Content area
Water quality is an extremely important factor as it affects the ecological balance of ecosystems and the development of the social and economic wellbeing of the countries bordering it. Remote sensing multiconcept helps to understand the natural environment, managing water resources and assessing water pollution on local and regional levels. Landsat 8 data were used to monitor coastal water quality in the region of Tangier-Ksar Sghir. The main purpose of the current study is to establish a mathematical relationship between the amount of light emitted from the water bodies and the measured water parameters. The results permit to create a spatial distribution maps for the water quality parameters. The present work study three water parameters: total suspended solids (TSS), dissolved oxygen (DO), and total dissolved sediments (TDS). Thirty-four sampling points were used to represent water parameters measurements along the coastline. The 75% of the in situ measurements were used to build the statistical models by using the spectral characteristics obtained from the sensors, while the remaining 25% were used for testing the accuracy of the developed equations. For the correlation analysis and the regression development, the Statistical Package of the Social Sciences (SPSS) software was used. The final results of the statistical analysis showed a high correlation between the calculated data and the observed ones with R2 ˃ 0.713 and p value ˂ 0.001. The obtained values showed a high accuracy as well (RMSE ranging between 0.23 and 0.69 and SEE ranging between 0.01 and 0.47). SNAP software and Qgis were used to do the image processing and to create the spatial distribution maps for the water parameters in the coastline of Tangier-Ksar Sghir region.
Introduction
The pollution of waters in coastal areas has increased significantly in recent years. In addition to commercial and industrial activities, over 40% of the world’s population lives in coastal regions and lake or river shores (United Nations, 2017). Any degradation of the coastal waters quality can affect human beings and their social and economic activities. Besides, coastal areas present an important natural ecosystem for the survival of a variety of species as they provide a habitat for many of living organisms. Therefore, chemical, physical, and bacteriological coastal water properties including concentrations and distribution need to be under continuous assessment. In most cases, water quality indicators are measured using traditional methods by collecting samples from the terrain and then proceeding to the analysis at the laboratory. The in situ sample measurements can offer a high veracity, yet the process can be exhaustive, expensive and time consuming, thus it is not possible to afford water quality data simultaneously especially for large areas (Duan, 2013). With advances in remote sensing techniques, many researchers started to study the relationship between water quality variables and space borne and airborne sensors data (Gholizadeh et al., 2016; Giardino, 2014; Hellweger et al., 2004). Since 1970s, remotely sensed approaches have been in use widely in water monitoring and continue to be used in the present (X. J. Wang & Ma, 2001). Sensors measure the amount of the reflected light coming from the water surface. Many physicochemical and bacteriological water parameters can be identified in different ways from the reflectance of various electromagnetic radiations at different wavelengths reflected from the water’s surface, such as chlorophyll-a (Giardino et al., 2014), turbidity (Alparslan et al., 2009), salinity (Font, 2013), TSS (Ritchie, 1976), pH (Abdelmalik, 2018), temperature (Ahn et al., 2006), conductivity (Mallick et al., 2014), Secchi disk depth (Allan et al., 2011), total dissolved solids (Aljoborey et al., 2019), total phosphorus (Wu, 2010), etc. Water components and characteristics are the factors determining the spectral characteristics of water substance (Seyhan & Dekker, 1986).
The present study aims to develop mathematical relations between Landsat 8 data and the three water parameters: TSS, DO, and TDS by developing several statistical equations using modeling techniques and then testing the regressions in order to find the appropriate models for water quality monitoring in the coastline of Tangier-Ksar Sghir region in North Morocco. Finally, the algorithms will be presented in form of spatial distribution of TSS, DO, and TDS over the shore of the study area.
Study area
The region of Tangier-Ksar Sghir is a coastal area located in the northwest of Morocco (Fig. 1). The capital city Tangier covers an area of 11,570 km2 representing 1.6% of the total surface of the Moroccan Kingdom (Bourouhou et al., 2018). It is bordered by the Mediterranean Sea to the north, the Atlantic Ocean to the west, the region of Taza-Al Hoceima-Taounate to the east, and the Gharb-Chrarda-Beni Hssen to the south (Laghzal et al., 2016). The region of Tangier Ksar-Sghir is characterized by a variability of relief manifested by a large number of mountain ranges of the Rif chain and the coastal plains as well as, a varied weather conditions. According to the Directorate-General of Local Authorities, the oceanic influence that characterizes the basins in this area gradually weakens in the northern coastal basins with increasing aridity from west to east. According to a previous study conducted by the Directorate-General of Local Authorities, the oceanic influence that characterizes the basins in this area gradually weakens in the northern coastal basins with increasing aridity from west to east (Ministry of the Interior & Directorate General of Local Authorities of Morocco, 2015).
[See PDF for image]
Fig. 1
Study area (Ministry of the Interior & Directorate General of Local Authorities of Morocco, 2015)
Divers economic activities are taking place in the coastal bays of Tangier-Ksar Sghir region mostly located in Tangier Med port. Geographically, the port is positioned on the east–west world shipping trade route between Asia, Europe, and North America, which is considered the second busiest sea lane in the world, with more than 100,000 boats per year.
Methodology and data
In the current study, a semi-empirical approach is chosen for estimating the concentration of water quality parameters. Both physical and spectral information are utilized to develop the algorithms. Region and time of calibration can mostly determine the statistical coefficients of the algorithm (Gholizadeh et al., 2016). Figure 2 represents the diagram of the process used in the present study.
[See PDF for image]
Fig. 2
Flowchart of the water parameters estimation process
[See PDF for image]
Fig. 3
Samples location in study area
[See PDF for image]
Fig. 4
Relational plot if the measured parameters values vs. the estimated ones
Sampling and water analysis
A campaign of 34 sampling points was conducted to have a representation of the coastline water bodies (Fig. 3).
The sampling program and techniques were executed in accordance with the rules given in the ISO 5667–1, ISO 5667–2, and ISO5667-3. The water analysis were realized during June, July and September 2017 based on standards methods (American public health association & Water Environment Federation, 1992) for water examination. The b75% of the in situ measurements and sampling site’s data were used to calibrate the spectral characteristics obtained from the sensors while the remaining 25% were used for the validation and for testing the accuracy of the developed models.
Remotely sensed data preparation
The Landsat 8 images were used to provide the required data for the correlation to the measured constituents. The images were ordered and downloaded from the US Geological Survey (USGS) through Earth Explorer site (https://earthexplorer.usgs.gov/), obtained as showed in Table 1 with Path and Row 201/35. The satellite images were level 2 Surface Reflectance GeoTIFF type and have 30 m spatial resolution on a Universal Transverse Mercator (UTM) or Polar Stereographic (PS) mapping grid (Jenkerson, 2019). The used images were acquired in June, July and September with a ratio scene cloud cover of 6.76%, 8.32%, and 2.83%, respectively.
Table 1. Landsat 8 OLI images data
Product ID | Data provider | Satellite | Instruments | Path | Row | Acquisition date |
|---|---|---|---|---|---|---|
LC08_L1TP_201035_20170626_20170714_01_T1 | USGS/EROS | Landsat 8 | OLI | 201 | 35 | 2017–06-26 |
LC08_L1TP_201035_20170728_20170810_01_T1 | 2017–07-28 | |||||
LC08_L1TP_201035_20170914_20170928_01_T1 | 2017–09-14 |
The observatory of Landsat 8 is designed for 705 km, Sun-synchronous orbit, with a 16-day repeat cycle, completely orbiting the Earth every 98.9 min (Department of the Interior & U.S. Geological Survey, 2019). Landsat 8 monitors the Earth’s surface by recording multispectral images using eleven spectral bands with different spatial resolution in order to obtain information about land cover change and land use dynamics (Table 2). Two-sensor payload is carried by Landsat 8: the operational land image (OLI) multispectral sensor and the thermal infrared sensor (TIRS). The Landsat 8 has an improved radiometric resolution of 12 bit and a greater data coverage comparing to Landsat 5 TM and Landsat 7 ETM + (Roy, 2014).
Table 2. OLI and TIRS spectral bands (Department of the Interior & U.S. Geological Survey, 2019)
Landsat 8 | |
|---|---|
Band description (30 m native resolution unless otherwise denoted) | Wavelength (µm) |
Band 1-coastal/aerosol | 0.43–0.45 |
Band 2-blue | 0.45–0.51 |
Band 3-green | 0.53–0.59 |
Band 4-red | 0.64–0.67 |
Band 5-near infrared | 0.85–0.88 |
Band 6-shortwave infrared | 0.85–0.88 |
Band 7-shortwave infrared | 2.11–2.29 |
Band 8-panchromatic (15 m) | 0.50–0.68 |
Band 9-cirrus | 1.36–1.38 |
Band 10-thermal infrared (100 m) | 10.60–11.19 |
Band 11-thermal infrared (100 m) | 11.50–12.51 |
The image processing chain is based on both radiometric and atmospheric corrections. Principally, radiometric correction aims to convert digital number (DN) values to spectral reflectance or to top of atmosphere (TOA) planetary reflectance. In this current study, OLI band level 2 data were converted to surface reflectance using rescaling coefficients provided in the product metadata file of Landsat 8.
The atmospheric correction aims to minimize and reduce the influence of atmosphere compounds on the recorded signal in order to avoid the under or over estimation of spectral information. Several atmospheric correction methods have been used by many authors in order to get the actual reflectance from the surface as De Keukelaere, (2018), Li et al. (2019), Vermote et al. (2016) and Wei (2018).
In this study, Landsat Surface Reflectance products are generated from specialized software called Landsat Surface Code (LaSRC) and they are delivered atmospherically corrected. In order to map the water line, the normalized difference water index (NDWI) which was developed by McFeeters (1996) was applied on the image. The output of the NDWI values ranges between + 1 and − 1, where positive values indicate the water and the negative ones indicate non-water. The following equation shows the generated mask applied by the NDWI:
Statistical analysis
In order to develop the statistical models qualified to express the relationship between water parameter measurements and spectral bands, the outcome variables used are the measured samples and the regressors that represent the causes of variation and the remote sensed data. Model analysis tested the correlation between the independents variables and the dependents ones. For the generation of the mathematical equations and the correlation analysis, the Statistical Package of the Social Sciences (SPSS) software was used to evaluate the data of this current study.
Distribution maps generation
After elaborating and testing the statistical regressions, the models were applied on Landsat 8 data in order to convert pixel values of different spectral bands to water quality parameters values using the satellite image processing software SNAP_Landsat section. The SNAP is an architecture program for image processing; it can be downloaded from Science Toolbox Exploitation Platform (STEP: http://step.esa.int/. STEP is a free open-source toolbox exploitation platform developed by the European Space Agency in order to access to software and all documents.
Spatial distribution maps of TSS, DO and TDS as well as water samples map were produced using SNAP Survey Version 7.0.0 and Qgis software Version 3.12.0.
Results and discussion
Statistical models for water parameters
The study of the potential of remote sensing in estimating water quality parameters has attracted many authors such as Gholizadeh et al. (2016), Khattab and Merkel (2013), Coskun et al. (2008), Abdelmalik (2018), and El Saadi et al. (2014).
In the current study, multiple reflectance bands of Landsat 8 were used in order to investigate three parameters of coastal waters in the region of Tangier-Ksar Sghir namely: TSS, DO, and TDS.
The statistical factors chosen to specify the best regression models that express water quality properties are the adjusted R-squared (R2), the standard error of estimation (SEE), the root mean square error (RMSE), and the p value for the predictors. The adjusted R-squared changes the number of predictors in a model comparing to normal R-squared; it presents the percentage of explained variation considering that all the independent variables act on the dependent variables. Standard error of estimation indicates the accuracy of a regression model in prediction. The root mean square error is used to measure the difference between predicted values by a model and observed values. The models selected should have the higher values of adjusted square coefficient and the lower values of Standard error of estimation and Root mean square error. The p value is a statistical measure that can determine the significant terms in a developed model. In other words, p value (or significance sig.) indicates the possible variability of the outputs following changes in the predictors. A lower p value (< 0.001) can be, a better model can result.
The SPSS statistics version 25.0 was used to perform the regression analysis and to test the accuracy of the models for a final validation. Table 3 and Table 4 gave the results of the developed equations for correlating water parameters using OLI data and in situ measurements.
Table 3. Results of Statistical regression for the water parameters
Models | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Water parameters | Multiple regression | Linear | Logarithmic | Inverse | Quadratic | Cubic | Compound | Power | Exponential | |||||||||||
Indep. variable | R2 | SEE | Indep. variable | R2 | SEE | R2 | SEE | R2 | SEE | R2 | SEE | R2 | SEE | R2 | SEE | R2 | SEE | R2 | SEE | |
Total suspended solids | Band 1 to band 7 | 0.769 | 0.110 | Band 4 | 0.718 | 138.921 | 0.681 | 147.684 | 0.617 | 161.827 | 0.726 | 140.083 | 0.736 | 140.558 | 0.681 | 0.284 | 0.695 | 0.278 | 0.681 | 0.284 |
Dissolved oxygen | Band 1 to band 7 | 0.732 | 0.268 | Band 4 | 0.667 | 0.293 | 0.630 | 0.309 | 0.540 | 0.345 | 0.667 | 0.300 | 0.718 | 0.282 | 0.674 | 0.038 | 0.636 | 0.040 | 0.674 | 0.038 |
Total dissolved solids | Band 1 to band 7 | 0.659 | 2.582 | Band 4 | 0.713 | 2.479 | 0.684 | 2.485 | 0.681 | 2.498 | 0.668 | 2.604 | 0.713 | 2.479 | 0.619 | 0.085 | 0.630 | 0.084 | 0.619 | 0.085 |
Table 4. Regression equations for water parameters
Water quality parameter | Statistical model | Regression | Validation | |||||
|---|---|---|---|---|---|---|---|---|
R2 | SEE | p value | RMSE | Tolerance | SEE | RMSE | ||
TSS(mg/l) | y = 71.715 + 4619.467(band 4) + 3958.101(band 5) | 0.769 | 0.110 | ˂ 0.0001 | 0.603 | 0.274 | 0.067 | 0.229 |
DO(mg/l) | y = 7.98 + 16.625(band 1) − 23.839(band 4) | 0.732 | 0.219 | ˂ 0.0001 | 0.252 | 0.239 | 0.473 | 0.690 |
TDS(103mg/l) | y = 57.382—1023.837(band 4) + 13,434.125(band 4)2 − 61,438.988(band 4)3 | 0.713 | 0.020 | ˂ 0.0001 | 0.114 | - | 0.011 | 0.064 |
TSS is a very important parameter to study water quality using remote sensed data by reason of its direct connection with reflectance of solar radiation. Many studies present algorithms for estimating TSS based on a single band as Onderka (2014) using the near infrared band of Landsat 7-ETM + , Papoutsa et al. (2014) using the Red band of the same sensor and more recently Caballero et al. (2018) that used the Red band of the Landsat 8-OLI, and also algorithms based various spectral bands such as Van thao et al. (2018) and Nurgiantoro and Jaelani (2017) based on multitemporal Landsat 8-OLI images, and Wang et al. (2006) developing a model based on Landsat 5-TM data.
In the current study, TSS is correlated with reflectance of both the Red and the NIR bands (correlation coefficient ˃ 0.85 and sig ˂ 0.001) therefore, the best fit model found for the bands 4 and 5 was the multiple linear regression using the stepwise method with R2 = 0.718 and sig ˂ 0.001. However, by examining other developed models, the TSS has a considerable relation with the Red band; the cubic model showed a high determination coefficient (R2 = 0.736). Generally, TSS has an optical activity which makes it absorb the light spectrum rather the transmit it. Studies (Wu et al., 2010; Nguyen, 2012) reveal that the decrease of TSS in water (case of clear water) cause high reflectance in the green region of the visible spectrum, and it decreases in red and NIR region. Still, according to Ritchie et al. (1976) using the spectrum range between 700 and 800 nm is more advantageous to retrieve Suspended matters parameter. In fact, the examination the developed relational equations for TSS showed that band 4 (0.63–0.68 µm) and band 5 (0.84–0.88 µm) of Landsat 8 are well correlated with suspended solids. By applying and testing the model, we obtained RMSE = 0.229, SEE = 0.067.
In order to demonstrate the linear correlation between calculated values of TSS, DO, and TDS and the estimated ones, Fig. 1 Study area (Ministry of the Interior & Directorate General of Local Authorities of Morocco, 2015).
Figure 4 illustrates the graphical relation.
According to Shmeis (2018) and Grant and McLimans (2016), DO is a critical parameter for many organisms living in the water that should be observed continually. Authors have studied the DO using remotely sensed data generated from different satellites (Landsat, SPOT, MODIS, Sentinel, etc.) as Kim et al. (2020), Arief (2017), Karaoui et al. (2019), and Qiu et al. (2006). Although, many studies specified that the sensor TM in Landsat 5 was the most useful for the determination of DO in surface water (Gholizadeh et al., 2016).
The final results of the statistical analysis showed a high correlation of DO with the validated equation (correlation coefficient ˃ 0.84 and sig. ˂ 0.001). The multiple linear model illustrated the best relation between DO values and the reflectance of coastal (0.43–0.45 µm) and red bands (R2 = 0.732, SEE = 0.268). The other models showed a good SEE values but determination coefficients lower than 0.718. An examination of the tolerance coefficient in order to check the multicollinearity between the independent variables used in the equation, a value higher than 0.2 was found (absence of collinearity). By testing the resulting equation, we obtained RMSE = 0.690 and SEE = 0.473. Both coefficients presented good results.
A review of the literature confirmed that the atmospheric correction of images used for the estimation of DO is a crucial phase for the calibration and the validation of DO prediction algorithms.
The variable TDS is one of the parameters that indicate the measure of the mass of solid material dissolved in water. Various visual spectral bands and their combination were used to quantify TDS (Aljoborey et al., 2019; Ferdous et al., 2019; Mushtaq & Mili, 2016). Aljobory et al. (2019) in their article pointed that the absorbance of dissolved matters in water decreases in longer wavelengths, especially above 0.50 µm which goes with results found in the current study. In fact, a high correlation has been shown between the red region of the spectrum of OLI data and the measured TDS values resulting a mathematical equation based on a single band. All the regression models (linear, logarithmic, inverse, quadratic, compound, power, and exponential) presented close values of R2 (between 0.619 and 0.684). Yet, the highly relational combination found was the cubic equation (third order), as correlation coefficient was found ˃ 0.84, p value ˂ 0.001, R2 = 0.713, and SEE = 2.479. By applying and testing the obtained model, the RMSE = 0.064 and SEE = 0.011.
In the present study, a probability test called P-P plot was conducted in order to assess graphically the normality of data distribution. The P-P plot aims to compare the observed cumulative distribution function of the standardized residual to the expected cumulative distribution function of the normal distribution (Institute for Digital Research & Education, 2019). The test is basically examining the normality of the residuals and not the predictors. By observing the normal p-p plots of TSS regression (Fig. 5b), the DO regression (Fig. 5a), the data fall along the diagonal line and it shaped like a normal curve which means statistically it normally distributed. For the total dissolved Solids probability plot (Fig. 5c), the scatter lied not far from the line with no obvious pattern coming away from the line; thus, the TDS data is approximately normally distributed.
[See PDF for image]
Fig. 5
Probability p-p plot for a TSS residuals, b DO residuals, and c TDS residuals
Spatial distribution for water parameters
Multiple regression methods were applied on Landsat 8-OLI data in order to examine the relationship between spectral reflectance and three of water quality parameters. The results enabled assessment of spatial distribution of TSS, DO, and TDS in the study region. Figure 6 illustrates the spatial distribution maps of the water quality parameters for the study region divided in three sub-areas (a, b, and c) during June, July, and September 2017.
[See PDF for image]
Fig. 6
Spatial distribution of water quality parameters in the region of Tangier-Ksar Sghir during 2017. a September, b June, c July
The spatial distribution maps of estimated values of TSS during June and September show generally an increase along the coastline and a decrease in the direction of the Mediterranean Sea (lower values at the northwest). The month of June indicates higher values then September, whereas during July, an irregular distribution appears in the estimated TSS map. This irregularity may be attributed to scene cloud cover ratio that reaches 8.32%.
The spatial maps for DO show the same distribution pattern. DO has a range of values of 6 to 9 mg/l for the entire study area during June, July, and September. The lower values were found along the shore line. However, during July, DO distribution map shows a high DO value (up to 8.15 mg/l) recorded in the bay of Dalia Beach. In general, DO is one of the parameters used to calculate water quality index. The high value may be related to the good quality of Dalia beach since it has obtained the eco-label “Blue Flag” given by the Mohammed VI Foundation for the Protection of the Environment as a part of Clean Beach program 2017 (FM6e for the protection of the environment, 2017).
TDS shows approximately the same distribution pattern during June, July and September as they show minimum values along the coastline. The TDS concentrations decrease toward the sea (up to 24,000 mg/l), and this value is a bit lower than the TDS value estimated in normal seawater (about 35,000 mg/l) (Saeed et al., 2019). The obtained values can be used as a water indicator of salinity level, and it can be very useful in studying seawater intrusion (Rusydi, 2018).
Conclusion
The current work shows that satellite sensors can monitor water quality parameters with reliable precision depending on severable criterions. The results prove the existence of a compelling relationship between Landsat 8 OLI sensor image and surveyed water variables. The semi-empirical models based on remote sensing data in conjunction with in-situ measurements represent a reliable, inexpensive and more effective tool for detecting water quality parameters. Therefore, the managements of water resources and associated services can be more efficient. The mathematical equations developed for TSS, DO, and TDS were based on a regression analysis, and they were statistically tested.
As mentioned in this study, a high special and spectral resolution of the sensor can provide a better result with more accuracy depending on the variables studied. This paper has focused on three water parameters, yet there are several chemical and physical water characteristics that can be retrieved from remotely sensed data depending on their optical properties (active or inactive). Further efforts should be made in order to develop more approaches for determining the concentrations of water quality parameters for a larger area and with more accuracy, especially those who present a real challenge in the field of water quality monitoring.
Acknowledgements
We appreciate the efforts of all researchers who have worked hard to advance knowledge and improve outcomes of water quality assessment using remote sensing. We also thank the Department of Analytical Chemistry, Physical Chemistry and Chemical Engineering at Alcalà de Henares University (UAH) in Madrid as well as the National Office of Food Safety (ONSSA) in Tangier for providing us with materials and equipment for this work.
Data availability
The datasets analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Conflict of interest
The authors declare no competing interests.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
Abdelmalik, KW. Role of statistical remote sensing for inland water quality parameters prediction. The Egyptian Journal of Remote Sensing and Space Science; 2018; 21,
Aljoborey, A; Dherar, A; Abdulhay, HS. Estimating total dissolved solids and total suspended solids in Mosul Dam Lake in situ and using remote sensing technique. Periodicals of Engineering and Natural Sciences (PEN); 2019; 7,
Ahn, YH; Shanmugam, P; Lee, JH; Kang, YQ. Application of satellite infrared data for mapping of thermal plume contamination in coastal ecosystem of Korea. Marine Environmental Research; 2006; 61,
Allan, MG; Hamilton, DP; Hicks, BJ; Brabyn, L. Landsat remote sensing of chlorophyll a concentrations in central North Island lakes of New Zealand. International Journal of Remote Sensing; 2011; 32,
Alparslan, E; Coskun, G; Alganci, U. Water quality determination of Küçükçekmece Lake, Turkey by using multispectral satellite data. The Scientific World Journal; 2009; 9, pp. 1215-1229.[COI: 1:CAS:528:DC%2BC3cXhtl2qtLY%3D] [DOI: https://dx.doi.org/10.1100/tsw.2009.135]
American public health association, et Water Environment Federation. (1992). APHA Method 9221: Standard methods for the examination of water and wastewater. 18th edition: 13.
Arief, Muchlisin. (2017). Development of dissolved oxygen concentration extraction model using Landsat data case study: Ringgung coastal waters. International Journal of Remote Sensing and Earth Sciences 12.
Bourouhou, Imane, Farida Salmoun, et Yusuf Gedik. (2018). Characteristics of Mediterranean Sea water in vicinity of Tangier region, north of Morocco. Proceedings,2(20), 1291.
Caballero, I; Steinmetz, F; Navarro, G. Evaluation of the first year of operational Sentinel-2A data for retrieval of suspended solids in medium- to high-turbidity waters. Remote Sensing; 2018; 10,
Coskun, HG; Tanik, A; Alganci, U; Cigizoglu, H. Determination of environmental quality of a drinking water reservoir by remote sensing, GIS and regression analysis. Water, Air, and Soil Pollution.; 2008; 194, pp. 275-285.[COI: 1:CAS:528:DC%2BD1cXhtV2qt73N] [DOI: https://dx.doi.org/10.1007/s11270-008-9716-x]
De Keukelaere, L et al. Atmospheric correction of Landsat-8/OLI and Sentinel-2/MSI data using ICOR algorithm: Validation for coastal and inland waters. European Journal of Remote Sensing; 2018; 51,
Department of the Interior, et U.S. Geological Survey. (2019). Landsat 8 (L8) Data Users Handbook.
Duan, W et al. Spatiotemporal evaluation of water quality incidents in Japan between 1996 and 2007. Chemosphere; 2013; 93,
EarthExplorer. (2017). Web site: earthexplorer.usgs.gov. https://earthexplorer.usgs.gov/
El Saadi, Aiman, Mohsen Yousry, et Hossam Jahin. (2014). Statistical estimation of Rosetta branch water quality using multi-spectral data. Water Science 28.
ESA SNAP Plateforme. (2017). Web site: http://step.esa.int/main/toolboxes/snap/
Ferdous, J., Rahman, MTU, Ghosh, S. K. (2019). Detection of total dissolved solids from landsat 8 OLI image in Coastal Bangladesh. Proceedings of the 3 International Conference on Climate Change, ISSN 2513-258X online, Vol. 3, pp. 35–44.
Foundation Mohammed VI for the Protection of the Environment (FM6e). (2017). Rapport. Plages propres.
Font, J et al. SMOS first data analysis for sea surface salinity determination. International Journal of Remote Sensing; 2013; 34,
Gholizadeh, MH; Melesse, AM; Reddi, L. A comprehensive review on water quality parameters estimation using remote sensing techniques. Sensors; 2016; 16,
Giardino, C et al. Evaluation of multi-resolution satellite sensors for assessing water quality and bottom depth of Lake Garda. Sensors; 2014; 14,
Grant, CJ; McLimans, CJ. A simple approach to manipulate dissolved oxygen for animal behavior observations. Journal of Visualized Experiments; 2016; 112, 54430.
Hellweger, FL; Schlosser, P; Lall, U; Weissel, JK. Use of satellite imagery for water quality studies in New York harbor. Estuarine, Coastal and Shelf Science; 2004; 61,
Institute for Digital Research & Education. (2019). Introduction to regression with SPSS lesson 2: SPSS regression diagnostics. https://stats.idre.ucla.edu/spss/seminars/introduction-to-regression-with-spss/introreg-lesson2/
Karaoui, I., et al. (2019). Evaluating the potential of Sentinel-2 satellite images for water quality characterization of artificial reservoirs: The Bin El Ouidane Reservoir case study (Morocco). Meteorology Hydrology and Water Management, 7(1). http://www.journalssystem.com/mhwm/,95087,0,2.html (11 avril 2020).
Jenkerson, Calli. (2019). Landsat 8 Surface Reflectance Code (LASRC) Product Guide.
Khattab, Mohammed, et Broder Merkel. (2013). Application of Landsat 5 and Landsat 7 images data for water quality mapping in Mosul Dam Lake, Northern Iraq. Arabian Journal of Geosciences 7.
Kim, Yong Hoon et al. (2020). Application of satellite remote sensing in monitoring dissolved oxygen variabilities: A case study for coastal waters in Korea. Environment International 134: 105301.
Laghzal, A et al. (2016). Evaluation of physico-chemical and bacteriological quality of water springs by using a principal component analysis (PCA): A case study of Tingitane Peninsula (Morocco). 8.
Li, Fuqin, David Jupp, Stephen Sagar, et Thomas Schroeder. (2019). Atmospheric correction and water surface issues for aquatic applications of Landsat 8. In IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium, Yokohama, Japan: IEEE, 6899‑6902. https://ieeexplore.ieee.org/document/8899180/ (9 avril 2020).
Mallick, J; Hasan, MA; Alashker, Y; Ahmed, M. Bathymetric and geochemical analysis of Lake Al-Saad, Abha, Kingdom of Saudi Arabia using geoinformatics technology. Journal of Geographic Information System; 2014; 6,
McFEETERS, SK. The use of the normalized difference water index (NDWI) in the delineation of open water features. International Journal of Remote Sensing; 1996; 17,
Ministry of the Interior, and Directorate General of Local Authorities, et Morocco. (2015). Monographie générale de la région de Tanger Tétouan Al Hoceima.
Mushtaq, F., & Mili, G. (2016). Remote estimation of water quality parameters of Himalayan Lake (Kashmir) using Landsat 8 OLI imagery. Geocarto International,32.
Nguyen, D. D. (2012). Water body extraction from multi spectral image by spectral pattern analysis. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XXXIX-B8, pp. 181–186.
Nurgiantoro, N., & Jaelani, L. M. (2017). Monitoring of total suspended solid in coastal waters due to conventional gold mining using multi temporal satellite data, case study: Bombana, Southeast Sulawesi. IPTEK Journal of Proceedings Series, 0(2), 83.
Onderka, M. (2014). Remote sensing and identification of places susceptible to sedimentation in the Danube River. 8.
Papoutsa, C; Retalis, A; Toulios, L; Hadjimitsis, DG. Defining the Landsat TM/ETM+ and CHRIS/PROBA spectral regions in which turbidity can be retrieved in inland waterbodies using field spectroscopy. International Journal of Remote Sensing; 2014; 35,
Qiu, Y. et al. (2006). Monitoring the water quality of water resources reservation area in upper region of Huangpu River using remote sensing. In 2006 IEEE International Symposium on Geoscience and Remote Sensing, Denver, CO, USA: IEEE, pp. 1082–85. http://ieeexplore.ieee.org/document/4241426/
Ritchie, Jerry C. (1976). Remote sensing of suspended sediments in surface waters. Photogrammetric Engineering, 7.
Roy, DP et al. Landsat-8: Science and Product Vision for Terrestrial Global Change Research. Remote Sensing of Environment; 2014; 145, pp. 154-172. [DOI: https://dx.doi.org/10.1016/j.rse.2014.02.001]
Rusydi, A. F. (2018). Correlation between conductivity and total dissolved solid in various type of water: A review. IOP Conference Series: Earth and Environmental Science, 118, 012019.
Saeed, Mohamed O., Mohammed A. Al-Nomazi, et Ahmed S. Al-Amoudi. (2019). Evaluating suitability of source water for a proposed SWRO plant location. Heliyon, 5(1), e01119.
Seyhan, E; Dekker, A. Application of remote sensing techniques for water quality monitoring. Hydrobiological Bulletin; 1986; 20,
Shmeis R. M. A (2018). Water chemistry and microbiology. In Comprehensive Analytical Chemistry, Elsevier, pp. 1–56. https://linkinghub.elsevier.com/retrieve/pii/S0166526X18300023
Van Thao, N; Vinh, VD; Huong, DTT; Gouramanis, C. Remote sensing data analysis with validation by numerical model for detecting suspended particulate matter concentration in coastal waters of the Red River Delta, Vietnam. Tạp Chí Khoa Học Và Công Nghệ Biển; 2018; 18,
United Nations. (2017). Factsheet, the ocean conference. In New York, pp. 1–7. https://oceanconference.un.org
Vermote, E; Justice, C; Claverie, M; Franch, B. Preliminary analysis of the performance of the Landsat 8/OLI land surface reflectance product. Remote Sensing of Environment; 2016; 185, pp. 46-56. [DOI: https://dx.doi.org/10.1016/j.rse.2016.04.008]
Wang, F; Han, L; Kung, H-T; Van Arsdale, RB. Applications of Landsat-5 TM imagery in assessing and mapping water quality in Reelfoot Lake, Tennessee. International Journal of Remote Sensing; 2006; 27,
Wang, XJ; Ma, T. Application of remote sensing techniques in monitoring and assessing the water quality of Taihu Lake. Bulletin of Environmental Contamination and Toxicology; 2001; 67,
Wei, J et al. An assessment of Landsat-8 atmospheric correction schemes and remote sensing reflectance products in coral reefs and coastal turbid waters. Remote Sensing of Environment; 2018; 215, pp. 18-32. [DOI: https://dx.doi.org/10.1016/j.rse.2018.05.033]
Wu, C et al. Empirical estimation of total phosphorus concentration in the mainstream of the Qiantang River in China using Landsat TM data. International Journal of Remote Sensing; 2010; 31,
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2021.