Menghua Wang, NOAA/NESDIS/STAR Remote Sensing of Water Properties Using the SWIR- based Atmospheric...

Menghua Wang, NOAA/NESDIS/STAR

Remote Sensing of Water Properties Using the SWIR-based Atmospheric Correction Algorithm

Menghua Wang Wei Shi and SeungHyun Son

NOAA/NESDIS/STARE/RA3, Room 102, 5200 Auth Road

Camp Springs, MD 20746, [email protected]

MODIS Science Team MeetingMarriott Inn/Conference Center, College Park, Maryland, May 18-20, 2011

AcknowledgementsSupports from NOAA/NASA funding, NASA SeaBASS in situ data, in situ data from Lake Taihu (J. Tang and Y. Zhang), and MODIS Level 1B data.


Atmospheric Correction: SWIR Bands(Wang & Shi, 2005; Wang, 2007)

At the shortwave infrared (SWIR) wavelengths (>~1000 nm), water has much stronger absorptions. Thus, atmospheric correction (Gordon & Wang, 1994) can be carried out for coastal regions with the black pixel assumption at the SWIR bands.

Water absorption for 869 nm, 1240 nm, 1640 nm, and 2130 nm are 5 m-1, 88 m-1, 498 m-1, and 2200 m-1, respectively.

MODIS has three SWIR bands at 1240, 1640, and 2130 nm, designed for atmosphere and land applications.

We use the SWIR band for the cloud masking. This is necessary for coastal region and inland lake waters.

Require sufficient SNR characteristics for the SWIR bands and the SWIR atmospheric correction has slight larger noises at the short visible bands (compared with those from the NIR algorithm).

2


nLw(443)Scale: 0.-3.0

(mW/cm2 m sr)

Wang, M., S. Son, and W. Shi (2009), “Evaluation of MODIS SWIR and NIR-SWIR atmospheric correction algorithms using SeaBASS data,” Remote Sens. Environ., 113, 635-644.

July, 2005NIR-SWIR Data

Processing

July, 2005NIR-SWIR Data

Processing

Chlorophyll-a0.01-10 (mg/m3)

(Log scale)


Using the SWIR algorithm, we have derived the water optical properties over the Lake Taihu using the MODIS-Aqua measurements during the spring of 2007 for monitoring a massive blue-green algae bloom, which was a major natural disaster affecting several millions residents in nearby Wuxi city.

Wang, M. and W. Shi (2008), “Satellite observed algae blooms in China’s Lake Taihu”, Eos, Transaction, American Geophysical Union, 89, p201-202, May 27.

Wang, M., W. Shi, and J. Tang (2011), “Water property monitoring and assessment for China’s inland Lake Taihu from MODIS-Aqua measurements”, Remote Sens. Environ., 115, 841-854.

The work was featured in the NASA 2008 Sensing Our Planet (http://nasadaacs.eos.nasa.gov/articles/2008/2008_algae.html)

Results from Inland Lake Taihu

5


The third largest fresh inland lake in China (~2,250 km2). Located in one of the world’s most urbanized and heavily populated areas. Provide water resource for several million residents in nearby Wuxi city.

Geo-location of Lake Taihu Geo-location of Lake Taihu Extremely

Turbid Waters


The SWIR atmospheric correction algorithm (Wang, 2007; Wang & Shi, 2005) is used for the water property data processing.

Since MODIS 1240 nm band is not always black for the entire Lake Taihu, we have developed three-step method in the data processing for each MODIS-Aqua data file: First, regions for the black of 1240 nm band are determined using the

SWIR data processing. Second, a dominant aerosol model from the region with black of

1240 nm band is obtained, and Finally, with the derived aerosol model, the SWIR atmospheric

correction algorithm is run using only 2130 nm band (with fixed aerosol model).

The Lake Taihu water property data are then derived.

Methodology (1) Methodology (1)

Wang, M., W. Shi, J. Tang (2011), “Water property monitoring and assessment for China’s inland Lake Taihu from MODIS-Aqua measurement,” Remote Sens. Environ., 115, 841-854.

Menghua Wang, NOAA/NESDIS/STAR(Wang and Shi, 2008)


Validation Results for MODIS-derived Water-leaving Radiance Spectra


10-1

100

101

400 500 600 700 800 900

In Situ DataMODIS-Aqua

( )Wavelength nm

, 10, 2007Lake Taihu in China June (a)

Location: (120.20oE, 31.43oN)

10-1

100

101

400 500 600 700 800 900


( )Wavelength nm

, 12, 2007Lake Taihu in China June

: (120.29Location o , 31.27E o )N

(b)

10-1

100

101

400 500 600 700 800 900


( )Wavelength nm


: (120.31Location o , 30.95E o )N

(c)

10-1

100

101

400 500 600 700 800 900


( )Wavelength nm


: (120.30Location o , 30.94E o )N

(d)


Black pixel assumption at the SWIR 1640 and 2130 nm is generally valid for Lake Taihu.

The SWIR atmospheric correction algorithm using bands 1640 and 2130 nm (Wang, 2007) can be used for the water property data processing.

However, for MODIS-Aqua, four out of ten detectors for the SWIR 1640 nm band are inoperable (dysfunctional).

We focus on deriving seasonal results for the lake using the SWIR 1640 and 2130 nm atmospheric correction algorithm.

More in situ data (five seasonal cruises in 2006-2007 in the lake) are also available to us now.

Methodology (2) Methodology (2)


Validation Results in Lake Taihu (1)Validation Results in Lake Taihu (1)

0

2

4

6

8

10

0 2 4 6 8 10

412 nm443 nm488 nm555 nm645 nm859 nm1240 nm

In Situ nLw(λ) ( mW cm-2 m

-1 sr-1)

Lake Taihu in China1640 & 2130 nm Method

λ ( ) nm Ratio Std 412 1.068 0.132 443 1.033 0.119 488 1.038 0.161 555 0.946 0.128 645 0.942 0.180 859 1.381 0.4561240 0.913 2.903

5 5 >x with 12 valid pixels

(a)



0

2

4

6

8

10

0 2 4 6 8 10

412 nm443 nm488 nm555 nm645 nm859 nm


-1 sr-1)

Lake Taihu in China1240 & 2130 nm Method( ., 2011)Wang et al

λ ( ) nm Ratio Std 412 0.832 0.125 443 0.885 0.111 488 0.968 0.116 555 0.937 0.090 645 0.930 0.141 859 1.495 0.532

5 5 >x with 12 valid pixels

(b)



0

2

4

6

8

10

0 2 4 6 8 10

412 nm443 nm488 nm555 nm645 nm859 nm


-1 sr-1)

Lake Taihu in China1640 & 2130 nm Method

5 5 x with≥5 valid pixels

(c)

Extreme in situnLw(859)


Spatial Non-uniformity in Lake Taihu


Spring Summer Fall Winter


Seasonal Histograms in nLw(1240)Seasonal Histograms in nLw(1240)

0.1

1

10

-0.1 -0.05 0 0.05 0.1 0.15 0.2

nLw(1240) (mW cm-2

m-1 sr

-1)

Lake Taihu in China- (2002-2010)MODIS Aqua

Winter

Summer

Spring

Fall

(a)

TOA Radiance at 1240 nm ~0.1 mW cm-2 um-1 sr-1


Climatology Histogram in nLw(1240)Climatology Histogram in nLw(1240)

0

5

10

15

0

5

10

15

20

25

-0.1 -0.05 0 0.05 0.1 0.15 0.2

MODIS-Aqua

In Situ Data

nLw(1240) (mW cm-2

m-1 sr

-1)

Lake Taihu in China- (2002-2010)MODIS Aqua

All Data

# : 135Total of In Situ Data

(b)

TOA Radiance at 1240 nm ~0.1 mW cm-2 um-1 sr-1


Conclusions

For the turbid waters in coastal regions and inland lakes, it has been demonstrated that the SWIR bands can be used for atmospheric correction.

Future ocean (water) color satellite sensor needs to include the SWIR bands with sufficient SNR values for cases with turbid waters.

20


Thank You!Thank You!


Comparisons Between MODIS and In Situ (SeaBASS) Data

NIR-SWIR Method Product Data #

Slope Int† R‡ Mean Ratio€ Medium Ratio&

nLw(412) 116 0.827 0.198 0.873 1.120 1.084

nLw(443) 128 0.952 0.073 0.891 1.040 1.025

nLw(488) 104 1.057 -0.050 0.951 0.997 1.015

nLw(531) 32 1.066 -0.069 0.961 1.018 1.040

nLw(551) 116 1.106 -0.046 0.968 1.047 1.055

nLw(667) 97 0.822 0.014 0.727 1.296 1.129

Overall nLw( )

593 0.998 0.024 0.935 1.090 1.030

nLw(443)/ nLw(551)

43 1.021 0.025 0.879 1.052 1.032

nLw(488)/ nLw(551)

43 1.034 -0.023 0.953 1.017 0.996

†Intercept for lin efi t ‡Correlatio ncoefficient €Mean ratio of MODIS v .s in situ dat a &Medium ratio of MODIS v .s in situ data

Menghua Wang, NOAA/NESDIS/STAR Remote Sensing of Water Properties Using the SWIR- based Atmospheric...

Documents

Transcript of Menghua Wang, NOAA/NESDIS/STAR Remote Sensing of Water Properties Using the SWIR- based Atmospheric...