Response of remotely-sensed sea surface properties to Hurricane Dennis in the Gulf of Mexico
Chuanmin Hu and Frank E. Muller-Karger
Institute for Marine Remote Sensing, College of Marine Science, University of South Florida
140 Seventh Avenue, South, St. Petersburg, FL 33701
(727)5533987, hu@marine.usf.edu
Abstract
Sea Surface Reflectance (SSR) and Sea Surface temperature (SST) were assessed in the eastern
Gulf of Mexico in early July 2005 after passage of hurricane Dennis, a category 4 storm, using
data from the MODIS, SeaWiFS, and AVHRR satellite sensors, historical hydrographic and
nutrient surveys, and a simple vertical water column mixing model. The hurricane induced large
SSR and SST changes in waters shallower than 50 m. Here, a strong sediment resuspension
event was accompanied by an increase in reflectance in the visible MODIS ocean color channels,
but chlorophyll concentration also increased as observed in MODIS fluorescence line height
(FLH) imagery. The short response time (< 2 days) implied that the blooming of phytoplankton
may have been due to injection of nutrients from near-bottom waters and seeding of the water
column by resuspended benthic and/or near-bottom algae, as opposed to simply new primary
production in the overlying water column. SSR changes in < 50-m waters recovered to normal
conditions after about 10 days. For waters > 50 m deep, where significant surface cooling was
observed, there was no apparent change in SSR due to the deep nutricline. The SST changes
occurred almost entirely along the right side of the hurricane track (–1.960.66oC for an area of
158,600 km2)
Introduction
Dramatic surface cooling and color changes in the ocean have been documented after a
hurricane’s passage (Stramma et al., 1986; Shay et al., 1992; Hoge and Lyon, 2002; Babin et al.,
2003; Lin et al., 2003; David and Yang, 2004; Walker et al., 2005; others). In the coastal ocean,
hurricanes can also cause sediment resuspension/transport as well as coastal erosion and flooding.
Changes in sea surface reflectance (SSR) may be due to increased surface primary productivity
resulting from nutrients supplied by deep mixing or upwelling (Babin et al., 2003) and/or
entrainment of the deep chlorophyll maxima (Walker et al., 2005). Some investigators have
proposed that SSR changes are more likely due to enhanced production of colored dissolved
organic matter (CDOM) (Hoge and Lyon, 2002). Over shallow waters, sediment resuspension
and river plume injections have been observed (Yuan et al., 2004). Yet, to date there are no
studies of concurrent in situ and satellite observations of physical and bio-optical properties in
the immediate aftermath of a hurricane over a shelf environment.
In early July 2005, Hurricane Dennis (category 4) swept across the eastern Gulf of Mexico
(EGOM) and the west Florida shelf (WFS). In addition to sediment resuspension and surface
cooling, other unique features were observed in satellite-derived SSR and SST. Specifically, 1) a
sharp contrast in SST response on either side of the 30-m isobath on the shelf; 2) a large increase
1
in the solar stimulated fluorescence over the shallow portions of the shelf; and 3) negligible
changes in SSR over deeper shelf waters, where surface cooling was maximum.
The upper ocean heat and temperature response to Hurricane Dennis has been modeled by Morey
et al. (2006), yet the unusual physical and biological features have not been documented. Here
we report on these unique changes and use concurrent satellite data, historical field surveys, and
a simple mixing model to explain the underlying driving mechanisms. We hypothesize that
hurricane perturbation over shallow waters reaches the summertime thermocline and nutricline
depths, and indeed affects nutrients and algae lving on the bottom in waters shallower than about
50 m.
Data Sources
SST data were obtained from the Advanced Very High Resolution (AVHRR on NOAA Polar
Orbiters 12, 15 and 17) and the Moderate resolution Imaging Spectrometer (MODIS on the Terra
and Aqua satellites) sensors. SSR and related products (chlorophyll concentration or Chl,
fluorescence line height or FLH) were obtained from the Sea-viewing Wide-Field-of-view
Sensor (SeaWiFS) and MODIS. All satellite data were captured and processed using the most
updated software and algorithms at University of South Florida.
Historical CTD and nutrient data in the western Gulf of Mexico were obtained from the
Northeastern Gulf of Mexico (NEGOM) program (1998-2000, Hu et al., 2003), the EcoHAB
program (1998-2001), and a field survey conducted to the Florida Current in August 2004 (Hu et
al., 2005a). Data for Hurricane Dennis, including the storm category, eye location, and sustained
and maximum wind speeds were obtained from the National Hurricane Center (National Oceanic
and Atmospheric Administration, U.S.A.).
Results
SST and SSR Response to Hurricane Dennis
Fig. 1 [NOTE: In Fig 1a the arrow showing the direction of the eddy above the LC in the SE
GOM is a bit too much extended to the east - I would actually extend the LC boundary to the
north and show that arrow going the other way - I think this may be more real. Otherwise there is
no reason the filament should show the edge of a cyclonic eddy. This would only happen in an
anticyclone]shows that the surface cooling is almost entirely located to the right of the hurricane
track. The average cooling in near the core of the cold plume (area outlined in Figure 1) was –
1.960.66oC (minimum was < – 4oC). In total, an area of ~158,600 km2) was defined by the
difference between SST on July 11 (26.470.58oC) and the 1-week average SST derived for the
week before Dennis (28.430.37oC). However, for the < 30 m shelf waters between Cape San
Blas and Dry Tortugas (an area ~ 36,000 km2), the SST response was smaller (-1.660.60oC,
from 29.230.44 to 27.570.59). Before Dennis the <30-m waters were ~0.8oC warmer than the
deeper area. This likely led partially to the sharp SST gradient across the 30-m isobath after
Dennis (Fig. 1b).
2
The significant sediment resuspension event over the entire shallow (< 50 m) shelf (Fig. 2b)
illustrates the intensity with which hurricane Dennis mixed shelf waters. Accompanying the
resuspended sediment we observed a significant enhancement of the MODIS FLH signal (Fig.
2d). The highest FLH values were limited to waters shallower than 50 m. The MODIS FLH data
are most sensitive to detect larger changes in chlorophyll-a (Chl > 0.1 mg m-3; Hu et al., 2005b).
[was the SeaWiFS CHL screwed up here due to the sediment?] Except for a small filament along
the edge of a Loop Current eddy near 26°N, 84-86W (Fig. 2d), there was no detectable change in
FLH for waters > 50 m depth, even in the area of significant surface cooling. SeaWiFS Chl timeseries in deep waters, not affected by sediment resuspension, did not reveal any changes after
Dennis except for the filament around the edge of the eddy in the southern Gulf. Further, there
was no significant change in either Rrs(443) or Rrs(555) (Fig. 3). The slight decrease in Rrs(443)
on July 13 (about –0.001 sr-1) is perhaps primarily due to an imperfect atmospheric correction,
because of the inherent atmospheric correction uncertainty of 0.00064 sr-1 at this wavelength
(Gordon, 1997). Clearly, for deep waters (> 50 m depth), bio-optical and biological responses are
minimal. This is in contrast to previous reports suggesting that hurricanes at these latitudes
stimulate phytoplankton blooms in their wake.
Forcing
How much force is neeed to cause the unusual SST and ocean color response in the wake of
Dennis? The cyclonic wind forcing will result in deep mixing and upwelling, particularly to the
right-side of the hurricane's track. If the water is already well mixed down to the bottom prior to
the passage of the storm, SST changes should be small. Data collected in the past around the
same time of the year (Fig. 4) suggest that the surface layer due to summer stratification is about
20-30 m for the eastern Gulf of Mexico, at least over most of the shelf and out to the 1000 m
isobath. If mixing were to occur to greater depths, a change should be expected in SST over
these regions. Indeed, significant cooling was observed in the outlined area in Fig. 1b. Where the
bottom is shallower than 30 m, hurricane perturbation of the SST should be relatively small.
The upward displacement of the seasonal thermocline in waters deeper than 30 m by mixing
required to make a change in SST can be approximated as (Price et al., 1994; Babin et al., 2004):
 = /(wfUh)
[what is this formula? How is this lifting - at what depth - any depth? Or is this just the
sea surface elevation-eta is normally used to compute sea surface elevation? How do you know
this will be the lifting of isopycnals at 30 m?? If the formula were for a mixed layer depth, it may
be clearer possibly?] where  is the wind stress derived as a function of wind speed (U10, m s-1)
and air density (a = 1.26 kg m-3) as =a(0.49+0.065 U10)10-3 U102, w is the water density
(1020 kg m-3), f is the Coriolis parameter, and Uh (m s-1) is the hurricane transit speed.
Using data from the National Hurricane Center, we estimated that for the average wind speed of
56.17.9 m s-1 during Dennis’ passage, the isopycnal displacement in the eastern GOM was
38.319.3 m. The associated average surface cooling, estimated as the difference between the
average temperature for the water depth of 38.3 m and the pre-hurricane 30 m (Fig. 4), is 26.9 –
29.7 = 2.8oC. Considering the size of the cooling area (158,600 km2) and factors such as heat
loss, inhomogenous vertical mixing, which we don't estimate but which also affect SST, this
result agrees reasonably well with that observed from the satellites.
3
Why was there no apparent change in SSR in the area of maximum cooling? Nutrient profiles
collected by the NEGOM program along a west-east transect from Tampa Bay to the 1000-m
isobath suggest that the nutricline for this time of year in the eastern GOM is typically deeper
than 50-60 m. Above this depth, nitrate concentration is below 0.5 M and frequently
undetectable (Belabbassi et al., in press). The mixing did not penetrate the nutricline and
therefore little nutrients were available to stimulate phytoplankton growth near the surface. One
may similarly infer that CDOM concentration above the nutricline was low and homogeneously
mixed, since if there had been an upward injection of CDOM we would have observed a drop in
Rrs(443).
Discussion
Walker et al. (2005) observed average surface cooling of about –1.9oC on both sides of a
hurricane track in the NGOM near the Mississippi Delta, and a filament of enhanced SeaWiFS
Chl only 1.5 days after the hurricane passage (from 0.36 to 0.50 mg m-3). Two days later the Chl
in the patch reached a maximum concentration of Chl=0.81 mg m-3. Similarly, Davis and Yan
(2004) observed Chl-rich coastal filaments after a hurricane’s passage using SeaWiFS data.
More generally, Chl enhancement in the wake of a hurricane in the open ocean was reported by
Babin et al. (2004). However, Hoge et al. (2002) concluded that the color change in the open
ocean was due to increased CDOM rather than Chl. Here, in the deep Gulf of Mexico we
observed a change in Chl as enhanced MODIS FLH and SeaWiFS Chl only in a filament along
an edge of an eddy (Fig. 2d). SeaWiFS Chl showed an increase from ~0.12 to 0.25 mg m-3 in this
filament. We believe that the mixing due to the hurricane reached the uplifted thermocline and
nutricline that typically wraps around the edge of the anticyclonic eddy cloacted immediately to
the north of the Loop Current in the southeastern Gulf. [NOTE: (same comment as above on Fig
1a and eddy direction - can you change?] We did not observe other changes in the SSR or Chl in
deep water, even in the area of maximum SST response. This, again, is likely the result of
mixing to depths shallower than about 60 m[check that this is correct by mixing formula above;
this is how I interpret what you said with the formula]. Clearly, hurricanes do not necessarily
induce color or chlorophyll changes every time they pass over deep waters as implied in previous
studies. Further research is needed to verify the bio-optical responses, particularly in terms of
reliable bio-optical algorithms and timely field surveys to examine biogeochemical changes in
the wake of a hurricane.
Similar to the CDOM interference to the Chl bio-optical algorithm for the open ocean, sediment
resuspension over the shallow shelf (0-50 m, Fig. 2b) can also cause overestimates of Chl if a
band-ratio algorithm is used. The MODIS FLH provides an alternative in the form of a reliable
index of chlorophyll. In highly turbid waters, significant amount of sediments may cause
overestimates in the FLH signal. Overestimates for sediments concentrations of 40-50 mg L-1
(based on Fig. 3b - [in Fig 3b I don't see any sediment concentrations or red color estimates for
FLH estimates - only the green reflectance; what are you referring to? Reword sentence.]) are
expected to be much smaller and therefore could not account for the observed significant
increase in FLH (Fig. 2d). Then, does the increased FLH over the <50 m shelf (Fig. 2d) suggest
increased phytoplankton biomass?
4
The EcoHAB data showed that surface waters where the shelf is deeper than 30 m are relatively
nutrient depleted (N, P, Si). Slightly higher nutrient levels are found in the bottom layer [how
much?]. In waters between the 0-20 m isobaths both P and Si are typically higher (P ~ 0.05 – 0.3
M; Si ~ 1 – 8 M) and vertically homogenous. However, there may be significant amount of
nutrients in bottom sediment pore waters. These may be released into the overlying water
column by hurricane perturbation. The EcoHAB data also showed high Chl (~ 0.3 – 1 mg m-3)
are typical over the shelf in the bottom layer where the shelf is shallower than 40-60 m,
especially when the surface layer has low Chl (~ 0.1 mg m-3). The day after the passage of
hurricane Dennis (July 11), high FLH was observed (image not shown due to partial sun glint).
We believe that the bottom Chl layer and possibly benthic algae that is prevalent throughout the
shallow shelf (Gabe Vargo, USF, personal comm.) were resuspended and stimulated, leading to
the increased FLH signal after Dennis.
Sediment resuspension to the surface should be largely restricted to bottom depths < 50 m for the
average wind speed of 56.17.9 m s-1 (category 4 hurricane) and average storm translation speed
of 7.73.0 m-1. Similar observations were made during three hurricanes that affected the west
Florida shelf during summer 2004 (Charley, Frances, and Jeanne). Sediment resuspension events
were always restricted to bottom depths < 50 m.
Why did the surface cooling occur almost entirely to the right side of Dennis’s track (Fig. 1b)? In
the past, much emphasis has been placed on the asymmetric distribution of wind and physical
forcing around the periphery of these storms. However, there is also spatial heterogeneity in the
hydrographic properties in the eastern GOM over the shelf. Over deeper waters, most of the
waters to the left of the hurricane track are in the Loop Current and other deep Gulf waters (Fig.
1a). Historical data (e.g., Fig. 4b, profile labeled FC) shows that the water in the Florida Current
is warmer to much deeper depths than outside the current, and that the thermocline is at least 60
m deep. Therefore it is more difficult to mix to depths that will lead to surface cooling. The
shape of the cooled area (outlined at black line on both Figs 1a and 1b) show that differences in
the hydrographic properties inside and outside the LC and its eddies are an important factor
leading to spatial changes in the observed cooling patterns. These patterns are clearly of interest
in defining the heat content of surface waters that can provide heat to sustain storm strength.
Conclusion
Unusual patterns in hurricane-induced surface cooling and SSR changes occurred after Hurricane
Dennis’ passage in the eastern Gulf of Mexico in early July 2005. These include significant
surface cooling to the right of the hurricane track, minimal cooling in waters where the bottom is
< 30 m over the shelf, and significant sediment resuspension over bottom depths <50 m shelf
accompanied with a strong increase in the solar stimulated fluorescence. There was negligible
changes in SSR and Chl in the area of maximum cooling over deep shelf and Gulf waters. These
responses are different from those reported in the recent literature examining SST and
chlorophyll changes in the wake of hurricanes. The differences can be explained by the threedimensional hydrographic and nutrient structure in waters near the west Florida shelf.
Acknowledgement
5
Funding was provided by NASA through Grants NNS04AB59G and NAG5-10557 to F. MullerKarger and C. Hu. Dr. Douglas Biggs and Ms. Leila Belabbassi (Texas A&M Univ.) provided
historical hydrographic and nutrient data from the MMS-sponsored NEGOM project. We thank
Dr. Robert Masserini (USF) for providing historical EcoHAB data on the WFS.
Reference
Babin, S. M., J. A. Carton, T. D. Dickey, and J. D. Wiggert (2004). Satellite evidence of
hurricane-induced phytoplankton blooms in an oceanic desert. J. Geophys. Res. 109, C03043,
doi:10.1029/2003JC001938.
Belabbassi, L., P. Chapman, W. D. Nowlin, Jr., A. E. Jochens, and D. C. Biggs. (in press).
Summertime nutrient supply to near-shore waters of the Northeastern Gulf of Mexico: 1998,
1999, and 2000. Gulf of Mexico Science.
Davis, A., and X. Yan (2004). Hurricane forcing on chlorophyll-a concentration off the northeast
coast of the U.S., Geophys. Res. Lett., 31, L17304, doi:10.1029/2004GL020668.
Gordon, H.R. (1997). Atmospheric correction of ocean color imagery in the Earth Observing
System era. J. Geophys. Res., 102, 17081-17106.
Hoge, F. E., and P. E. Lyon (2002). Satellite observation of Chromophoric Dissolved Organic
Matter (CDOM) variability in the wake of hurricanes and typhoons. Geophys. Res. Lett. Vol.
29(19), 1908, doi:10.1029/2002GL015114.
Hu, C., Muller-Karger, F. E., Biggs, D. C., Carder, K. L., Nababan, B., Nadeau, D., and
Vanderbloemen, J. (2003), Comparison of ship and satellite bio-optical measurements on the
continental margin of the NE Gulf of Mexico, Int. J. Remote Sens. 24:2597-2612.
Hu, C., J. Nelson, E. Johns, Z. Chen, R. Weisberg, and F. E. Muller-Karger (2005a). Mississippi
River water in the Florida Straits and in the Gulf Stream off Georgia in summer 2004.
Geophys. Res. Lett., 32, L14606, doi:10.1029/2005GL022942.
Hu, C., F. E. Muller-Karger, C. Taylor, K. L. Carder, C. Kelble, E. Johns, and C. Heil (2005b).
Red tide detection and tracing using MODIS fluorescence data: A regional example in SW
Florida coastal waters. Remote Sens. Environ., 97:311-321.
Lin, I-I, W. T. Liu, C-C Wu, G. T. F. Wong, C. Hu, Z. Chen, W-D Liang, Y. Yang, and K-K Liu
(2003). New evidence for enhanced ocean primary production triggered by tropical cyclone.
Geophys. Res. Lett. 30(13), 1718, doi:10.1029/2003GL017141.
Morey, S. L., M. A. Bourassa, D. S. Dukhovskoy, and J. J. O’Brien (2006). Modeling studies of
the upper ocean response to a tropical cyclone. Ocean Dynamics, DOI 10.1007/s10236-0060085-y.
Price, J. F., T. B. Sanford, and G. Z. Forristall (1994). Forced stage response to a moving
hurricane. J. Phys. Oceanogr. 24:233-260.
Shay, L. K., P. G. Black, A. J. Mariano, J. E. Hawkins, and R. L. Elsberry (1992). Upper ocean
response to hurricane Gilbert. J. Geophys. Res. 97:20227-20248.
Stramma, L., P. Cornillon, and J. F. Price (1986). Satellite observations of sea surface cooling by
hurricanes. J. Geophys. Res. 91, 5031-5035.
Yuan, J., R. L. Miller, R. T. Powell, and M. J. Dagg (2004). Storm-induced injection of the
Mississippi River plume into the open Gulf of Mexico. Geophys. Res. Lett., 31, L09312,
doi:10.1029/2003GL019335.
6
(a)
LC Eddy
Loop
Current
(a) July 2-8, 2005
Cape
San Blas
(b)
28oN
July 10, 2005
05:00 GMT
26oN
24oN
Dry
Tortugas
88oW
86oW
(b) July 11, 2005
Figure 1. Sea surface temperature (SST, oC) before (a) and after (b) Hurricane Dennis. The
weekly mean SST in (a) shows dominant current directions associated with Loop Current and LC
eddies (outlined in dashed red lines with arrows). The overlaid numbers indicate the location of
the hurricane eye and the category, at approximately 2-hour steps, starting from the south around
9 July 2005, 05:00 GMT. Overlaid green lines show bathymetry contours for 30, 40, 50, 100,
200, 500, and 1000 m. In (b), the maximum surface cooling is outlined by the black dashed line,
and the outline is overlaid in (a) to show its relationship with the pre-hurricane LC and eddies.
An area of about 20,000 km2 between 100- and 500-m in the center of the cooling (blue outline
in (b)) was chosen to study the changes in several bio-optical properties. The color scales used to
code the SST in the images were adopted from a fixed omni-scale to cover the entire SST range
(0 – 31oC) to assist time-series analysis for the global ocean.
7
(a)
(b)
July 7, 2005
18:55 GMT
July 11, 2005
18:30 GMT
(c)
(d)
July 7, 2005
18:55 GMT
July 12, 2005
19:11 GMT
Figure 2. MODIS imagery showing changes in SSR (a and b) and FLH (c and d) from before to
after Dennis, where the entire <50-m shallow shelf experienced significant sediment
resuspension and increase in FLH. Elevated FLH values were also found along an LC eddy edge,
as annotated with the black arrows in (d).
8
29.0
0.4
28.0
5.0
27.0
0.3
0.2
3.0
26.0
0.1
1.0
30.0
25.0
0.0
20.0
Rrs(443)
Rrs(555)
SST
Chl
-3
0.5
o
7.0
30.0
Chl (mg m )
9.0
(a)
SST ( C)
Rrs (x1000 sr-1)
Rrs (x1000 sr-1)
Dennis
(b)
10.0
0.06/30
7/5
7/10
7/15
7/20
Date (2005)
Figure 3. [the legend is in Fig 3b but it applies to Fig 3a also - this may be confusing]SST and
SSR parameters before and after Hurricane Dennis in the eastern Gulf of Mexico: (a) for the area
between 100- and 500-m isobath and between 26.5 and 28.5oN (outlined in blue in Fig. 1b); (b)
for the entire shallow (<50 m) shelf region from Cape San Blas to Dry Tortugas. The dashed
lines denote the one-week mean values before hurricane Dennis. SSR parameters were derived
from SeaWiFS from the dates when at least half of all pixels in the area were valid.
9
0
0
10
10
20
28
30
30
Depth(m)
Depth(m)
20
30
26
28
30
26
24
24
40
Temperature
2000
22
40
Temperature
2001
20
50
20
50
-83.4
-83.2
-83
-82.8
-83.4
-83.2
Longitude
0
Depth (m)
22
20
-83
-82.8
Longitude
NEGOM station L11S16
7/27/1998; 8/21/1999; 8/7/2000
o
o
27.50 N, 85.23 W
40
60
FC station MR5
8/22/2004
o
o
23.90 N, 82.76 W
80
100
20.0
25.0
30.0
o
Temperature ( C)
Figure 4. (a) Temperature profiles along a west-east transect south of Tampa Bay along about
27oN in July 2000 and 2001 EcoHAB cruises; (b) Temperature profile (meanstandard deviation)
for the NEGOM station L11S16 (near the center of the maximum cooling area outlined in Fig.
1b), and temperature profile from a station in the Florida Current (labeled FC).
10