Airborne lidar sensing of massive stony coral colonies on patch reefs in the northern Florida reef...

ent 104 (2006) 31–42www.elsevier.com/locate/rse

Remote Sensing of Environm

Airborne lidar sensing of massive stony coral colonieson patch reefs in the northern Florida reef tract

John C. Brock a,⁎, C. Wayne Wright b, Ilsa B. Kuffner a, Raquel Hernandez c, Philip Thompson a

a U.S. Geological Survey, 600 4th Street South, St. Petersburg, FL 33701, United Statesb NASA Goddard Space Flight Center, Wallops Island Flight Facility, Wallops Island, VA 23337, United States

c National Coral Reef Institute, Nova Southeastern University, 8000 North Ocean Drive, Dania Beach, FL 33004, United States

Received 2 February 2006; received in revised form 20 April 2006; accepted 22 April 2006

Abstract

In this study we examined the ability of the NASA Experimental Advanced Airborne Research Lidar (EAARL) to discriminate cluster zonesof massive stony coral colonies on northern Florida reef tract (NFRT) patch reefs based on their topographic complexity (rugosity). Spatiallydense EAARL laser submarine topographic soundings acquired in August 2002 were used to create a 1-m resolution digital rugosity map foradjacent NFRT study areas characterized by patch reefs (Region A) and diverse substratums (Region B). In both regions, sites with lidar-sensedrugosities above 1.2 were imaged by an along-track underwater videography system that incorporated the acquisition of instantaneous GPSpositions. Subsequent manual interpretation of videotape segments was performed to identify substratum types that caused elevated lidar-sensedrugosity. Our study determined that massive coral colony formation, modified by subsequent physical and biological processes that breakdownpatch reef framework, was the primary source of topographic complexity sensed by the EAARL in the NFRT. Sites recognized by lidar scanningto be topographically complex preferentially occurred around the margins of patch reefs, constituted a minor fraction of the reef system, andusually reflected the presence of massive coral colonies in cluster zones, or their derivatives created by mortality, bioerosion, and physicalbreakdown.© 2006 Elsevier Inc. All rights reserved.

Keywords: Coral reefs; Remote sensing; Lidar; Videography

1. Introduction

The Florida Keys extends over 360 km from just south ofMiami to the Dry Tortugas, and protects the broad offshorecarbonate platform from inshore bay waters that fluctuate intemperature, salinity, and turbidity (Ginsburg & Shinn, 1964).Variation in the geographic continuity of these Pleistocenelimestone islands has controlled Holocene reef development inthe Florida reef tract (FRT) by creating three regions of differinglong-term water quality, the Upper, Middle, and Lower Keys.Shelfal patch reefs and bank-barrier reefs are most abundant inthe Upper Keys, and secondarily in the Lower Keys, seaward ofcontinuous island complexes (Ginsburg & Shinn, 1994). Incontrast, limited Holocene reef development off the Middle

⁎ Corresponding author.E-mail address: [email protected] (J.C. Brock).

0034-4257/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.rse.2006.04.017

Keys, cut by broad tidal channels, is attributed to the influx oflagoonal waters from Florida Bay, and the mobility of bottomsands (Ginsburg et al., 2001; Ginsburg & Shinn, 1994; Halley etal., 1997). The northern Florida reef tract (NFRT), off the UpperKeys where reef development is most extensive, is separatedfrom Barnes Sound, Card Sound, and Biscayne Bay by KeyLargo and Elliot Key, and is frequently bathed by the FloridaCurrent. Tidal and wind-driven exchange with Biscayne Baythrough the Safety Valve, a wide pass just north of Elliott Keyand the Ragged Keys, sets the northern limit of the FRT(Ginsburg et al., 2001; Marszalek et al., 1977; Shinn, 1988).

Nearly 4000 of the roughly 6000 patch reefs found in theFRT are concentrated off Elliot Key in 1 to 12 m water depthamid bare sand and seagrass meadows on the 5 to 10 km widecarbonate platform between Hawk Channel and the shelf-margin bank-barrier reefs (Marszalek et al., 1977; Shinn et al.,1989). These patch reefs typically have vertical relief andbreadth ranging from 5 m to 100 m, respectively, and based on

mailto:[email protected]

http://dx.doi.org/10.1016/j.rse.2006.04.017

32 J.C. Brock et al. / Remote Sensing of Environment 104 (2006) 31–42

core-sampling, were built by large massive stony corals overroughly the last 6000 years (Shinn et al., 1989).

Although FRT coral cover has declined in general over thelast several decades (Jaap et al., 2001), coral cover on NFRTpatch reefs has consistently been found to be among the highestobserved in the FRT (Burns, 1985; Chiappone, 1996;Chiappone & Sullivan, 1997; Jaap, 1984; Miller et al., 2002).Recent surveys of these reefs found mean stony and meanoctocoral cover in excess of 8% and 35%, respectively, haverecorded the greatest coral species richness presently found inthe FRT, and have documented particularly high gorgonianspecies diversity (Miller et al., 2002; Moulding & Patterson,2002).

The maturation–senescence cycle (Jones, 1977) and thephysical disturbance by storms (Lirman & Fong, 1995, 1997) ofthese patch reefs create an associated internal zonation(Ginsburg et al., 2001). Well developed composite NFRTpatch reefs are typically divided into a “cluster zone” composedof one or more giant massive coral colonies, a “campus zone”with massive corals scattered among abundant octocorals, andin some cases, a zone dominated by coral rubble and skeletalsands (Ginsburg et al., 2001).

1.1. Objectives

Topographic complexity, or rugosity, is an element of habitatcomplexity, and is a fundamental ecological factor on coral reefsthat is relevant to species diversity and richness, herbivoreshelter, predation, recruitment, metabolic processes, hydrody-namics, and nutrient fluxes (McCormick, 1994; Sale, 1991;

Fig. 1. Map of the Florida reef tract dep

Sebens, 1991; Szmant, 1997). NFRT patch reef internalheterogeneity and zonation are likely to be reflected by thespatial pattern of topographic complexity, because the clusterand campus zones represent two different size populations andspatial densities of predominately 9 species of massive stonycorals (Ginsburg et al., 2001).

The NASA Experimental Advanced Airborne ResearchLidar (EAARL) is a temporal waveform-resolving, blue-greenwavelength lidar (light detection and ranging) designed tomeasure the sub-meter scale topography of shallow reefsubstratums (Wright & Brock, 2002). Recently, we demon-strated that this airborne laser altimeter is capable of detectingoverall rugosity differences between NFRT bank-barrier reefsand patch reefs (Brock et al., 2004). The objectives of this paperare to:

(1) evaluate the capability of the NASA EAARL lidar todiscriminate cluster zones of stony coral colonies onNFRT patch reefs, and

(2) identify other NFRT substratums that also display hightopographic complexity as sensed by the NASA EAARLairborne blue-green laser altimeter.

2. Materials and methods

The study area is located in the NFRT within BiscayneNational Park, on White Bank, the inner ridge of the platformseaward of Hawk Channel, and landward of Ajax Reef, anelongated bank-barrier shoal mainly composed of coral rubbleand skeletal sand (Halley & McPherson, 1997) (Fig. 1). The

icting the location of the study area.

Fig. 2. Conceptual depiction of the operation of the spatial lidar rugosityalgorithm. The area of the DEM surface kernel (shown in pink on the uppersurface) is divided at each grid point by the area of an imaginary perfectlyhorizontal and flat kernel (shown in pink on the lower surface).

33J.C. Brock et al. / Remote Sensing of Environment 104 (2006) 31–42

broad shelf that hosts the study area is covered by Thalassiatestudinum seagrass meadows and bare calcareous sand regions,and is studded by numerous patch reefs, transitional reefs, andhardground communities (Jaap, 1984; Jones, 1963). Althoughquite numerous, patch reefs on White Bank appear to surviveunder marginal environmental conditions, and typically occur in2 to 9 m base water depth as roughly circular domes. Thesepatch reefs have internal topographic relief up to 3 m, and rangefrom several meters to over 700 m in diameter (Jaap, 1984). Theupper surfaces of these reefs may be close to emergent at lowtide, and support various stony corals, notably Montastraeaannularis, M. faveolata, M. cavernosa, Porites porites, P.astreoides, Siderastrea radians, S. siderea, Diploria clivosa,and Acropora cervicornis. Acropora palmata is lacking onthese patch reefs, but octocorals are abundant, many octocoralskeletons are covered with the hydrozoan Millepora alcicornis,and M. complanata is common (Jaap, 1984; Jones, 1963).

2.1. Lidar survey and analysis

The NASA Experimental Advanced Airborne ResearchLidar (EAARL), a temporal waveform-resolving, airbornegreen wavelength lidar, was designed to sense the submarinetopography and morphologic habitat complexity of shallow reefsubstratums (Brock et al., 2004). During a two-week timeperiod in August 2002, EAARL missions collected densesubmarine topographic measurements over a broad swath of thenorthern Florida reef tract extending from north of TriumphReef to south of Carysfort Reef. Following the use of theNASA-USGS Airborne Lidar Processing System (ALPS) tocreate a regional substratum spot elevation data set based on thelaser soundings, a random point gridding operation was appliedto generate a digital elevation model (DEM) at 1-m cellresolution for the surveyed portion of the NFRT. Next, thislidar-derived DEM was analyzed using commonly availablealgorithms to create a 1-m resolution digital substratum slopemap for the study region.

A 1-m resolution digital substratum rugosity map wascreated for the study region, using an approach loosely basedupon Dahl's surface index, defined as the ratio of the actualsurface area to that of a flat, horizontal two-dimensional planewith similar boundaries (Dahl, 1973). An image processingalgorithm that convolves an imaginary perfectly flat kernel withthe study region lidar DEM was created to implement thisanalysis throughout the study region. This spatial lidar rugosityalgorithm first determined the surface area of a 4 cell kernelaround each lidar DEM point, and then divided that value by thekernel surface area of a hypothetical perfectly flat, horizontal2 m square kernel surface, 4 m2 (Fig. 2). The calculation thatyielded the lidar rugosity values that were assigned to each lidarDEM point was:

Ak

Ai

where Ak and Ai are the area of the DEM kernel about the pointand the area of an imaginary flat surface around the point,respectively. By definition, rugosity as determined from a lidar

DEM must always be greater than or equal to 1. For a perfectlyflat, smooth surface, the value of this ratio is 1. Higher values ofthis ratio correspond to greater departure from smooth flatness.

2.2. Substratum observations

Substratum type was determined by the interpretation ofobservations acquired by an along-track underwater videogra-phy system mounted on a 26 ft long vessel (Fig. 3). This systemcombined the following components: (1) a GPS receiver, (2) anavigation computer, (3) a pole-mounted underwater videocamera, and (4) a deck-mounted video cassette recorder andmonitor. The GPS unit is a CSI MiniMax high accuracy receiverthat received corrections from the Wide Area AugmentationSystem (WAAS) (Ormand, 1999). During operation and whilereceiving the satellite-broadcast WAAS-correction, the CSIMiniMax provided real-time positioning at finer than 3 maccuracy. This instantaneous WAAS GPS position wasprovided simultaneously to a navigation computer and a videoannotator through a two-legged cable.

Fig. 3. The along-track underwater videography system used to acquire substratum observations. System components include: 1) a GPS receiver, 2) a navigationcomputer, 3) a pole-mounted underwater video camera, and 4) a deck-mounted video cassette recorder and monitor.


The navigation computer was a Dolch NotePAC ruggedizedmachine running Windows 2000 and the Hypack Version 2.12navigation software. The Hypack software allowed transectlines to be pre-planned, and also provided guidance for the pilotin following transect lines during image acquisition. Duringfield operations, both the predetermined survey lines and theactual vessel track were overlaid on a background navigationalchart and a lidar DEM image of the area under study.

The video camera was a Model UWC-325/p fixed focuscolor camera that used a 1/3ʺ CCD sensor and outputs astandard NTSC video signal at a resolution of 480 lines. Thecamera was mounted to the base of a vertical pole attached tothe transom of the vessel that supported a GPS antenna on theopposite upper end. A topside Outland Technology UWS-7010Portable Color Video System composed of a JVC VHSvideotape recorder and a Sharp flat screen monitor was usedto control the video camera, and captured GPS-annotated videoimagery of the substratum. The monitor displayed what thevideo camera was viewing at all times, and the recorder wasused as desired during transects. The annotator was an OutlandTechnology Model 5000A RS-232/Video Overlay box thatinterrupted the video camera signal and allowed annotation textto be superimposed on the GPS-derived location as a text stringon the video image. The annotator box automatically extractedthe GPS-provided latitude, longitude, and Universal Coordi-nated Time (UTC), and posted this information on the image,providing continuous video image geolocation that was used in

subsequent analysis. During operation, the 3 m vertical pole thatcarried the underwater video camera and GPS antenna wasraised and lowered to track the bottom terrain at a nominalcamera-to-substratum distance of about 2 m. The vertical polesetting for each reef visited was determined by inspecting theassociated lidar topographic map to estimate the approximateaverage depth of the upper surface of that particular reef. Thepole was then raised or lowered just prior to the survey of eachreef, in order to generally achieve the desired 2 m camera-to-bottom range on that individual reef. Thereby, video images thathave roughly uniform spatial scale were acquired at a nominalforward speed of about 0.5 m/s.

In September 2003, this along-track underwater videographysystem was used to acquire video imagery of varioussubstratums within 2 regions on the NFRT platform (Fig. 4).Region A covers an area known to contain numerous patch reefsscattered across a seagrass meadow, based on prior mapping bythe authors (Brock et al., 2006). Region B is larger andimmediately to the north of Region A, and covers diversesubstratums, including patch and transitional reefs, sand,rubble, and hardgrounds. The benthic observations in RegionA were acquired to evaluate the capability of the EAARL todiscriminate cluster zones of stony coral colonies on NFRTpatch reefs. The substratum videography in the heterogeneousRegion B was undertaken to identify other NFRT substratumsthat also display high topographic complexity as sensed by theEAARL.

Fig. 4. Study area map (QuickBird image) depicting Region A, a seagrass meadow studded with patch reefs, and the more heterogeneous and larger Region B.


2.3. Analysis of high rugosity substratums

Following the vessel-based along-track videography survey,the digital lidar rugosity, elevation and slope maps weresampled under the GPS-determined vessel tracks in Regions Aand B (Fig. 5). In order to select for interpretation only theportions of the track that correspond to elevated topographiccomplexity, a threshold filter set at 1.2 were applied to thetrackline lidar rugosity values. This operation identified thesegments along the vessel track where lidar rugosity exceeded1.2, and subsequent analyses dealt with only these segments(Fig. 6). The maximum lidar rugosity, and the mean elevationand slope were calculated for each high lidar rugosity (HLR)track segment, and the start and end UTC time for each segmentwas used to identify the corresponding segment of thevideotape.

The videotape segments for the sites exhibiting elevatedlidar-sensed rugosity found along the vessel track were

manually interpreted, and thereby assigned to one of 7substratum classes. Three of these classes, live stony coralcolony (CC), dead stony coral colony (DCC), and deadovergrown stony coral colony (DOCC) are subdivisions of thepatch reef cluster zone (Ginsburg et al., 2001) (Fig. 7). The CCclass is composed of living colonies of massive stony corals,mostly Montastraea species, typically exhibiting partialmortality, and diseased in some cases. The DCC and DOCCclasses were recognized as degraded versions of the CC class.The DCC class was visually recognized as massive stony coralwith greater than 50% dead surfaces, and with overgrowthlimited to encrusting algae. The DOCC class was alsorecognized as massive coral heads with greater than 50%dead surfaces, but in this case overgrowth included gorgoniansand sponges in addition to encrusting algae.

The HB2 (Fig. 7C) class is morphologically irregular mostlybare pavement that we inferred to be derived from thecementation of the large coral rubble fragments created by


physical damage, recognized in uncemented form as class R2.The HB1 (Fig. 7D) class is hardbottom with over 25%octocoral, sponge, and macroalgae cover, and is equivalent tothe patch reef campus zone described by Ginsburg et al. (2001).The remaining two HLR classes, R1 and R2, are both composedof rubble skeletal remains of branching or massive stony corals.Class R1 is characterized by fragments generally less than15 cm in diameter, and class R2 is composed of rubblefragments generally larger than 15 cm that we inferred to bederived from the physical breakdown of massive coral colonies.

3. Results and discussion

Field observations of high rugosity sites were divided intotwo regions, with the aims of assessing the origins oftopographic complexity on just patch reefs (Region A), and tosearch broadly for similar lidar rugosity signatures associated

Fig. 5. QuickBird image maps depicting A) the patch reef Region A and B) the heterogthe high lidar rugosity videography interpretation sites are shown in red. The blue box

with transitional reefs and carbonate shoals (Region B).However, although roughly 10 times larger in area and muchmore diverse in habitat type, the investigation of Region Byielded results that closely resembled those obtained for RegionA, and accordingly, the two regions were combined and treatedtogether in our analyses.

45% of the high rugosity sites included in the analysis wereidentified as belonging to the massive coral colony cluster zoneclasses, with 11%, 17%, and 17% assigned to CC, DCC, andDOCC, respectively (Table 1). Hardbottom classes totaled 30%of sites, split unevenly between HB2 (27%) and HB1 (3%), andrubble accounted for 6% of these sites. 15% of the sites wereclassified as unknown (UKN), and were not included insubsequent analyses. Classes with an n<4 (SGR and R1) wereexcluded from the analyses as well.

Of the three massive coral colony cluster substratum classes,live colonies (CC) had the highest mean lidar rugosity (1.42),

eneous substratum Region B. The videography vessel track is plotted in green, andshown in b shows the location of the QuickBird image map depicted in Fig. 6B.

Fig. 5 (continued).


followed by dead coral (DCC) at 1.31 and dead overgrown coral(DOCC) at 1.28 (Table 1, Fig. 8A). Hardbottom (H2) had amean rugosity very similar to dead coral (DCC) (Table 1). AKruskal–Wallis one-way nonparametric analysis of variance(data did not meet assumptions of parametric ANOVA) forrugosity found statistical differences among classes (K–Wstatistic=12.7, p=0.026 using chi-squared approximation). Asubsequent all-pairwise comparisons test found that CC had asignificantly higher mean rugosity compared to HB1.

Live colony (CC) sites were found to be the most shallow(highest mean substratum class elevation relative to theNAVD88 vertical datum) at −4.5 m, with progressively lowervalues for DCC (−5.3 m), HB1 (−5.6 m), DOCC (−5.9 m),HB2 (−6.3 m), R2 (−6.7 m), and R1 (−7.9 m) (Table 1, Fig.

8B). A parametric one-way analysis of variance found statisticaldifferences among substratum classes (F=9.87, p<0.0001),and Tukey Honestly Significant Difference (HSD) all-pairwisecomparisons test found that the CC class was shallower thanDOCC, HB2 and R2. Slope was not significantly differentbetween substratum classes (parametric ANOVA F=1.45,p=0.21) (Table 1, Fig. 8C).

In both Region A and Region B, high rugosity sitescomprised only a minor fraction of the total area, and nearlyalways occurred around the margins of patch reefs (Fig. 9).Elevated substratum rugosity in both regions reflected clustersof massive coral colonies, or their derivatives created bymortality, physical degradation, and subsequent colonization.Our results indicating that 45% of all high rugosity sites

Fig. 6. A) Lidar substratum elevation (top) and rugosity (bottom) along the videography survey track over a single patch reef, plotted against vessel track time (s). B)Map view of the videography vessel track, with portion plotted in a shown in blue, and annotated with start and end track time (s). Location of this QuickBird imagemap is shown in blue in B.


Fig. 7. Underwater photographs of the visually recognized high lidar rugosity substratum types. The depicted classes are A) live stony coral colony (CC), B) deadcolonized stony coral colony (DOCC), C) irregular hardbottom (HB1), and D) campus zone hardbottom (HB1). The foreground width of the view shown on theoblique orientation photographs (A, B, and D) is roughly 2 m, and the horizontal width of the nadir orientation image shown in C is about 1 m. Not all class variants arepresented in the interest of brevity.


examined were recognizable as living, dead, or dead overgrownmassive coral colonies are an important finding because ithighlights the importance of these massive coral species as reefframework builders on the FRT. In addition, three more HLR

Table 1Substratum classes in study area of Biscayne National Park with respectivesample sizes, means, and standard deviations for rugosity, elevation, and slope

Class N Rugosity Elevation (m) Slope (degrees)

Mean S.D. Mean S.D. Mean S.D.

CC 14 1.4209 0.1871 −4.4661 0.8520 35.620 13.638DCC 22 1.3099 0.1036 −5.2660 1.1262 27.651 11.271DOCC 22 1.2833 0.0723 −5.9360 0.8725 26.592 9.2753HB1 4 1.2229 0.0144 −5.6275 1.2464 26.245 8.1380HB2 34 1.3061 0.1904 −6.2626 0.8136 28.001 11.458R1 2 1.2521 0.0301 −7.9403 0.6167 14.690 2.1513R2 6 1.3021 0.0538 −6.7044 0.9314 24.814 12.021SGR 3 2.1080 0.5588 −8.4033 0.9666 57.593 9.2477UKN 19 1.3945 0.2467 −6.4274 1.5146 33.151 20.559

substratum types, coarse rubble (R2), coarse rubble colonizedby octocorals and macroalgae (HB2), and bare medium rubbleon uneven pavement (R1) were interpreted to be the degradationproducts of cluster zone massive colonies. We inferred thatthese three HLR substratums are created by physical damage tocolonies caused by storms (Lirman & Fong, 1995, 1997) andother factors, followed by bioerosion and colonization. Aeriallidar mapping of these substratums about the margins of patchreefs provides evidence for the theory of Shinn (1988) thatNFRT patch reefs grow laterally by shedding rubble, andthereby creating expanded rocky bottom available for subse-quent colonization by hermatypic corals and other reeforganisms.

If the substratum classes HB2, R1 and R2 are considered asmassive cluster-zone derivatives, and are combined with parentcluster zone live or dead substratum classes, 94% of the HLRvideography sites can be accounted for, and the analogous resultin Region B is 70%. Note that all Region A HLR sites along thevessel track were evaluated, but in Region B only 71% of the

Fig. 8. Plots of high lidar rugosity substratum type versus mean lidar rugosity(A), mean NAVD88 elevation in ms (B), and mean slope in degrees (C) forRegion A and Region B combined. Vertical bars centered on substratum depictstandard errors.


HLR sites could be seen on the video imagery, leaving 29%classed as unknown. Therefore, all but 1% of the evaluatedRegion B HLR sites were associated with a cluster zone orderivative class.

The hypothesis that massive coral colony formation,modified by subsequent physical and biological processes thatbreakdown those colonies, is the primary source of NFRTtopographic complexity was supported by the results ofanalyses by class for HLR lidar rugosity (Fig. 8). Mean livemassive coral colony (CC) lidar rugosity exceeded 1.4,followed by dead coral colony (DCC) class with a mean of1.31, and dead overgrown colony (DOCC) with a mean of 1.28.Further evidence for this hypothesis was provided by therankings of mean HLR class elevation for both regions. Theserankings reveal that live massive coral colonies (CC) aregenerally situated at the highest geomorphic elevations, andshowed a progressive drop in elevation mirrored by a general

trend in HLR class towards greater mortality (DCC, DOCC),disintegration (R2, R1), and colonization (HB2).

HB1, the hardbottom class that represents the campus zoneof patch reefs, accounted for 6% of HLR sites observed alongthe videography track in Region A. This result is anomalous, asno pronounced hard substratum topographic variation exists atthese sites. We suggest that the relatively rare occurrence ofHLR over patch reef campus zones is due to the spatiallyvariable reflection of neighboring laser pulses off elements ofthe octocoral canopy and the underlying sediment or pavement.Algorithm tuning could likely suppress this effect, byconstraining the “bottom-picking” procedure to be moreconservative, and thereby minimizing the incorrect recognitionof a gorgonian branch as hard substratum. Alternatively, thedemonstrated capability to sense the elevation of gorgonianbranches suggests that the EAARL could be applied to mapoctocoral campus canopy metrics, analogous to the use ofairborne lidar to map terrestrial forest structure (Harding et al.,2001).

Two decades ago, Jaap (1984) reported mean stony coralcover of 25.7±10.6% and 14.3±9.7%, for NFRT patch reefstony coral-dominated cluster zones and interior octocoral-dominated campus zones, respectively. More recent studies(Miller et al., 2002; Moulding & Patterson, 2002) generallyreport lower stony coral cover on NFRT patch reefs, perhapsreflecting the well documented decline of the FRT over the lastseveral decades (Jaap et al., 2001). Our results suggest that itmay be possible to track massive colony cluster decline (orrecovery) over time with lidar technology, as there weresignificant differences among substratum classes for rugosityand elevation. Our results were consistent with past work,showing that the colony cluster zones are characteristicallyfound along the steepest margins of NFRT patch reefs (Jones,1977), and consist of 1 to 5 m wide coalesced colonies of M.annularis and M. faveolata (Ginsburg et al., 2001). Clusterzones were described as reaching to within 3 m of mean lowwater (Ginsburg et al., 2001), as similarly shown in our study(mean of 4.7 m below NAVD88 vertical datum). The campuszone normally covers the bulk of the patch reef, and consists ofmassive corals usually less than 0.5 m in diameter scatteredamong abundant octocorals. The campus zone is very similar toa hardground community, but exists on a topographicallypositive site (Ginsburg et al., 2001). As our present workfocused on only high rugosity sites, an expanded study over alarger gradient in rugosity values may produce more informa-tion on hardround/campus zone communities.

4. Conclusions

Nearly 4000 patch reefs occur off Elliot Key in the northernFlorida reef tract (NFRT), creating topographic variabilityacross a 10 km wide carbonate platform. The resultingtopographic complexity, or rugosity, ranges from centimetersto kilometers in spatial scale, and either influences or reflectsmany ecological variables, including community composition,herbivory, predation, metabolic processes, hydrodynamics, andnutrient fluxes. The NASA Experimental Advanced Airborne

Fig. 9. Airborne lidar rugosity map for Region A, filtered and edited to remove false bottom retrievals.


Research Lidar (EAARL) was designed to measure the sub-meter scale topography of shallow reef substratums. This studyevaluated the capability of the NASA EAARL lidar todiscriminate cluster zones of massive stony coral colonies onNFRT patch reefs, and identified other NFRT substratums thatalso display high topographic complexity as sensed by theEAARL.

Using dense EAARL laser submarine topographic sound-ings, a 1-m resolution digital substratum rugosity map wascreated for 2 adjacent NFRT study areas, a region studded withpatch reefs on a seagrass meadow (A), and a larger region withdiverse substratums (B). The results obtained for both regionssupport the hypothesis that massive coral colony formation,modified by subsequent by physical and biological processesthat breakdown patch reef framework, is the primary source ofNFRT topographic complexity. Sites with lidar-sensed rugosityabove 1.2 comprised only a minor fraction of the total area onthe NFRT platform, and nearly always occurred about themargins of patch reefs (Fig. 9). Elevated substratum rugosity inboth regions mostly reflected the presence of massive coralcolonies in cluster zones, or their derivatives created bymortality, degradation, physical breakdown, and secondaryovergrowth. In both regions combined, rubble and hardgroundsubstratums derived from massive coral colony cluster-zones,combined with parent cluster zone live or dead colonies,accounted for 79% of the HLR videography sites. Morespecifically, in the heterogeneous Region B, these substratumtypes accounted for all but 1% of the HLR substratumsrecognizable on the along-track video images. Thus, the along-track videography and lidar methodology examined here show

great promise for identifying, mapping, and perhaps even themonitoring of massive coral colony clusters and derivative highrugosity substratum.

Acknowledgements

The authors gratefully acknowledge Richard Curry, ShayViehman, and Todd Kellison of Biscayne National Park forproviding a “test bed” for the development of new techniquesfor coral reef assessment, and for their very significant efforts toaid our field investigations. J. Brock thanks C. Rogers for athoughtful and worthwhile review of an earlier version of thismanuscript. The authors acknowledge the U.S. GeologicalSurvey's Coastal and Marine Geology Program for funding ourinvestigation of coral reef ecosystems.

References

Brock, J. C., Wright, C. W., Clayton, T. D., & Nayegandhi, A. (2004). LIDARoptical rugosity of coral reefs in Biscayne National Park, Florida. CoralReefs, 23, 48−59.

Brock, J. C., Wright, C. W., Patterson, M., Nayegandhi, A., Patterson, J., Harris,M., & Mosher, L. (2006). NASA EAARL Airborne Lidar SubmarineTopographic Maps for the Northern Florida Reef Tract in Biscayne NationalPark, USGS Open File Report 2006-1118.

Burns, T. P. (1985). Hard-coral distribution and cold-water disturbances in SouthFlorida: Variation with depth and location. Coral Reefs, 4, 117−124.

Chiappone, M. (1996). Coral Watch Program Summary: A report on volunteerand scientific efforts to document the status of reefs in the Florida KeysNational Marine Sanctuary, The Nature Conservancy, Florida andCaribbean Marine Conservation Science Center, c/o University of Miami,Coral Gables, Florida, 95 pp.


Chiappone, M., & Sullivan, K. M. (1997). Rapid assessment of reefs in theFlorida Keys: Results from a synoptic survey. Proceedings of the EighthInternational Coral Reef Symposium, 2, 1509−1514.

Dahl, A. L. (1973). Surface area in ecological analysis: Quantification of benthiccoral-reef algae. Marine Biology, 23, 239−249.

Ginsburg, R. N., Gischler, E., & Kiene, W. E. (2001). Partial mortality ofmassive reef-building corals: An index of patch reef condition, Florida reeftract. Bulletin of Marine Science, 69(3), 1149−1173.

Ginsburg, R. N., & Shinn, E. A. (1964). Distribution of the reef-buildingcommunity in Florida and the Bahamas. American Association of PetroleumGeologists Bulletin, 48−527.

Ginsburg, R. N., & Shinn, E. A. (1994). Preferential distribution of reefs in theFlorida Reef Tract: The past is the key to the present. In R. N. Ginsburg(Ed.), Proceedings of the colloquium on global aspects of coral reefs:Health, hazards, and history (pp. 21−26). Miami: University of Miami.

Halley, R., & McPherson, B. F. (1997). The South Florida environment—aregion under stress. U.S. Geological Survey Circular, 1134, 22−24.

Halley, R. B., Vacher, H. L., & Shinn, E. A. (1997). Geology and hydrogeologyof the Florida Keys. In H. L. Vacher, & T. Quinn (Eds.), Geology andhydrology of carbonate islandsDevelopments in Sedimentology, vol. 54.(pp. 217−248).

Harding, D. J., Lefskey, M. A., Parker, G. G., & Blair, J. B. (2001). Laseraltimetry canopy height profiles—methods and validation for closed canopy,broadleaf forests. Remote Sensing of Environment, 76, 283−297.

Jaap, W. C. (1984). The ecology of the South Florida coral reefs: A communityprofile. U.S. Fish and Wildlife Service, FWS/OBS-82/08, 138 pp.

Jaap, W. C., Porter, J., Wheaton, J., Hackett, K., Lybolt, M., Callahan, M. K., etal. (2001). Environmental Protection Agency—Florida Keys NationalMarine Sanctuary Coral Reef Monitoring Project: Updated executivesummary 1996–2000, Washington, DC, 22 pp.

Jones, J. A. (1963). Ecological studies of the southeastern Florida patch reefs:Part I. Diurnal and seasonal changes in the environment. Bulletin of MarineScience of the Gulf and Caribbean, 13(2), 282−307.

Jones, J. (1977). Morphology and development of southeastern Florida patchreefs. Proceedings of the Third International Coral Reef Symposium,University of Miami, Miami, Florida (pp. 232−235).

Lirman, D., & Fong, P. (1995). The effects of Hurricane Andrew and tropicalstorm Gordon on Florida reefs. Coral Reefs, 14, 172.

Lirman, D., & Fong, P. (1997). Susceptibility of coral communities to stormintensity, duration and frequency. Proceedings of the Eighth InternationalCoral Reef Symposium, 1, 361−366.

Marszalek, D. S., Babashoff, G., Noel, M. R., & Worley, D. R. (1977). Reefdistribution in south Florida. Proceedings of the Third International CoralReef Symposium, University of Miami, Miami, Florida (pp. 224−229).

McCormick, M. I. (1994). Comparison of field methods for measuring surfacetopography and their association with a tropical reef fish assemblage.Marine Ecology. Progress Series, 112, 87−96.

Miller, S. L., Chiappone, M., Swanson, D. W., Vermeij, M., & Eaken, D. (2002).Rapid assessment and monitoring of coral reef habitats on the Florida reeftract: Quick look report: Summer 2002. NOAA NURC.

Moulding, A., & Patterson, M. (2002). Biscayne National Park coral reefinventory and monitoring project, final report. Homestead: National ParkService.

Ormand, D. (1999). Wide Area Augmentation System (WAAS). Proceedings ofthe Thirty-first Annual Precise Time and Time Interval (PTTI) Systems andApplications Meeting, Dana Point, CA, December 7–9 1999 (pp. 155−160).

Sale, P. F. (1991). Habitat structure and recruitment in coral reef fishes. In S. S.Bell, E. D. McCoy, & H. R. Mushinsky (Eds.), Habitat structure, thephysical arrangement of objects in space (pp. 197−210). New York:Chapman and Hall.

Sebens, K. P. (1991). Habitat structure and community dynamics in marinebenthic systems. In S. S. Bell, E. D. McCoy, & H. R. Mushinsky (Eds.),Habitat structure, the physical arrangement of objects in space(pp. 211−234). New York: Chapman and Hall.

Shinn, E. A. (1988). The geology of the Florida Keys. Oceanus, 31(1), 47−53.Shinn, E. A., Lidz, B. H., Kindinger, J. L., Hudson, J. H., & Halley, R. B. (1989).

Reefs of Florida and the Dry Tortugas: A guide to the modern carbonateenvironments of the Florida Keys and the Dry Tortugas: InternationalGeological Congress Field Trip Guidebook T176 Washington, DC:American Geophysical Union, 53 pp.

Szmant, A. M. (1997). Nutrient effects on coral reefs: A hypothesis on theimportance of topographic and trophic complexity to reef nutrient dynamics.Proceedings of the Eighth International Coral Reef Symposium, 2,1527−1532.

Wright, C. W., & Brock, J. (2002). EAARL: A lidar for mapping shallow coralreefs and other coastal environments. Proceedings of the SeventhInternational Conference on Remote Sensing for Marine and CoastalEnvironments, Miami, 20–22, May 2002.

Airborne lidar sensing of massive stony coral colonies on patch reefs in the northern Florida reef...

Documents

Transcript of Airborne lidar sensing of massive stony coral colonies on patch reefs in the northern Florida reef...