| GISdevelopment.net ---> Application ---> Natural Resource Management |
Water Quality Mapping Using Landsat TM imagery over Penang Island, Malaysia
H. S. Lim, M. Z. MatJafri, K. Abdullah, A. N. Alias and N. Mohd. Saleh
School of Physics,
Universiti Sains Malaysia,
11800 Penang, Malaysia
E-mail: hslim@usm.my, mjafri@usm.my, khirudd@usm.my
Tel: +604-6533888, Fax: +604-6579150
Abstract
This study uses an empirical model, based on actual water quality of total suspended solids (TSS) measurements from the Penang Strait, Malaysia to predict TSS based on optical properties of satellite digital imagery. The study area is Penang Island, Malaysia which is situated within latitudes 5o 12 N to 5o 30 N and longitudes 100o 09 E to 100o 26 E. The proposed algorithm is based on the reflectance model that is a function of the inherent optical properties of water, which can be related to its constituents concentrations. Concentrations of total suspended sediment (TSS) can be estimated by establishing an optical model that correlates the TSS with the radiance. Seawater reflectance in particular is dependent on the coefficient of absorption and backscattering of their inherent optical properties. The digital numbers for each band corresponding to the sea-truth locations were extracted and then converted into radiance values and reflectance values. The reflectance values were used for calibration of the water quality algorithm. The efficiency of the proposed algorithm was investigated based on the observations of correlation coefficient (R) and root-mean-square deviations (RMS) with the sea-truth data. The proposed algorithm is considered superior to other tested algorithms based on the values of the correlation coefficient, R=0.93 and root-mean-square error, RMS=9 mg/l. This algorithm was then used to map the TSS concentration over Penang, Malaysia. The TSS map was color-coded and geometrically corrected for visual interpretation.
Introduction
Water pollution problem becomes increasingly critical in this present-day, whether in the developed or developing countries. Water management is one of the important issues in this 21st century. Remote sensing is a useful and advanced technique for mapping water quality. Determination of water quality parameters using regression algorithm technique has been adopted by in many workers [Dekker, et al., (2002), Tassan, (1993) and Doxaran, et al., (2002)]. The objective of this study is to evaluate the feasibility of using spatial digital satellite images for TSS mapping over Penang by using a newly developed algorithm.
Water quality parameter used in this study is the Total Suspended Solids (TSS). We used high spatial digital camera image, with a pixel size of 1.1 m, for the purpose. Water samples were collected simultaneously with the image taken from an aircraft at an altitude of 8000ft.; the water samples were used in the training analysis.
Study Area
The study area is Penang Island, Malaysia which is situated within latitudes 5o 12 N to 5o 30 N and longitudes 100o 09 E to 100o 26 E (Figure 1). The corresponding TSS measurements were collected at several selected locations. Laboratory analysis for the determination of TSS follows the procedure as suggested by Strickland and Parsons 1972.
The location of the study area
Optical model of water
A physical model relating radiance from the water column and the concentrations of the water constituents provides the most effective way of analyzing remotely-sensed data for water quality studies. Reflectance is particularly dependent on inherent optical properties: the absorption coefficient and the backscattering coefficient. Remote sensing reflectance given by Doxaran et al. (2002) as
(1)
where = internal Fresnel reflectance
= air-water Fresnel reflection at the
interface
= water-air reflection
n = refractive index (1.34)
Q = p
R = reflectance
Equation 1, according to Doxaran et al.(2002) can also be written as
(2)
So, equation (2) can be written as remote sensing reflectance as
(3)
For estuaries water, backscatter is much less significant than absorption (Gohin et al., 2002). The irradiance reflectance just below the water surface, R(l), is given by Morel dan Prieur (1977), Siddorn et al. (2001) and Kirk (1984) as
(4) where
t = constant
b = backscattering coefficient
a = absorption coefficient
The inherent optical properties are determined by the contents of the water. The contributions of the individual components to the overall properties are strictly additive (Gallegos and Correl, 1990). For a case involving two water quality components, i.e. chlorophyll, C, and suspended sediment, P, the simultaneous equations for the two channels given by Gallie and Murtha (1992) can be expressed as
(5) (6) where
bbw(i) = backscattering coefficient
bbc* = chlorophyll coefficient
bbp = sediment coefficient
aw(i) = absorption coefficient
ac* = chlorophyll specific absorption coefficient
ap* = sediment specific absorption coefficient
C = chlorophyll
P = suspended sediment
Regression Algorithm
TSS concentration can be obtained by solving the two simultaneous equations to get the series of terms R1 and R2 that is given as P = ao+a1R1+a2R2+a3R1R2+ a4R12+ a5R22+ a6R12R2+ a7R12R22+a8R12R22+ (7)
where aj, j = 0, 1, 2, are the coefficient for equation (7) that can be solved empirically using multiple regression analysis. This equation can also be extended to the three-band method given as
P = eo+e1R1+e2R2+e3R3+ e4R1R2+ e5R1R3+ e6R2 R3+ e7R1 2+e8R22+e9R32 (8) where the coefficient ej, j = 0, 1, 2, can also be solved empirically.
Data analysis and results
An Landsat TM satellite scene of the study area captured on 24 February 2007 was used in the present investigation. Figure 2 show the raw Landsat TM satellite scene. All image-processing tasks were carried out using PCI Geomatica version 10.1 digital image processing software at the School Of Physics, Universiti Sains Malaysia (USM).
Raw satellite image.
The digital numbers (DN) for each band corresponding to the sea-truth locations were determined. The satellite image was then geometrically corrected by second order polynomial equation using the nearest neighbor method. The DNs values extracted using the window size of 3 by 3 was used due to the higher correlation coefficient (R) with the sea-truth data. A simple atmospheric correction, namely darkest pixel technique was performed in this study. This is a very simple correction, based on 2 assumptions:
The first assumption is that in the darkest water pixel of the image there is total light absorption and the radiation light recorded by this pixel comes from the atmospheric path radiance.
Secondly it is assumed that the atmospheric path radiance is uniform all over the image. The radiation of the darkest water pixel (assumed to represent the atmosphere) is subtracted from the whole image. The darkest pixel is found by searching for the lowest values over water for all wavelengths. The pixel with the lowest value for each band was selected as the darkest pixel.
The proposed algorithm produced the correlation coefficient of 0.9345 between the predicted and the measured TSS values and RMS value of 9.12 mg/l. A map of the chlorophyll parameter was then generated using the calibrated proposed algorithm. Then the generated chlorophyll map was geometrically corrected using the cubic convolution method to produce a smoother map. The generated map was filtered using 3 by 3 pixels averaged to remove random noise and then colour-coded for visual interpretation as shown in Figure 3.
Map of TSS around Penang Island, Malaysia (Blue < 50 mg/l, Green = (50-10) mg/l, Yellow = (100-150) mg/l, Orange = (150-200) mg/l, Red = (>200) mg/l and Black = Water and cloud area)
Conclusion
This study produced a promising result for TSS measurement retrieval form Landsat TM image. The proposed technique can be used for the determination of the TSS values from the satellite image with a reasonable accuracy. This preliminary analysis showed that this technique could be applied to generate a water quality map of a study area.
Acknowledgements
This project was carried out using USM short term grants and Science Fund. We would like to thank the technical staff who participated in this project. Thanks are extended to USM for support and encouragement.
References
Dekker, A. G., Vos, R. J. dan Peters, S. W. M. (2002). Analytical algorithms for lakes water TSM estimation for retrospective analyses of TM dan SPOT sensor data. International Journal of Remote Sensing, 23(1), 15-35. Doxaran, D., Froidefond, J. M., Lavender, S. dan Castaing, P. (2002). Spectral signature of highly turbid waters application with SPOT data to quantify suspended particulate matter concentrations. Remote Sensing of Environment, 81, 149-161. Gallegos, C. L. and Correl, D. L. (1990). Modeling spectral diffuse attenuation, absorption and scattering coefficients in a turbid estuary. Limnology and Oceanography, 35, 1486-1502.
Gallie, E. A. and Murtha, P. A. (1992). Specific absorption and backscattering spectra for suspended minerals and chlorophyll-a in Chilko Lake, British Columbia. Remote Sensing of Environment, 39, 103-118.
Gohin, F., Druon, J. N. and Lampert, L. (2002). A five channel chlorophyll concentration algorithm applied to SeaWiFS data processed by SeaDAS in coastal waters. International Journal of Remote Sensing, 23(8), 1639-1661. Kirk, J. T. O. (1984). Dependence of relationship between inherent and apparent optical properties of water on solar altitude. Limnology and Oceanography, 29, 350-356.
Morel, A. and Prieur, L. (1977). Analysis of variation of ocean color. Limnology and Oceanography, 22, 709 722. Siddorn, J. R., Bowers, D. G. dan Hoguane, A. M. (2001). Detecting the Zambezi river plume using observed optical Properties. Marine Pollution Bulletin, 42 (10), 942-950.
Tassan, S. (1993). An improved in-water algorithm for the determination of chlorophyll and suspended sediment concentration from Thematic Mapper data in coastal waters. International Journal of Remote Sensing, 14(6), 1221-1229.
H. S. Lim, M. Z. MatJafri, K. Abdullah, A. N. Alias and N. Mohd. Saleh
School of Physics,
Universiti Sains Malaysia,
11800 Penang, Malaysia
E-mail: hslim@usm.my, mjafri@usm.my, khirudd@usm.my
Tel: +604-6533888, Fax: +604-6579150
Abstract
This study uses an empirical model, based on actual water quality of total suspended solids (TSS) measurements from the Penang Strait, Malaysia to predict TSS based on optical properties of satellite digital imagery. The study area is Penang Island, Malaysia which is situated within latitudes 5o 12 N to 5o 30 N and longitudes 100o 09 E to 100o 26 E. The proposed algorithm is based on the reflectance model that is a function of the inherent optical properties of water, which can be related to its constituents concentrations. Concentrations of total suspended sediment (TSS) can be estimated by establishing an optical model that correlates the TSS with the radiance. Seawater reflectance in particular is dependent on the coefficient of absorption and backscattering of their inherent optical properties. The digital numbers for each band corresponding to the sea-truth locations were extracted and then converted into radiance values and reflectance values. The reflectance values were used for calibration of the water quality algorithm. The efficiency of the proposed algorithm was investigated based on the observations of correlation coefficient (R) and root-mean-square deviations (RMS) with the sea-truth data. The proposed algorithm is considered superior to other tested algorithms based on the values of the correlation coefficient, R=0.93 and root-mean-square error, RMS=9 mg/l. This algorithm was then used to map the TSS concentration over Penang, Malaysia. The TSS map was color-coded and geometrically corrected for visual interpretation.
Introduction
Water pollution problem becomes increasingly critical in this present-day, whether in the developed or developing countries. Water management is one of the important issues in this 21st century. Remote sensing is a useful and advanced technique for mapping water quality. Determination of water quality parameters using regression algorithm technique has been adopted by in many workers [Dekker, et al., (2002), Tassan, (1993) and Doxaran, et al., (2002)]. The objective of this study is to evaluate the feasibility of using spatial digital satellite images for TSS mapping over Penang by using a newly developed algorithm.
Water quality parameter used in this study is the Total Suspended Solids (TSS). We used high spatial digital camera image, with a pixel size of 1.1 m, for the purpose. Water samples were collected simultaneously with the image taken from an aircraft at an altitude of 8000ft.; the water samples were used in the training analysis.
Study Area
The study area is Penang Island, Malaysia which is situated within latitudes 5o 12 N to 5o 30 N and longitudes 100o 09 E to 100o 26 E (Figure 1). The corresponding TSS measurements were collected at several selected locations. Laboratory analysis for the determination of TSS follows the procedure as suggested by Strickland and Parsons 1972.
The location of the study area
Optical model of water
A physical model relating radiance from the water column and the concentrations of the water constituents provides the most effective way of analyzing remotely-sensed data for water quality studies. Reflectance is particularly dependent on inherent optical properties: the absorption coefficient and the backscattering coefficient. Remote sensing reflectance given by Doxaran et al. (2002) as
(1)
where = internal Fresnel reflectance
= air-water Fresnel reflection at the
interface
= water-air reflection
n = refractive index (1.34)
Q = p
R = reflectance
Equation 1, according to Doxaran et al.(2002) can also be written as
(2)
So, equation (2) can be written as remote sensing reflectance as
(3)
For estuaries water, backscatter is much less significant than absorption (Gohin et al., 2002). The irradiance reflectance just below the water surface, R(l), is given by Morel dan Prieur (1977), Siddorn et al. (2001) and Kirk (1984) as
(4) where
t = constant
b = backscattering coefficient
a = absorption coefficient
The inherent optical properties are determined by the contents of the water. The contributions of the individual components to the overall properties are strictly additive (Gallegos and Correl, 1990). For a case involving two water quality components, i.e. chlorophyll, C, and suspended sediment, P, the simultaneous equations for the two channels given by Gallie and Murtha (1992) can be expressed as
(5) (6) where
bbw(i) = backscattering coefficient
bbc* = chlorophyll coefficient
bbp = sediment coefficient
aw(i) = absorption coefficient
ac* = chlorophyll specific absorption coefficient
ap* = sediment specific absorption coefficient
C = chlorophyll
P = suspended sediment
Regression Algorithm
TSS concentration can be obtained by solving the two simultaneous equations to get the series of terms R1 and R2 that is given as P = ao+a1R1+a2R2+a3R1R2+ a4R12+ a5R22+ a6R12R2+ a7R12R22+a8R12R22+ (7)
where aj, j = 0, 1, 2, are the coefficient for equation (7) that can be solved empirically using multiple regression analysis. This equation can also be extended to the three-band method given as
P = eo+e1R1+e2R2+e3R3+ e4R1R2+ e5R1R3+ e6R2 R3+ e7R1 2+e8R22+e9R32 (8) where the coefficient ej, j = 0, 1, 2, can also be solved empirically.
Data analysis and results
An Landsat TM satellite scene of the study area captured on 24 February 2007 was used in the present investigation. Figure 2 show the raw Landsat TM satellite scene. All image-processing tasks were carried out using PCI Geomatica version 10.1 digital image processing software at the School Of Physics, Universiti Sains Malaysia (USM).
Raw satellite image.
The digital numbers (DN) for each band corresponding to the sea-truth locations were determined. The satellite image was then geometrically corrected by second order polynomial equation using the nearest neighbor method. The DNs values extracted using the window size of 3 by 3 was used due to the higher correlation coefficient (R) with the sea-truth data. A simple atmospheric correction, namely darkest pixel technique was performed in this study. This is a very simple correction, based on 2 assumptions:
The first assumption is that in the darkest water pixel of the image there is total light absorption and the radiation light recorded by this pixel comes from the atmospheric path radiance.
Secondly it is assumed that the atmospheric path radiance is uniform all over the image. The radiation of the darkest water pixel (assumed to represent the atmosphere) is subtracted from the whole image. The darkest pixel is found by searching for the lowest values over water for all wavelengths. The pixel with the lowest value for each band was selected as the darkest pixel.
The proposed algorithm produced the correlation coefficient of 0.9345 between the predicted and the measured TSS values and RMS value of 9.12 mg/l. A map of the chlorophyll parameter was then generated using the calibrated proposed algorithm. Then the generated chlorophyll map was geometrically corrected using the cubic convolution method to produce a smoother map. The generated map was filtered using 3 by 3 pixels averaged to remove random noise and then colour-coded for visual interpretation as shown in Figure 3.
Map of TSS around Penang Island, Malaysia (Blue < 50 mg/l, Green = (50-10) mg/l, Yellow = (100-150) mg/l, Orange = (150-200) mg/l, Red = (>200) mg/l and Black = Water and cloud area)
Conclusion
This study produced a promising result for TSS measurement retrieval form Landsat TM image. The proposed technique can be used for the determination of the TSS values from the satellite image with a reasonable accuracy. This preliminary analysis showed that this technique could be applied to generate a water quality map of a study area.
Acknowledgements
This project was carried out using USM short term grants and Science Fund. We would like to thank the technical staff who participated in this project. Thanks are extended to USM for support and encouragement.
References
Dekker, A. G., Vos, R. J. dan Peters, S. W. M. (2002). Analytical algorithms for lakes water TSM estimation for retrospective analyses of TM dan SPOT sensor data. International Journal of Remote Sensing, 23(1), 15-35. Doxaran, D., Froidefond, J. M., Lavender, S. dan Castaing, P. (2002). Spectral signature of highly turbid waters application with SPOT data to quantify suspended particulate matter concentrations. Remote Sensing of Environment, 81, 149-161. Gallegos, C. L. and Correl, D. L. (1990). Modeling spectral diffuse attenuation, absorption and scattering coefficients in a turbid estuary. Limnology and Oceanography, 35, 1486-1502.
Gallie, E. A. and Murtha, P. A. (1992). Specific absorption and backscattering spectra for suspended minerals and chlorophyll-a in Chilko Lake, British Columbia. Remote Sensing of Environment, 39, 103-118.
Gohin, F., Druon, J. N. and Lampert, L. (2002). A five channel chlorophyll concentration algorithm applied to SeaWiFS data processed by SeaDAS in coastal waters. International Journal of Remote Sensing, 23(8), 1639-1661. Kirk, J. T. O. (1984). Dependence of relationship between inherent and apparent optical properties of water on solar altitude. Limnology and Oceanography, 29, 350-356.
Morel, A. and Prieur, L. (1977). Analysis of variation of ocean color. Limnology and Oceanography, 22, 709 722. Siddorn, J. R., Bowers, D. G. dan Hoguane, A. M. (2001). Detecting the Zambezi river plume using observed optical Properties. Marine Pollution Bulletin, 42 (10), 942-950.
Tassan, S. (1993). An improved in-water algorithm for the determination of chlorophyll and suspended sediment concentration from Thematic Mapper data in coastal waters. International Journal of Remote Sensing, 14(6), 1221-1229.