Resource Description: NODC Accession Number 0001753
Observations are reported as a height range. Observers ignore the smaller waves. As a simplified example, assume a given day has dominant wave energy in the 14-17 second wave period range with negligible energy outside this band. Assume five waves catch the eye of an observer every four minutes, or 100 waves every 80 minutes. This takes into account the time periods of varying length without waves arriving. The upper end of the reporting range is approximately equivalent to the H1/10, the average of the highest 10 waves, which if evenly distributed in time, would occur every 8 minutes. The lower end of the observing range is near the H1/3, or the average of the highest 1/3 waves, which if evenly distributed in time, would occur every 3 minutes. The highest wave over this nominal 80-minute period with 100 waves, or H1/100, would be equivalent to the observer's use of "occasional" heights in their reports.
A digital database of surf observations, referred to as the Goddard-Caldwell (GC) set, dates back to August 1968 for the north shore, and to March 1972 for the south shore of Oahu. It is described in more detail in Caldwell (2005). Data are recorded in HSF. The daily values in the GC database refer to the surfing location along the given coast with the highest reported breakers. For the north shore, most observations are taken at Sunset Point, which is usually one of the highest surf spots along the coast under the dominant northwest swell direction. For days of extreme surf with heights greater than roughly 15 HSF, visual observations are reported at Waimea Bay, where breakers are much closer to shore relative to most of the north shore, which is fringed with offshore reefs. The surf reports are typically made several times per day. The daily value in the GC set represents the upper end of the reported height range for the observing time with the highest breakers. This number aims to be equivalent to H1/10. Comparisons of the GC database to 1981-2002 data from NOAA buoy 51001, which is located roughly 400 km west-northwest of Oahu, show the surf observations are temporally consistent with the shoaling-only, buoy-estimated breaker heights and have an uncertainty of 10 to 15% of the surf height (Caldwell, 2005).
Photograph analysis =================== A breaker or surf is defined at the moment in time when some portion of the front face of a wave becomes vertical and unstable due to a decrease in water depth. The trough to crest surf height is defined as the vertical distance between the crest and the preceding trough at the moment and location along the wave front of highest cresting, which has been shown in models and observations to be at the time and location of breaking. For locations with high refraction, such as Sunset Point, where most of the visual observations are made, the breaker often forms an A-shape. The trough to crest height refers to the center of the A, i.e., the point along the wave front with the highest height.
Photographs were obtained from Internet sites or directly from photographers. Location and date was a prerequisite. Photographs showing the highest waves of a given day were chosen from the available pool of pictures. Pictures were sorted by size in HSF matching the date to the GC database. Typically, 15 images for each size category were selected (Table 1).
Each photograph requires a surfer or some other identifiable object to use as a benchmark in estimating wave height. Dashed lines were superimposed on each photograph to indicate the approximate trough and crest. An arrow was overlaid next to each benchmark to denote a 5' unit. The benchmark arrow was duplicated and subsequent arrows were stacked from trough to crest to gauge the wave size (Figure 3).
Photographs capture a two-dimensional image of a three-dimensional world and distortions of shapes and sizes are inherent. Shots taken from a high vantage point, such as a cliff or helicopter, make detection of the wave trough difficult. Wave size is distorted in pictures taken by a swimming photographer near the surfer. Priority in selecting shots was given to images taken by a photographer standing close to mean sea level either on shore or on a floating craft. Distortion of perception decreases as the distance between the camera and the surfer/wave increases.
There are various sources for errors in this exercise. The error associated with trough identification has been estimated at 10% of the wave height. The surfer's height is not known in most images. It is assumed that the average surfer height is 5'9" and a typical surfer stance is roughly 5' with a 6" uncertainty, which leads to an error of 10% in the surf height estimate. For both cases, the errors average out as the number of photographs increases. Since the photographs were selected from still images, it is not certain that any given picture represents the highest height reached by that wave during breaking, or if these few select waves represent H1/10, which is assumed in the GC database. With the small number of available pictures per day, the translation based on these pictures likely underestimates the heights in the GC database.
Each scaled photograph was examined to estimate the height to the nearest tenth of a foot. For each size category, a mean and standard deviation of the estimated trough to crest heights were computed (Table 1). Using two standard deviations as a proxy for uncertainty, the margin of error is within 10-20% of a given size for categories with at least 15 photographs, disregarding the lowest and highest percentage values. For days with surf heights of 15 HSF or less, most photographs are taken at spots from Log Cabins to Sunset Beach, which typically has the highest surf on the north shore (Figure 2a). For days with surf heights greater than or equal to 15 HSF, photographs were further sorted by location: Waimea Bay, Oahu outer reefs, and Jaws (Peahi), Maui. Under northwest swell with 17-20 second wave periods, the travel time from Oahu to Maui is roughly three hours, which makes comparisons of daily data appropriate. Fewer photographs were available for the Oahu outer reefs than for Waimea Bay. The paired HSF and trough to crest heights are plotted in a scatter diagram (Figure 4).
For surf heights of two to six HSF the translation shows the trough to crest heights are more than double the HSF observations (Table 1, last column). From six to twelve HSF, the translation is close to double. An inadequate supply of photographs were available of Sunset Point for heights in the 13-15 HSF range, when the offshore-most breaking point is roughly one km from shore and strong currents impede water photography. The available images suggest the translation of 15 HSF to trough to crest heights is slightly less than double. For the entire range from two to fifteen HSF, the translation can simply be defined as double within the margin of error.
For surf above 15 HSF, the wave energy at Sunset Point becomes overwhelming and the resultant breakers occur unpredictably over a wide area both parallel and perpendicular to shore. This makes surfing dangerous due to difficulties in maintaining a safe wave-entry point. Under such extreme conditions, surfers historically challenged Waimea Bay, where the take-off zone is narrower and the proximity to shore allows landmark referencing for more precise wave-entry positioning.
Numerous photographs are available for Waimea Bay during surf in the 12-30 HSF range. The surfers enter the wave about 50-100 m outside the point on the northeast side of the bay and surf at an angle toward the safety of the deep waters in the center of the bay. The wave-entry point shifts northwest of the northeast point of the bay with increasing wave size. At approximately 30 HSF, the entire wave front cascades nearly simultaneously across the breadth of the bay, ending a surfer's chance for a safe ride.
The photographs at Waimea Bay suggest the trough to crest heights are roughly 1.5 times HSF during days with observations in the 15-30 HSF range. Within the collection of photographs, there are several occasions when images were available for the same day from both Waimea Bay and outer reefs of Oahu and Maui, where tow-in surfing (motor-powered-watercraft-assisted breaker entry) has gained popularity over the past decade. Over the submerged ridges of the offshore reefs to either side of Waimea Bay, the SWAN output (Figure 2b) shows increased heights due to convergence of wave rays, i.e. refraction. Photographs of tow-in surfers on outer reefs validate the larger heights relative to Waimea Bay.
In summary, the translation of HSF to trough to crest heights is a factor of two within the 10-20% margin of error for the full range of breaker sizes encountered in Hawaii. This assumes the height is defined as the highest height reached in the vertical from the trough to crest at any point along the wave front during breaking and zones of high refraction (outer reefs) are included for extreme days when Waimea Bay was the reporting location. The HSF, or simply dividing trough to crest height by two, has been adopted by other big wave enthusiasts around the globe as seen in pictures and dialogue from extreme surf contests in California, Peru, and South Africa. It is important for scientists and the general public to understand this relationship for utilizing surf observations reported in HSF.
Acknowledgments. Larry Goddard is recognized for establishing the surf observation database from 1968-87. Great appreciation is given to the north shore surf observers, surfers, and photographers. Special thanks are given to photographers Claudia Ferrari (claudiaferrari.com), Jamie Ballenger, (HawaiianWaterShots.com), and Ian Masterson. Warm regards go to photographers of Oahu outer reefs (Justin Langlais, John Bilderback, and Hankfotos.com) and Jaws, Maui (Eric Aeder, Steve Kornreich, Les Walker, C. Levy, and C. Oreve).
The scaled photographs are sorted by size in Subdirectories, which have names as xx_Hsf or xx_xx_Hsf, where xx denotes a size and xx_xx denotes a size range.
Within each subdirectory, filenames of each photograph vary yet all contain the date (mmddyy) and the recorded size in the Goddard-Caldwell dataset as _Cxx, where xx is the size in Hsf.