Fernando Gilbes1, William Hernández, Natlee Hernández, José Martínez, and Vilmaliz RodriguezGeological and Environmental Remote Sensing LabDepartment of GeologyUniversity of Puerto Rico at Mayagüez1gilbes@cacique.uprm.edu

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)