The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200

Contents lists available at ScienceDirect

The Egyptian Journal of Remote Sensing and Space Sciences

journal homepage: www.sciencedirect.com

Research Paper

Role of statistical remote sensing for Inland water quality parameters

prediction
Abdelmalik K.W.
Department of Geology, Faculty of Science, Ain Shams University – Cairo, Egypt

a r t i c l e i n f o a b s t r a c t

Article history: Understanding the statistical relations among the Advanced Space borne Thermal Emission and
Received 8 October 2016 Reflection Radiation (ASTER) data and observed water quality parameters, in order to develop a mathe-
Revised 5 December 2016 matical relation for the precise prediction of the missing data in a given area, is the main aim of the pre-
Accepted 10 December 2016
sent study. This should enable to establish a spatial distribution map for each parameter of water quality
Available online 29 December 2016
for the area. The method was applied to Qaroun Lake in the Fayoum depression of Egypt.
The water quality parameters obtained from ASTER data used in the present work are: Temperature,
Keywords:
Turbidity, Hydrogen ion concentration (pH), Salinity, Total Dissolved Solids (TDS), Electrical
Remote sensing
Regression
Conductivity (EC), Total alkalinity, Total Organic Carbon (TOC) and Ortho-phosphorus.
Inland water quality 18 water sample data were used in the study: 15 sample data for mathematical model construction,
ASTER giving the relation between the ASTER values and the water quality parameters, while 3 samples data
were used to test the obtained model.
The SPSS software of IBM was also used in the present research for the applied statistical analysis.
The analysis showed a significant correlation between the observed values and the remotely sensed
data with R2 > 0.94 and sig. < 0.01 in most cases. The calculated values resulting through the obtained
equation showed a high accuracy: Root mean square error (RMSE) ranging from 0.8 to 0.014 and
Standard Estimated Error (SEE) ranging from 0.9 to 0.0116.
ERDAS Imagine and ArcGIS packages were used for applying the obtained mathematical model and
spatial distribution map to the Qaroun Lake.
Ó 2016 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-
nd/4.0/).

1. Introduction Recently many authors as Gholizadeh et al. (2016), Matthews

(2011) and Kutser (2009) and many others; provided a detailed
Water quality is an extremely important environmental factor review of available types of remotely sensed data, which can be
as it affects human beings and their economic activities (e.g. useful and effective in water quality studies. Authors mainly clas-
Chen et al., 2009 and He et al., 2008). The major factor that affects sify these data into 3 categories based upon their spatial resolution.
the physical aspect (such as color) of the water is the concentration Overall these categories are:
and distribution of fine suspended components and dissolved
matter (Clark et al., 1980). 1. High resolution data: such as the data of IKONOS, Quick bird
The traditional methods of measuring and studying water qual- (4 m) and the Worldview Series (0.5–2 m).
ity parameters of water bodies are expensive, both in cost and 2. Moderate resolution data: such as EO-1 Hyperion, ALI (10–30 m),
time, when compared to remote sensing techniques, especially if Landsat series (15–120 m) and ASTER (15–90 m).
they deal with large areas. 3. Regional/Global resolution data: such as NOAA-16 AVHRR
Many authors have studied parameters affecting water quality (1000 m), ATSR-1,2 (1000 m), terra MODIS (250-1000 m) and
by using remotely sensed data. seaWiFs (1130 m).

Broadly speaking many parameters have been considered by

using a remote sensing approach, such as: Physiochemical parame-
Peer review under responsibility of National Authority for Remote Sensing and
Space Sciences.
ters (pH, Temperature, EC, Salinity, TDS, TSS, Turbidity and Total
E-mail address: Karim.abdelmalik@gmail.com alkalinity), Organic parameters (Biochemical oxygen demand BOD,

http://dx.doi.org/10.1016/j.ejrs.2016.12.002
1110-9823/Ó 2016 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
194 K.W. Abdelmalik / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200

TOC, DOC, TIC) and Microbiological parameters (Total Colloform TC, 3.1. Remote sensing data preparation
Cholorophyll Chl-a). Of course the relationship between the water
quality parameters and the radiance or reflection is affected by 3.1.1. Atmospheric correction
changing in one or more of these parameters. Moreover, it also In any satellite imagery that observes the surface of the earth in
affected by seasonal changing. So, any investigation deals with the visible/infrared regions of the electromagnetic spectrum, the
mentioned relation must be addressed to particular season and recorded signal contains two brightness types; the first due to
conditions. the actual reflectance from the surface while the second one is
Several authors have used and applied the regression statistical due to the atmospheric effect and is considered as noise.
techniques in order to estimate and calculate water quality param- So, the atmospheric correction is an essential step of the image
eters as Khattab and Merkel (2014), Guan (2009), Hirthle and processing chain. Therefore, to obtain accurate and precise quanti-
Rencz (2003), Coskun et al. (2001) and Choubey and tative data using remote sensing, it is necessary to perform atmo-
Subramanian (1990). spheric correction (Tyagi and Bhosle, 2011; Chander et al., 2009;
The main aim of the present study was to explore several algo- Liang et al., 2001; Lillesand et al., 2004).
rithms in order to determine the most suitable ones for the estima- In the current study the Aster (V/VNIR and SWIR) regions atmo-
tion of temperature, Turbidity, potential Hydrogen ions (pH), spheric correction process was executed using the fast line-of-sight
Electric Conductivity (EC), Total Dissolved Solids (TDS), Salinity, atmosphere analyses of hypercubes (FLAASH) modules of ENVI 5.1,
Total alkalinity, Orthophosphorous, Total Organic Carbon (TOC), while for the thermal (TIR) region, the thermal atmospheric correc-
for the Qaroun lake using the statistical approach and ASTER Data. tion of calibration utilities of the same software version was used.
An additional goal was to map the water quality parameters by
using estimated ones instead of conventional methods. 3.1.2. Precisely mapping of water lines
In order to extract and map the water line, the Normalized Dif-
ference Water Index (NDWI) which was developed by McFeeters
2. Study area and general geology (1996) was applied on the image. The output of the NDWI was
found to range between 1 and +1, the threshold is, thus, zero.
The Fayoum is an enormous depression located in the north So, the water has positive values while non-water gets the negative
eastern part of the Western Desert of Egypt, The Qaroun Lake occu- ones. The effect of wet soil in Qaroun Lake’s edge could be
pies the lowest part of it. Qaroun Lake is laying between latitudes neglected specially in the season of study. Therefore, the lake bor-
29° 250 & 29° 300 N and longitudes 30° 250 & 30° 500 E extending over der line can be obtained by vectorization of the water/non-water
an area exceeding 200 km2. It has an elongated shape, about image. The raster data of the Qaroun Lake water body has been
42 km., east-west and is about 10 km. in width. Qaroun Lake was subset by the water line final result shape file.
declared a natural reserve by virtue of the Prime Minister decision
No. 348 of 1989, with an aim to protect and conserve the biological,
3.2. Statistical analysis
archaeological and geological diversity of the area.
The geology of the Fayoum area has attracted many authors as
The correlation and regression analysis techniques are widely
Beadnell (1905), Said (1962), Tamer et al. (1975), Swedan (1986,
used for studying and investigating the strength of the relationship
1992), Abd El Motaal (1989), Abdel Hafez (1991), Nabawy (1998),
– if any – between two quantitative variables. The correlation anal-
El-Younsy et al. (1999), Azab (2001), Issawi et al. (2001) and
ysis quantifies the strength of the relation between a variable pars
Kusky et al. (2011). According to these previous works, the Fayoum
while the regression analysis develops the mathematical equation
depression occurs in a thick sedimentary series of Middle Eocene
which express such relation.
to Quaternary age.
In the present study, SPSS software package of IBM was used for
Most of joints and faults observed by Kusky, strike NW-SE
executing both correlation and regression analyses in order to
(Clysmic Trend) and E-W (Tethyan Trend).
investigate and study the relationship between scoped parameters
and remotely sensed data of the study area.
3. Data and methodology
4. Results and discussion
In the present study, both water samples and ASTER terra data
were used in order to develop and drive an accurate simple algo- 4.1. Regression analysis
rithm to retrieve the water quality parameters for Qaroun Lake.
The sampling and water analysis were executed during May/ Various algorithms have been used by many authors to examine
June 2007, based on the APHA (1992) in sampling and testing characteristic relationships between remote sensed data and water
methodology (Gad, 2009). quality physiochemical, Organic and Microbiological parameters.
The current study uses water samples parameters of 18 loca- (Khattab and Merkel, 2014; Elsaadi et al., 2014; Yüzügüllü and
tions representative to the Qaroun lake water body (Fig. 1). The Aksoy, 2011; Coskun et al., 2008).
parameters of 15 samples were used for processing the regression The present study includes the elaboration of regression analy-
analysis to build up the equations expressing the relations sis algorithms to examine and use the relations between the ASTER
between the parameters and the ASTER readings, while the data reflectance bands and the water quality parameters of Qaroun
remaining 3 samples were used for testing the accuracy of the Lake. The parameters used in such analysis are: Temperature, Tur-
developed equations. bidity, pH, Salinity, TDS, EC, Total alkalinity, TOC and
The remotely sensed image, used in the present study, were Orthophosphorus.
those of ASTER Terra obtained from the US Geological Survey Statistically, the created equations and relations were chosen
(USGS) via Earth Explorer site (Earth explorer.usgs.gov). The used according to three factors, which were: Adjusted Square Correction
scene were obtained with Path and Row 177/39 at an acquisition Coefficient (R2: Which is the percentage of the response variable vari-
data of 24/6/2007 with zero cloud coverage. ation that explained by the model, the higher value the better fitting of
(Fig. 2) illustrates the flowchart of the process cycle for the pre- the model), Standard Error of Estimation (SEE: Standard Error of
sent work. Estimation, which measurers the accuracy of the model in prediction,
 K.W. Abdelmalik / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200 195

Fig. 1. Location map of water samples.

Fig. 2. Work flowchart organizing and illustrating the processing cycle.

the lower the value the higher the accuracy) and significance or prob- The EC was highly correlated with bands b3 to b6 and b8 (cor-
ability value (sig.: or p-value, which is primary output of statistical relation coefficient > 0.9 and sig < 0.001) therefore, the selected
tests which indicates the probability of observing error in prediction, band ratios [(b4/b3) * b8] clearly highly correlated with EC values,
the lower the value the better the model ‘‘p” must less than 0.05). the best fit curve equation model was the quadratic model
The Statistical Package for Social Sciences version 20.0 (SPSS) of (R2 = 0.995, SEE = 0.617 and sig < 0.001), the other models show a
IBM was used to execute the correlation studies as well as the good R2 values but high SEE values (up to 4.9).
regression analysis and to calculate the mathematical parameters Generally, the water salinity is significantly related with the
of the developed relational equations of the reflectance values of reflectance radiations within visible and near infrared regions
ASTER data and the water parameters (Tables 1 and 2). (Pegau et al., 1997). Statistically, it is strongly correlated with
The final results of the analysis shows a valid high correlation of ASTER bands b3 to b8 and it was extremely correlated with
the obtained models with measured environmental parameters, as selected band ratio [(b5/b3) * b7]. The best fit equation to repre-
most R2 were found to be >0.94, and most sig. were found to be sent the equation was the quadratic model (R2 = 0.985, SEE = 0.7
<0.01 and SEE indicates an extremely small expected error (0.9– and sig. < 0.0001).
0.01). According to Duong (2012) and Wua et al. (2008), clear water
The pH correlated with reflectance of SWIR (Khattab and has more reflectance within the green band and it decreases in
Merkel, 2014), Thus, it was found, in the present study, that pH the red as well as in the NIR bands. By examining the correlation
has a considerable correlation with band 9 (the longest SWIR band coefficient of ASTER bands with observed turbidity values, this
in ASTER data). However, by examining the relation with band coefficient shows a strong value with bands b3, b4, b5, b6 and
ratio b9/b8, a still higher Correlation coefficient (=0.7) was found, b8. Moreover, it shows also a high value with band ratio
while the significance of the computed values was =0.004. The [(b3 * b2)/b1] (correlation coefficient 0.96). Statistically, the best
most and highly relational parameters found was the Quadratic models illustrating the relation between turbidity values and the
(second order polynomial) relation (R2 = 0.95, SEE = 0.04 and sig. selected band ratio were the quadratic equation (R2 = 0.998,
< 0.001), All other relations have an R2 < 0.75 and sig. > 0.05. SEE = 0.95) and the growth equation (R2 = 0.75, SEE = 0.58). But
196 K.W. Abdelmalik / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200

Table 1
The statistical parameters of the regression analysis for all studied models.

Water quality Independent Linear Logarithmic Inverse Quadratic Cubic Compound Power Growth
parameter 2 2 2 2 2 2
R SEE R SEE R2 SEE R SEE R SEE R SEE R SEE R2 SEE
(dependent)
pH Band 9/Band 8 0.692 0.117 0.733 0.109 0.771 0.101 0.951 0.049 0.943 0.051 0.690 0.014 0.731 0.013 0.690 0.014
EC (Band4/Band3) * Band8 0.903 2.808 0.818 3.840 0.702 4.920 0.996 0.617 0.995 0.645 0.895 0.127 0.807 0.172 0.895 0.127
Turb (Band3 * Band2)/Band1 0.908 6.497 0.878 7.487 0.841 8.544 0.998 0.956 0.998 0.833 0.722 0.584 0.681 0.626 0.752 0.584
TDS Band3 0.861 2.154 0.827 2.402 0.787 2.660 0.997 0.330 0.997 0.331 0.877 0.137 0.845 0.154 0.877 0.137
Salinity (Band5/Band3) * band7 0.780 2.815 0.681 3.391 0.570 3.939 0.985 0.744 0.987 0.687 0.780 0.199 0.679 0.240 0.780 0.199
temp 1/2 Band8  [(Band 11+Band13)/ 0.245 0.561 0.299 0.541 0.353 0.520 0.535 0.334 0.530 0.333 0.246 0.018 0.300 0.018 0.246 0.018
band10]
Alkalinity (b4/b5) + b6 0.726 0.026 0.673 0.028 0.603 0.031 0.822 0.023 0.822 0.022 0.643 0.091 0.597 0.097 0.643 0.091
Orthophosphorus Band 5 0.782 0.022 0.715 0.025 0.636 0.028 0.944 0.012 0.936 0.012 0.730 0.119 0.668 0.132 0.730 0.119
TOC Band 3 0.847 3.133 0.813 3.462 0.774 3.806 0.985 0.909 0.983 1.058 0.874 0.143 0.841 0.160 0.874 0.143

Table 2
The regression and testing analysis results of selected model for each parameter.

Water quality Equation Regression Testing

parameter
R2 RMSE SEE sign RMSE SEE
2
pH y = 10.083(band 9/Band 8) + 26.022(band 9/Band 8)  8.3195 0.9506 0.045 0.0488 0.0001 0.0810 0.0247
EC y = 0.1252[(Band4/Band3) * Band8]2 + 4.1531[(Band4/Band3) * Band8] + 10.527 0.9956 0.575 0.617 0.0001 0.8780 0.9000
Turb y = 0.4121[(Band3 * Band2)/Band1]2  31.74[(Band3 * Band2)/Band1] + 612.17 0.998 0.891 0.956 0.0001 0.9160 0.1290
TDS y = 0.1204(Band3)2 + 10.663(Band3)  207.21 0.9972 0.296 0.559 0.0001 0.3020 0.2850
Salinity y = 0.0967 [(Band5/Band3) * band7]2 + 4.4921[(Band5/Band3) * band7]  22.997 0.9857 0.693 0.7443 0.0001 0.0820 0.7520
temp y = 0.0584[(1/2) Band8  ((Band 11 + Band13)/band10)]2 + 1.9456[(1/2) Band8  ((Band 11 0.5347 0.426 0.3336 0.012 0.7296 0.2863
+ Band13)/band10)] + 15.443
Alkalinity y = 2E05[(b4/b5) + b6]3  0.0013[(b4/b5) + b6]2 + 0.0297[(b4/b5) + b6] + 0.0363 0.8217 0.117 0.0226 0.0001 0.1339 0.0251
Orthophosphorus y = 0.0006(b5)2  0.0388(b5) + 0.7684 0.944 0.014 0.0116 0.012 0.0117 0.001
TOC y = 0.1676 (b3)2 + 14.857 (b3)  289.86 0.9852 0.87 0.9092 0.012 0.5676 0.3266

by testing the two equations the results were RMSE = 0.9, However, in the present study, Correlation of calculated concentra-
SEE = 0.12 and RMSE = 0.82, SEE = 0.68 respectively, so the quadra- tion values of phosphorus from remote sensing data with analyti-
tic model was preferably applied (Table 2). cal data shows higher correlation with bands b4, b5, b6, b7 and b8
A clear correlation has been shown between TDS measured val- in addition to ASTER TIR bands b13, b14. Moreover, the simplest
ues and ASTER bands (band 3 to band 6 and band 8). Mostly the and highly precise results of regression analyses was found in
higher correlation was with band 3 (correlation coefficient = 0.97). the quadratic model for b5, R2 = 0.944, SEE = 0.012 (Table 1).
By applying a regression analysis for the different models, the val- By applying and testing the obtained equation, the
ues of R2 raised in the quadratic, cubic, compound, power and RMSE = 0.014, SEE = 0.0116 and RMSE = 0.0117, SEE = 0.001
growth models (Table 1), but by testing the equations the best respectively (Table 2).
model was found to be the quadratic one (RMSE = 0.302 and Chang et al. (2014) indicated that the range 0.459–0.9 lm had
SEE = 0.285) the highest contribution in determining the concentration of
Commonly, the mathematical combination of the Thermal TOC. Statistically, the highest correlation of TOC value was found
Infrared (TIR) bands (band 10 to band 14, in ASTER data) are gen- to be with band 3 (0.76–0.86 lm). Moreover, by applying and test-
erally used for estimating the water surface temperature (Wubet, ing the developed equation model, we found: RMSE = 0.87, SEE
2003 and Matsunaga, 1996). Unfortunately, in the present study 0.909 and RMSE = 0.567, SEE = 0.326 respectively (Table 2), which
the bands were found to be moderately correlated with the could be considered as a good result for the estimation of missing
observed temperature values (correlation coefficient just exceed- data.
ing 0.5) and the selected band ratio gives R2 = 0.54 and SEE = 0.33 For further validation, fig. 3 illustrates the graphical relation
in the quadratic model (Table 1). However, by applying and testing between the actually measured parameter and calculated one for
the developed equation, the results were found to be RMSE = 0.426, each studied water quality parameter. The R2 value greater than
SEE = 0.333 and RMSE = 0.72, SEE = 0.286 respectively (Table 2), 0.95 for all plots except those for alkalinity (0.82) and temperature
which was considered as a good result for estimation. (0.53) and RMSE value for all parameters were ranging between
The measured alkalinity has higher correlation with ASTER 0.014 up to 0.89.
bands b4, b5 and b6 (correlation coefficient = 0.92, 0.89 and 0.86
respectively). A regression analysis was applied for selected band 4.2. Spatial distribution maps
ratio [(b4 + b5)/b6] for all regression models. The result (Table 1)
shows the highest R2 value (0.822) and lowest SEE value (0.022) The developed models and created equations for each water
for the cubic model. By applying and testing the resulting model quality parameter were applied on ASTER data in order to establish
we obtained: RMSE = 0.117, SEE = 0.022 and RMSE = 0.133, a spatial distribution map for each studied parameter. Fig. 4 illus-
SEE = 0.025 respectively (Table 1). trates the produced maps for the nine parameters used.
According to Khattab and Merkel, 2014, Sridhar et al., 2009, The spatial distribution of estimated Temperature generally
Bistani, 2009 and Alparslan et al., 2007, The phosphorous has shows an increase along the shore line (higher values at the north-
strong correlation with the reflectance within wavelength interval ern edge) and decrease of temperature toward the central part of
1.6–2.36 lm which is covered by ASTER bands b4 through b9. the lake body (Fig. 4a).
K.W. Abdelmalik / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200 197

Fig. 3. Relational plot of the calculated water parameter values vs. the measured ones.
198 K.W. Abdelmalik / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200

Fig. 4. Spatial distribution map for estimated water quality parameters.

 K.W. Abdelmalik / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200 199

The higher values of pH values (up to 8.47) are found at the Chander, G., Markham, B.L., Helder, D.L., 2009. Summary of current radiometric
calibration coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI sensors.
northern part of the lake (Fig. 4b), decreasing towards the southern
Remote Sens. Environ. 113 (5), 893–903.
side, especially at the outlets of drains (minimum value 7.5). Chang, N.B., Vannah, B.W., Yang, Y.J., Elovitz, M., 2014. Integrated data fusion and
Fig. 4(c, d) clearly illustrates the same spatial distribution pat- mining techniques for total organic carbon concentrations in lake. Integr. J.
tern for both Salinity and TDS respectively. The concentration of Remote Sens. 35 (3), 1064–1093.
Chen, S., Fang, L., Zhang, L., Huang, W., 2009. Remote sensing of turbidity in
ions in the middle part of the water body may be the main reason seawater intrusion reaches of Pearl River Estuary—a case study in Modaomen
for this pattern. This is closely related to the distribution of EC water way, China. Estuar. Coast. Shelf Sci. 82 (1), 119–127.
(Fig. 4e), as the maximum values are found in the middle of the Choubey, V.K., Subramanian, V., 1990. Nature of suspended solids and IRS-IA-LISS-I
data: a case study of Tawa reservoir Narmada basin. Remote Sens. Environ. 34,
lake. This is due to ions mobility, whereas conductivity is also 207–215.
affected by heavy metal ions released into water by acid drainage Clark, D.K., Baker, E., Strong, A.E., 1980. Upwelled spectral radiance distribution in
(Gad, 2009), this is obvious, as recorded pH in this region shows relation to particulate matter in sea water. Boundary-Layer Meteor. 87, 287–
298.
minimum values. Coskun, G., Ekercin, S., O ztopal, A., 2001. Monitoring and water pollution modelling
Alkalinity shows the same spatial distribution pattern of of the Bosphorus by regression analysis using Landsat-TM data. In: EARSeL 21th
Orthophosphorous (Fig. 4f, g), as they show mainly higher values Symposium Observing our Environment from Space/New Solutions for a New
Millennium, ENSG, 14–16 May, Marne-la-Vallee (Paris), France.
in the western region of the lake and minimum values at drain out- Coskun, H.G., Tanik, A., Alganci, U., Cigizoglu, H.K., 2008. Determination of
lets, due to absence of buffering materials which neutralize the environmental quality of a drinking water reservoir by remote sensing, GIS
slightly acidic inflow. and regression analysis. Water Air Soil Pollut. 194 (1–4), 275–285. http://dx.doi.
org/10.1007/s11270-008-9716-x.
Orthophosphorous distribution in Qaroun Lake shows the same
Duong, N.D., 2012. Water body extraction from multi spectral image by spectral
pattern of distribution as Alkalinity, where it is found at the bor- pattern analysis. In: International Archives of the Photogrammetry,
ders of the lake, where plants are conspicuous. Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012,
The increase of TOC (Fig. 4h), mainly at the western as well as at XXII ISPRS Congress, 25 August–01 September 2012, Melbourne, Australia,
pp. 181–186.
the northern regions of the lake, may be attributed to planktonic El Saadi, A.M., Yousry, M.M., Jahin, H.S., 2014. Statistical estimation of Rosetta
organisms and organic matter concentrations in open basin area branch water quality using multi-spectral data. Water Sci. J. 28, 18–30.
far from the drains. El-Younsy, A.M., Mansour, H.H., Youssef, M.M., Makram, M.H.M., 1999. Facies,
paleoenvironments and mineralogy of the Upper Eocene-Oligocene sequence
In general the bulk of the lake has a very low turbidity (Fig. 4i), northwest Birket Qarun, Fayum depression, western desert, Egypt. In: The First
except near the shore (up to 100 NTU) and intermediate values of International Conference on the Geology of Africa, Assiut, Egypt, vol. 1, pp. 137–
turbidity near drains outlets. This is due to the presence of total 159.
Gad, A., 2009. Geoenvironmental Assessment of Qarun Lake, Western Desert of
suspended solids (TSS). Egypt (M.Sc. thesis), Ain Shams Univ., Cairo, Egypt, 222 P.
Gholizadeh, M.H., Melesse, A.M., Reddi, L.A., 2016. Comprehensive review on water
quality parameters estimation using remote sensing techniques. Sensors 16,
5. Conclusion 1298.
Guan X., 2009. Monitoring Lake Simcoe water quality using Landsat TM images (M.
Sc. thesis), Univ. Waterloo, Canada.
The present paper illustrates and proves the presence of signif- He, W., Chen, S., Liu, X., Chen, J., 2008. Water quality monitoring in a slightly-
icant empirical relations between ASTER data and observed water polluted inland water body through remote sensing: case study of the Guanting
Reservoir in Beijing, China. Front Environ. Sci. Eng. China 2, 163–171.
quality parameters. The calculated parameters obtained from
Hirthle, H., Rencz, A., 2003. The relation between spectral reflectance and dissolved
ASTER reflectance bands enables an accurate prediction for these organic carbon in lake water: Kejimkujik National Park, Nova Scotia, Canada.
parameters based on the mathematical equations derived through Int. J. Remote Sens. 24 (5), 953–967.
regression analysis. Also, the current research discloses that the Issawi, B., Osman, R., Meibed, A., 2001. The Fayium depression-Qarun Lake: a
geological study. In: 2nd Int. Conf. Geol. Afr., Eg., vol. 2, pp. 307–322.
future study investigating the effect of seasonal changing on the Khattab, M.F., Merkel, B.J., 2014. Application of Landsat 5 and Landsat 7 images data
radiance and/or reflectance values is a must, in order to be more for water quality mapping in Mosul Dam Lake, Northern Iraq. Arab. J. Geosci. 7,
accurate and precise of absolute prediction. 3557–3573.
Kusky, T.M., Ramadan, T.M., Hassaan, M.M., Gabr, S., 2011. Structural and Tectonic
The current study demonstrates that the use of the statistical Evolution of El-Faiyum Depression, North Western Desert, Egypt Based on
techniques discussed in this paper for establishing spatial distribu- Analysis of Landsat ETM+, and SRTM Data. J. Earth Sci. 22, 75–100.
tion maps for significant environmental parameters in water bod- Kutser, T., 2009. Passive optical remote sensing of cyanobacteria and other intense
phytoplankton blooms in coastal and inland waters. Int. J. Remote Sens. 30,
ies should be a rapid and inexpensive convenient method to obtain 4401–4425.
useful environmental data for the environmental sciences. It can be Liang, S., Fang, H., Chen, M., 2001. Atmospheric correction of Landsat ETM+ land
used efficiently for the routine monitoring of lakes specially those surface imagery—part I: methods. IEEE Trans. Geosci. Remote Sens. 39 (11),
2490–2498.
covering large geographic areas.
Lillesand, T.M., Kiefer, R.W., Chipman, J.W., 2004. Remote Sensing and Image
Interpretation. John Wiley and Sons, New York.
Matsunaga, T., 1996. Water surface temperature estimation using linear equations
References of brightness temperature observed by advanced spaceborne thermal emission
and reflection radiometer/thermal infrared radiometer (ASTER TIR) –
Abd El Motaal, E., 1989. Neotectonics in El-Faiyum area, Western Desert, Egypt (M. preliminary evaluation of estimation error with atmospheric temperature and
Sc. thesis), Al-Azhar Univ., Cairo, Egypt, 168 P. humidity data around Japan. J. Remote Sens. Soc. Jpn. 16 (5), 404–415.
Abdel Hafez, N., 1991. Geological and geochemical studies on the Pleistocene old Matthews, M.W., 2011. A current review of empirical procedures of remote sensing
lacustrine deposits in the Fayium depression, Western Desert, Egypt. Sci. Bull. in inland and near-coastal transitional waters. Int. J. Remote Sens. 32 (21),
Minia. 6855–6899.
Alparslan, E., Aydoner, C., Tufekci, V., Tufekci, H., 2007. Water quality assessment at McFeeters, S.K., 1996. The use of normalized difference water index (NDWI) in the
Omerli Dam using remote sensing techniques. Environ. Monit. Assess. 135, delineation of open water features. Int. J. Remote Sens. 17, 1425–1432.
391–398. Nabawy, B.S., 1998. Sedimentological, magnetic and petrophysical characteristics of
APHA ‘‘American Public Health Association”, 1992. Standard methods for the the Upper Eocene-Oligocene sequence in Fayoum depression and the
examination of water and wastewater. 18th ed. surrounding area, W.D., Egypt (M.Sc. thesis), Cairo Univ., Egypt, 219 P.
Azab, M.A., 2001. Geological and geomorphological studies of the area around Lake Pegau, W.S., Gray, D., Ronald, J., Zaneveld, V., 1997. Absorption and attenuation of
Qarun, Faiyum, Egypt using remote sensing data and geographic information visible and near-infrared light in water: dependence on temperature and
system (Ph.D. thesis), Fac. of Sci., Al-Azhar Univ., Egypt. salinity. Appl. Opt. 36 (24), 6035–6046.
Beadnell, H.L., 1905. The topography and Geology of Fayum province of Egypt, Said, R., 1962. Geology of Egypt. El Sevier publishing co., Amsterdam, New York, pp.
Survey Dept., Egypt, 101 P. 99–106.
Bistani, L.F.C., 2009. Identifying total phosphorus spectral signal in a tropical Sridhar, B.B., Maruthi, Vincent R.K., Witter, J.D., Spongeberg, A.L., 2009. Mapping
estuary lagoon using a hyperspectral sensor and its application to water quality total phosphorus concentration of surface soils with LANDSAT TM data. Sci.
modeling (Ph.D. thesis), Civ. Eng., Univ. of Putro Rico, Mayagüez Campus. Total Environ. 407 (8), 2894–2899.
200 K.W. Abdelmalik / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 193–200

Swedan, A.H., 1986. Contribution to the Geology of Fayum area (Ph.D. thesis), Cairo Wua, G., Leeuw, J.D., Skidmore, A.K., Liu, Y., Prins, H.H.T., 2008. Performance of
Univ., Egypt. Landsat TM in ship detection in turbid waters. Int. J. Appl. Earth Obs. Geoinf. 11,
Swedan, A.H., 1992. Stratigraphy of the Eocene sediments in the Fayum area. Ann. 54–61.
Geol. Surv. Egypt 13, 157–166. Wubet, M.T., 2003. Estimation of Absolute Surface Temperature by Satellite Remote
Tamer, A., El-Shazly, M., Shata, A., 1975. Geology of El-Fayoum Beni Suef region Part. Sensing Master’s thesis. International Institute for Geo-Information Science and
2: stratigraphy. Desert Inst. Bull. Egypt 25(1), 27–47. Earth Observation (ITC), Enschede, The Netherlands, p. 121.
Tyagi, P., Bhosle, U., 2011. Atmospheric correction of remotely sensed images in Yüzügüllü, O., Aksoy, A., 2011. Determination of Secchi Disc depths in Lake Eymir
spatial and transform domain. Int. J. Image Proc. 5 (5), 564–579. using remotely sensed data. Proc. Soc. Behav. Sci. 19, 586–592.