Spatio Temporal Analysis of Water Quality at Tasek Bera

Bera Lake is Malaysia’s largest natural fresh water reservoir. It has critical environmental and ecological significance for both humans and wildlife. Nonetheless, the water quality of this lake has deteriorated during the previous few decades as a result of land development projects in the river basin area. This study aims to analyse the spatio temporal pattern of water quality in Tasek Bera and to identify and evaluate the degradation grade for water quality, as well as to identify the environmental crucial factor for Tasek Bera and its surrounding area. This study also implements remote sensing technology to determine the water quality parameters in Tasek Bera. The data for this study was obtained from Landsat 8 OLI and Landsat 5 MSS from 3 different years which are year 2000, 2010 and 2020. The landset image were processed and integrated into the Geographical Information System (GIS) software to undergo classification, principal component analysis and weighted overlay process. Based on the weighted overlay mapping and Malaysia’s water quality index, the result showed that in year 2000, the water quality in Tasek Bera is mostly being in the second class (fair water quality) but in 2010, it shows a prominent increasing performance when the lake is mostly being in the first class (excellent water quality). However, in the year 2020, we had observe a significant decreasing of water quality especially on the upper section and the middle section of Tasek Bera. Through this study, researchers will be able to further analyse Tasek Bera’s current environmental state, and eventually bring evidence in the environmental deterioration that Tasek Bera has endured. This is also to make people realise the ecological importance of wetlands in Malaysia.


Introduction
Environmental degradation, encompassing concerns like pollution, biodiversity loss, animal extinction, deforestation and desertification, and global warming, is a highly significant issue on a global scale.[1].The term "environmental degradation" refers to the deterioration of the environment caused by the depletion of resources.This includes all biotic and abiotic elements, such as the air, water, soil, plants, animals, and other living and non-living things on the planet of Earth.Tasik Bera, a RAMSAR site of 6,870 hectares of wetlands, is made up of freshwater and peat swamp forest (5,440 ha, 79%) is located in Bera District, Pahang [2].This is also the home of several important ecosystem elements which included freshwater organisms and humans but because of human activities Tasek Bera has suffered substantial environmental degradation every year.[3] also reported on the Tasek Bera's destruction, emphasising the decline of water quality and quantity.Based on research by [4] they found out that the primary sources of Tasek Bera's deterioration are causes by surface run-off carrying nutrient-rich water from neighbouring plantations, sewage from communities living around the lake, logging activities, and oil discharges from motorboats.Moreover, shifting agriculture, possible pollution, destruction of watersheds, logging operations, erosion, and siltation are all reported to have degraded habitats at Tasek Bera, to the detriment of numerous valuable species [5].
A major factor in ensuring the conservation and sustainable use of water resources is the identification of water pollution sources in river basins and the study of their spatiotemporal variation [6].The spatial and temporal pattern of the river's water quality must be studied since it may significantly help the professionals in responsible of making decisions about water environment management in combating against environmental problems, particularly water pollution in Malaysia.Remote sensing (RS) and GIS technology can undoubtedly provide advantages for larger coverage data retrieval and massive data mining purposes [7].A study conducted by [8] showed that remote sensing techniques along with Geographic Information Systems (GIS) can help researchers to evaluate water quality and monitor changes in numerous parameters related to water quality over temporal and spatial scales which they are not always observable from in situ measurements.
Traditionally, water quality is monitored using spot measurements in situ and laboratory analysis of samples but the issue with this method is, this is an expensive and time-consuming technique for implementation, and it is not suitable for observing spatial and temporal variations in large regions in most circumstances [9].It is difficult to determine water quality in short periods of time using traditional monitoring methods, because its required systematic planning and evaluation of water bodies [10].Given that variations in parameter concentration can affect the optical properties of water, an algorithm have been developed that have favoured the accessibility and efficiency of monitoring water quality parameters via RS techniques using Landsat satellite imagery [11].The primary advantage of RS through satellite for water quality monitoring is the generation of synoptic views without the requirement for costly on-site surveys [12].Tasek Bera is utilised as a medium in this study to identify and evaluate the degradation grade for water quality, as well as to identify the environmental crucial factor for Tasek Bera and its surrounding area.

Study area
Tasek Bera is one of the natural freshwater lake systems that is located in Bera district, Pahang, Malaysia.Tasek Bera, located in the southwest of Pahang, is a freshwater lake system 35 kilometers long and 20 kilometers broad that feeds into Sungai Pahang via Sungai Bera.It is Malaysia's largest freshwater lake, as well as being distinctive and secluded.The lake's ecosystem reaches sections of peat swamp forest while being surrounded by dry lowland dipterocarp woodland.The Tasek Bera habitat is rich in fauna and flora, and it is also home to the indigenous Semelai tribe near the marshes.
Given that some species need specialized habitats to exist, it appears that Tasek Bera's flora and wildlife are threatened by widespread wetlands destruction and river pollution.Furthermore, according to [13], there are still human activities taking place at the Tasek Bera RAMSAR site, which has caused the area that has been demolished to increase from 47.14 km 2 in 1994 to 340 km 2 in 2009 due to the establishment of oil palm and rubber plantations, increasing the amount of cleared land.Tasek Bera's habitats are said to have been harmed by shifting agriculture, potential pollution, logging operations, wetlands degradation, erosion, and siltation, all to the detriment of numerous treasured species.More crucially, human-caused changes such as the conversion of peat swamp forests to agriculture and rural

Image Data Collection and pre-processing
In this study, we used open-access satellite data to improve estimates of water clarity in an optically complex coastal water body.Specifically, we created a remote sensing water clarity product by compiling Landsat-8 and Landsat 5 MSS reflectance data.The data that will be used in this study is from Landsat Imagery.All the data will be downloaded from United State Geological Survey (USGS) website, called Earth Explorer (EE).United States Geological Survey (USGS) is a science bureau that works within the United States Department of the interior.USGS provides science about natural hazards that could endanger lives and livelihoods of others, such as water, energy, minerals, and other natural resources.This website also provides real-time data for researchers to collect various data for their own researcher.
In this study, the data will consist of Landsat imagery.Landsat is a program of Earth-observing series that is jointly managed by NASA and U.S. Geological Survey.Landsat program is one of the longest running enterprises for acquisition of satellite imagery of earth.There are a total of 7 different Landsat satellites that have been launched, from Landsat 1 to Landsat 9.The Landsat that will be used in this research is Landsat 8 OLI, and Landsat 5 MSS.
Landsat 8 has two main sensors that are created to capture imagery of the surface of the earth.The main two sensors are Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS).OLI sensors produce up to 9 spectral bands from band 1 until band 9 with the resolutions of 15 meters, 30 meters, and 60 meters while TIRS produce 2 thermal bands at 100 meters resolution.The Thermal Infrared Sensor (TIRS) and the Operational Land Imager (OLI), two high-resolution sensors on Landsat 8 that were introduced in 2013 can more precisely monitor biophysical changes in the Tasek Bera.The decrease of sea surface temperatures (SST) associated with the surfacing wastewater plume can be detected by Landsat 8 TIRS [15].
Landsat 5 was launched on March 1,1984 with Multispectral Scanner (MSS) and the Thematic Mapper (TM) instrument to capture imagery of the earth's surface.One of the scanners that was installed in Landsat 5, Multispectral Scanner (MSS) have 4 spectral bands, 6 detectors that was used for each spectral band and the ground sampling size of 57 x 79 m.Besides that, Landsat 5 was equipped with Thematic Mapper that have 7 spectral bands with added mid-range infrared and the ground sampling of 30 m reflective and 120 m thermal.Landsat 5 was awarded for being the longest operating earth observation satellite in Guinness World Record, being the only satellite to orbit and delivered earth imaging data for 29 years.Landsat 5 was launched with Multispectral Scanner (MSS) and the Thematic Mapper (TM) instrument to capture imagery of the earth's surface.
Band combination is regularly used in doing any remote sensing or Geographical Information System study.It is usually used to identify the geological information of the surface, either on land, agriculture or wetlands, the features of the surface, either visible or hidden and the faults of the surface, either fractures or zones.There are a lot of band combinations that can be used but, in this study, the purpose is to do water quality.Hence, the band combination will be about water.
The data will also consist of three different years, that is year 2000, year 2010 and year 2020.It would help on seeing the difference of water quality on Tasek Bera.Selecting three different imageries from three different years with limited cloud coverage and clear scan is essential to get an accurate analysis.After acquiring the imageries, image pre-processing, such as image subset and image error is also essential to focus the study on the area and remove or reduce the effects of atmospheric, scattering and absorption that will improve the quality of the image.

Image Processing
Image processing is used in order to process and analysis many images to aid the interpretation of remote sensing and extract as many details and information that could be taken from the images that have been collected and processed.
In this stage, the image will undergo a couple of processing steps, that is layering, classification, principal component, and weighted overlay.In this study, the processing will only take 2 different bands from the downloaded imageries, that is band 2 and Band 3. In Landsat 8, Band 2 and Band 3 in for visible blue and visible green spectral where the wavelength of band 2 is 0.450 to 0.515 µm and wavelength for band 3 is 0.525 to 0.600 µm with both of their resolution is 30 meters.While in Landsat 5, band 2 and band 3 are defined as visible green and visible red with the wavelength of 052 to 0.60 µm and 0.63 to 69 µm with both bands have a resolution of 30 meters.

Layering
Layering is the first processing where the three imageries that have been collected will be processed using raster calculator to create three different layers.In achieving water quality assessment, parameters for water quality are needed.
For this study, the parameters that are used were Total Phosphorus (TP), Total Nitrogen (TN) and Biochemical Oxygen Demand (BOD).Raster Calculator is a tool where it will allow the user to create and execute a map algebra expression using Phyton syntax since raster calculator interface is built like a more scientific calculator.Using raster calculator, three different layers will be created using the formula.

Classification
Image classification is the process of assigning or categorizing classes within an image that is based on their own content or characteristic of the layer.In this study, using classification process can help on providing valuable information for the water quality classes.For this research, the image classification will be used for Total Nitrogen (TN) layer, Total Phosphorus (TP) layer and Biochemical Oxygen Demand (BOD) layer.Layering process will use Arc toolbox to use Reclassify option under Spatial Analyst Tools to create 3 classes based on their specifications.

Principal Component
Principal component is the process of identifying duplicate data over several datasets.Due to the unavailable data of percentage to each layer, principal component is used to get the weightage for each layer before the layers undergo weighted overlay process.The layer that will be used to process principal component analysis is the reclassified image of Total Phosphorus (TP) layer, Total Nitrogen (TN) layer, and Biochemical Oxygen Demand (BOD) layer.
It is a complex and involved diligent effort to fully understand the current state of the river basin's water quality, to identify the factors that affect it, and to improve the river basin's water environment quality [16].Principle component analysis (PCA) was used on the Tasek Bera to provide a comprehensive perspective of all the variables involved in the system [17].The PCA method consists of the following five key operational steps: (1) The original data matrix is listed: where   in the matrix is the observational data, n represents the monitoring station, and p represents each water quality parameter in this study.
(2) To diminish the effects of dimension, standardize the original data using the Z-score standardization formula (Equation (1) (Belkhiri 2015).
where   * is the standard variable,  is the average value for jth indicator, and sj is the standard deviation for jth indicator.
(4) Calculate the eigenvalues and eigenvectors of the correlation coefficient matrix, R, to determine the number of principal components.The eigenvalues of the correlation coefficient matrix, R, are represented by ⋌  ( = 1,2 ••• ) and their eigenvectors are   (  =  1,  2 ⋯   )( = 1,2 ⋅⋅⋅ ).The primary component's variance is represented by the λ value.Additionally, there is a positive correlation between variance value and primary component contribution rate.Additionally, the cumulative contribution rate of the first m primary components must be greater than 80%, which indicates:∑ =1  ⋌ j / ∑ =1  ⋌j≥0.80.The principal component is represented by Equation (3).
where   * is the standardized indicator variable.  * (  −   )/  (5) Equation ( 4) shows how the acquired principal components are weighted and added to produce a thorough evaluation function.
Principal Component is done by selecting the spatial analyst options in Arc Toolbox and inserting the data that is needed.In this study, the three parameters' layers will be the input raster bands.The set of raster bands will then generate a single multiband raster with the percent variance that will identify the amount of the variance each eigenvalue captures.Eigenvalues in principal component refer to the coefficient that is applied to eigenvectors that give the vectors their length or magnitude.The weightage will then be calculated to get the exact percentage to be used in weighted overlay.

Weighted Overlay
The last step in the process is performing weighted overlay.Weighted overlay is used to make the finishing map of defining the water quality for Tasek Bera.The layer that will be included in this process is the TN layer, TP layer and BOD layer.
Weighted Overlay is a process that helps users to overlay several layer rasters in accordance with their importance using their own measurement scale and weight.In weighted overlay, the raster layer will be assigned with percentage influence to complete the process.In this study, each raster has its own weight that will be used in this process to get the water quality for each year, that is year 2000, year 2010, and year 2020.The process will use ArcToolbox under Spatial Analyst tools.The output of the process will be the water quality map of the area that has been highlighted.

Results and discussions
The results and analysis for this study can be seen in Figure 2, Figure 3, and Figure 4 where the Water Quality Index for the year 2000, year 2010 and year 2020 is shown.For the analysis, the Tasek Bera maps will be divided into three different sections, the upper section of Tasek Bera, the middle section of Tasek Bera and the lower section of Tasek Bera.All the water quality classification is based on the Malaysian Water Quality Index.In the year 2000, the state of water quality at most regions of Tasek Bera were classified as class ⅡA.This is because, as we can see in Figure 2, it shows that the majority of the upper section and the majority of the lower section of the lake is green in color which indicate class ⅡA with the edge of the upper section can be considered as class Ⅲ.The middle section of the map is mostly categorized as class Ⅰ for water quality index.
Based on the findings, the greenery that is surrounding Tasek Bera is because of urbanization and bare land.Urbanization is used in terms of the infrastructure in the area, such as housing, building and other structures.Bare land is used in terms of the land is not used for anything and there is no greenery that has been detected.The greenery is more prominent in year 2000 since urbanization is not yet fully developed and only one spot can be seen close to Tasek Bera.However, there are many areas of bare land that is surrounding the lake, indicating that bare land has more effect on the water quality of Tasek Bera than vegetation and urban.Therefore, in year 2000, it has been concluded that 36.13% of IOP Publishing doi:10.1088/1755-1315/1316/1/01200410 Tasek Bera is in excellent water quality, 50.223% is fair water quality and 13.765% is poor water quality, it can be deduced that the water quality of Tasek Bera on year 2000 is not polluted nor it is healthy.Different from year 2000, based on water quality index, this study has determined that in the year 2010, the state of water quality at Tasek Bera was classified under class Ⅰ.This is because, as we can see from Figure 3, the lower section of the lake, most of the region are green in color which indicated to class Ⅰ water quality whereas the upper section and the middle section is scattered with majority red color which shows that the water quality in that section lies under class Ⅲ.
The reasons of the decreasing water quality are because the vegetation area in Tasek Bera has increased significantly, as well as urban areas while bare land has been decreased.This could also lead to the difference in the map of water quality of year 2000 and year 2010.With the percentage of the land use as 58% of the map is vegetation, 27% is water bodies, 9% urban and 6% is bare land.With vegetation as the majority of the land use map, it can be seen that having more vegetation could also affect the water quality of Tasek Bera, with the water quality of year 2010 effect more on the vegetation side.The lower section of the map shows green area at most of the region and this is because the area is surrounded by vegetation and bare land.Hence, we can come to a conclusion that the percentage of water quality at Tasek Bera in the year 2010 according to the overlaying map with 55.389% of excellent water quality, 29.865% of fair water quality and 14.746% of poor water quality, it can be concluded that the water quality of Tasek Bera on year 2010 is excellent water quality.Figure 4 shows that the state of water quality at Tasek Bera worsen in 2020, with the upper half of the map being more apparent on class III water quality and the middle section of the lake more prominent on class IIA water quality.However, Class I water quality may be found at the upper, middle, and lower sections of Tasek Bera, where the color of the map is predominantly green.According to the map in Figure 4, the percentage of water quality in Tasek Bera in 2020 is 48% excellent water quality, 34% fair water quality, and 18% poor water quality, it can be determined that Tasek Bera's water quality in 2020 is excellent water quality.Based on the process and result that has been shown, it can be stated that the water quality for Tasek Bera in 2020 is fairly polluted.The percentage of poor water quality can be seen increasing by year and the amount of excellent water quality is decreasing rapidly.The cause for Tasek Bera to be increasingly polluted can be various, the most specific one is the human's activities that surrounds the lake.

Conclusions and Recommendations
Water quality can vary from decade to decade, depending on their habitat and aquatic organisms, and it is our responsibility to nourish the lake and the environment that surrounds it to the best of our ability.Based on the study that has been conducted, three main maps for water quality for Tasek Bera have been created to analyse the degradation of water quality in Tasek Bera, Pahang.Based on the maps and finding that has been analysed, the water quality index for Tasek Bera has been observed.Tasek Bera's pollution level has drastically decreased between the years 2000 and 2010.However, from 2010 to 2020, the degree of pollution in Tasek Bera has increased, resulting in a deterioration in water quality in 2020.
Besides that, using Remote sensing and GIS can make it easier for government to keep an eye and make a plan on improving the well-being of the lake.This is because, using remote sensing and GIS, researchers don't have to do any in-situ monitoring to obtain the data of water quality parameters.This method not only saves the researcher time and energy, but it also speeds up the process of determining the lake's water quality.Some recommendations for future research include downloading images with less cloud coverage, identifying the procedure of the study, and completely understanding the process.Aside from that, we can utilize supervised classification to generate a land use/land cover map, and we may further investigate the formula and band that will be utilized.

Figure 2 .
Figure 2. Water Quality Index for the year 2000

Figure 3 .
Figure 3. Water Quality Index for the year 2010.

Figure 4 .
Figure 4. Water Quality Index on year 2020

Table 1 .
Statistic for colour coordination in year 2000, year 2010 and year 2020 imageries for Band 2

Table 2 .
Statistic for color coordination in year 2000, year 2010 and year 2020 imageries for Band 3

Table 3 .
The table of formulas for TN, TP, and BOD

Table 4 .
Table of classification and range for the year 2000

Table 5 .
Table of classification and range for the year 2010

Table 6 .
Table of classification and range for the year 2020

Table 7 .
Table specification for the year 2000 before and after classification is done

Table 8 .
Table for layers for the year 2000

Table 9 .
Table for layers for the year 2010

Table 10 .
Table for layers for the year 2020