P A P E R

A Critique of Alternative Power Generationfor Florida by Mechanical and Solar Means

A U T H O R SRobert H. WeisbergYonggang LiuClifford R. MerzCollege of Marine Science,University of South Florida

Jyotika I. VirmaniFlorida Institute of Oceanography

Lianyuan ZhengCollege of Marine Science,University of South Florida

12 Marine Technology Society Journa

A B S T R A C T

l

Using observations of surface winds, solar radiation, ocean currents and wavescollected by the University of South Florida, Coastal Ocean Monitoring and Predic-tion System (COMPS), augmented by other data and numerical model simulations,we address the potential for electrical power generation for Florida by harnessingthe natural energy sources of wind and solar, along with ocean currents and waves.We begin by identifying what nature offers. For wind and solar, we use specificationsfrom existing, commercially available devices to convert nature’s bounty to power-generation estimates. In the absence of mature, commercially available devices forocean currents and waves, we draw upon physical principles to arrive at power-generation estimates for these potential sources. On the basis of what nature offersand what machinery may be capable of producing, we then make reasonable extra-polations onwhat these estimationsmaymean in a practical sense for supplying energyto society. Power generation from these naturally occurring, alternative energy sources,particularly wind and solar,may provide ameans for supplementing power generationby conventional fuels but does not provide a replacement for conventional fuels.Keywords: alternative power generation, ocean observations, windmills, watermills,waves, solar

seem to be relatively good for Florida, a

subtropical peninsula, which is nearly

Introduction

When compared with other locales,the potential for electrical power gener-ation by alternative energy sources may

surrounded by water and bathed insunlight. Herein, we critically assessthis potential using observations ofwinds, incoming short-wave radiation,ocean currents and waves, supple-mented by other data and model simu-lations. The Ocean Circulation Groupat the College of Marine Science, Uni-versity of South Florida (OCG-CMS-USF), through the CMS CoastalOcean Monitoring and PredictionSystem (COMPS), began collectingsuch serial observations on the WestFlorida Continental Shelf (WFS) in1998. We analyze these data to deter-mine the energy fluxes (energy per unitarea per unit time) that are availablethrough natural processes, then trans-form these energy fluxes into practicalpower-generation time series based oneither commercial literature or physi-cally reasonable assumptions, and

then compare the results with con-sumptive metrics. The purpose is todemystify the concept of alternativepower generation by mechanical andsolar means and to place realistic ex-pectations on what may be achievablefor the state of Florida under typical,natural conditions. While our work isspecific to Florida (and primarily westcentral Florida), the findings, withsome modifications, are expected toalso apply elsewhere.

The article is organized as follows.Each of the subsequent four sectionsdeals with power-generation poten-tial by winds, ocean currents, oceanwaves, and incident solar radiation, re-spectively. For each medium, we useeither WFS observations collected bythe COMPS program, or model simu-lations, along with specifications fromcommercially advertised devices, or

reasonable assumptions, for convertingthe natural energy fluxes to power gen-eration. Given nature’s bounty andhow much of this may be convertedto electrical power, the Discussion sec-tion then presents these findings rela-tive to consumptive metrics, such asthe requirement for powering a house-hold and the economics of doing this.Not included, however, are any discus-sions on other complicating matterssuch as electrical transmission, storage,or daily to seasonal variations in peakor base loads that must be accommo-dated on a utility scale. Conclusionsfollow in the last section.

WindThe Data

An offshore array of WFS COMPSmoorings (Figure 1) was initiated in

1998. In addition to measurements ofvelocity, temperature, and salinity overthe water column, as many as fivesurface moorings also collected me-teorological data. Water velocities(currents) were measured using RDInstruments (now Teledyne) acousticDoppler current profilers (ADCP),andmeteorological variables weremea-sured using either Coastal Environ-mental Systems (CES) Weatherpaksor Woods Hole Oceanographic Insti-tution (WHOI)-designed ImprovedMETeorological/Air-Sea InteractionMETeorological (IMET/ASIMET)sensor suites. The surface mooringsall measured air and sea surface tem-peratures, relative humidity, barometricpressure, and wind speed and direction.The IMET/ASIMET system and oneof the Weatherpaks also measureddownward long-wave and short-waveradiation. The IMET/ASIMET sam-pling consisted of 12; 5-s intervalsformed into a 1-min average every20 min. The Weatherpak data werecollected every second for 15 minand averaged to provide 15-min sam-ples. After quality control, these (either15- or 20-min) samples were then

formed into hourly averages for furtheranalysis. A review of these observationsis provided by Weisberg et al. (2009).

Wind speed and direction weremeasured using RM Young 5103Wind Monitor sensors at either 2.8or 3.2 m above the sea surface on theIMET/ASIMET or the Weatherpakbuoys, respectively. In either configu-ration, these observations were ad-justed to a standard 10 m heightabove sea level using a log boundarylayer scaling under neutral stability(e.g., Large & Pond, 1981). For thepurpose of applying such buoy windobservations to large-scale commercialwind turbines, a further adjustmentwas needed to account for the turbinehubs being located some 80 to 100 mabove the sea surface. Using a 100-mhub height, we estimated the windspeed there (U100) from the windspeed at the standard 10-m level(U10) by

U100 ¼ U10log z=z0ð Þlog 10=z0ð Þ ð1Þ

where z0 is the surface roughness,which, for open water exposure, was

September/Octo

taken to be 0.015 m. Recognizingthat such log layer scaling (yieldingan amplification factor of 1.35) ismerely an approximation, we acquiredNational Centers for EnvironmentalPrediction (NCEP) North AmericanMesoscale (NAM) modeled wind re-sults as a check. Downloaded fromhttp://nomads.ncdc.noaa.gov/data /namanl/ for six sites along a shore-normal line intersecting Sarasota, FL,for the period 1/1/12 to 4/12/12(such multiple level results are notavailable for earlier times), a linearregression between winds modeled at10 and 80 m heights (the lowestlevels available) for the offshore openwater exposure sites yielded a co-efficient of 1.16, less than the loglayer scaling result. Thus, the use of1.35 as a conversion factor from 10to 100 m winds is offered as a con-servative estimate, overestimating,versus underestimating the windsaloft, on average.

On the basis of these (hourly aver-aged) wind observations scaled to ahub height of 100 m, and using moor-ing C10, located about 25 nm offshorefrom Sarasota, FL, Figure 2 showswhat was available for wind power gen-eration at this site over the 8-year inter-val, from 1999 to 2007. The valuesrange from zero to around 20 m s−1,but with the higher end occurringonly on rare occasions during the pas-sage of tropical storms.

Whereas our analysis is limited to asingle point off the west central Floridacoast, these winds are representative ofwinds elsewhere in Florida with somecaveats. On both long-term and sea-sonal averages, the winds tend to in-crease from north to south by virtueof the trade winds’ meridional struc-ture and Florida peninsular land effects(e.g., Weisberg et al., 2009; Liu &Weisberg, 2005), which also results

FIGURE 1

Map of theWest Florida Continental Shelf COMPS stations. Observations frommooring C10, locatedapproximately 25 nm offshore from Sarasota, FL, are used herein.

ber 2012 Volume 46 Number 5 13

in the trade winds being a little stron-ger on the east coast. For synoptic scaleweather events, the entire state ofFlorida is similarly affected. On thediurnal time scale, the east coast seabreeze tends to be more regular thanthat on the west coast.

Converting Wind Speedto Electrical Power-Generation Potential

Commercially available wind tur-bines are discussed with respect to theirnameplate-rated power-generationcapacity. This can be misleadingbecause the actual power output de-pends on wind speed. As a representa-tive example we consider a GeneralElectric (GE) 3.6 MWOffshore SeriesWind Turbine, with specificationsthat are available in a brochure, whichmay be downloaded from the manu-

14 Marine Technology Society Journa

facturer’s Internet site. From thewind load-power curve, we see thatthe turbine does not begin to produceelectrical power until the wind speedexceeds 3.5 m s−1 (the cut-in windspeed). Power generation then in-creases with increasing wind speed,reaching the nameplate-rated capacity(3.6 MW) at a wind speed of 14 m s−1

(the rated wind speed), and the deviceceases power generation and shutsdown when the wind speed exceeds27 m s−1 (the cut-out wind speed).The lower and upper horizontal lineson Figure 2 represent the cut-in andthe nameplate-rated capacities for theGE 3.6 MW turbine, respectively.From 8 years of WFS data, we seethat the winds at 100-m hub heightfail to drive the turbine some 20% ofthe time and that rarely do the windsreach the nameplate rated capacity.

l

To determine what the power out-put may be when the turbine is run-ning, we fitted a polynomial to thepower curve provided by the brochure,and we used this to convert wind speedto power output for speeds betweenthe cut-in and rated capacity values.For wind speeds between 14 and27 m s−1, the output was held constantat 3.6MW (Figure 3). The results, fur-ther averaged to provide daily values,are shown in Figure 4, from whichseveral points are clear. First, thenameplate-rated capacity is rarelyachieved. Second there are many days(excluding the larger interval data gaps)when the cut-in speed of 3.5 m s−1 isnot exceeded, and hence no power isgenerated. Third, when looking atthe climatological monthly meantime series (the lowest panel) obtainedby averaging all Januarys, all Februarys,etc., we see that minimum and maxi-mum power generation occurs in sum-mer and fall months, respectively. Themonthly mean minimum is about0.6MW, themonthlymeanmaximumis about 1.8 MW, and the grand aver-age across all years and months is about1 MW. The minimum in summermonths is troublesome for Florida be-cause that is when the demand for airconditioning is the largest.

Ocean CurrentsThe Physics

Unlike windmills, where commer-cial maturity provides known power-generation potential, watermills drivenby ocean currents remain in develop-ment. Estimating the potential forpower generation by ocean currents re-quires that we begin from first princi-ples. Available power, P, for potentialextraction from ocean currents is thekinetic energy flux, 1

2 ρV3, times the

area, A, of the device used for extracting

FIGURE 2

Hourly averagedwinds scaled to 100-mhub height using C10 observations from 1999 (top) through2007 (bottom). The lower and upper red lines indicate the turbine cut-in value of 3.5 m s−1 and therated power-generation wind speed of 14 m s−1, respectively.

September

.

/Octo

this flux, or P ¼ 12 ρV

3A, where theunits for P are watts (W). Watermills,like windmills, are subject to the samehydrodynamic limitations embodiedin Betz’s law (Betz, 1920), which statesthat the maximum power that may beextracted is 59% of the kinetic energyflux offered by nature. Additional losescome from the efficiency of the deviceitself, such that the expected power-generation potential for either a wind-mill or a watermill is in the approximaterange of 40-50%.

Watermills, like windmills, arealso expected to have a cut-in speedthreshold below which they will notfunction. In other words, a baselinetorque is necessary to drive an electricalgenerator. Given that torque equalsforce times distance, it is proportionalto the pressure on a turbine blade timesboth the area and the length of theblade. A dimensional analysis, makinguse of Bernoulli’s theorem, suggeststhat the cut-in speed may scale as

Sw ¼ Saffiffiffiffiffiffiffiffiffiffiffiffiρa=ρw

pLaLw, where S denotes

the cut-in speed, ρ the density and Lthe length, and the subscripts a andw denote air and water, respectively.Using the cut-in speed for the GE3.6 MW turbine and its length scaleand assuming that a watermill mayhave a length scale about an order ofmagnitude smaller than the windmill,we arrive at a cut-in speed estimate ofaround 1 m s−1 for the watermill.Granted, this is a very crude estimate,but what it does suggest, even if off bya factor of two to four, is that typicalcoastal ocean current speeds on thecontinental shelf (away from tidal inlets),which are of order 0.2 to 0.5 m s−1

(e.g., Weisberg et al, 2009; Liu &Weisberg, 2005), are too small todrive watermills. Nevertheless, Floridadoes have a strong western boundarycurrent seaward of its continental

FIGURE 4

Daily-averaged power and climatology of power using C10 wind speed observations scaled to100 m hub height, from 1999 (top) through 2007 (bottom).

FIGURE 3

A polynomial fit to a power curve provided for a GE 3.6 MW wind turbine (see ge_36_brochurePDF available at http://www.gepower.com). The power curve gives the machine output in kW asa function of wind speed in m s−1.

ber 2012 Volume 46 Number 5 15

shelf, for which we can examine thepotential utility of power generationby watermills. For the west coast ofFlorida this western boundary currentis the Gulf of Mexico Loop Current,which feeds into the Gulf Stream onthe east coast of Florida. Being thatthe Gulf Stream is tightly constrainedto flow between Florida and theBahamas, this is the most practicalplace for considering power genera-tion by watermills.

Application to the Gulf StreamThe Gulf Stream, as it flows

through the channel between Floridaand the Bahamas, is highly baroclinic,with maximum speeds generally lo-cated at the surface toward the westernside of the channel (see Figure 5, afterLeaman et al., 1987). Speeds near thesurface, and approximately within abaroclinic Rossby radius of deforma-tion from the Florida side, are as highas 2 m s−1 (4 kts), diminishing rapidlywith depth to about 1 m s−1 at 300 mdepth and then to less than 0.5 m s−1

below 500 m depth. The total volumeflux through the Florida Straits haslong been recognized to be around30 × 106 m3 s−1 (or Sverdrups) (e.g.,

16 Marine Technology Society Journa

Stommel, 1965; Niiler & Richardson,1973; Leaman et al., 1987).

For our purposes, we use velocitytime series simulated by a numericalcirculation model to estimate thepower-generation potential for theGulf Stream. The model chosen isthe global Hybrid Coordinate OceanModel (HYCOM) run by the U.S.Navy Research Laboratory and theHYCOM consortium (e.g., Chassignetet al., 2007, 2009). Figures 6 and 7show volume and kinetic energy trans-ports, respectively, for transects at thelatitudes of Miami, FL, and PalmBeach, FL. In each of these figures, thetransports were calculated over three dif-ferent depth intervals, 0 to 50 m, 50 to

l

300 m, and 300 m to the bottom,plus the total transports across the en-tire cross sections. The first of these fig-ures provides a check on the HYCOMsimulation. From it, we see that thetotal transports properly represent theobservations towithin reasonable boundson natural variability and measurementerror. This provides justification forusing these model simulation resultsto discuss the kinetic energy flux andhow these integrate to provide esti-mates on power-generation potential.

Two practical considerations cometo play. First, with sea state undernortherly winds being very large attimes across the Florida Straits, itwould be difficult, if not impossible,

FIGURE 6

Gulf Stream volume transport across two sections between Florida and the Bahamas at the latitudesof Miami, FL, and Palm Beach, FL, calculated for the calendar year 2008 using a Global HYCOMsimulation. Transports are shown for three different depth intervals, plus the total transport acrossthe entire cross sections. The dashed lines on the bottom panel provide the year-long averages.(Color versions of figures available online at: http://www.ingentaconnect.com/content/mts/mtsj/2012/00000046/00000005.)

FIGURE 5

A Gulf Stream cross section for the northcomponent of velocity sampled at 27°N(from Leaman et al., 1987).

to operate turbines within the upper50 m of the water column. For in-stance, the wavelength for a deep-water wave of 8 s period is 101 m;hence, particle speeds for waves oflonger period would impact the flowfield at 50-m depth. Second, with cur-rent speeds generally less than 0.5 m s−1

below 300 m there is likely little po-tential for energy generation by water-mills below that depth. Given thesepractical considerations, it is reason-able to limit our attention to a depthinterval of 50-300 m.

For the depth range of 50-300 m,Figure 7 shows that the annual meanpower that may potentially be tappedby watermills is about 7.4 and 8.3 GWat the Miami and Palm Beach trans-

ects, respectively. These potential val-ues are less than the correspondingtotal Gulf Stream cross-sectional an-nual mean power estimates (surfaceto bottom) of 11.4 and 13.2 GW, re-spectively. The approximate 2.7 and4.3 GW of the upper 50 m, respec-tively, are not available, nor are theapproximate 1.3 and 1.4 GW below300-m depth, respectively, for thepractical reasons just provided. More-over, the application of Betz’s law,plus additional mechanical loses re-duces the power potential by at least50% to about 3.7 and 4.1 GW forthe Miami and Palm Beach transects,respectively. Further recognizing theimpossibility of filling the entire crosssection depth range with turbines re-

September/Octo

duces these numbers by at least an-other order of magnitude. The endresults are the more realistic potentialsof 370 and 410 MW for these Miamiand Palm Beach sections, respectively.But even these numbers are likely to beoverestimates because any watermillwill have a water load-power curvejust like a windmill (e.g., Figure 3) soits power output will be further re-duced from that potentially providedby nature. Given what an actual water-mill cut-in speed and water load-powercurve may be, it is probably reasonableto reduce the above by a factor oftwo and settle on about 200 MW as apotential estimate. This would be equiv-alent to what may be obtained fromabout 200 (GE 3.6 MW) windmills ifoperated under West Florida wind con-ditions (Converting Wind Speed toElectrical Power Generation Potential).

Because windmills installed oneither land or offshore and watermillsinstalled within the Gulf Stream (themost powerful of the ocean currentsadjacent to the United States) tend tohave similar promise for power gener-ation, it is useful to consider two addi-tional factors. The first is a matter ofscale. Being that windmills and water-mills rely on products of fluid densitytimes fluid velocity cubed times area(ρV 3A), it is apparent that (with airbeing approximately a thousandtimes less dense than water while airvelocity is about ten times faster thanwater) the area necessary to generatecomparable amounts of power in airand water are the same. This begs thequestion: if the size of the machinerymust be the same in air and water,why would we choose to work in afluid medium (water) that is techno-logically so much more challengingthan air? The second consideration iseven more daunting. Windmills de-ployed in air operate within the lower

FIGURE 7

Gulf Stream kinetic energy transport across two sections between Florida and the Bahamas atthe latitudes of Miami, FL, and Palm Beach, FL, calculated for the calendar year 2008 using aglobal HYCOM simulation. Transports are shown for three different depth intervals, plus the totaltransport across the entire cross sections. The dashed lines on the bottom panel provide theyear-long averages.

ber 2012 Volume 46 Number 5 17

100 m of the atmosphere or in thelower part of the frictional boundarylayer, which is driven by and continu-ally replenished by the geostrophic in-terior that extends to the tropopause,some 10 km aloft. Watermills, in con-trast, operating across a major portionof the water column, are in the geo-strophic interior itself and hence, un-like windmills, have no natural meansfor replenishing the energy that theymay extract. It is for this reason thatwindmill farms may have closelyspaced units with additional windmillseven distributed downwind from oneanother. Watermills, in contrast, can-not share this deployment strategy.Once power is removed from a crosssection, it cannot be readily replen-ished. For these two reasons, even inthe swiftest of currents, like the GulfStream, the notion of power genera-tion by watermills tapping the kineticenergy flux seems very limited whencompared with what may be achievedby windmills.

Ocean WavesThe Physics

Waves, like currents, also possessan energy flux that may be tapped bymechanical devices. The availablepower, PW, for potential extractionfrom ocean gravity waves is the totalmechanical energy flux per unitwave crest width, 12 ρga

2CG , times thecrest width length, L, of the deviceused for extracting this flux, orPW ¼ 1

2 ρga2CGL, where g is the accel-

eration of gravity, a is the wave ampli-tude, CG is the group velocity, and theunits for PW are watts. An alternativeexpression for PW, based on significantwave height and application of thedeep water dispersion relation, isPW ¼ 1

2H2s TL, where T is the wave

period and significant wave height Hs

18 Marine Technology Society Journa

is defined as the average of the highestthird of the waves.

Application to the WFSSeveral point measurements of

surface gravity waves are available forthe WFS, either from COMPS obser-vations or a NOAA NWS weatherbuoy. As it is more instructive to lookat the entire field of waves and how thisvaries throughout the year, we opted tobase our estimates on numericalmodel simulations, of which thereare several. The longest of these is

l

from the NOAA application of theWaveWatch III model, e.g., Tolman(1991, 1999, 2010). Figure 8 showsa monthly mean wave energy fluxper unit crest width climatology for theWFS with contour units of kW m−1

calculated from an 8-year analysis ofWaveWatch III model results inclusiveof 1999 to 2007. A robust annual cycleis seen with minimum wave energy insummer andmaximumwave energy inwinter, similar to that of the significantwave height measured at NDBC Buoy42036 (Liu et al., 2010). This finding

FIGURE 8

Monthly mean surface gravity wave energy flux per unit crest width calculated for the WFS usingthe WaveWatch III reanalysis from 1999 to 2007. The contour units are kW m−1, and each ofthe monthly climatologies is calculated by averaging all similar months, i.e., all Januarys, allFebruarys, etc.

is consistent with the prevailing winddirections for the WFS varying fromsoutheasterly in summer to north-easterly in winter (e.g., Liu &Weisberg,2005).With low wave energy, theWFSis not very promising for alternativepower generation by tapping the energyflux of surface gravity waves, except per-haps for running low power instrumentsin situ.

The east coast of Florida does havea larger wave climate, especially northof the Bahamas where the entire fetchof the north Atlantic comes to play.But even there the energy flux per unitcrest width remains small comparedwith other higher energy coastlinesworldwide (Figure 9), where tens toeven a hundred kWm−1 are potentiallyavailable. As with electrical generationpotential using ocean currents, thequestion becomes one of feasibility.Is there enough energy potentiallyavailable to justify the costs for extrac-tion and can the technical challengesbe met?

The commercial literature andcommercial advocate group studies(e.g., McGowen et al., 2005) suggestthat wave energy extraction is econom-ically feasible for regions with energyflux per unit crest width greater than

15 kW m−1. Two examples of devicesunder development are (1) a largesnake-like set of linked cylinders thatextract wave energy via undulationsacross a large linked cylinder length(e.g., 180 m for the 4-m diameterPelamis WavePower device; http://www.pelamiswave.com) and (2) a4-m diameter buoy that tracks thevertical motion associated with wavepropagation past a fixed point (e.g.,Ocean Power Technologies, Inc.,http://www.oceanpowertechnologies.com). The first of these uses the lengthof the device to extract a major portionof the wave energy flux past the 4-mdiameter cross section; the second ofthese is more limited in the percent-age of the flux that can be extracted.Without questioning any of thedetails (see the industry brochures), itis clear that for either of these devicesthe total flux that may potentially betapped is whatever passes the 4-mdevice width. So in either case, thequestion becomes: can some frac-tion of 15 kW m−1 times 4 m pro-vide an economically feasible sourceof power (assuming that 15 kW m−1,as promoted, is a viable level)? Atbest, assuming complete extractionwith no loses (itself impossible),

September/Octo

devices such as these can garner atmost 60 kW.

SolarThe Observations

Two of the COMPS buoys carrieddownward short-wave radiation sen-sors. Here we will use the most com-plete of these from mooring C10,located approximately 25 nm offshorefrom Sarasota, FL, where an EppleyPSP sensor was mounted as part ofthe IMET/ASIMET suite of air-sea in-teraction sensors. The veracity of thesemeasurements for our purposes herefollows from previous studies thatutilized these data to diagnose the netsurface heat flux and the associated var-iations on water column temperature(e.g., Virmani & Weisberg, 2003,2005). Figure 10 provides these hourlysampled, solar insolation data thatwere collected approximately 2.5 mabove the sea surface for the 8-year in-terval, from 1999 to 2007. On anhourly basis, we see that the daily max-imum insolation varies from as high as1,000Wm−2 in spring and summer toas low as 500Wm−2 in fall and winter.Averaging diurnally to account for thefact that there is no incoming short-wave radiation at night, we findhighest daily mean values rangingfrom about 300 W m−2 in summerto 150 W m−2 in winter.

Equivalent Energy GenerationPotential Using Solar Panels

As with our analysis of wind energygeneration, which recognized thematurity of the industry and hence al-lowed for us to choose a vendor withestablished product specifications, wedo the same here for solar power.The example used is a Siemens SP75solar panel. The method employed todetermine the output of such a solar

FIGURE 9

Global wave energy flux per unit crest width (after McGowen et al., 2005).

ber 2012 Volume 46 Number 5 19

panel is to calculate the equivalentnumber of hours for which suchpanel would capture insolation at a1,000 W m−2 level. Thus, we com-puted an annual mean, hourly insola-tion curve for 2001 (the year for whichour data are most complete) and inte-grated the area under that curve to ar-rive at an equivalent area (insolation ×t ime) a t an inso la t ion leve l o f1,000 W m−2 (Figure 11). Given thatthe Siemens SP75 solar panel is ratedto output 75 W at an insolation of1,000 W m−2, we then determinedthe daily mean output for the deviceto be 404W h (based on the equivalentsolar production day of 5.39 h). Divid-ing by 24 h gives 16.8W as the averageoutput for this annually averaged day,and dividing by the area of the solarpanel (0.63 m2) gives 26.7 W m−2

as the daily averaged power output

20 Marine Technology Society Journa

F

Tbb

l

per unit area under WFS daily aver-aged insolation for this particularsolar panel.

This technique (just demonstratedfor an annually averaged diurnal cycle)was applied on a daily basis to convertobserved incoming short-wave radia-tion to solar panel output normalizedto a square meter. The results werethen averaged diurnally to providedaily averaged power generation timeseries for the entire 8 years of record.Figure 12 shows the daily results,along with a climatological monthlymean time series (the bottom panel)obtained by averaging all of the Januarys,all of the Februarys, etc.

The daily averages range fromessentially zero on strongly overcastdays to 40 W m−2 on clear summerdays. When averaged over the monthwe see a minimum in winter ofabout 20 W m−2 and a maximumin spring of about 38 W m−2. The

FIGURE 10

Hourly sampled time series of incoming short-wave radiation measured 2.5 m above mean sealevel at mooring C10 located approximately 25 nm offshore from Sarasota, FL. Shown are8 years of observations spanning 1999 (top) through 2007 (bottom).

IGURE 11

he 2001 annually averaged daily insolation curve observed at C10 (blue) with equivalent num-er of hours at 1,000 W m−2 (black), where the areas under either of these distributions (blue orlack) are the same.

average across the entire year is about25 W m−2.

DiscussionElectrical power generation by

windmills, solar panels, watermills(ocean current turbines), and wave de-vices have been topics of discussion forseveral decades. Wind and solar appli-cations are mature, and commercialdevices may be purchased and operated.Currents and waves applications havenot reached a similar level of maturity.Devices exist and some have under-gone field tests, but none are commer-cially viable yet. That in itself speaks tothe relative utility of these concepts.The application of over a decade ofcoastal ocean observations from theWFS (augmented by model simula-tions) supports this viewpoint that al-ternative power generation from windand solar sources appears to be more

promising than that from ocean cur-rents and waves.

Further appreciation of this findingfollows from a few simple economicsconsiderations. We begin with theknown cost for powering a modesthome. As an example, consider a2,000 feet2 Florida apartment thatconsumes about 1.7 kW of electricalpower on annual average at a cost ofabout $140 per month or $1,680 peryear. If we were to transition from con-ventional fuels to wind power by de-ploying a machine equivalent to a GE3.6 MW turbine, we would average1 MW of production, sufficient (onannual average) to power 588 suchhomes. The present cost for poweringthese homes based on conventionalfuels and rounded is $1,000,000. Acost effectiveness transition to windpower would therefore require thatthe combined amortization, mainte-nance, energy storage, transmission

September/Octo

and distribution, salaries, general andadministrative costs, plus shareholderprofits be about $1,000,000 per yearfor such a windmill. Whereas the spe-cific costs for turbine purchase and in-stallation are not readily available,anecdotal, nonrefereed literature sug-gests that these range between $1.2and $2.6 million per MW of name-plate capacity. Thus, depending onthe amortization of these capitalcosts, wind power may be approachingcost effectiveness. Support for thiscomes from an Associated Press articlesuggesting that the electrical deliverycost for a wind farm proposed offshoreof Cape Cod, MA, will be roughlytwice that by conventional fuels. Sowhile the costs for electrical power gen-eration by wind is higher than by con-ventional fuels, wind power generationmay be economically feasible in the fu-ture, consistent with the fact that thewind power industry is indeed a ma-ture one.

As with wind, it is difficult to findstraightforward information on solarpanel costs. Systems installation costs areestimated at about $1,000 m−2, andwith an annually averaged solar panelpower production of 26.7 W m−2 byobserved Florida insolation, the costper watt would be about $38. Thus, thesolar panel cost for a modest 1.7 kWhouse would be about $64,000, some38 times the present annual cost ofelectricity by conventional fuels. Thisof course does not include any consid-erations of metering or storage strategiesor properly sizing a system to actuallymeet user needs. While the above esti-mate may not be out of the realm ofwhatmay be reasonable based on amorti-zation costs, any individual home ownerwould be hard pressed to justify suchinvestment without a major subsidy.

The economics take a rapid turn forthe worse when considering either

FIGURE 12

1999 through 2007 time series of daily average solar panel output based on observed insolationfrom C10 and climatology.

ber 2012 Volume 46 Number 5 21

ocean currents or waves. Ocean cur-rents, as discussed in Ocean Currents,require similar sized machines (water-mills) as for wind, and the energyextraction potential is much more lim-ited than for wind, even when consid-ering a massive current like the GulfStream.Moreover, it would be imprac-tical to deploy watermills of the samesize as windmills; hence many more,much smaller watermills would be nec-essary than for windmills, greatly com-pounding the costs for an equivalentamount of energy. This is even beforeany consideration is given to the tech-nical challenges of deploying andmaintaining a large number of me-chanical devices in a swift ocean cur-rent. From these arguments, we mustconclude that alternative electricalpower generation by ocean currentsfor any regional utility applicationwould be both prohibitively costlyand impractical. It is therefore not sur-prising that this industry remains in adevelopmental versus a mature stage.

Waves, in our opinion, are evenmore impractical than currents forutility scale power generation. Con-sider, for instance, either the ap-proximate 4-m buoy or 180-m-longcylinder discussed in Ocean Waves.Even for seas with energy flux perunit crest width of 15 kW m−1, thesemachines would have the potential togenerate no more than 60 kW permachine, much less when necessaryloses are considered. Using presentelectric costs as documented for a1.7 kW house, 60 kW is worth about$59,000 per year. Considering theneed to swap out machines for servic-ing perhaps twice yearly, if not morefrequently, it might cost more just forthe ship time necessary to deploy andrecover these machines than the valueof the electricity that they could po-tentially generate and that does not

22 Marine Technology Society Journa

include any of the costs for purchasingthe machines and establishing the in-frastructure for their use.

In summary, the economics for al-ternative power generation by windand solar means may result in cost-effective strategies in the future, whereasthose for ocean currents and offshorewaves will not, at least for projects inwhich large quantities of power are re-quired, such as powering a majorurban area. Other inhibiting consid-erations also come into play. Althoughour cost estimates are based on an-nually averaged quantities, it must berecognized that it is not uncommonfor winds to be low enough to fail com-pletely as an energy source (about 20%of the time on the WFS) and similarcan be said of solar insolation. Thus,neither of these two potential alterna-tive power-generation sources can fullyreplace power generation by conven-tional fuels; they can only supplementthe use of conventional fuels.

Windmills of the type required tosupplement large power needs arealso massive in size. To provide asense of spatial scale, a professionalfootball stadium stood up on end pro-vides an analogy to the equivalent crosssectional area that would be occupiedby a large windmill. Accommodatingsuch structures along highly populatedcoastlines would be difficult. Florida,with its highly developed, tourist-oriented, coastline would likely notbe amenable to situating offshore utility-scale wind farms in sight of land. Theuse of solar panels in an array largeenough to supplant a conventionalfacility also suffers from such mattersof spatial scale. For instance, based onFlorida insolation, replacing a 1.8-GWpower plant, such as the TampaElectric Company Big Bend facility(located on the east shore of TampaBay) would require 90 km2 of solar

l

panels, plus other storage devices re-quired to accommodate evening houror extended days of low insolation.Lacking installations of such massivescale, it remains unknown whether ornot their maintenance would even befeasible.

ConclusionsBased on analyses using coastal

ocean observing system data forwinds, incoming short-wave radiationand ocean currents, supplemented, asneeded, by numerical model simula-tions, the following conclusions maybe drawn. First, there are good reasonswhy industries pertaining to electricalpower generation by the alternativemeans of wind and solar are muchmore mature than those pertaining toocean currents and waves. Wind andsolar sources of energy do providepromise for alternative power genera-tion on a utility scale, whereas formany physical and practical reasons,ocean currents and waves do not. Sec-ond, even if wind and solar power inFlorida are eventually produced in aneconomically competitive way (whichrequires a much more detailed eco-nomics analysis), these power sourcesmay only be able to supplementpower generation by conventionalfuels or other means. They cannot re-place presently required base loadpower-generation capacity.

AcknowledgmentsSustained observations made this

work possible. Initiated in 1998 withUSF support from the Florida Legisla-ture, COMPS has operated ever sinceby marshaling support from a varietyof sources, including the U.S. Geolog-ical Survey, Minerals ManagementService, Office of Naval Research,

National Oceanic and AtmosphericAdministration, the Florida Depart-ment of Community Affairs and theFlorida Wildlife Commission, recentsupport (in addition to USF) beingthrough the Office of Naval Researchgrants N00014-05-1-0483 andN00014-10-1-0785, NOAA ECOHABgrant NA06NOS4780246, NOAAIOOS grant NA07NOS4730409,and NSF grant OCE-0741705. Specificto this application is also a grant fromthe Florida Energy Systems Consor-tium (FESC). Data collection successis owed to COMPS staff, with seagoingoperations by J. Law and (formerly)R. Cole, data analysis, managementand quality assurance by J. Donovan,D. Mayer and P. Smith, and engineer-ing assistance through the CMSCenterfor Ocean Technology, in particularR. Russell. Prof. S. Shin assisted withthe NAM boundary layer winds.

Corresponding Author:Robert H. WeisbergCollege of Marine Science,University of South Florida140 7th Ave. S., St. Petersburg,FL 33701Email: weisberg@usf.edu

ReferencesBetz, A. 1920. Das Maximum der theoretisch

möglichen Ausnutzung des Windes durch

Windmotoren, Zeitschrift für das gesamte

Turbinenwesen. 20. September 1920.

Chassignet, E.P., Hurlburt, H.E., Smedstad,

O.M., Halliwell, G.R., Hogan, P.J., Wallcraft,

A.J.,… Bleck, R. 2007. The HYCOM (HYbrid

Coordinate Ocean Model) data assimilative sys-

tem. J Marine Syst. 65:60-83. http://dx.doi.org/

10.1016/j.jmarsys.2005.09.016.

Chassignet, E.P., Hurlburt, H.E., Metzger, E.J.,

Smedstad, O.M., Cummings, J., Halliwell,

G.R., … Wilkin, J. 2009. U.S. GODAE:

Global ocean prediction with the HYbrid

Coordinate Ocean Model (HYCOM). Ocean-

ography. 22:48-59. http://dx.doi.org/10.5670/

oceanog.2009.39.

Large, W.G., & Pond, S. 1981. Open ocean

momentum flux measurements in moderate to

strong winds. J Phys Oceanogr. 11:324-36.

http://dx.doi.org/10.1175/1520-0485(1981)

011<0324:OOMFMI>2.0.CO;2.

Leaman, K.D., Molinari, R., & Vertes, P.

1987. Structure and variability of the Florida

vurrent at 27N: April 1982–July 1984. J Phys

Oceanogr. 17:565-83. http://dx.doi.org/

10.1175/1520-0485(1987)017<0565:

SAVOTF>2.0.CO;2.

Liu, Y., & Weisberg, R.H. 2005. Patterns

of ocean current variability on the West

Florida Shelf using the self-organizing map.

J Geophys Res. 110:C06003. http://dx.doi.org/

10.1029/2004JC002786.

Liu, Y., Weisberg, R.H., Merz, C.R.,

Lichtenwalner, S., & Kirkpatrick, G.J. 2010.

HF radar performance in a low energy envi-

ronment: CODAR SeaSonde experience on

the West Florida Shelf. J Atmos Oceanic

Technol. 27(10):1689-710. http://dx.doi.org/

10.1175/2010JTECHO720.1.

McGowen, C., Bedard, R., & Lenssen, N.

2005. Ocean tidal and wave energy, renewable

energy technical; assessment guide-TAG-RE,

EPRI, Palo Alto, CA, 2005, 1010489.

Niiler, P.P., & Richardson, W.S. 1973.

Seasonal variability of the Florida current.

J Mar Res. 31:144-67.

Stommel, H.M. 1965. The Gulf Stream:

a physical and dynamical description.

Berkeley, CA: University of California Press.

248 pp.

Tolman, H.L. 1991. A third-generation

model for wind waves on slowly varying,

unsteady and inhomogeneous depths and

currents. J Phys Oceanogr. 21:782-97.

Tolman, H.L. 1999. User manual and system

documentation of WAVEWATCH-III ver-

sion 1.18. NOAA / NWS / NCEP / OMB

Technical Note 166. 110 pp. http://dx.doi.org/

September/Octo

10.1175/1520-0485(1991)021<0782:

ATGMFW>2.0.CO;2.

Tolman, H.L. 2010. WAVEWATCH III (R)

development best practices Ver. 0.1. NOAA /

NWS / NCEP / MMAB Technical Note 286,

19 pp.

Virmani, J.I., & Weisberg, R.H. 2003. Fea-

tures of the observed annual ocean-atmosphere

flux variability on the West Florida Shelf.

J Climate. 16:734-45. http://dx.doi.org/

10.1175/1520-0442(2003)016<0734:

FOTOAO>2.0.CO;2.

Virmani, J.I., & Weisberg, R.H. 2005.

Relative humidity over the West Florida

Continental Shelf. Mon Weather Rev.

133:1671-86. http://dx.doi.org/10.1175/

MWR2944.1.

Weisberg, R.H., He, R., Liu, Y., & Virmani,

J.I. 2005. West Florida shelf circulation on

synoptic, seasonal, and inter-annual time

scales. In Circulation in the Gulf of Mexico,

W. Sturges and A. Lugo-Fernandez, eds.,

AGU monograph series. Geophysical

Monograph. 161:325-47.

Weisberg, R.H., Barth, A., Alvera-Azcárate, A.,

& Zheng, L. 2009. A coordinated coastal

ocean observing and modeling system for

the West Florida Shelf. Harmful Algae.

8:585-98. http://dx.doi.org/10.1016/

j.hal.2008.11.003.

ber 2012 Volume 46 Number 5 23