fernando 1 william hernández, - uprm · fernando 1gilbes, william hernández, natlee hernández,...
TRANSCRIPT
Fernando Gilbes1, William Hernández, Natlee Hernández, José Martínez, and Vilmaliz RodriguezGeological and Environmental Remote Sensing LabDepartment of GeologyUniversity of Puerto Rico at Mayagü[email protected]
MOTIVATION FOR THIS WORKMODIS (Moderate Resolution Imaging Spectroradiometer) is considered an improved generation of ocean color sensors. However, its validation in coastal waters is still underway. There is a great expectation toward MODIS due to its temporal, spatial, and spectral resolutions.Therefore, the testing and validation of MODIS in coastal waters of Puerto Rico is necessary for an implementation of continuous monitoring of water quality parameters using remote sensing.
MODIS TEMPORAL RESOLUTION
EOS AM~10:30 AMLOCAL TIME
EOS PM~1:30 PMLOCAL TIME
MODIS SPATIAL RESOLUTION
Bands 1-2 = 250 m
Bands 3-7 = 500 m
Bands 8-36 = 1000 m
MODIS SPECTRAL RESOLUTIONPrimary Use Band Bandwidth
Land/Cloud/AerosolsBoundaries
1 620 - 6702 841 - 876
Land/Cloud/AerosolsProperties
3 459 - 4794 545 - 5655 1230 - 12506 1628 - 16527 2105 - 2155
Ocean Color/Phytoplankton/Biogeochemistry
8 405 - 4209 438 - 448
10 483 - 49311 526 - 53612 546 - 55613 662 - 67214 673 - 68315 743 - 75316 862 - 877
AtmosphericWater Vapor
17 890 - 92018 931 - 94119 915 - 965
Primary Use Band Bandwidth
Surface/CloudTemperature
20 3.660 - 3.84021 3.929 - 3.98922 3.929 - 3.98923 4.020 - 4.080
AtmosphericTemperature
24 4.433 - 4.49825 4.482 - 4.549
Cirrus CloudsWater Vapor
26 1.360 - 1.39027 6.535 - 6.89528 7.175 - 7.475
Cloud Properties 29 8.400 - 8.700Ozone 30 9.580 - 9.880Surface/CloudTemperature
31 10.780 - 11.28032 11.770 - 12.270
Cloud TopAltitude
33 13.185 - 13.48534 13.485 - 13.78535 13.785 - 14.08536 14.085 - 14.385
Chl-a
SS
NASAChl-a
MAYAGUEZ BAY
CLIMATOLOGY OF AÑASCO RIVER DISCHARGE
(20 years)
DRY SEASON
RAINY SEASON
Añasco River
Yaguez River
Guanajibo River
Sewage Outfall
Deep and ClearWaters
Shallow and ClearWaters withCoral Reefs
MAIN OBJECTIVE
To validate the accuracy of the MODIS sensor for measuringphytoplankton Chlorophyll‐a (Chl‐a)suspended sediments (SS)
in Mayagüez Bay, PuertoRico.
TECHNICAL APPROACH
Chl‐aTesting of NASA OC3 algorithm
using the 1 km bands
Develop local algorithm using 500 m bands
SSTesting the
Miller & Mckeealgorithm using 250 m bands
Develop local algorithm using 250 m bands
20 SAMPLING DATESApril 24-26, 2001October 2-4, 2001February 26-28, 2002August 20-22, 2002February 25-27, 2003October 7-9, 2003January 12-14, 2004February 12, 2004August 19, 2004March 10, 2005
July 19, 2005August 17, 2005September 20, 2005October 19, 2005December 6, 2005March 8, 2006April 21, 2006September 26, 2006October 26, 2006May 1-2, 2007
GRID OF SAMPLING STATIONS
The SeaWiFS Data Analysis System (Sea DAS) was used to process the raw 1km images and generate Chl‐a images using the standard OC3 algorithm.
TESTING THE NASA ALGORITHM
STANDARD MODIS OC3 ALGORITHM
)550490R
550443(RlogR
)403.1659.0457.1753.22830.0(10
103M
43
33
233
>=
−++−=
where
RRRRC MMMM
EXAMPLE OF MODIS CHL‐A
1ST CHALLENGE‐GET GOOD IMAGESDate Satellite Percent of Missing Data Comments
April 24-26, 2001 Aqua & Terra 100% Satellite not available
October 2-4, 2001 Aqua & Terra 100% Satellite not available
February 26-28, 2002 Aqua & Terra 100% Satellite not availableAugust 20, 21, 22, 2002 Aqua & Terra 100% Clouds
February 25, 26, 27, 2003 Aqua 92% CloudsFebruary 25, 26, 27, 2003 Terra 33% Clouds
October 7, 8, 9, 2003 Aqua 100% Not image available for PROctober 7, 8, 9, 2003 Terra 33% Clouds
January 12, 13, 14, 2004 Aqua 56% CloudsJanuary 12, 13, 14, 2004 Terra 33% Clouds
12-Feb-04 Aqua 100% Clouds12-Feb-04 Terra 30% Clouds19-Aug-04 Aqua & Terra 100% Clouds10-Mar-05 Aqua 0 % Good Image10-Mar-05 Terra 67% Clouds19-Jul-05 Aqua & Terra 100% Clouds17-Aug-05 Aqua 100% Clouds17-Aug-05 Terra 83% Clouds20-Sep-05 Aqua & Terra 100% Clouds19-Oct-05 Aqua 100% Clouds19-Oct-05 Terra 17% Clouds6-Dec-05 Aqua 100% Clouds6-Dec-05 Terra 33% Clouds8-Mar-06 Aqua 100% Clouds8-Mar-06 Terra 33% Clouds21-Apr-06 Aqua & Terra 100% Not image available for PR26-Sep-06 Aqua & Terra 100% Clouds26-Oct-06 Aqua 100% Clouds26-Oct-06 Terra 50% Clouds
FIELD VERSUS MODIS CHL‐A
R² = 0.028
0.00
0.50
1.00
1.50
2.00
2.50
0 5 10 15 20
Fie
ld c
hlo
rop
hyll (
ug
/l)
MODIS Chlorophyll (ug/l)
Terra Data
R² = 0.026
0.000.200.400.600.801.001.201.401.601.80
0.00 0.50 1.00 1.50 2.00Fie
ld c
hlo
rop
hyll
(u
g/
l)
MODIS Chlorophyll (ug/l)
Aqua Data
Georeferenced B3/B4 Ratio (500 m) MODIS Terra image ‐ February 12, 2004
NEW ALGORITHM VERSUS OC3
[Chlorophyll-a] = OC3 MODIS algorithm1 km resolution
[Chlorophyll-a] = Empirical algorithm500 m resolution
[Chl-a]= -42.12*(B3/B4)+1.8219
SUSPENDED SEDIMENTSMiller and McKee (2004)
Study of northern Gulf of Mexico‐Mississippi DeltaLinear relationship: in situSuspended Matter vs. MODIS Terra Band 1 (620 –670 nm)Provided evidence of sediment transport
Calibrated images of Mississippi River delta derived from MODIS Terra Band 1 (Miller & McKee, 2004)
Miller and McKee (2004) Algorithm
RESULTS IN MAYAGUEZ BAY
2ND APPROACH‐A NEW ALGORITHM
Image preprocessing with ENVI (Environment for Visualization of Images)
Geo‐referencing: UTM 19 (Datum NAD 83) 17 GPS points corroboratedAtmospheric correction:
Dark Subtract
GPS points used in geo-reference validation
All Stations: (2001 – 2006)
y = 105.19x + 2.6373R2 = 0.1443
0
5
10
15
20
25
30
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14
MODIS Terra Band 1 Reflectance (%)
Susp
ende
d Se
dim
ent C
once
ntra
tion
(mg/
l)
FIELD SS AND MODIS‐B1 %REFLECTANCE
y = 80.703x + 3.614R2 = 0.0695
0
5
10
15
20
25
30
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
MODIS Band 1 Reflectance (%)
Susp
ende
d Se
dim
ent C
once
ntra
tion
(mg/
l)
y = 136.58x + 0.8668R2 = 0.2788
0
5
10
15
20
25
30
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14
MODIS Terra Band 1 Reflectance (%)
Susp
ende
d Se
dim
ent C
once
ntra
tion
(mg/
l)
Dry Season (2001 – 2006) Rainy Season (2001 – 2006)
FIELD SS AND MODIS‐B1 %REFLECTANCE
y = 63.781x + 5.962R
2= 0.0473
0
5
10
15
20
25
30
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14
MODIS Terra Band 1 Reflectance (%)
Susp
ende
d Se
dim
ent
Con
cent
ratio
n (m
g/l)
y = 55.796x + 3.2736R2 = 0.0468
0
2
4
6
8
10
12
14
16
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
MODIS Terra Band 1 Reflectance (%)
Susp
ende
d Se
dim
ent C
once
ntra
tion
(mg/
l)
In-shore Stations (2001 – 2006) Off-shore Stations (2001 – 2006)
FIELD SS AND MODIS‐B1 %REFLECTANCE
REFLECTANCE VS. SEDIMENTS
ReflectancePrevious findings
Increase in SS = Increase in reflectanceRed & Near IR wavelengths show lower slopes in high SS
Mayagüez BayConcentration too low to provide strong response
From Witte et al. (1982)
A NOVEL APPROACH FOR SSVILMALIZ RODRIGUEZ POSTER
0.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
A1
A2 G2
A1
A2AA
A1 Y1 G1
G2
A1
A2AA
A1 G2
A1
A2AA
A1 Y1 A1
A2 Y1 G1
G2
A1
A2AA
A1 Y1 G2
A1
A2AA
A1 G2
A1
A2AA
A1 Y1 G1
July 1705
August 17 05 September20 05
October 1905
December 6 05 April 21 06 September26 06
October 26 06Susp
ende
d Se
dim
ents
con
c (m
g/l)
TSS (Observed)TSS (Estimated)
SS = 337.26 *(Band 1) + 854.12 *(Band 2)
0.00
5.00
10.00
15.00
20.00
25.00
30.00
A1
A2
AAA Y
1G
1G
2S0
5S0
9S1
1S1
3S1
5S1
7S1
9S2
1S2
3S0
1S0
2S0
3S1
3S1
4S1
5S2
1S2
2S2
3S0
1S0
4S0
5S0
7S1
7S1
9S2
1S2
3
Mar-06 Jan-13-04 Jan-14-04 Feb-12-04 Oct-07-03 Feb-27-03
Susp
ende
d S
edim
ents
con
c (m
g/l)
ObservedEstimated
y = 0.4033 * band 1 – 0.0006SS (mg/l) = 452.41 * y + 2.9603
SS WITH NEW ALGORITHM
October 7, 2003 January 14, 2004
CONCLUSIONSImage processing and analyses clearly demonstrated that MODIS is not the most appropriate ocean color sensor for Mayagüez Bay. Another sensor with better temporal, spatial, and spectral resolutions is still needed for the estimation of Chl‐a and SS in tropical coastal waters.However, MODIS Band 1 gives good qualitative results of SS when a site‐specific algorithm is applied in Mayagüez Bay.Future work with MODIS in this region must include (but not limited):
Improve the atmospheric correctionConsider other sources of error like bottom and land signalContamination by CDOM (Patrick Reyes Presentation‐Tomorrow)