EMPACT 1ST YEAR REPORT

Document Sample
EMPACT 1ST YEAR REPORT Powered By Docstoc
					                  EMPACT 1ST YEAR REPORT
                      November 2000

  Satellite Remote Sensing of Surface Water Temperature, Surface
Reflectance and Chlorophyll a Concentrations: Southeastern Louisiana

                                by

         Nan D. Walker*, Adele Hammack* and Soe Myint+
                     Coastal Studies Institute*
              Dept. of Geography and Anthropology+
                    Louisiana State University
                     Baton Rouge, LA 70803
                   Nwalker@antares.esl.lsu.edu




                                 1
I. Objectives and Background Information
The remote sensing component of the EMPACT project in the first year had two main
objectives:

1. to produce, using NOAA satellite data, quantitative information of surface water temperature
   and reflectance in the Davis Pond Diversion region and the water bodies of the Greater New
   Orleans area for one full year prior to the diversion opening and to provide these information
   in near-real-time to the public via an LSU web site

2. to obtain and analyze “ground truth” measurements coincident with overpasses of the
   Orbview-2 SeaWiFS sensor to assess the accuracy and usefulness of SeaWiFS data and
   algorithms for estimating chlorophyll a within the EMPACT region.

The satellite coverage included the region from 28 20’ to 30 42’ N latitude and 88 to 91 W
longitude. The satellite image of 3/21/97, with the Bonnet Carre Spillway open into Lake
Pontchartrain, is shown as an example of the study region (Figure 1). The satellite data used in
this project were received directly from polar orbiting satellites via antenna at the Earth Scan
Laboratory, Coastal Studies Institute, Louisiana State University. Data from the NOAA
Advanced Very High Resolution Radiometer (AVHRR) and the Orbview-2 Sea-viewing Wide
Field of View Sensor (SeaWiFS) were used. The channels/bands and the spatial resolution of
these data are shown in Table 1. Three NOAA satellites (12,14 and 15) were transmitting data
during most of the study period resulting in as many as six temperature images/day over the
study region. The NOAA AVHRR afternoon reflectance image was used for the regional
detection of suspended sediments. The orbitography of Orbview-2 SeaWiFS usually yields one
good image of the central Gulf of Mexico every second day.

The NOAA data were processed each morning using Earth Scan Laboratory facilities and posted
to the web site (http://www.esl.lsu.edu/research/empact.html). The Orbview-2 SeaWiFS data has
an embargo period of at least 14 days and, therefore, was not available in real-time and was not
posted to the web site. It was used only for research on the retrieval of chlorophyll a from space.




                                                 2
Figure 1. NOAA AVHRR Reflectance image of the water bodies in the Greater New Orleans
Region obtained on March 21, 1997, during flooding of the Mississipi River and opening of the
Bonnet Carre Spillway.


II. Methodology

A. NOAA AVHRR Satellite Data Processing Summary

Sea surface temperatures were computed with NOAA AVHRR satellite data using a
modification of the MCSST technique described by McClain et al (1985). This technique uses
two thermal infrared channels, 10.3-11.3 m and 11.5-12.5 m (Table 1). The accuracy of this
technique in the study area will be discussed in the Results section.



                                              3
Table 1. Bands/channels of the NOAA AVHRR and Orbview-2 SeaWiFS sensors.


       Channel/Band                  NOAA-14 AVHRR                         SeaWiFS
                                      Wavelength (µm)                    Wavelength (nm)
             1                          0.58-0.68                           402-422
             2                          0.725-1.1                           433-453
             3                          3.55-3.93                           480-500
             4                          10.3-11.3                           500-520
             5                          11.5-12.5                           545-565
             6                                                              660-680
             7                                                              745-785
             8                                                              845-885


Surface reflectance was determined using the visible channel, 0.58-0.68 m, of the NOAA
AVHRR. A modification of the Stumpf atmospheric correction technique (1992) was used
(Walker and Hammack, in press) that corrects for downwelling solar irradiance, aerosols,
sunglint and Rayleigh scattering, and is valid over 200 km in the east-west direction. The
modified technique yields more accurate reflectances of the turbid waters along the Louisiana
coastline. Background information on the NOAA satellites, their orbits and calibration can be
found at http://140.90.207.25/noaasis.html.

The steps performed in the image processing of temperature and reflectance are outlined below.


           Calibration of visible and thermal infrared data from count values to science units


                             Initial screening of the data for image quality


                              Calculation of temperatures & reflectances


                   Navigation/registration of images to a rectangular map projection


                                  Scaling of temperatures/reflectances


                                       Production of GIF images


             Posting of images to web site (http://www.esl.lsu/edu/research/empact.html)

                                                   4
B. SeaWiFS “Ground Truth” Summary

On April 26, 2000, a SeaWiFS ground truth experiment was undertaken in Barataria Bay and the
coastal ocean, seaward of the bay. Water samples were collected within two hours of satellite
overpass (1200 CST). The USGS collected water samples at the eight EMPACT stations near
mid-day of that same day. Twenty-four stations were sampled with the collection of two 500 ml
water samples at each station. Twenty-two of the stations lay along a north-south transect line
and two additional stations were sampled to the east of this line within Barataria Bay.
Fortunately, the sky in this region was cloud-free and the wind conditions were calm.

The offshore stations were sampled first. The water on the inner shelf was exceptionally clear at
the offshore stations. Within a few kilometers of the coast, a large number of shrimping vessels
were operating, and the disturbance of the bottom sediments had a marked influence on the water
clarity and color.

The water samples were kept cool and in the dark until they were processed the following day in
Dr. Eugene Turner’s lab at LSU. Several analyses were performed including the determination
of chlorophyll a concentration, concentration of suspended solids, the concentration of
suspended sediments and the dissolved organic carbon. Using a dual path spectrophotometer
(belonging to the Wetland Biogeochemistry Institute), estimates of the absorption coefficients for
selected water masses were obtained using the EMPACT samples and the “ground truth”
samples. The results of these analyses are shown in the Results section. The reader is referred to
the Turner and Swenson section (this report) for a discussion of the laboratory techniques used.

The SeaWiFS image data were atmospherically corrected using the TerascanTM software that is
based on the NASA algorithm in use in their SEADAS package. The primary algorithm
components are discussed in Gordon and Wang (1994). The NASA OC2 algorithm was used to
estimate chlorophyll a concentrations with the 490 and 555 nm bands (O’Reilly et al., 1998).
For more details on the Orbview-2 SeaWiFS sensor, the reader is referred to
http://seawifs.gsfc.nasa.gov//seawifs.html.

C. Additional Datasets

Wind measurements from the monitoring station and from the Burrwood C-Man station,
Southwest Pass, were used to interpret the image patterns and write the monthly text provided on
the LSU Earth Scan Lab EMPACT web site. The time-series measurements obtained hourly at
the Lake Salvador monitoring station were obtained from USGS and were used in the
interpretation of the satellite data. In addition, the water sample data from the eight EMPACT
stations were used in the “ground truth” experiment. River discharge information from the
Tarbert Landing station were obtained from the U.S.A.C.O.E. (New Orleans District).




                                                5
III. RESULTS AND DISCUSSION

A. NOAA Monitoring of Surface Water Temperature and Reflectances

i.     Comparisons of Satellite Data with Data from the Monitoring Station in Lake Salvador

Clear-sky satellite-derived surface water temperatures and reflectances were averaged over an
area in northern Lake Salvador that corresponded with the position of the monitoring station
(station 8) (Figure 2). This subset of time-synchronous satellite and in-situ measurements were
subsequently compared.




         90 33, 30 04




                                                        4           5
                                                                            6
                                                         3                  7
                                                                2       8
                                                                1

                                                 Lake Salvador




                      Real-time continuous monitoring                               89 58, 29 33

                      station
                      Weekly data collection
                      stations                                        Landsat TM
                                                             Band 7, 5, 3 (1992- 93 Winter)



Figure 2. LANDSAT TM 3 satellite image of Lakes Salvador and Cataouatche, showing the
location of the monitoring station (#8) in Lake Salvador and the other stations where surface
water samples were obtained weekly during the study period.
                                                            6
 The water temperature comparison is shown in Figure 3. The satellite and field measurments of
temperatures were very similar. Linear regression of the temperature data-sets revealed a strong
correlation with R2 of 0.951 using 173 data points. This linear relationship is shown in Figure 4.
The satellite reflectances are compared with the YSI turbidity values in Figure 5. The satellite
reflectance values and the turbidity measured at station 8 with the YSI probe were not similar in
many cases and the statistical relationship was low with an R2 of 0.43. The scatter-plot of these
two data-sets reveals no slope to the line indicating little correlation (Figure 6). This result was
not a total surprise as similar observations have been made in the Atchafalaya region. The
differences are thought to result from several factors including the fact that the YSI turbidity
measures back-scatter from particles suspended in the water column (5 ft below surface) in the
830-890 nm region. The satellite reflectance measurements were made at 580-680 nm and are
related to light reflected from near the water surface by suspended material in the water column.
The satellite reflectance is effected by the concentration of inorganic and organic material, type
of inorganic sediment (clay, silt, sand) and additional pigments in the water column (from
chlorophyll, colored dissolved organic matter) (Curran and Novo, 1988). Clay particles yield
high reflectance, whereas organic particles such as detritus yield low reflectance in the
wavelength region being used. If there are numerous particles of organic origin suspended in the
water, the YSI back-scatter will be high, but the satellite reflectance will be low. In addition, the
YSI turbidity sensor data exhibited much noise and, therefore, the data quality may be in
question. The satellite reflectance measurements are better suited to regional detection and
assessment of suspended sediment sources and physical processes (see ii below).


Table 2. Conversion from Temperatures in Celsius to Fahrenheit (rounded to nearest whole
number.

ºC    6     8      10    12     14    16    18     20    22     24    26    28     30    32
ºF    43    46     50    54     57    61    64     68    72     75    79    82     86    90




                                                  7
                                                                    Comparison of Field and Satellite Measurements:
                                                                                 Water Temperature
                   35




                   30




                   25
Temperature ( C)
o




                   20




                   15




                   10




                   5
                   8/1/99   8/31/99   9/30/99 10/30/99 11/29/99 12/29/99 1/28/00                   2/27/00   3/28/00   4/27/00   5/27/00   6/26/00   7/26/00   8/25/00


                                                                                         Time (days)

                                                                                        NOAA sst             YSI sst

         Figure 3. Time-series graph of satellite-derived and in-situ water temperatures at the monitoring
         station in Lake Salvador from 8/1/1999 through 8/25/00. The satellite temperatures were
         obtained from the polygonal area 3, shown in Figure 7.



                                                                     Linear Model (R2=0.951) for InSitu vs. Satellite Water Temperature
                                                                                                Relationship


                                                                    35
                                        In-Situ Water Temperature




                                                                    30

                                                                    25
                                                    (oC)




                                                                    20

                                                                    15

                                                                    10
                                                                         5        10          15                 20         25              30             35
                                                                                       Satellite-derived Water Tem perature ( oC)




         Figure 4. Linear relationship between satellite and in-situ water temperatures, northern Lake
         Salvador, using data from 8/1/99-8/25/00.


                                                                                                             8
                                                              Comparison of Field and Satellite Measurements:
                                                                        Reflectance and Turbidity
                    20


                    18
  Turbidity (ntu)




                    16


                    14


                    12


                    10
  Reflectance (%)




                    8


                    6


                    4


                    2


                    0
                    8/1/99   8/31/99                 9/30/99 10/30/99 11/29/99 12/29/99 1/28/00   2/27/00   3/28/00   4/27/00   5/27/00    6/26/00   7/26/00   8/25/00


                                                                                      Time (days)
                                                                               NOAA reflectance             YSI turbidity


Figure 5. Time-series graph of satellite reflectances and YSI turbidity from the monitoring
station in Lake Salvador from 8/1/99 through 8/25/00. The satellite reflectances were obtained
from the polygonal area 3, shown in Figure 7.


                                                        Scatterplot of satellite reflectance values and YSI turbidity


                                             6

                                             5
                             YSI turbidity




                                             4

                                             3

                                             2

                                             1

                                             0
                                                 0                2               4                6                  8               10
                                                                              Satellite Reflectance (%)




Figure 6. Scatter-plot of satellite reflectances and YSI turbidity values, northern Lake Salvador,
using data from 8/1/99-8/25/00.
                                                                                                       9
Since a very strong relationship was found between the satellite and in-situ temperatures, the
satellite temperatures have been extracted from several regions in the study area including North
and South Lake Salvador, Lake Cataouatche, Lac des Allemands, East and West Lake
Pontchartrain and Barataria Bay. The polygonal areas for which temperatures were extracted are
shown in Figure 7.




Figure 7. Polygonal areas over which satellite temperatures and satellite reflectances were
extracted.



                                               10
The water temperature changes exhibited by these relatively shallow lakes and bays in the New
Orleans areas were overall very similar and caused primarily by air-sea fluxes of heat and water
vapor (through evaporation). Barataria Bay is most effected by tidal and wind-related fluxes of
water from the coastal ocean. The satellite-derived surface temperatures are shown in Figure 8
for four areas within the study region: north Lake Salvador, Lake Cataouatche, Lac des
Allemands and Lake Pontchartrain. Maximum temperatures occurred in the months of July and
August. Temperatures in August 2000 were higher than in August 1999. The highest
temperatures were measured in Lac des Alemands and Lake Cataouatche, perhaps due to their
relatively shallow depths and proximity to a major urban area (i.e. the heat island effect).
Temperatures between 33 and 35 C (91.4-95 F) were measured on several occasions during
summer 2000 (Figure 8). Minimum temperatures of 8-9.5  C (46.4-49 F) were measured at all
stations in December, January and February. The lowest temperatures were experienced in
northern Lake Salvador and Lac des Allemands. The low temperatures resulted from the passage
of severe winter storms and the loss of heat from the shallow bays due to evaporation, sensible
heat loss and back radiation (Huh et al., 1984; Walker et al., 1987). The prolonged period of low
water temperature between late January and mid February 2000 corresponded in time with a
prolonged period of strong winds from the north.




                                               11
                                                                                     North Lake Salvador
                                    35



                                    30
Satellite water temperature ( oC)




                                    25



                                    20



                                    15



                                    10



                                     5



                                     0
                                    08/01/99 08/31/99 09/30/99 10/30/99 11/29/99 12/29/99 01/28/00 02/27/00 03/28/00 04/27/00 05/27/00 06/26/00 07/26/00 08/25/00

                                                                                             Time (days)


                                                                                                  mean




                                      Figure 8a. Satellite-derived water temperatures from Northern Lake Salvador, area 3.




                                                                                                 12
                                                                                       Lake Cataouatche
                                    35



                                    30
Satellite water temperature ( oC)




                                    25



                                    20



                                    15



                                    10



                                     5



                                     0
                                    08/01/99 08/31/99 09/30/99 10/30/99 11/29/99 12/29/99 01/28/00 02/27/00 03/28/00 04/27/00 05/27/00 06/26/00 07/26/00 08/25/00

                                                                                             Time (days)


                                                                                                  mean




                                    Figure 8b. Satellite-derived water temperatures from Lake Cataouatche, area 4.




                                                                                             13
                                                                                      Lac Des Allemands
                                    35



                                    30
Satellite water temperature ( oC)




                                    25



                                    20



                                    15



                                    10



                                     5



                                     0
                                    08/01/99 08/31/99 09/30/99 10/30/99 11/29/99 12/29/99 01/28/00 02/27/00 03/28/00 04/27/00 05/27/00 06/26/00 07/26/00 08/25/00

                                                                                             Time (days)


                                                                                                  mean




                                         Figure 8c. Satellite-derived water temperatures from Lac des Allemands, area 5.




                                                                                                 14
                                                                                      Lake Pontchartrain
                                    35



                                    30
Satellite water temperature ( oC)




                                    25



                                    20



                                    15



                                    10



                                     5



                                     0
                                    08/01/99 08/31/99 09/30/99 10/30/99 11/29/99 12/29/99 01/28/00 02/27/00 03/28/00 04/27/00 05/27/00 06/26/00 07/26/00 08/25/00

                                                                                             Time (days)


                                                                                                  mean




                                         Figure 8d. Satellite-derived water temperatures from Lake Pontchartrain, area 7.




                                                                                                 15
ii.    Processes effecting coastal systems, based on imagery analyzed

From August 1, 1999-August 31, 2000, 200 NOAA AVHRR temperature images and 89 images
of water reflectance were posted to the web site (See
http://www.esl.lsu.edu/research/empact.html. Descriptions of the temperature and reflectance
patterns were provided with each month of imagery. These data can be viewed on the web site
but are not all included in this report. A subset of the imagery will be discussed in the report,
chosen to be representative of some of the important processes effecting coastal circulation and
coastal systems in this area. The temperature images on the web site were color enhanced for the
range of temperatures in the image. Adjustments to the temperature/color enhancement were
made every 2-3 months. In this report, all images were color enhanced using the same
temperature/color conversion to enable the reader to better compare the image patterns. Features
within the individual images are often better revealed, however, with the enhancements used on
the web page.

A sequence of eight NOAA SST images show the annual cycle of SST patterns across the study
region (Figure 9). These regional views of surface SST patterns reveal that the interior lakes and
bays gain and lose heat in a similar fashion. Surface temperatures of Lake Cataouatche, Lake
Salvador, Lac des Allemands, Lake Pontchartrain and Lake Maurepas are usually within a few
degrees of one another in temperature. The shallowest lakes and bays change temperature more
quickly as they have less capacity for heat storage. The coldest surface waters are usually those
of the Mississippi River as it is discharged onto the continental shelf of the northern Gulf of
Mexico. The warmest waters, especially in the winter months, are the deep waters in the
southeast corner of the study region. Filaments of warm Loop Current water intermittently move
up onto the shelf (< 200 m in the study region). Circulation of water around the Mississippi delta
region changes rapidly due to changes in wind direction and speed as well as the intrusion of
deepwater currents within cyclonic and anticyclonic eddies or northward surges of the Loop
Current (Walker, 1996; Walker et al., 1996).

The first image of the sequence, 8/5/99, shows that the interior lakes and bay waters are a few
degrees warmer than waters of the continental shelf. However, by 9/17/99 the opposite is true
and the coastal waters including the interior lakes and bays are colder than shelf waters. The
cooler river water is observable in the image of 9/17/99 where streamers of cooler river water are
discharged from the passes east of the bird-foot delta as well as from South Pass and Southwest
Pass. The nearshore flow of water is from east to west, driven by a wind from the northeast. By
10/27/99, additional cooling had taken place and the effluent plume of the Mississippi River was
much more extensive than on 9/17/99. The 10/27/99 image reveals a clockwise gyre in the
Louisiana Bight (west of the bird-foot delta) with cold river water approaching the Barataria Bay
region. The clockwise gyre is usually present when winds blow out of the east along the coast
(Walker, 1996).

Note also the large and rapid changes in temperature seaward of the river water where it
encounters the warmer outer shelf and slope waters. The offshore waters maintained their heat
and were 28 C in contrast to the lake and bay waters of 17º C( 62º F). The image of 12/25/99
depicts regional temperature patterns during one of the coldest events of the winter. Interior
                                                16
lakes and bays were similar in temperature (near 8º C, 46º F). The river plume reversed direction
due to a change in wind direction from easterly to westerly. A long streamer of river water
extended far east of the bird-foot delta. This feature is thought to have formed from the
combined influence of the wind and from eastward currents within a clockwise eddy offshore.
Notice that the Chandeleur-Breton Sound is much colder than the inner shelf waters west of the
bird-foot delta, as the Sound is much shallower. The coldest water temperatures of the year were
recorded on 2/9/2000. In this image, the Mississippi plume is seen as a spatially extensive lens of
cold water around the bird-foot delta with relatively warm water along its seaward margin. Note
the movement of cold water from the eastern side of the delta into Breton Sound. Waters
warmed quickly during February 2000 and then more gradually in March April and May. The
images of 3/10/2000 and 5/14/2000 illustrate the gradual increase in temperatures. The
Mississippi River water stands out as very cold in comparison to the ambient shelf waters. The
annual maximum temperature in the shallow lakes was experienced on 7/20/2000, when these
interior water bodies were again warmer than the coastal and shelf waters.

An interesting event occurred in early October 1999 that deserves mention in this report. The
satellite image of 10/5/99 shows unusually cold water in Lakes Salvador and Cataouatche, as
well as along the south shore of Lake Pontchartrain. This regional chilling probably reflects a
local rainfall event and the discharge of colder water into this limited area. The image obtained
about 24 hours later revealed a return to more normal temperatures. The water sample data of
total pigments revealed an increase in chlorophyll a within a week of this event, indicating a
strong and rapid response to the influx of nutrients from land runoff. When the river diversion is
opened, major pulses in nutrients will be introduced into the system, in addition to those entering
from local runoff.




                                                17
a




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on
a) 8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, 0/2000.




                                                18
b




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on a)
8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, 7/20/2000.




                                               19
c




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on a)
8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, 7/20/2000.




                                               20
d




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on a)
8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, 7/20/2000.




                                               21
e




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on a)
8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, 7/20/2000.




                                               22
f




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on a)
8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, 7/20/2000.




                                               23
g




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on a)
8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, 7/20/2000.




                                               24
h




Figure 9. NOAA AVHRR satellite imagery showing regional surface temperature structure on a)
8/5/99, b) 9/17/99, c) 10/27/99, d) 12/25/99, e) 2/9/2000, f) 3/10/2000, g) 5/14/2000, h)
7/20/2000.

                                            25
a




Figure 10. NOAA AVHRR satellite imagery showing regional surface temperature structure on
a) 10/5/99 and b) 10/6/99.




                                            26
b




Figure 10. NOAA AVHRR satellite imagery showing regional surface temperature structure on
a) 10/5/99 and b) 10/6/99.




                                            27
Along the Louisiana coast where river discharge is large, the satellite reflectance measurements
in the 580-680 nm wavelength region (NOAA AVHRR Ch1) are generally related to the amount
of suspended sediments in the surface waters. The presence of certain pigments such as those
found in plankton (mainly chlorophyll a) or those in colored dissolved organic matter can alter
the reflectance signal. At times, these pigments substantially lower the reflectance especially in
the Lake Salvador and Cataouatche region, due to the potentially large amounts of tannic acids
entering the system. The overall assessment of the reflectance observations from these lakes is
that little sediment entered the system from other sources during the year of monitoring. Previous
imagery analyzed during flooding of the Mississippi River demonstrated that turbid water can
enter through the Intracoastal Waterway. However, as Mississippi River discharge was
abnormally low during 1999/2000, this process was not observed in the imagery analyzed during
this year of study. In addition, no distinct seasonal change in the level of reflectance was
observed in these interior lakes.

The highest levels of reflectance in the inland lakes were observed during strong wind events,
and attributed to wind-wave re-suspension of bottom sediments. Four reflectance images have
been chosen to represent the range of reflectances under various wind conditions (Figure 11).
The 10/22/99 image demonstrates low values of surface reflectance in Lakes Salvador and
Cataouatche as well as the other inland lakes in the New Orleans area. The wind was light to
moderate northwest and southeast around the time of image acquisition. The highest reflectances
were obtained around the bird-foot delta where the river continuously discharges suspended
sediments into the Gulf of Mexico. Reflectances were also relatively high along the Mississippi
and Alabama coasts. The images of 12/28/99 and 12/29/99 both depict conditions of high
reflectance in the interior lakes that can be explained by strong winds and the wind-wave
resuspension of unconsolidated bottom sediments into the water column. As long as the wind
speed remains high, the sediments can remain suspended in the water column where the satellite
can detect them. The image obtained on 12/29/99 demonstrates an extreme case where surface
reflectance was high in all nearshore regions. This image was obtained near the transition from
strong northwest winds to strong southwest winds. The Mississippi River sediment plume is
particularly impressive in that it extends at least 100 km from the bird-foot delta towards the
southwest. The image on 1/10/2000 shows the regional reflectance patterns under conditions of
a moderate southeast wind. It is interesting to note that the reflectance of Lake Salvador and
Cataouatche became substantially elevated under these wind conditions. The reader is referred
to the web site to view the full range of reflectance variations within the study region.




                                                28
a




Figure 11. NOAA-14 satellite reflectance imagery showing regional surface suspended sediment
distribution on a) 10/22/99, b) 12/28/99, c) 12/29/99 and d) 1/10/2000.




                                             29
b




Figure 11. NOAA-14 satellite reflectance imagery showing regional surface suspended sediment
distribution on a) 10/22/99, b) 12/28/99, c) 12/29/99 and d) 1/10/2000.




                                             30
c




Figure 11. NOAA-14 satellite reflectance imagery showing regional surface suspended sediment
distribution on a) 10/22/99, b) 12/28/99, c) 12/29/99 and d) 1/10/2000.




                                             31
d




Figure 11. NOAA-14 satellite reflectance imagery showing regional surface suspended sediment
distribution on a) 10/22/99, b) 12/28/99, c) 12/29/99 and d) 1/10/2000.




                                             32
B. Ground Truth Experiments

A very successful ground truth experiment was performed on April 26, 2000. It was planned to
coincide with a clear sky day and high quality SeaWiFS image around mid-day. The Coastal
Studies Institute coastal research vessel “Changes in Latitude” was used for the survey in the
coastal ocean and Barataria Bay. The USGS weekly water sample survey was timed to
correspond with the sampling in Barataria Bay on April 26. Twenty-four stations were occupied
in the lower Barataria system and 8 stations were occupied in Lakes Salvador and Cataouatche.
The samples were run early on April 27 in the LSU laboratory of Dr. Eugene Turner by Erick
Swenson, Charles Milan and Nan Walker. The parameters of interest for the ground truth effort
were chlorophyll a, suspended sediments, suspended solids, dissolved organic carbon and an
estimate of colored dissolved organic matter (CDOM), also referred to as yellow substance
(Bricaud et al., 1981). More details on the analysis techniques are provided in the Methodology
Section and also in other sections of this report.

The satellite estimate of chlorophyll a obtained during the field sampling is shown in Figure 12.

a                                                b




Figure 12. NOAA-14 satellite imagery obtained on April 26, 2000 1800 UTC (1300 CDT)
showing a) chlorophyll a estimates, and b) water leaving radiance in the 670 nm channel.
Stations 1-22 lay along the transect and E1 and E2 were located to the east of the station 1.

                                                33
Two water samples were obtained at the most seaward white dot (21,22) and one sample was
taken at each of the other white dots (Figure 12). Exceptionally clear water was encountered at
the outermost stations on the continental shelf. Waters became more turbid close to the coast,
partially as a result of the presence of many shrimping vessels especially east of the transect line.
The water samples within the bay were visibly enhanced with chlorophyll and suspended
sediments.

The satellite-derived chlorophyll a estimates are compared with the chlorophyll a concentrations
from the field samples in Figure 13. The close correspondence between the satellite and field
measurements is noteworthy, particularly when the difference in scale between the two
measurements is considered. The satellite pixel represents an area at least 1100 m x 1100 m,
whereas the field measurement was obtained from 500 ml of water! Various linear and non-
linear models were investigated and the best fit was obtained with a cubic model, with an R2 of
0.92 (Figure 14).



                                            Comparison of Field and Satellite-estimated Chlorophyll a


                                   25
           Chloroph yll a (µg/l)




                                   20

                                   15

                                   10

                                   5

                                   0
                                        X1 X2 2   3   4   5   6   7 12 13 14 15 16 17 18 19 20 21 22
                                                                    Stations

                                                                      In-Situ        SeaWiFS




Figure 13. Comparison of field and satellite estimates of chlorophyll a on April 26, 2000 for the
Barataria Bay and coastal ocean stations (See Figure 12 for station locations).


These results were rather surprising as there are potentially many pigments in the Louisiana
coastal waters (mostly CDOM) that could alter the chlorophyll estimation from the SeaWiFS
satellite data. The conditions under which the ground truth data were obtained were abnormal
for spring. The discharge of the Mississippi River was abnormally low during the entire study
period (Figure 15) and particularly during winter and spring 2000. Compared with a 30 year
mean (1970-1999), the April river discharge was 25% lower than the long-term April mean.
Discharge was less than ½ of normal during the previous winter and spring months. In addition,
southern Louisiana was in the middle of a 2 year drought. Thus, the supply of land-derived
                                                                                34
pigments was minimal during this ground truth experiment. As noted earlier, abnormally clear
water was present at the seaward end of the ground truth transect line. These waters would have
entered Barataria Bay with the incoming tide and diluted further the land-derived pigments.

                                                                   Cubic Model (R2 = 0.92) for Field/Satellite Chlorophyll Relationship
      Field measured Chlorop hyll a




                                      25

                                      20

                                      15
                  (µg/l)




                                      10

                                                    5

                                                    0
                                                              0                5               10                    15                20                  25
                                                                                   SeaWiFS Chlorophyll a estim ate (µg/l)




Figure 14. Cubic model for the field/satellite chlorophyll relationship (R2 = 0.92).


                                                                                       Mississippi River Discharge- Tarbert Landing, MS

                                                             900


                                                             800


                                                             700


                                                             600
                                      Discharge (1000 cfs)




                                                             500


                                                             400


                                                             300


                                                             200


                                                             100


                                                              0
                                                               Aug      Sep    Oct     Nov     Dec      Jan         Feb    Mar      Apr        May   Jun        Jul   Aug
                                                                                                                Months

                                                                                                 1970 - 1999 Mean           1999 - 2000 Mean




Figure 15. Comparison of long-term monthly-averaged Mississippi River water discharge at
Tarbert Landing (using 1970-1999 data) with monthly-averaged discharge from August 1999
through August 2000. Discharge is in cubic feet/sec.


                                                                                                                      35
The absorption coefficients obtained for selected stations using the spectro-photometer indicated
a gradient in absorption from the offshore stations to the interior bay stations in Lakes Salvador
and Cataouatche with highest absorption values in the ultra-violet region of the spectrum (< 400
nm) at the four EMPACT stations (Figure 16). Lowest absorption values were obtained in the
offshore stations. Color dissolved organic matter (CDOM) absorbs strongly in the ultra-violet
and, thus, these results indicate that relatively high levels were found in the Davis Pond diversion
region. During a more normal rainfall year, higher CDOM values would be expected. The
CDOM absorbs strongly in the UV wavelengths and also in blue wavelengths, used to compute
chlorophyll a concentrations. Thus, high levels of CDOM will produce chlorophyll a estimates
that are excessively high. Unfortunately, the standard NASA SeaWiFS atmospheric correction
and chlorophyll algorithm did not yield acceptable chlorophyll values for Lakes Cataouatche and
Salvador. Many variations of the atmospheric correction technique were applied to the satellite
data. One attempt, using the Straylight correction (Yeh et al, 1997) yielded chlorophyll values in
these interior lakes, however, the values that resulted were too high in Lakes Cataouatche and
Salvador and, at the same time, too low in Barataria Bay. The preliminary conclusion based on
the ground truth information is that chlorophyll a estimation with the SeaWiFs ocean color
sensor may not provide sufficient accuracy to be useful in Lakes Salvador and Cataouatche due
to the high levels of CDOM. The SeaWiFS estimates of chlorophyll a in Barataria Bay and the
coastal ocean were excellent in the April 2000, a case of low river discharge and rainfall. Future
ground truth experiments may reveal how the algorithm holds up under different environmental
conditions.




Figure 16. Absorption coefficients (m-1) for selected stations in the EMPACT region. E1, E4, E5
and E8 are in Lakes Salvador and Cataouatche. Stations 1 - 19 are in Barataria Bay and the
coastal ocean (see Figure 2 for locations). A south to north gradient (high in north) in absorption
in the ultraviolet region (< 400 nm) is shown that relates to concentration of CDOM.
                                                36
The ground truth measurements of suspended sediments and suspended solids were compared
with measurements made in the 555 nm and 670 nm SeaWiFS channels. The 670 nm channel
yielded the highest statistical relationships between the satellite and field measurements (Figure
16, 17). The relationship was not as strong as was that for chlorophyll a. A non-linear power
relationship yielded the best fit to the data with an R2 of 0.84 (Figure 18). The relationship
between the suspended solid concentration and satellite measurements at 670 nm were lower.
Again, we are comparing a huge satellite area with a tiny water sample. The statistical
relationship between suspended sediments/solids and satellite reflectance in the 555 nm channel
may have been lower due to the relatively shallow sampling depth (< 0.5 m).




                                     Com parison of Field-m easured Suspended Sedim ent and
                                                           Satellite data

                             50                                                                                        1.8
                             45                                                                                        1.6




                                                                                                                             Radiance (mW cm -
                                                                                                                              SeaWiFS 670 nm
                             40                                                                                        1.4
           Sediment (mg/l)




                             35
             Suspended




                                                                                                                       1.2




                                                                                                                                   µm -1sr -1)
                             30
                                                                                                                       1
                             25
                                                                                                                       0.8
                             20
                                                                                                                       0.6




                                                                                                                                 2
                             15
                             10                                                                                        0.4

                              5                                                                                        0.2
                              0                                                                                        0
                                  X1 X2   1   2   3   4   5   6   7   8   9   12 13 14   15 16   17 18   19 20 21 22

                                                                      Stations

                                                              In-Situ             SeaWiFS-670




Figure 17. Comparison of field-measured suspended sediment and SeaWiFS radiance
measurements in the 670 nm channel for the Barataria Bay and coastal ocean stations.




                                                                                 37
                                        Power Model (R2 = 0.844) for Field Suspended Sediment vs.
                                                     Satellite 670 nm Relationship

                                    50.00
         Field measured Suspended




                                    40.00
               Sediment (mg/l )




                                    30.00

                                    20.00

                                    10.00

                                     0.00
                                            0.4    0.5        0.6       0.7        0.8          0.9   1
                                                                                     -2   -1   -1
                                                     SeaWiFS 670 nm Radiance (m W cm µm sr )




Figure 18. Power model developed for the relationship between field measurements of
suspended sediments and SeaWiFS radiance measurements in the 670 nm channel.


IV. CONCLUSIONS

This one year study showed that remote sensing can provide a powerful tool for measuring
regional spatial and temporal changes in surface temperature, surface reflectance (primarily
surface suspended sediments) and chlorophyll a. Surface temperatures within the bays and lakes
of southeastern Louisiana exhibited similar heating and cooling cycles with an annual
temperature range of 8º C (º 46 F) to 35º C (95º F). Temperature changes on the continental
shelf were slower than bay changes due to the deeper water. Mississippi River waters remained
cooler than other water masses throughout the year. Satellite-derived water temperature were
closely correlated with field measurements of water temperature.

NOAA AVHRR satellite-derived reflectance did not correlate well with field measurements
obtained by the YSI backscatter sensor. However, the SeaWiFS reflectance values at similar
wavelength to the NOAA AVHRR showed a positive and significant relationship with in-situ
measurements of suspended sediment concentrations. Thus, the satellite measurements are
considered a good approximation of spatial and temporal variations in near-surface suspended
sediments. In an attempt to better understand the satellite reflectance values, the weekly water
sample data of suspended sediments should be compared with clear sky reflectance values.

The SeaWiFS “ground truth” experiment yielded noteworthy results, namely that accurate
chlorophyll estimates can be made using the SeaWiFS channels and the NASA OC2 algorithm

                                                                        38
along the Louisiana coast, at least under conditions of low river discharge and land drainage,
when CDOM is minimized. Chlorophyll estimates were not obtained with the standard NASA
software package in the Lake Cataouatche and Lake Salvador region. The absorption
measurements made across the region, however, lead the authors to conclude that chlorophyll
estimation in these inland lakes, containing relatively high concentrations of CDOM, will be less
accurate than in the coastal ocean. The use of additional channels in the ultraviolet portion of the
spectrum (e.g. future data available from GLI on Adeos II) should enable the regional detection
of CDOM effects.




                                                 39
V. REFERENCES


Bridaud, A., A. Morel and L. Prieur, Absorption by dissolved organic matter of the sea (yellow
substance) in the UV and visible domains, Limnology and Oceanography, 26, 43-53, 1981.

Curran, P.J. and E.M.M. Novo, The relationship between suspended sediment concentration and
remotely sensed spectral radiance: A Review, J. Coastal Research, 4, 351-368, 1988.

Gordon, H.R. and M. Wang, Retrieval of water-leaving radiance and aerosol optical thickness
over the oceans with SeaWiFS: a preliminary algorithm. Applied Optics, 33, 433-452, 1994.

Huh, O.K., L.J. Rouse and N.D. Walker, Cold air outbreaks over the northwest Florida
continental shelf: Heat flux processes and hydrographic changes, J. Geophys. Res. 89, 717-726,
1984.

McClain, E.P., W.G. Pichel and C.C. Walton, Comparative performance of AVHRR-based mult-
channel sea surface temperatures, J. Geophys. Res. 90, 11,587-11,601, 1985.

O’Reilly, J.E, S. Marisorena, B.G. Mitchell, D.A. Siegel, K.L. Carder, S.A. Garver, M. Kahru
and C. McClain, Ocean color chlorophyll algorithms for SeaWiFS, J. Geophys. Res., 103,
24,937-24,953, 1998.

Stumpf, R.P. 1992, Remote sensing of water quality in coastal waters, in Proceedings of the First
Thematic Conference on Remote Sensing for Marine and Coastal Enviornments, New Orleans,
LA, 15-17 June, SPIE 1930, Environmental Research Inst. of Michigan, Ann Arbor, MI, pp. 293-
305.

Walker, N.D., L.J. Rouse, and O.K. Huh, Response of shallow sub-tropical environments to cold
air outbreak events: Satellite radiometry and heat flux modeling, Continental Shelf Research, 7,
735-757, 1987.

Walker, N.D., Satellite assessment of Mississippi River plume variability: causes and
predictability, Remote Sensing of Environment, Vol. 58, no. 1, 21-35, 1996.

Walker, N.D., O.K. Huh, L.J. Rouse, and S.P. Murray, Evolution and structure of a coastal squirt
off the Mississippi River delta: Northern Gulf of Mexico, Journal of Geophysical Research, Vol.
101, no. C9, 20,643-20,655, 1996.

Yeh, E.-N., M. Darzi, L. Kumar, 1997: SeaWiFS Stray Light Correction Algorithm. Chapter 4.
SeaWiFS Technical Report Series, Volume 41, Case Studies for SeaWiFS Calibration and
Validation, Part 4. NASA TM 105566.




                                               40

				
DOCUMENT INFO