oom /20 -:0pm
Figure 6.19: Sufc wahrmasfrJuy2: a )700mETan )700mET
Figure~ 6.2 shw o siae aeaglscostruhzr sthlo
pressure~~~ ~ 6 sytmmvsars h ahnl.Wnddrcin ee unofhr ic
the~ ~ ~ ~ ~~8*":1 low sty ot fFoid.Tecntn nsoewn wcusdteosoewv
field~~~~~~~~~~~~T seen inf Fiue62. aehiht erae stelw oeothadwn
Auus 3s hd or iprlaeddowinsin Pescoa Fiue62 hw
stationary ~ ~ ~ ~ ~ ~ ~ ~~~~~~~' frn reeti tesutesen ..duigtistm. o rssr n
thgrough zero uring te wearlyr map orng ofl the 31t, before resue occrrd.B :0m
Figure 6.21: Surface weather maps for the period A) August 30th, 7:00am through D)
August 31s~t, 7:00pm. Eastern standard time.
Estimated Wave Conditions for Bay County: July 1 July 3, 2003
Prob ble hours s drowrng
) 10 20 30) 40 50 60 70 80
) 10 20 30 401 sO 60 70 80
jj OlIY~ i rl i
0 10 20 30) 40
50 60 70 80)
Figure 6.20: Wave conditions estimated from NDBC buoy #42039. Two drownings
occurred on July 2nd
x- u~i~ ~I~PP~;X:
,1 I I~
ir, I;, ~I
ih I;B /7
s/sotos -7:oopm d ~
I 74?~~~~ 18
a- -I b
j 8/3/0 -7:0a a /10 70p
Figur 6.1 otne
Esimtd ae odiinsfo enaol;Auut 0- Setebe 1, 00
0, 10 2t 30 40 C 6 7
0 0T 040S 07
Figure 6.22: WavineodiinesiaefrmN B buy#24.F rdowng
occurred onv Augutiost 31st scaa ~i~~ 0- atmer,2
As Lascody (1998) noted, spec~ific weather: paten semt eascitdwt i
curn ubek nFoia hs atrsaeuulyascae ihlwpesr
sytm n eae id.Th atr o ubek n oui onyivle h
passage of low pressure systems and associated fronts as well as the presence of cold
fronts to the north. Low pressure systems that move into the Gulf create long period
swell which decreases in height as the low moves farther from shore. This signature is
associated with rip current outbreaks. An onshore flow of wind is associated with cold
fronts that push through the state. When fronts stay north of the state weak wind patterns
do not appear well correlated with rip outbreaks.
Rip currents in the Panhandle are associated with cold fronts that often stay north
of the state. The presence of a cold front to the north creates strong onshore winds and
wave directions. Afternoon winds may lighten, but directions remain onshore instead of
becoming calm or turning offshore as they do in Volusia. Onshore winds causing choppy
ocean conditions may not be as great a deterrent for entering the water to Panhandle
beachgoers as they are to bathers in the Atlantic. As stated earlier, the relative small size
and shallow depth of the Gulf of Mexico keeps wave heights and periods lower than in
the Atlantic. Swimmers may be more prone to go in the water during onshore wind
waves in the Panhandle where clear water and waves breaking close to shore make
conditions appear more benign than similar conditions on the Atlantic Coast.
MODIFIED ECFL LURCS SCALE
Further analysis of the Modified ECFL LURCS (Engle, 2003) scale was carried out
in this study using data from Volusia County during the two-week period in May, 2002.
The analysis was run as closely as possible to Engle's (2003) experiment in Volusia
County. Wave heights and directions taken at the Sontek wave gauge were shoaled and
refracted out to deep water. No modifications were made to the parameter ranges used by
the Modified ECFL LURCS. Engle selected these ranges based on rescue statistics for
the period of April 1996 through September 1996. This study, by using a new data set, is
a blind test of the Modified ECFL LURCS scale.
Statistics used by the National Weather Service are employed in order to evaluate
the performance of the scale. The POD (Probability of Detection) represents the
accuracy of the scale. It equals the sum of rescues during a day that was forecast to have
rip currents, normalized by the total number of rescues on all days. The FAR (False
Alarm Ratio) is a measure of over-warning and equals the percentage of days that rips
were predicted but had no rescues. As in Engle's (2003) study, an Alarm Ratio (AR) was
also computed. This is the percentage of days that the scale predicted rip currents.
A representative value for each parameter was computed for each day. An average
daily value for deep water wave height, wave period and directional spreading was used.
The median value for wave direction and the minimum tide level were used. An inherent
problem with using median and average values is that if patterns vary significantly
throughout the day, the median or average may not be relevant to conditions at the time
rescues occurred. However, Engle (2003) found that analyzing data for periods shorter
than a day (hourly, for example) was difficult due to rescue data being too noisy on that
The Modified ECFL LURCS assigns an index value greater than zero to specific
ranges of each parameter. Figure 7. 1 is an example computation of the Modified ECFL
M~odifietd$ ECFL LURCS Checklist
Example computations; appear in bold.
Wave Periodl Wave Direction
Period T [s Factor Direcron. aa deg Facror
T < 6 0 0 < -35 or 0 m 20 0
6 <= T < 9 0.5 -35
9<4=T< 11 1 -30
11 <= T <12 2 -25 <= 0 < -15 or 10 >= B > 5 3
T > 1 2 3 -156 < 0 < 5 4
Wave Period Factor = 0.5 Wave Direction FactoEr = 4
Hight Hopt Factori
Ho <: 1 0 Tide
1<~= Ho < 2 0 5 Tide, h (m) Factor
2 <5= ~Ho < J 1 h > -0.2 0
Wave Height Factor = f idal Factor = 1
Dsp. E Fac[or
6 < 30 4
DsrFactor = 3
yrnle acors The Modlfied ECFL LURCS rip cu1rrent threa1 1.
Figure 7.1: Example computation of the Modified ECFL LURCS checklist.
Ranges with higher positive correlation to rip current rescues are given higher
index values. Index values are added for each parameter, and the sum is the rip current
threat. A rip current warning is issued if the threat is greater than five. A "very high
threat" warning would be issued for a value of nine or greater.
Figure 7.2 depicts the Modified ECFL results for the two-week period in May,
2002. The dark bars represent the rip current threat. Light bars represent the amount of
daily rescues. Ideally, if there are rescues on a particular day, the threat index value
should be above the threshold value of Hyve. This indicates that a rip current threat
warning would be issued.
Modified LURCS Index and Rescues
7 8 9 10 11 12 13 14 15 16 17 18 19 20
Figure 7.2: Modified LURCS rip current predictive scale results for May 7th-20th, 2002,
Volusia County, FL.
A POD of 0.595 was computed for the period. This means that over 50% of the
rescues were predicted. An AR of 0.643 relays that the scale predicted rip currents on 9
out of 14 days, and a FAR of0. 1 11 means the scale predicted rips on one day that had no
rescues. Both the AR and FAR are low which is important for the applicability of the
scale. A scale that falsely predicts rip currents on many days is of no use to beach rescue
staff. These low values are comparable to those from Engle's (2003) study. However,
the large rip current events (from May 27th through July 5th) were better predicted in his
study which resulted in a POD of 0.971. The relatively low POD of this study is the
result of the Modified ECFL LURCS not predicting the large rescue day on the 11th I
this day had been predicted, the POD would jump up to 0.919.
Table 7.1: Parameter values for hours 10:00am through 5:00pm used by the Modified
ECFL LURCS. Wave height and direction are deep water values shoaled and
refracted from the Sontek wave gage for May 7th-20th, 2002, Volusia County,
Day #of Rip Beach Wave Ht. Wave Dir. Wave Dspr Low Tide Threat Warning
Rescues Pop (m D. Period (s)D. Level (m Index Value Issued?
07 Tues. 1 2836 3.33 -14.50 9.88 35 -0.40 11 yes
08Wed. 1 2938 3.00 -16.66 8.82 35.4 -0.43 6.5 ye
09Thurs. 2 2996 2.71 -26.73 9.15 33.8 -0.46 8 ye
10 Fri. 5 3842 2.51 -26.67 10.15 33.2 -0.49 8 ye
11 Sat. 12 6980 3.05 -37.50 8.31 36.9 -0.50 3.5 no
12 Sun. 9 8197 3.67 -17.65 5.69 38 -0.53 7 ye
13 Mon. 2 2576 2.64 -29.09 7.97 36.5 -0.52 5.5 ye
14 Tues. 1 1488 3.30 39.00 5.73 40.9 -0.52 4 no
15 Wed. 1 2625 5.20 35.16 6.42 39.4 -0.52 5.5 yes
16 Thurs. O 2602 4.74 -20.00 5.69 41.1 -0.46 6 ye
17 Fri. O 3500 3.43 -35.97 7.32 36.6 -0.31 3.5 no
18 Sat. 2 6216 2.43 -39.34 7.05 34.5 -0.06 4.5 no
19 Sun. O 654 6.22 63.00 7.67 37.2 -0.15 3.5 no
20Mon. 1 754 8.14 -13.00 10.22 38 -0.39 10 ye
The actual POD increases to 0.686 when rescues are normalized by population
(Figure 7.3). As stated earlier, dividing daily rescues by daily population reduces the
effect that population has on the number of rescues each day. Normalized rescues are a
better representation of rip current risk than un-normalized rescues. Days with low
population and a high amount of rescues may have more "risk" associated with them than
days with high population and a high number of rescues.
Modified LURCS Index and Normalized Rescues
16C I normalized rescues
7 8 9 10 11 12 13 14 15 16 17 18 19 20
Figure 7.3: Modified LURCS rip current predictive scale results, using normalized
rescues, for May 7th-20th, 2002, Volusia County, FL.
The reasons that the 11Ith WaS not predicted to have rescues are non-shore normal
waves (-37.5 degrees) and directional spreading of 36.9 degrees, which is just outside the
limit of 35 degrees. The day would have been predicted had the spreading been 2
degrees lower. Directional spreading for the hours of 10:00am through 5:00pm on the
11Ith WaS fairly constant, shifting from 35.8 to 37.8 degrees. Figure 5.4 depicts deep water
wave directions for the entire period. Wave direction fluctuates widely on the 11Ith with a
spike, due to one point, reaching -53 degrees.
x 10 D~iredianpl spectrum estimated using IMLM rnethod
x10' Directional spectrum edlimated using ILtM method
Figure 7.4: Spectra depicting A.) wave direction at 4m depth and B.) wave frequency for
May 11Ith, 2002 in Volusia County, FL.
Figure 7.4 shows that the fluctuation may have been caused by a confused sea state
where swell with a period of around 10 seconds was relatively shore normal, while higher
frequency waves (possibly locally generated wind swell) of around 4-5 seconds came
from the southeast.
From the Eigure, it is apparent that on the 11Ith the greatest amount of energy was
from nearly shore normal waves of around 10 seconds, which would undoubtedly aid in
the forcing of rip currents.
Large fluctuations in wave direction during the second week of the study period are
also indicative of a confused sea state. Figure 7.4 shows hourly spectra from the 14th and
16th. From the Eigure, it is apparent that energy was coming from different directions.
A. .S~, B.
Figure 7.4: Spectra at 4m water depth depicting A.) May 14th, 2002, 11:00am and B.)
May 16th, 2002, 12:00pm. Volusia County, FL.
Figure 7.4(A) shows some shore normal wave energy with a 10-second period and
greater energy from around 75 degrees with a 3-4 second period. May 16th at 12:00pm
saw strong energy from -50 to 50 degrees with periods of 4 through 9 seconds. These
conditions of fluctuating wave direction would not be conducive to rip current
development. This is supported by the lack of rescues during the second week even
though the population was similar to the first week (with the exception of the 19th and
Also of note is that the only two days were designated with a very high threat
(index value of 9 or greater), the 7th and the 20th. Only one rescue occurred on both of
these days. Both days had shore normal wave directions under -15 degrees. This
combined with directional spreading just under 35 degrees were the main reasons the
index produced such a high threat level on the 7th
The average daily beach population during the period was 3,443 cars. The 7th(
Tuesday) had 2,836 cars (see Table 6.3) while the 20th (a Monday) had 754. No days
with below average population had more than two rescues. It is possible that strong rip
currents were present on the 7th and 20th, but due to low population, only one rescue
occurred. Figure 7.5 shows mid-day conditions around the time of the rescue on the 7th,
which occurred just after low tide. Rip channel formation is evident on the bar and
supports the index's claim that there was a rip current threat on this day. Possibly the
"high threat" wave conditions on this day were the beginnings of the event that reached
its peak on the 11Ith
Figure 7.5: Time-lapse photo of Ormond Beach, Volusia Co. May 7 ', 2002 at 1:01pm.
One rescue occurred between 12:00pm and 1:00pm.
Due to a high wave height, long average period, and shore normal wave directions,
it is very likely that there were rip currents on the 20th. However, high precipitation kept
the population very low and only one rescue occurred.
There was not such a marked relation between directional spreading and rescues in
this study as there was in Engle's. He noted a peak in rescues whenever directional
spreading values dropped below 30 degrees. However, there was not as wide a
fluctuation in spreading during this study as in Engle's. Values during his study varied
between 60 degrees to around 25 degrees. Directional spreading during this study
generally fluctuated between 42 and 30 degrees.
Figure 7.6 shows the correlation between directional spreading and rescues. The
large rescue events occur at and below 38 degrees, but rescues are not dominantly
grouped at low points in the series as they were in Engle's study.
It should be emphasized that Engle created the Modified ECFL LURCS using data
collected during summer months when waves generally come from the southeast. This is
reflected in the weighting of the wave direction factor being biased toward negative
(southerly) wave angles. In Figure 7.1, it can be seen that an index value greater than 0 is
given to wave angles from -35 to 20 degrees. Since there are fewer bathers (and
therefore rescues) in the winter months when wave directions are more northerly, this
bias is valid as long as rescue data is being used in place of actual rip current data.
However, the scale would have to be altered for analysis on in situ rip current
measurements which may include data from any time during the year.
Overall, the Modified ECFL LURCS performed well on the short data set
available. Two out of the three high rescue days (the 10th, 1th, and 12th) WeTO issued a
rip current warning. Data from a longer period of time would obviously facilitate a better
analysis of the scale. With a longer time series, errors (such as the scale not predicting
one large rescue day) would not have such a great impact on the evaluation.
Directional Spreading (deg.) and Rescues
05/08 05/09 05/10 05/11 05/12 05/13 05/14 05/15 05/16 05/17 05/18 05/19 05/20
Figure 7.6: Entire record of directional spreading for Volusia County, FL and 10:00am
through 5:00pm rescues. May 7 May 20, 2003.
SUMMARY AND CONCLUSIONS
Rip current outbreaks may have certain "signature" parameters that can be used to
identify their occurrence. These signatures involve: wave height, wave period, wave
direction, directional spreading, and tidal stage. Patterns involving these parameters are
evident before outbreaks in Volusia County, Florida, and similar patterns are evident in
the Panhandle of Florida. These patterns include: decreasing wave energy, shore normal
wave direction, and low directional spreading. Tidal stage and surf-zone topography are
also important factors.
Results from this study were similar to those found by Engle (2003). Outbreaks in
both Volusia County and the Panhandle often occurred during times of decreasing wave
height and relatively high wave period. In Volusia County a correlation was apparent
between relatively onshore wave directions of -20 to -39 degrees. However, the high
correlation that Engle found between shore normal (0 degrees) waves and high rip rescue
occurrence was not found here. In the Panhandle wave direction often crossed through
zero and remained shore normal before the outbreaks.
In the relatively two-dimensional topography of Volusia County beaches, low tide
had a strong correlation with rescues. However, tidal stage may not play as important a
role along Panhandle beaches where a more three-dimensional topography appears to aid
in the forcing of rip currents. A strong correlation between directional spreading and
rescues was not found in this study. However, Engle's general conclusion was evident in
the results: that rip currents occur at lower values of directional spreading.
Large-scale weather patterns such as pressure systems and associated frontal
systems were usually found in the proximity of areas where rip current outbreaks
occurred. These systems seemed to have the greatest effect on wind direction which
would affect wave direction and, in the case of onshore winds, possibly augment (but not
drive) mass transport of breaking waves, thereby increasing rip current strength.
However, all the parameters mentioned that force rip currents could be affected by
weather conditions. Rip current occurrence may be better understood with further study
of such meteorological systems, and predictive indexes may be improved by including a
factor indicating their approach or presence.
A new data set was used to further analyze the Modified ECFL LURCS. The
ranges of parameters determined by Engle (2003) were effective in predicting rip currents
given the fact that the two-week period of study was relatively short. The scale predicted
most maj or rip current rescue days and predicted over half of all rescue days. Rip
currents were predicted only 64% of the total period, and only one false alarm was given.
These are important points for beach patrol staff; A predictive scale is of less use to
lifeguards if it greatly overpredicts rip current days and gives a high number of false
With further analysis demonstrating its reliability, the Modified ECFL LURCS will
become a practical tool for beach rescue staff. Such a tool would greatly aid in the
preparation for rip current events, which would reduce the number of rip-related rescues
and drownings. This would be accomplished by the index alerting beach rescue staff to
the presence of conditions favorable for rip current occurrence, thereby allowing them to
increase staff numbers, frequency of beach patrols, and other preventative measures. The
index would also allow government agencies to warn the public of unsafe conditions.
The ECFL LURCS is currently used by the National Weather Service (NWS) to forecast
rip currents along the east coast of Florida in order to issue warnings through the media.
Future studies may further analyze how tidal stage affects rip currents at different
sites. The correlation found between rescues and falling tide in this study may also be of
importance. Wind velocity and direction may be important site-dependent parameters
and should be given further consideration. As mentioned above, stronger correlations
between rip currents and meteorological events may be drawn from additional study.
The next step involving the use of the Modified ECFL LURCS should incorporate
in-situ directional wave data in order for real-time rip current threat predictions to be
made and compared to actual rescue data.
LIST OF REFERENCES
Ahrens, C. Donald, M~eteorology Today. An Introduction to Weather, Climate, and the
Environment, West Publishing Company, St. Paul, MN, 1994.
Charles, L., R. Malakar, R.G. Dean, Sediment Data for Florida 's East Coast,
UFL/COEL-94/014, Coastal Engineering Department, University of Florida, 1994.
Dean, Robert G., Dalrymple, Robert A., Coastal Processes nI ithr Engineering
Applications, Cambridge University Press, Cambridge, UK, 2002.
Engle, Jason A., Formulation of a Rip Current Forecasting Technique through Statistical
Analysis of Rip Current-Related Rescues, Master of Science Thesis, University of
Hauserman, Julie, Currents Deadly for Fla. Tourists, St. Petersburg Times, Sunday, July
Lascody, L.L., East Central Florida Rip Current Program, Natl. Wea. Dig., Vol.22,
No.2, pp.25-30, 1998.
Lushine, J.B., A Study of Rip Current Drownings and Related Weather Factors, Natl.
Wea. Dig., Vol. 16, pp.15-30, 1991.
McKenzie, P., Rip Current Systems, J. of Geology, 66(2), pp. 103-1 13, 1958.
National Data Buoy Center (NDBC), Station 42039 Historical Data, 2003,
NOAA/NDBC, Available [On-Line]
http ://www.ndbc.noaa.gov/station_page.phtml? station=4203 9, January 10, 2004.
National Data Buoy Center (NDBC), Station 42040 Historical Data, 2003,
NOAA/NDBC, Available [On-Line]
http ://www.ndbc.noaa.gov/station_page.phtml? station=42040, January 11i, 2004.
Shepard, F.P., K.O. Emery, and E.C. LaFond, Rip Currents: A Process of Geological
importance, J. Geol., 49(4), 337-369, pp.337-369, 1941.
Sonu, C.J., Field Observations of Nearshore Circulation and Meandering Currents, J.
Geophys. Res., 77, 3232-3247, pp.244-258, 1972.
Unisys Weather, Archived Surface Weather Maps, 1996, 2002, 2003, Unisys, Available
[On-Line] http://weather.uni sys.com/archive/, March 5, 2004.
Volusia County Council, Volusia County, Florida Beach Patrol, pamphlet, Orlando, FL,
WWW Tide and Current Predictor, Tidal Data, 1996, 2002, 2003, Available [On-Line]
http://tbone.biol. sc.edu/tide/index.html, March 12, 2004.
Matthew Schrader' s earliest beach memory involves the smell of Coppertone
suntan lotion as his mother smeared globs of it across his face while visiting his
grandparents' condo in Boca Raton, Florida. He was only three or four years old at the
time, with pale white skin (since his family had relocated from his 1975 birthplace of
Falls Church, Virginia, to Sayre, Pennsylvania). So, his mother was right to slather him
in an SPF overcoat. Once released from her grasp, he immediately tripped and, as he
rolled across the beach, applied a thorough layer of gritty sand to the freshly applied
By the age of 13, his family was living in Tampa, Florida, where Matthew began
skimboarding on west coast beaches and surfing whenever his parents would make the
two-hour drive to the east coast. After graduating from high school in 1994, he began
undergraduate studies at the University of South Florida in Tampa. He decided to study
civil engineering because he liked to draw, and he liked to build. He specialized in
environmental engineering, not yet understanding that this meant "wastewater
engineering." He also enjoyed writing and took extra classes in order to minor in creative
writing. At different times during his studies, he worked part time as a lifeguard,
swimming instructor, outdoor educator (in Colorado), and as an engineering intern for the
Southwest Florida Water Management District (SWiFtMuD).
It was while working for SWiFtMuD that he became interested in engineering
applications for habitat restoration and creation. After graduating from USF, he
immediately began postbaccalaureate studies at the University of Florida in Gainesville
with the environmental engineering and sciences program, studying ecological
engineering. Under the direction of Dr. Mark Brown, he came to the conclusion that he
wanted to apply ecological principles to the coastal zone. With the helpful advice of
others, including his parents, Dr. Brown, and Jason Engle, Matthew decided to begin his
Master of Science degree in the coastal and oceanographic engineering program at UF.
His future will involve many more beach memories and the pungent smell of Coppertone,
without the grit.