Response of subtropical shallow-water environments to cold-air outbreak events: Satellite radiometry and heat flux modeling
N. D. WALKER,* L. J. ROUSE, Jrt and O. K. HUHt
(Received 19 March 1986; accepted 19 June 1986)
Abstract--Cold-air outbreak induced chilling of shallow bay, bank, and shelf waters of southern Florida and the northern Bahamas was examined using satellite thermal infra-red measurements, in situ measurements, and a shallow-water heat flux model. Vast expanses of shallow waters are rapidly modified by the cold, dry continental air and high wind speeds characteristic of cold-front passages. Although water mass modifications are more rapid in shallow areas, prolonged effects are experienced in deeper shelf regions. Northerly winds accompanying the cold-air outbreak induce a net offshelf circulation, subjecting deeper regions to an inflow of chilled waters generated in shallower areas. Absence of coral reef development along preferential routes for offshelf water movement suggests that these winter processes are a limiting influence to southern Florida and northern Bahamas reef distribution.
INTRODUCTION
EFFECTS of cold-air outbreaks on subtropical shallow estuarine and continental shelf waters have been described, measured, and modeled by previous researchers (NowLIN a n d PARKER, 1974; GARWOOD et al., 1981; HUH et al., 1978, 1984). Farther to the south, shallow subtropical waters respond in a similar manner, but chilling can be more rapid and extreme. The impact on local biota is often more dramatic, as subtropical fauna are less adapted to temperature fluctuations.
This paper presents results of a study on cold-air outbreak induced water mass transformations in bay and bank environments of the southern Florida and northern Bahamas region (Fig. 1). The results obtained, however, could apply to any shallow- water area subjected to outbreaks of cold, dry continental air. Over the northern Gulf of Mexico, the cold-air outbreak cycle repeats itself every 5-6 days (ANGELOVIC, 1976). Such weather systems reaching the study region are less frequent and of somewhat reduced intensity, having been modified en route by relatively warm deep waters of the Gulf of Mexico and the Florida Current (in the northern Bahamas case).
The objectives of this paper are to: 1. Describe the temporal evolution and spatial extent of chilling of shallow bay/bank
and shelf waters during cold-air outbreak episodes. 2. Identify the heat flux processes responsible for the observed water mass transforma-
tions in order of their importance. 3. Estimate the residence times and circulation of the resulting chilled water masses in
the southern Florida region.
* National Research Institute of Oceanology, Stellenbosch, South Africa. t Coastal Studies Institute, Louisiana State University, Baton Rouge, LA 70803, U.S.A.
735
736 N . D . WALKER et al.
I 7 8 °
Bahama Bank REEF TRACT
NW Providence Channel
Key W e s t
Contours in meters
Gulf of Mexico
~81
/
7DA ENT
/ '20
~Cay~Sal;~ the
C U B A
Fig. 1. The southern Florida and northern Bahamas study region showing primary water masses and measurement sites. The stars with dates adjacent to them are sites of reported coral
mortalities and the years of occurrence.
A record cold-air outbreak episode during January 1981 will be discussed in detail; however, episodes during the 1976-1977 and 1979-1980 winters were also investigated (WALKER, 1982).
S A T E L L I T E D A T A A N A L Y S I S
Temperature-corrected thermal infra-red measurements from the Advanced Very High Resolution Radiometer (AVHRR) of NOAA-6 provided the means for detecting synoptic sea surface temperature (SST) variability over the southern Florida and northern Bahamas region. All available 1 km resolution winter data during 1976-1977, 1979-1980, and 1980-1981 were obtained as photographic prints, from which selected digital data were acquired from the Satellite Data Services Division, National Climatic Data Center. These digital data were calibrated to brightness temperatures using internal sensor reference parameters (KOFFLER, 1976; KIDWELL, 1979; HUH and DIROSA, 1981) and were subsequently processed interactively at the Remote Sensing and Image
Subtropical shallow-water envi ronments 737
Processing Laboratory, College of Engineering, Louisiana State University. Geometric corrections of selected data were provided by Rosenstiel School of Marine and Atmos- pheric Sciences, University of Miami.
Satellite-acquired brightness temperatures in the 3.7, 11, and 12 lxm windows are lower than in situ sea surface temperatures primarily as a result of absorption and re- emission of energy by atmospheric water vapor. Additional major contributors to the in situ/satellite temperature offset are subpixel-sized clouds and the "cool skin" of the sea surface. Clear skies, low atmospheric humidity, and elevated wind speeds accompanying cold-air outbreaks minimize these sources of error (HUH, 1976; LEGECKIS, 1978; KATSAR- OS, 1978), thus maximizing spatial information on the thermal structure of the ocean surface. Although satellite-acquired temperatures are from the top few micrometers, previous research has shown cold-air outbreak induced mixing to 30 and 45 m (PRICE, 1976; Hun et al., 1984). For this study, satellite temperatures of bay/bank regions (<10 m) were considered representative of the entire water column, since salinity stratification is small.
AVHRR digital data from both channel 3 (3.7 lam) and channel 4 (11 ~tm) were processed. Although the 3.7 ~tm data should be closer to actual sea surface temperature (CuAHtNE, 1980), under conditions of this study the 11 ~tm channel gave the same values. Since channel 3 data have high levels of sensor noise, only channel 4 data are presented. Previous research under clear-sky, cold-air outbreak conditions has demonstrated an in situ/satellite offset of about 2°C when total precipitable water (TPW) was 1 cm or less (Hun and DIRosA, 1981). Comparisons of satellite temperatures with "surface truth" obtained from Naples, Florida, northeast Florida Bay, and a NOAA buoy indicated offsets of 1.3-3.0°C, corresponding to TPW measurements of 0.6-2.4 cm (total precipi- table water was calculated from radiosonde measurements made at Miami and Key West, Florida). A linear regression of atmospheric TPW and in situ temperature/satellite temperature (To) offset allowed more accurate estimation of actual surface temperatures for the data sets analysed:
To = 0.833 TPW + 1.075 (1)
(R = 0.79).
H E A T F L U X M O D E L C O M P U T A T I O N S
A one-dimensional heat flux model was used to provide an acurate time history of oceanographic changes throughout cold-air outbreak events and to allow identification of the dominant heat flux processes responsible for the resultant water mass modifications. The model is based on the well-known energy balance equation (MvNN, 1966).
OH Ot - Q~ - Q B - Q H - QF, (2)
where H is the heat content of the entire water column calculated from an arbitrary reference temperature, Qs is the input from solar radiation, QB is the net long-wave back radiation, QH is sensible heat flux, and QE is the latent heat flux. Since an advective term was not incorporated, episodic discrepancies (0.5-1.5°C) between in situ and model- predicted temperatures are detected and ascribed to advective influences.
Atmospheric inputs to the model include hourly dry bulb and dew point temperatures,
738 N.D. WALKER et al.
wind speed, solar radiation, and cloud cover. An initial surface temperature and salinity are also required. Hourly data from Miami and Key West were obtained from the National Climatic Center. Comparisons of these data with sporadic measurements and daily maximum/minimum temperatures available for Tavernier, Florida (Fig. 1), revealed that air and dew point temperatures over northeast Florida Bay were most similar to those measured at Miami; however, Key West wind speeds more accurately reflected wind conditions over Florida Bay during cold-air outbreaks. During 1981, a recording thermograph was placed at Lignumvitae Key (Fig. 1), providing an accurate air and water temperature time series during the 1981 episode.
The model incorporated the equations below to compute an hourly heat balance for Florida Bay waters, assuming a thoroughly mixed water column. Sensible (Q,) and latent (QE) heat fluxes were computed from the bulk aerodynamic equations (I-tsu, 1978):
QH = OCpCI4U_. (T~s- Tair), (3)
where p is air density, Cp is specific heat of air at constant pressure, CH is the sensible heat transfer coefficient, U: is wind speed at 10 m. 7~,s is sea surface temperature and T~,i, is dry bulb temperature.
QE = pCEUzL (Q~s- Qair), (4)
where CE is the latent heat transfer coefficient, L is latent heat of vaporization, Q~ is specific humidity of air at ocean surface and Qair is specific humidity of air. Long-wave back radiation was computed from the first part of Berliand's equation, as suggested by REED and HALPERN (1975):
where ea is vapor pressure of air (mb). Values were then corrected for mean cloudiness by the relationship:
QBC = QB(1.0 - 0.83C),
where C is mean cloudiness in tenths (COLON, 1963). Solar radiation measurements were obtained from Miami, Florida, and 93% of the
incoming radiation was assumed to penetrate the sea surface. The coefficients for sensible and latent heat flux were assumed to vary with atmos-
pheric stability (DEARDORFF, 1968; SMITH and BANKE, 1975; FRIEHE and SCHMITr, 1976; SMITH, 1980). Under stable conditions, CH and CE were set to 0.85 × 10 -3 (FRIEHE and SCHMrrr, 1976; Lxu et al., 1979; SMITH, 1980). Since unstable conditions prevailed during chilling episodes and considerable controversy exists over appropriate values for CH and CE, a sensitivity test was performed using two Florida Bay data sets, 19-27 January 1977 and 8-13 January 1981, with CH and CE values ranging from 1.2 to 1.6 × 10 -3. Values of 1.2 x l0 -3 for C/_/ and 1.5 × 10 -3 for CE were found to minimize both the mean temperature differences (measured - predicted = 0.01°C for both time periods) and the r.m.s, errors (0.33 and 0.5°C, respectively) (Fig. 2). Although the data sets were chosen to minimize advection for model tests, both thermograph sites were near tidal passes connecting to the deeper reef tract waters, possibly explaining the relatively high r.m.s. errors.
Subtropical shallow-water environments 739
(-,I
U.lo n- o
OE
W
o
TIHE(HOURSJ ;.00 120.
LI
bJo OCU~
r e - bA n o l l E o
k - - ~ -
=o
Fig. 2.
I I
I
\ ff ' ' fa, \
_'~_ " / V ~ I
I
TIME(HOURS]
Heat flux model test results. The darker line is the model-predicted water temperature and the lighter line is the measurement result. (a) 8 January (19h00) to 13 January (10h00) 1981. The hourly mean difference (measured/model-predicted) =0.01 _+ 0.5. (b) 19 January (20h00) to 27 January (10h00) 1977. Hourly mean difference (measured/model-predicted) = 0.01 _+ 0.33.
Through model tests, it was discovered that resultant temperatures were not very dependent on the initial surface temperature of the shallow water column. The model was run using input data for 8--13 January 1981 for a 1.6 m water column with initial temperatures ranging from 18 to 20°C. Resultant temperatures differed by only 0.3°C (9.3-9.6°C). The test results may have important consequences, since they indicate that reasonably accurate hindcasting can be performed without an accurate knowledge of initial shallow-water temperature conditions. The local atmospheric forcing parameters are, however, very crucial to the results obtained.
Fig. 3. Meteorological history over northeast Florida Bay during the record cold-air outbreak episode in January 1981 and corresponding Florida Bay water temperatures. The black triangle
marks the acquisition time of cloud-free digital data over the study area.
Fig. 4. Thermal infra-red image acquired by NOAA-6 on 19 January 1981 (07h54) illustrating mean brightness temperatures in shallow-water areas. Temperatures extracted along transects
AA' , BB', and CC' are shown in Fig. 5.
Subtropical shallow-water environments 743
THE SEVERE STORMS OF 1981
Air and water temperature data from Lignumvitae Key, northeast Florida Bay, were used to isolate the episodic cold-air outbreak event of 1981, which caused widespread fish mortalities in Florida Bay and patch reef coral deaths along the Florida Reef Tract. Between 8 and 22 January 1981, two major cold-air outbreak periods were distinguish- able; they were separated by a brief maritime air return period (Fig. 3). The high- pressure cells that invaded the study area during Period I (9-13 January) were accompa- nied by northerly winds attaining maximum speeds of 15 m s -I on 12 January, a dry bulb minimum near 0°C, and dew point temperature minimum of-6 .7°C on 13 January. Florida Bay waters responded rapidly to the atmospheric forcing, chilling from 18.8 to 8.7°C by 13 January. Highest wind speeds accompanied frontal passages, after which full effects of the cold, dry continental air masses were evident from the dry bulb minimum and pressure maximum on 13 January (Fig. 3). Soon after the peak atmospheric pressure was attained, winds abated and shifted to the southeast and cloud cover increased. A very rapid rise in Lignumvitae air and water temperatures occurred until the evening of 16 January. The heat flux model results suggest that, in contrast to temperatures near major tidal passes, the mass of Florida Bay water did not exceed 15°C during the maritime return phase.
Period II of the cold-air outbreak episode commenced abruptly as two frontal systems merged over southern Florida on 16 January, resulting in a wind shift through southwest to northwest and a dramatic increase in velocity. The associated cold, dry, high-pressure air mass caused Florida Bay water temperatures to plummet again, reaching 11.0 and 10.7°C on 18 and 19 January.
On 19 January, clear skies prevailed over most of the southern Florida and northern Bahamas region (Fig. 4), allowing an assessment of regional chilling effects. Coldest water temperatures were detected in shallow South Florida environments where mean
25 25
~ 2O
2 15
IO
~ s
o
-5 LAND
5'O
A'
LLES CURRENT ~ CURRENT
LITTLE BAHAMA BANK
DISTANCE ~0~ I
Fig. 5a.
20
15
10
744 N . D . WALKER et al.
251 , b
25 B'
z
2o I 15 A
UDS
10
5
0 ILAND
~o ,6o ~o 26o 2~o 36o 3~0 ,60 DISTANCE (KM)
20
15
10
22
20
18
16
IC I
|
6
C o
RENT"
I i
20 40 | i i
60 80 100
22
20
18
16
14
12
10
DISTANCE {IOA)
Fig. 5. Profiles of brightness temperatures (not corrected) from individual pixels along the line A A ' , BB', and CC'. Correction factors range from 1.6°C over southern Florida environments to
2.6°C over the Bahama Banks.
Subtropical shallow-water environments 745
brightness temperatures of 9.4°C were detected. Using the linear regression of equation (1), a correction of 1.6°C (corresponding to a TPW of 0.6 cm) can be added to give a calculated temperature of 11.0°C. Similar temperatures were observed in Biscayne Bay and near Naples, along the southwest coast of Florida. Mean brightness temperatures over the Little and Great Bahama Banks ranged from 12.5 to 14.0°C (corrected to 15.1 and 16.6°C for a TPW of 1.8 cm). Modification of the continental air mass by the Florida Current and the slightly deeper water columns overlying the Bahama Banks explain the warmer resultant temperatures in comparison with the southern Florida region. The nearly isothermal temperature structure of bay and bank waters and the dramatic thermal fronts generated are clearly illustrated by SST transects taken from the 19 January image (Fig. 5). The I°C km -1 front traversing the Florida Reef Tract was the most impressive (Fig. 5C).
As an additional case, during a record cold-air outbreak in January 1977, mean temperatures of 13.5 and 15.4°C (Little and Great Bahama Banks, respectively) were produced (WM, KER, 1982), resulting in both fish and coral mortalities east of Andros Island, Great Bahama Bank (ROBERTS et al., 1982; WALKER, 1982).
A T M O S P H E R I C F O R C I N G A N D W A T E R MASS T R A N S F O R M A T I O N S
High wind speeds, cold, low-humidity air, and cloud-free skies maximize heat loss to the atmosphere from all water surfaces during cold-air outbreak episodes. Heat flux model results demonstrate that the greatest percentage of heat loss resulted from latent heat flux (evaporation) (Fig. 6a). Since latent heat flux maxima occur primarily when conditions of high wind speeds and low humidity prevail, it is not surprising that peaks in latent heat flux coincided with spikes in the wind speed record during low humidity conditions, notably from 11 to 14 and 17 to 18 January (Figs 6a, c, d). As humidity increased, from 15 to 16 January and 20 to 22 January, latent heat flux diminished sharply. Evaporation is clearly the most efficient mechanism for cooling of the sea during a cold-air outbreak. Turbulence within the marine boundary layer augments negative oceanic heat fluxes by continually replacing modified air at the ocean-air interface with colder, drier air. Model-computed daily heat fluxes and radiation values for Florida Bay and Florida Current waters on 2 days during the January 198l episode are presented in Table 1. Heat loss due to back radiation (QB), a weak but still significant process, reaches its maximum efficiency only under conditions of low atmospheric humidity and reduced cloud cover. It constitutes a measurable part of the total heat flux under those conditions during periods of low winds. It is, however, probably more than compensated for by increased solar radiation during daylight hours. Sensible heat flux and back radiation values were similar over Florida Bay waters; however, over the Florida Current, heat loss from sensible heat was much greater.
Insolation was found to be a significant energy source in the subtropics even during winter, as shown in Fig. 6b and Table 1. When negative heat fluxes diminished as outbreak conditions waned, insolation caused rapid warming of shallow bay and bank waters.
The greatest 24-h heat loss during January 1981 occurred on 12-13 January. During this 24-h period the heat loss from Florida Bay was calculated to be 18.2 x 10 ~ J m -2 ( 4 0 0
cal cm-2), whereas the heat loss from Florida Current waters during this same time, assuming a 100-m surface mixed layer, was 75 x 106 J m -2 (1793 cal cm 2). These heat
746 N.D. WALKER et al.
5.0 I % 4 .0
3.0!
~ 2.0 } 1.0
0
"1.0
7.0
6.0
5.0
4.0 o x 3.0
=E 2,0
"1.0.
0
LIGNUMVITAE KEY, FLORIDA BAY
LATENT HEAT FLUX - - a
SENSIBLE HEAT FLUX
ll- -t
SOLAR RADIATION - -
BACK RADIATION - - -
/
t, lt-,,J 15 I" WIND SPEED c
0
I S SPECIFIC HUMIDITY ,.~
'~ i .o x
0 .5
25 TEMPERATURES AIR ( / .
Q
, ~ -. WATER ~ ' ' l \ ' \ J
I0
5
O!
10700] DAYS JANUARY 1981
Fig. 6. Heat fluxes, solar and back radiation, meteorological parameters, and resultant shallow- water temperature changes for northeast Florida Bay.
flux rates are similar to those obtained by NOWLIN and PARKER (1974) and HUH et al.
(1984) for coastal and offshore waters of the northern Gulf of Mexico. Although a greater quantity of heat was lost from deep offshore waters, shallow bay/
bank and coastal waters experienced the most rapid and dramatic temperature changes due to their depth-limited heat storage capacities (thermal inertia). A good demon- stration of this thermal inertia is the 8-13 January period, when waters of Florida Bay, the Reef Tract, and the Florida Current chilled 9.4, 6.0, and 0.6°C (Table 2). From 13 to 19 January the deeper Reef Tract and Florida Current waters continued to chill in contrast to those in Florida Bay. The net result was that shallow areas experienced rapid and extreme cooling but did not remain in this super-chilled state unless intense atmospheric forcing was sustained. Slightly deeper environments chilled more slowly but experienced prolongation of their modified states. Model-computer water transforma- tions, including temperatures salinity, density, and evaporative water loss, are given in Table 2 for the three different water masses.
Depth-dependent temperature changes coincident with and subsequent to cold-aik outbreak events are graphically illustrated in Fig. 7. Depths from 2 to 10 m were tested with the heat flux model, assuming an initial input temperature of 18.8°C (Florida Bay data). Temperature differences between the 2- and 10-m depths reached a maximum of 5.5°C on 13 January. They converged on 17 January as the shallower waters regained heat from solar radiation. After initiation of the second outbreak phase, temperature differences increased to 3.7°C by 18 January at 0800 local time. A thermal infra-red data set acquired at this time (Fig. 8) clearly illustrates depth-controlled chilling of southern Florida waters, since temperatures steadily increased in a westerly direction correspond- ing to the bathymetry. In this enhancement, isotherms in degrees Celsius are displayed by gray shades, ranging from lightest gray (11-12°C) along the Florida west coast to dark gray (22-23°C) in the Florida Current. The model-predicted temperature of 11.7°C for
\ \ ............ . / s ~ . - - \ ' ~ , " ~ "'" 6 re
I 4m \ / \
\ i I k
I i i i I I I I I i
9 10 II 12 13 14 15 16 17 18
JANUARY 1981
Fig. 7. Model-computed temperature changes as a function of water depth from 2 to 10 m during part of the 1981 outbreak episode. A Florida Bay temperature of 18.8°C was used as an input temperature. The model prediction ends on 18 January (08h00) coincident with image
acquisition time over south Florida (Fig. 8).
Fig. 8. Satellite data acquired by NOAA-6 on 18 January 1981 (08h17) of the southern Florida area illustrating depth-controlled chilling. The 20-m bathymetric contour is depicted by a solid black line. Isotherms ranging from 11-12°C to 22-23°C (corrected temperatures) are illustrated
with gray shades ranging from white to dark gray.
Fig. 9. NOAA-6 thermal infra-red data from (a) 4 February 1980 (07h45) and (b) 18 January 1981 (08h17). The enhancement is similar to that used for Fig. 8; however, uncorrected brightness temperatures are shown (add 1.6 for actual temperatures). Arrows indicate the
preferential routes for advection of chilled Florida Bay waters onto the Florida Reef Tract.
Fig. 10. NOAA-6 satellite data acquired on 30 January 1981 (20h00) over southern Florida environments. The same enhancement and geometric correction were applied as for 18 January 1981 data (Fig. 9). Warmest Florida Bay waters were 19-20°C and those north of Key West,
along the 20-m contour, were 15-16°C.
Subtropical shallow-water environments 753
Florida Bay waters (2 m) agrees closely with corrected satellite temperatures of 11-12°C. The 20-m bathymetric contour (solid black line) intersects the 15-16°C isotherm west of Florida Bay. An interesting feature to note is that the satellite-determined isotherms do not run exactly parallel to the 20-m contour along the Florida west coast. Water overlying this depth contour is 2°C colder in northern areas than in the south, near Key West. This temperature difference could result from more intense atmospheric forcing in the north and adjacent to the Florida land mass or from the existence of pre-outbreak meridional temperature gradients along the 20-m contour, or from both. Extensive cloud formations evident over the Florida Current on 18 January 1981 (Fig. 8) result from the large quantities of heat and water lost from deep ocean areas as a result of intense cold- air outbreak forcing.
C O L D - A I R O U T B R E A K I N D U C E D C I R C U L A T I O N AND E F F E C T S ON S I I A L L O W
R E E F E N V I R O N M E N T S
Discharge of Florida Bay waters through tidal passes onto the Florida Reef Tract was identified in several satellite images coinciding with cold-air outbreak events. Largest and coldest "plume" features were noted when strong northwesterly to northerly winds prevailed in conjunction with ebbing tide or low water in Florida Bay (Figs 9a,b). Cold "plumes" were particularly conspicuous on 4 February 1980, since the adjacent reef tract waters were several degrees warmer.
Having exited the shallow confines of Florida Bay, the relatively cold, dense water masses would have a tendency to sink and flow near the bottom, seeking their density compensation depths. Under conditions of light winds, satellite-observed surface tem- peratures could actually underestimate the areal extent of chilled waters beneath the surface. Current meter measurements acquired during January 1981 in Hawks Channel, along the Florida Reef Tract (Van Waddell, personal communication, 1982) showed that the prevailing currents during cold-air outbreak forcing flowed along the reef tract in a south-southwesterly direction (220°T), paralleling the depth contours. A patch reef coral community at Elliot Key in the northern Reef Tract suffered considerable thermal stress and death during January 1981 (WALKER et al. , 1982). The low-temperature water mass probably originated in Biscayne Bay, a few kilometers to the north. Cold-air outbreak processes most probably provide a major limiting influence to coral distribution and vitality in this region, as a temperature of 16°C is considered stressful to reef-building species (MAYOR, 1915) and temperatures below this were generated during the three winters studied (WALKER, 1982). Coral mortalities caused by thermal stress have been reported in the recent years of 1970, 1977, and 1981 (HuDsON et al. , 1976; DAVIS, 1982; ROBERTS et al. , 1982; WALKER et al. , 1982; HUDSON, personal communication). The sites of these reported mortalities are shown in Fig. 1.
On 30 January 1981, cloud-free satellite data over southern Florida were available, allowing an assessment of post-outbreak temperature structure (Fig. 10). The shallowest waters of southern Florida bays had warmed from 11 to 19.5°C by this time. Along the 20-m contour, however, water temperatures were 15-16°C, essentially unchanged since 18 January (Fig. 8). Elongation of the chilled surface features from the southwest tip of Florida to the west indicated net surface transport from north to south and from east to west. Where the shelf shoals, forming the Florida Key chain, bay and shelf waters were diverted to the west past Dry Tortugas. During January 1977, chilled Florida Shelf water
754 N . D . WALKER et al.
resulted in the demise of 90% of a 220-hectare coral reef stand down to its depth limit of 15 m (WALKER, 1982). An extremely cold water filament (narrow surface plume) was evident on 18 January 1981 (Fig. 9b) very close to the site of the coral kill. Recurrent modification of shelf waters by winter forcing in conjunction with alongshelf and offshelf movement induced by northerly winds produces a cold, dense water mass on the shelf that was observed during each winter investigated. VUKOVICH et al. (1979) showed that oceanic thermal front meanders in this area are not simply surface features, but usually extend vertically to the ocean floor, which in an extreme case was 700 m. The 30 January image also illustrates cold Florida Shelf water south of Dry Tortugas, where it has been entrained by the Florida Current.
C O N C L U S I O N S
Temperature-corrected thermal infra-red satellite data, a one-dimensional heat flux model, and supporting ground truth measurements allowed an evaluation of the thermal responses of shallow subtropical bay/bank and continental shelf waters to severe winter cold-air outbreak events. The most rapid temperature change occurred in the bays of southern Florida. From 9 to 13 January 1981, northeast Florida Bay waters (1.6 m) chilled from 18.8 to 8.7°C, an average rate of 2.24°C per day, although the rates were even higher for short periods within this episode. On 19 January, during Period II of the cold-air outbreak episode, satellite data revealed similar temperature structure in Florida Bay, Biscayne Bay, and along the southwest Florida coast. Downwind of the Florida Current, vast expanses of water overlying Little and Great Bahama Banks had reached mean temperatures of 15.5 and 17.0°C, respectively. Over the Bahama Banks colder temperatures developed during a record cold-air outbreak in January 1977.
A simple, one-dimensional heat flux model, used with locally available meteorological data, limited oceanographic measurements, and water depth information, successfully simulated the cooling/heating processes that created the sea surface temperature variabi- lity.
Heat loss to the atmosphere was maximized during the outbreak episodes by high wind speeds, cold, low-humidity air, and clear skies. Although the highest heat flux rates occurred over deep warm-water areas, such as the Florida Current, temperature changes were most rapid in shallow areas with limited thermal inertia. Latent heat transfer contributed most to the negative heat fluxes and was greatest when wind speeds increased under low-humidity cold-air outbreak conditions.
Modeling indicated that resulting water temperatures of shallow bays are almost independent of initial (boundary conditions of) temperature but strongly dependent on the severity (temperature, humidity, wind speed, and duration) of the cold-air outbreak event.
The strong northerly winds accompanying frontal passages are effective in moving chilled waters out of their shallow generation areas, thus prolonging their modified thermal states. The rapid and strong fluctuations of barometric pressure associated with winter storms result in sea level changes of about 1 cm mb -~ pressure change. This inverse barometer effect also forces surface water exchanges between shallow and deep areas. The southwest shelf of Florida is influenced considerably by cold-water advection during winter from both shallower Florida Bay to the east and the shelf to the north. Satellite imagery collected over several winters indicates a massive cold-water body in
Subtropical shallow-water environments 755
Cold A i r O u t b r e a k s over the
Southern Florida and Northern Bahama Region
1. High Wind Speeds 2. Cold, Low Humidity Air 3. Increased Back Radiation
Through Clear Skies
Heat Loss From Water Surfaces
Inhibition of Coral Growth on Platform Interiors
Rapid Cooling of
Shallow Lagoon and Shelf Waters
~!Bean 1 Movimenl
Thermohaline Induced
Circulation
Inverse Barometer Effect
1. D i s c o n t i n u o u s Ree f D i s t r i b u t i o n a l o n g E a s t e r n M a r g i n s .
2. Absence o f Reef Development a l o n g Western Platform M a r g i n s .
Fig. 11. Effects of winter cold-air outbreak processes on southern Florida/northern Bahama lagoon waters and coral reef distribution.
756 N.D. WALKER et al.
the vicinity of Dry Tortugas, which we infer could have resulted only from the accumulation of chilled shelf waters generated and forced south by recurrent winter cold- air outbreak processes.
These winter processes exert considerable influence on coral reef vitality and on their distribution within the carbonate depositional systems of south Florida and the northern Bahamas. Reefs within the region grow along eastern bank margins (Fig. 1). In these locations they are usually bathed with warm oceanic waters. If overexposure to chilled bay or bank waters occurs as a result of wind- or tide-induced circulation, a coral kill can result. A schematic summary of the interaction between atmospheric cold fronts and these reef-platform environments is presented in Fig. l 1.
Acknowledgements--This research was sponsored by the NASA Graduate Student Researcher's Program through the Jet Propulsion Laboratory in Pasadena, California. Additional support was provided by the Marine Sciences Department, Louisiana State University, and the Office of Naval Research, Arlington, Virginia. Many thanks are extended to Bob Evans of Rosenstiel School of Marine and Atmospheric Sciences, University of Miami, for geometric correction of selected satellite data, and to Don DiRosa of the College of Engineering, Louisiana State University, for invaluable assistance in image processing. Harold Hudson of USGS, Miami, Florida, and Jeannie Parks of Long Key State Recreational Area are thanked for supplying water temperature data.
R E F E R E N ( ' E S
ANGELOVIC J. W. (1976) Environmental studies of the south Texas outer continental shelf, 1975. Physical Oceanography, V. II, U.S. Department of Commerce, National Technical Information Service PB-283 872.
CHALtINE M. T. (1980) Infra-red remote sensing of sea surface temperatures. In: Remote sensing of atmospheres and ocean, A. DEEPAK, editor, Academic Press, New York, pp. 411-435.
COLON J. A. (1963) Seasonal variations in heat flux from the sea surface to the atmosphere over the Caribbean Sea. Journal of Geophysical Research, 68, 1421-1430.
DAV1S G. E. (1982) A century of natural change in coral distribution at the Dry Tortugas: a comparison of reef maps from 1881 and 1976. Bulletin of Marine Science, 32,608-623.
DEARDOREE J. W. (1968) Dependence of air-sea transfer coefficients on bulk stability. Journal of Geophysical Research, 73, 2549-2557.
FRIEHE C. and K. SCHMIrr (1976) Parameterization of air-sea interface fluxes of sensible heat and moisture by the bulk aerodynamic formulas. Journal of Physical Oceanography, 6, 801-809.
GARWOOD R. W., R. FETT, K. RABE and H. BRANDLI (1981) Ocean frontal formation due to shallow-water cooling effects as observed by satellite and simulated by a numerical model. Journal of Geophysical Research, 86, 11000-11012.
HstJ S. A. (1978) Micrometeorologieal fluxes in estuarines. In: Estuarine transport processes, B. KJFRFVE, editor, Belle W. Baruch Library in Marine Science, 7, 125-134.
HUDSON J. H., E. A. SHINN, R. B. HALLEY and B. LIDZ (1976) Autopsy of a dead coral reef (abst.). Btdh'tin oJ the American Association of Petroleum Geologists, 60, 682.
HUH O. K. (1976) Detection of oceanic thermal fronts off Korea with the Defense Meteorological Satellites. Remote Sensing of the Environment, 5, 191-213.
Hu~t O. K. and D. DIROSA (1981) Analysis and interpretation of TIROS-N AVHRR infra-red imagery, western Gulf of Mexico. Remote Sensing of the Environment, l 1,371-383.
HuH O. K., L. J. ROUSE, Jr and N. D. WALKER (1984) Cold air outbreaks over the northwest Florida continental shelf: heat flux processes and hydrographic changes. Journal of Geophysical Research, 89, 717-726.
HuH O. K., Win. J. WISEMAN, Jr and L. J. ROUSE, Jr (1978) Winter cycle of sea surface thermal patterns, northeastern Gulf of Mexico. Journal of Geophysical Research, 83, 4523-4529.
KATSAROS K. B. (1978) The aqueous thermal boundary layer. In: Passive radiometry of the ocean. F. J. R. GOWER, editor, D. Reidel Publishing Company, 359 pp.
KIt)WELl, K. B. (1979) NOAA Polar orbiter data (TIROS-n) users Guide, Preliminary version. Satellite Data Services Division, National Climatic Center, Washington, D.C., 118 pp.
KOFFLER R. (1976) Digital processing of NOAA's Very High Resolution Radiometer (VHRR) data. Proceedings of the 27th International Astronautics Congress Reference 76-209, International Astronau- tics Federation, Anaheim, CA, 7 pp.
Subtropical shallow-water environments 757
LEGECKIS R. (1978) A survey of worldwide sea surface temperature fronts detected by environmental satellites. Journal of Geophysical Research, 83, 4501-4522.
LIU W. T., K. B. KATSAROS and J. A. BUSINGER (1979) Bulk parameterization of air-sea exchanges of heat and water vapor including the molecular constraints at the interface. Journal of Atmospheric Science, 36, 1722-1735.
MAYOR A. G. (1915) The lower temperature at which reef corals lose their ability to capture food. Yearbook, Carnegie Institute, Washington, 14,212 pp.
MUNN R. E. (1966) Descriptive micrometeorology. Academic Press, New York, 245 pp. NOWHN W. D. and C. A. PARKER (1974) Effects of a cold-air outbreak on shelf waters of the Gulf of Mexico.
Journal of Physical Oceanography, 4, 467-486. PRI~'E J. F. (1976) Several aspects of the response of shelf waters to a cold front passage. Memoires de la Societe
Royale des Sciences de Liege, 6X, 201-208. REED R. K. and D. HALPERN (1975) Insolation and net-long wave radiation off the Oregon coast. Journal of
Geophysical Research, 80, 839-844. ROBERTS H. H., L. J. ROUSE, Jr, N. D. WALKER and J. H. HUDSON (1982) Cold-water stress in Florida Bay
and northern Bahamas: a product of winter cold-air outbreaks. Journal of Sedimentary Petrology, 52, 145-155.
SMmt S. D. (1980) Wind stress and heat flux over the ocean in gale force winds. Journal ~f Physical Oceanography, 10, 709-726.
SMITH S. D. and E. G. BANKE (1975) Variation of the sea surface drag coefficient with wind speed. Quarter(v Journal of the Royal Meteorological Society, 101,655-673.
VUKOVlCtt F. M., B. W. CRISSMAN, M. BUSHELL and W. KING (1979) Some aspects of the oceanography of the Gulf of Mexico using satellite and in situ data. Journal of Geophysical Research, 84, 7749-7768.
WALKER N. D. (1982) Physical responses of southern Florida and northern Bahama lagoon waters to severe cold air outbreaks and effects on hermatypic reef corals. MSc. Thesis, Louisiana State University, Baton Rouge, 114 pp.
WAIKER N. D., H. H. ROBERTS, L. J. ROUSE, Jr and O. K. HUH (1982) Thermal history of Ireef-associated environments during a record cold-air outbreak event. Coral Reefs, l, 83-87.
