Spatiotemporal Assessment of Water Pollution for Beira Lake, Sri Lanka

Abstract

Beira Lake, located in Colombo, Sri Lanka, has suffered severe anthropogenic impacts, with previous restoration attempts failing due to a limited understanding of pollutant dynamics. Aiming to fill this gap, a comprehensive study was conducted during dry and wet seasons to assess the spatiotemporal water pollution of Beira Lake, employing key physicochemical parameters, numerical indices, and remote sensing analysis. The water pollution index (WPI) results categorize Beira Lake as highly polluted, with WPI values ranging from 2.38 ± 0.92 in the wet season to 2.53 ± 1.32 in the dry season. Comparatively higher COD levels recorded in the Beira Lake network, especially for Gangarama Lake show significant pollution levels during both the dry and wet seasons, e.g., the highest COD levels, at 306.40 mg/L, were observed during the wet season. The Trophic State Index (TSI) results indicate eutrophic and hypereutrophic conditions in Beira Lake, which are particularly pronounced during the wet season. The heavy metal pollution index (HPI) results suggest elevated heavy metal concentrations in Beira Lake, especially in the wet season. Combined with field investigation results, a remote sensing data analysis between 2016 and 2023 reveals significant improvements in water transparency, suggesting positive effects of recent management interventions. Parameters demanding attention include COD, nitrate, and total phosphate levels due to their consistent exceedance of permissible limits. The PCA results of indices correlations between wet and dry seasons offer valuable insights into the complex dynamics of Beira Lake’s water quality. The study makes recommendations for restoring Beira Lake, including stringent pollution controls, regular dredging, green infrastructure implementation, implementing new rules and regulations, and community engagement.

1. Introduction

Freshwater resources are indispensable for sustaining both natural ecosystems and human progress, serving as a fundamental necessity for agriculture, industry, and overall human survival [1]. Among these resources, freshwater lakes are precious, playing essential roles within ecosystems globally [2]. Urban lakes, which are larger than ponds and nestled within urban environments [3], also hold significance in urban water systems [4]. These urban lakes serve various functions, from rainwater harvesting to groundwater recharge, providing recreational spaces, emergency firefighting water sources, and ecological balance [5]. Moreover, they play pivotal roles in flood control and low-impact development, which aims to reduce runoff and pollution during heavy rainfall through decentralized source control measures [6].

Unfortunately, most lakes are vanishing due to tremendous population growth and rapid urbanization [7]. The water quality issue in urban lakes is of significant concern, as it is closely tied to the increased likelihood of urban residents coming into direct contact with the water [8]. Urban lakes face environmental challenges stemming from urban development [9]. Rapid urbanization, especially in densely populated areas, poses risks to urban lakes due to the presence of untreated sewage and wastewater discharge [10]. This urban expansion negatively impacts water quality by increasing nutrient and pollutant inputs, altering organic matter transport, and changing water dynamics [11]. As a result, many urban lakes experience eutrophication, driven by excessive urban-related nutrients [12] and organic matter inputs [13,14]. This transformation often shifts water bodies from oligotrophic to mesotrophic and ultimately, to eutrophic states [15,16].

Beira Lake, located in the bustling urban center of Colombo, Sri Lanka’s primary economic hub, exemplifies these challenges. Comprising a network of four interconnected lakes, Beira Lake’s degradation began during the latter half of the 19th century, triggered by land reclamation for commercial purposes and the discharge of municipal and industrial effluents into the lake [17]. The lake’s eutrophication stems from human activities within its urbanized watershed, characterized by improper construction practices and wastewater discharge. This contamination risks both human health and the environment, as the lake remains stagnant and heavily polluted [18]. Beira Lake has also lost its economic and aesthetic value, impacting recreation, leading to foul odors, unsightly views, and health problems such as mosquito infestations in surrounding areas [19].

In recent years, several governmental efforts and numerous studies have investigated the water quality of Beira Lake across different basins and time periods. The study of Ref. [20] focused on the physiochemical parameters in the surface water of East Beira Lake over six months, assessing its suitability for fish and aquatic life. In 2017, the Beira Lake Restoration Study examined internal and external nutrient loadings, while Kamaladasa in 2007 [17] studied the restored Southwest and non-restored East Beira Lakes, assessing their trophic status and the impact of restoration on water quality parameters. A recent study of Dharmarathna in 2023 [21] aimed to fill this gap by examining all four basins simultaneously, but omitted seasonal variations and heavy metal pollution levels, which significantly influence the lake’s water quality. In addition, no investigation on water quality based on remote sensing has been carried out until now. Despite these investigations, challenges remain in identifying the causes of current pollution due to the lack of effective pollution mitigation strategies.

These identified gaps prompted this study to investigate the spatial and temporal variations in water quality across all four basins, evaluating the pollution state according to seasonal changes, using multiple indices, identifying potential pollution sources, and assessing heavy metal contamination. By addressing these gaps, the study not only enhances the understanding of Beira Lake’s water quality, but also contributes to more effective management and sustainable development strategies for the region. Hence, this study aims to comprehensively evaluate the water quality of Beira Lake, including all four basins simultaneously, based on both spatiotemporal variations, using physiological chemical parameters along with numerical indices, specifically the water pollution index (WPI), heavy metal pollution index (HPI), and the Trophic State Index (TSI), along with remote sensing analysis. These indices condense complex data into easily understandable scores, facilitating comparisons of water quality across different locations and time periods [22,23].

Satellite remote sensing technology is crucial for understanding surface water and is extensively used for monitoring water quality parameters [24], as these techniques have significantly evaluated water quantity and the quality in inland waters, facilitating technological advances in instruments/sensors and algorithms/image processing [25]. Despite its value, remote sensing data is considered complementary to in situ measurements, providing synoptic, spatiotemporal views of water characteristic changes and contributing to a more comprehensive understanding of water quality [26]. Notably, the application of water pollution indices and remote sensing in Beira Lake remains limited, and this study offers a detailed evaluation of the lake’s water quality, by identifying the seasonal fluctuations, to propose strategies for pollution control and lake restoration, ultimately contributing to more effective management of Beira Lake.

2. Materials and Methods

2.1. Study Area

The research focuses on Beira Lake, a distinctive landmark in Colombo. Beira Lake comprises four main basins including the East Lake (EL), the Galle Face Lake (GFL), the West Lake (WL), and the Southwest Lake/Gangarama Lake (GL) (Figure 1) [27]. The extent of the Lake is about 65.4 ha, and it depends on the runoff of its highly urbanized catchments [28]. The lake has been gradually reduced to its present state from an estimated original area of 162 Ha, mainly due to reclamation for warehouse construction and hotels [29] (Table S2).

As Sri Lanka is a tropical country, rainfall is primarily derived from two monsoons: the southwest monsoon, from May to September, and the northeast monsoon, from December to February. Additionally, the country experiences inter-monsoonal rains in between the monsoon periods, while receiving annual rainfall averages around 2400 mm. As Beira Lake is located in the wet zone of the urban heart of Colombo, it experiences the wet seasons from May to September and October to January and a dry season from February to April. Seasonal rainfall causes substantial variability in the lake’s hydrology and hydrological parameters across the four lakes, with increased urban runoff during the wet season bringing various pollutants.

Hydrogeologically, Beira Lake is shallow, with complex surface water–groundwater interactions. Urbanization has altered its hydrogeology, with industrial discharges and sewage inflows contributing to pollution. The catchment area, including commercial, residential, and industrial zones, impacts water quality, especially during heavy rains.

2.2. Sampling Process

Based on the criteria of the complete and even coverage of the study area, sampling locations in these four lakes of the Beira Lake network were chosen randomly (Figure 1), according to the Technical Guidelines for Investigation and Evaluation of Lake Ecological Security (trial), released by the Ministry of Environmental Protection, China, in 2014.

Sri Lanka exhibits obvious differences between wet and dry seasons, and we chose to collect samples in September 2022 (dry season) and March 2023 (wet season) to accurately analyze the water quality changes in the lakes. During the sampling process, factors such as outfalls, rivers, and streams of the lake, as well as pollution status, were taken into account, with 29 samples collected in 2022 due to the COVID-19 localized lockdown; 51 samples were collected in 2023, ensuring broad temporal and regional coverage. The process involved collecting triplicate samples from each sampling point at a 0–10 cm depth. The samples were collected between 8.00 a.m. and 12.00 noon, using a boat to ensure consistency in the sampling conditions. These samples were then preserved and transported to the laboratory as soon as possible to maintain their integrity for further analysis.

• In situ surface water quality monitoring

The in situ physicochemical parameters were measured using portable instruments. Dissolved oxygen (DO) was measured using a LEHERO LFWCS-2008 DO Water Quality Analyzer. Electronic conductivity (EC), surface water temperature, pH, and the total dissolved solids (TDS) were measured using a MultiLine Multi 3530 (WTW, Welheim, Germany). Secchi disc depth (SDD) was measured with a standard Secchi disk, 20 cm in diameter, with black and white quarters [1]. Collected samples were stored at 4 °C until testing, according to the American Public Health Association (APHA) standards (23rd edition) [30] for ex situ water quality parameters.

• Ex situ surface water quality analysis

The CODcr was determined using a prefabricated tube reagent from HACH, Loveland, CO, USA. NH3-N was measured using the Nessler reagent (HJ 535-2009). The total nitrogen (TN) concentration of the water samples was determined using a TN analyzer (Elementra, Langenselbold, Germany). Total phosphate (TP) was measured using the ammonium molybdate spectrophotometric method [30]. Heavy metals (Cd, Cr, Pb, Hg) were determined by an inductively coupled plasma mass spectrometry (ICP-MS) device (NexION 300X Perking Elmer, Houston, TX, USA). The anions (Cl⁻, F⁻, nitrate, and sulfate) were analyzed using ion chromatography (ICS 1000, Dionex, Sunnyvale, CA, USA).

2.3. Data Analysis

The analysis of variance (ANOVA) and an independent sample Kruskal–Wallis test were performed according to the normality distribution of the sample to investigate the differences between and within sampling sites at a 95% confidence interval using IBM SPSS statistic (version 23) software. Correlation analysis and principle component analysis were carried out using OriginPro 2024 (learning edition 10.1.0.170) software. The differences between sites were examined to determine the spatial variation, while the differences within seasons (wet and dry) addressed the temporal variation of the water sample. The Pearson correlation was chosen to evaluate the correlation, and probability values were significant at a level of 0.05.

A remote sensing data analysis was carried out using Sentinel-2 data. Sentinel-2 has 13 spectral bands, with a width of 290 km and a revisit period as short as 5 days (Table S4). Its spatial resolution is up to 10 m. It can be used to monitor water quality changes. In this study, the Sentinel-2_MSI_L2A images for March 2016 and 2023 were acquired from the European Space Agency (ESA).

2.4. Index Analysis

For a comprehensive evaluation of the water quality, pollution status, and ecological condition of Beira Lake, this study uses the Water Pollution Index (WPI), Heavy Metal Pollution Index (HPI), and Trophic State Index (TSI). The selection of these indices for assessing Beira Lake’s water quality is based on their comprehensive and widely recognized utility. The WPI provides a broad overview of pollution by integrating multiple water quality parameters, while the HPI focuses on heavy metal contamination, crucial for understanding ecological and health impacts. The TSI assesses the trophic status and eutrophication levels, offering insights into nutrient dynamics and ecological health. These indices provide a robust evaluation of Beira Lake’s water quality and pollution levels. Their comprehensive capacity to assess different aspects of water quality and pollution enhances the methodological framework of the study using a thorough and multifaceted approach.

2.4.1. Water Pollution Index

The water pollution index (WPI) offers a sensitive, practical, and versatile approach to assess water quality. It consolidates multiple parameters into a single index, is adaptable to diverse datasets, avoids theoretical ideal values, and provides efficient evaluation using standard guidelines, making it a valuable tool for water quality assessment. WPI was calculated using the method provided by Ref. [31], using Equations (1) and (2). The water quality parameters, EC, TDS, pH, COD, NO₃⁻-N, NH₃-N, sulfate, Cl⁻, F⁻, and heavy metals (Cd, Cr, Pb, and Hg) were selected and used for the calculation of WPI in this study, considering the standards of WHO (2022), SLS614-2013 [32], and AWQS (2019) [33] category C, as that section is only suitable for aquatic life, the only usage of Beira Lake water. Water pollution status was grouped into four categories, based on WPI—highly polluted, moderately polluted, good water, and excellent water.

In the first step, the pollution load ($PLi$) of ith parameter was calculated using the following formula [31].

$$PLi=\frac{Ci-7}{Sib-7}$$

(1)

where Ci indicates the observed concentration of ith parameter; Sib is the respective parameter’s standard or highest permissible limit. Ultimately, the water pollution index (WPI), with n number of variables (parameters), can be evaluated by aggregating all the pollution loads and finally dividing by n, as shown in the following formula [31].

$$WPI=\frac{1}{n} {{\displaystyle \sum }}_{i=1}^{n}PLi$$

(2)

The WPI values may be classified based on n number of parameters in our categories (Table S3).

2.4.2. Trophic State Index

The Trophic State Index (TSI), a vital tool in eutrophication assessment worldwide [34], was first developed by Ref. [35], based on TP, SDD, and Chl.a, then a Carlson-type TSI for TN was proposed by Ref. [36]. The TSI value corresponds to <40, oligotrophic; between 40–60, mesotrophic; from 60–80, eutrophic; and >80, hypertrophic.

The calculation of TSI(SDD), TSI(TN), and TSI(TP) are as follows [34,37]:

TSI(TP) = 10 × [6 − ln(48/TP)/ln2]

(3)

TSI(SDD) = 10 × (6 − lnSD/ln2)

(4)

TSI(TN) = 10 × (6 − ln(1.47/TN)/ln2)

(5)

where, TSI(TP) = Trophic State Index of total phosphorus, TP, in (μg/L); TSI(SD) = Trophic State Index of Secchi disk depth, SD, in (m); TSI(TN) = Trophic State Index of nitrogen, TN, in (mg); ln = natural logarithm.

2.4.3. Heavy Metal Pollution Index (HPI)

The heavy metal pollution index (HPI) shows the combined effect of individual HMs on surface water quality [38]. It is a widely used method for evaluating water quality in terms of metals. The developed method measures the total effect of more than one heavy metal on water quality. HPI was calculated according to Ref. [39] in two steps. In the first step, the sub-index (Qi) value is calculated using each parameter’s measured heavy metal values and the fixed limit values in the relevant standards. In the second step, HPI is calculated using each parameter’s weight values and sub-index values. Pollution indices are generally used for any purpose (drinking water, irrigation water, industry, etc.) or specific water uses. A value of HPI below 100 represents low pollution of HMs, while 100 is the threshold value at which harmful health consequences are probable. An HPI value greater than 100 indicates that the water is unsuitable for consumption.

$$Qi={\displaystyle \sum }_{i=1}^n\frac{\left|Mi-Ii\right|}{Si-Ii} 100$$

(6)

where Mi is the examined value of the ith HM. Si and Ii are the standards and ideal values, respectively, taken from Refs. [33,40], for the HMs (µg/L).

In the second step, HPI is calculated using each parameter’s weight values and sub-index values.

$$HPI=\frac{{{\displaystyle \sum }}_{i=1}^{n}Wi Qi}{{{\displaystyle \sum }}_{i=1}^{n}Wi}$$

(7)

where Wi is the unit weighting of the ith HM, Qi is the sub-index for the ith HM, and n is the number of HMs measured (Cr, Cd, Pb, and Hg) in this study.

2.5. Modeling of SDD Using QAA

The quasi-analytical algorithm (QAA) can precisely interpret the sighting of a Secchi disk by combining empirical and analytical methods using radiative transfer equations [41,42]. Spectral bands of the Sentinel-2 sensors used for water quality monitoring are shown in Table S4.

• Estimation of IOPs

Remote sensing reflectance was processed using the QAA_v6 algorithm to obtain the total absorption a(λ) and backscattering bb(λ) coefficients. This algorithm consists of seven steps, including three semi-analytical models and two models using empirical and analytical approaches [43]. The modeled Secchi disk depth (Z_SD) is constructed based on four bands of Sentinel-2 satellites, centered at wavelengths 443, 490, 560, and 665 nm, respectively.

In Step 1, the remote sensing reflectance $R_{rs}$ was processed to the subsurface remote sensing reflectance ($r_{rs}$) data using the following equation:

$$r_{rs}\left(\lambda \right)=\frac{R_{rs}}{0.52+1.7R_{rs}}$$

(8)

In Step 2, the ratio $u\left(\lambda \right)$ is formulated in terms of $r_{rs}$, $g_0$, and $g_1$, where $g_0$ and $g_1$ are 0.089 and 0.125 sr⁻¹, respectively [44].

$$u\left(\lambda \right)=\frac{b_b\left(\lambda \right)}{a\left(\lambda \right)+b_b\left(\lambda \right)}=\frac{-g_0\pm \sqrt{g_0^2+4g_1{r}_{rs}\left(\lambda \right)}}{2g_1}$$

(9)

In Step 3, the total absorption coefficient (a) is estimated at the reference wavelength (λ₀) using the following equation:

$$a\left({\lambda }_0\right)=a_w\left({\lambda }_0\right)+\mathsf{\Delta }a\left({\lambda }_0\right)$$

where $a_w$ is the absorption coefficient of pure water, assumed as a constant, Δa(${\lambda }_0$) is the contribution from non-water constituents, and ${\lambda }_0$ is empirically assessed as 665 nm. The equation is modified as follows.

$$a\left(665\right)=a_w\left(665\right)+0.39{\left(\frac{r_{rs}\left(665\right)}{r_{rs}\left(443\right)+r_{rs}\left(490\right)}\right)}^{1.14}$$

(10)

In Step 4, the backscattering coefficients of suspended particles ($b_{bp}$) is calculated from the analytical equation.

$$b_{bp}\left(670\right)=\frac{u\left({\lambda }_0\right)a\left({\lambda }_0\right)}{1-u\left({\lambda }_0\right)}-b_{bw}\left({\lambda }_0\right)$$

(11)

where $b_{bw}$ is the backscattering coefficient of pure water.

In Steps 5 and 6, b_bp values at 443, 490, and 560 nm are estimated using the following expression (Table S4).

$$b_{bp}\left(\lambda \right)=b_{bp}\left({\lambda }_0\right){\left(\frac{{\lambda }_0}{\lambda }\right)}^{\eta }$$

(12)

$$\eta =2.0\left(1.0-1.2\exp \left(-0.9\frac{r_{rs}\left(443\right)}{r_{rs}\left(560\right)}\right)\right)$$

(13)

In the final step, the total absorption coefficient a(λ) is derived from a combined equation.

$$a\left(\lambda \right)=\frac{\left(1-\mathrm{u}\left(\lambda \right)\right)\left(b_{bw}\left(\lambda \right)+b_{bp}\left(\lambda \right)\right)}{\mathrm{u}\left(\lambda \right)}$$

(14)

2.

Estimation of K_d

K_d is calculated by the following equation [45].

$$K_d\left(\lambda \right)=\left(1+m_0\times {\theta }_s\right)\mathrm{a}\left(\lambda \right)+\left(1-\gamma \frac{b_{bw}\left(\lambda \right)}{b_b\left(\lambda \right)}\right)\times m_1\times \left(1-m_2\times e^{-m_3\times \mathrm{a}\left(\lambda \right)}\right)b_b\left(\lambda \right)$$

(15)

where $m_0$, $m_1$, $m_2$, $m_3$, and γ are the model parameters, with respective values of 0.005, 4.26, 0.52, 10.80, and 0.27. θs is the solar zenith angle in degrees. a(λ) and b_b(λ) are the absorption and backscattering coefficients estimated by QAA_v6. b_bw(λ) refers to the backscattering coefficient of pure water.

3.

Estimation of $Z_{SD}$

$Z_{SD}$ is inversely proportional to the minimum value of the diffuse attenuation coefficient K_d in the range of 443–665 nm [45].

$$Z_{SD}=\frac{1}{2.5K_d^{tr}{ }_{\min }}\ln \left(\frac{\left|0.14-R_{rs}^{tr}\right|}{0.013}\right)$$

(16)

4.

Accuracy Assessment of the $Z_{SD}$ Model

In situ ZSD sample data are used to validate the $Z_{SD}$ model. The statistical indicators of root mean square error (RMSE), the mean absolute difference (MAD), the mean absolute percentage error (MAPE), and the determination coefficient (R2) provide a better understanding of the differences between the measured and remote sensing data. If the difference is small, the remote sensing monitoring result is considered accurate; otherwise, the results of the ZSD model must be calibrated using the measurements.

3. Results and Discussion

3.1. Spatial and Temporal Distribution Characteristics of Water Quality

3.1.1. Key Parameters of COD, SD, DO, pH, EC, and TDS

Major water quality parameters, namely COD, SDD, DO, pH, EC, and TDS, are shown, according to each lake, in the graphs in Figure 2. Temperature changes are shown in the Supplementary Materials (Figures S1 and S2). The lake’s temperature remains below 40 °C throughout the study period. The tolerance limit for pH is 6.0–8.5, which has been exceeded by most of the sampling points throughout the sampling sessions, and the maximum value of 10.97 is recorded in the dry season. Conductivity, SD, and TDS tolerance limits were not provided for fish and aquatic life in the National Environmental Act No—47 OF 1980 (2019 edition), Ambient Water Quality Standards for Sri Lanka [33]. According to the graphs, TDS, salinity, and conductivity show similar spatial and temporal variability patterns. Most of the sampling points indicate a suitable DO level for fish and aquatic life in the lake by exceeding 5 mg/L, the given tolerance limit of DO. High mean DO, pH, temperature, and water level values were recorded in the dry season, and EC, TDS, and SD values were increased in the wet season. EC, TDS, and SD values are high in the wet season, and the differences in those values between seasons are very significant (p < 0.05).

• Chemical Oxygen Demand (COD)

East Lake shows an increase in COD in the wet season, while other lakes show higher COD values in the dry season (Figure 2). Some areas in Gangarama Lake show higher COD values in both wet (340 mg/L) and dry (373.41 mg/L) seasons (Figure 2). The lowest value was recorded in East Lake (51.34 mg/L) during the dry season and in West Lake (117.83 mg/L) during the wet season.

All COD values exceeded the ambient water quality standard of 15 mg/L in both seasons, indicating severe organic pollution inputs into several parts of the Beira Lake network, with Gangarama Lake significantly exceeding the water quality standards year-round (Figure 3). The primary source of high COD levels in this area is untreated or partially treated sewage, which introduces large amounts of organic matter that increase oxygen demand during decomposition. Additionally, urban runoff carries both organic pollutants from residential and commercial establishments and inorganic pollutants from roads and surfaces. Industrial discharges also contribute chemicals and heavy metals, which further elevate COD levels due to the significant oxygen required for chemical reactions. These combined sources—untreated sewage, urban runoff, and industrial discharges—result in consistently high COD levels, exceeding water quality standards in both wet and dry seasons.

• SDD

Although there was a high surface water runoff during the wet season, the water clarity was high in the lake during the wet season, as it showed high SD values (0.21 m–0.40 m) (Figure 2) and also expressed low turbidity during the wet season. The highest mean SDD (0.33 m) was recorded in East Lake, while the lowest (0.26 m) was recorded in Galle Face Lake during the wet season. SDD reduced slightly during the dry season, as the highest level (0.27 m) was recorded in Galle Face Lake, and the lowest (0.14 m) was recorded in West Lake. With the rain, the visibility of the water also increased, as well as TDS and EC, which are also relatively high during the wet season, especially in Galle Face Lake and West Lake, possibly be due to the high rainfall runoff (Figure S3). SD also shows high values during the wet season compared to the dry season.

• Dissolved Oxygen (DO)

The DO concentration is greater than 11 mg/L in most of the Beira Lake network in the dry season, remaining consistent in most East Lake areas (Figure 4). In dry and wet seasons, the DO concentration is below 5 mg/L in the lower part of the West Lake area. Lower mean DO concentration values (7.20–9.58 mg/L) were observed for the lakes in the wet season than those obtained in the dry season. A higher DO concentration was observed in the wet season than in the dry season in the narrow canal between Gangarama Lake and West Lake (Figure 4).

DO shows significant spatiotemporal variation during the study period. The mean DO values of the lakes in the dry season are higher than those in the wet season. The highest mean DO values for the wet and dry seasons (9.58 mg/L and 14.68 mg/L) were recorded in East Lake, and the lowest mean values (7.2 mg/L and 7.58 mg/L) were recorded (Figure 3) in West Lake.

The DO levels in Beira Lake show significant spatial and temporal variation, with higher concentrations during the dry season compared to the wet season. This is mainly due to increased algal growth during favorable weather, leading to increased photosynthesis and higher surface DO values. Higher temperatures in the dry season also enhance metabolic rates in aquatic flora and fauna, contributing to elevated DO levels. Conversely, increased runoff during the wet season introduces more organic material and pollutants, which consume oxygen through microbial decomposition, reducing DO levels in certain areas.

• pH

The pH values in Beira Lake exhibit significant seasonal and spatial variations, consistently maintaining alkaline conditions throughout the year. During the wet season, the pH ranges from 7.5 to 8.2, aligning with the ambient water quality guideline for aquatic life (6.5–8.0). In contrast, the dry season sees pH values range from 9.4 to 10.4, significantly exceeding this guideline. This high alkalinity is primarily attributed to the increased photosynthetic activity of blue-green algae, which elevates pH levels through the uptake of carbon dioxide. The relative rates of photosynthesis and respiration contribute to these higher pH levels. Additionally, urban runoff introduces various pollutants that can influence pH levels.

During the wet season, rainfall dilutes these pollutants, resulting in lower pH values compared to those of the dry season. Notably, a significantly lower pH value of 5.1 was observed in the dry season, indicating localized variability. Overall, the study period revealed that Beira Lake maintains alkaline conditions year-round, with notable pH fluctuations are influenced by algal activity and urban runoff.

• EC and TDS

The spatial and temporal variation in TDS and EC shows significant differences. The lake exhibited a high rainwater surface runoff, which may cause high TDS (Highest—306.40 mg/L, Mean—206.20 mg/L) and EC (Highest—557.00 µs, Mean—374.90 µs) values during the wet season and low levels during the dry season. Spatial variation of EC and TDS is significantly different across the lakes during the wet season, and high TDS and EC mean values are shown in the West Lake and Gall Face Lake areas (Figure S3). The elevated EC and TDS values during the wet season in Beira Lake are primarily due to increased surface runoff, carrying various pollutants. The runoff from urban areas and construction sites introduces dissolved ions such as chloride, sulfate, and other anthropogenic contaminants into the lake. The rainfall runoff carries residues from domestic activities that mostly end up in the rivers or surface waters, which are major sinks of pesticides and other human-related contaminants [46]. These areas contain many construction fields that may contribute more dissolved ions than suspended particles in the wastewater during the wet season. Also, lake TDS pollution mainly depends on industrial effluents, water balance changes (limiting inflow, increasing water use, and increasing rainfall), seawater contamination, etc. [20]. This fact is confirmed, as East Lake and Galle Face Lake show significant differences (p < 0.05) between the two seasons for EC and TDS.

3.1.2. Spatiotemporal Variation in Nitrogen, Phosphorus, Sulphate, Chloride, and Fluoride

Figure 5a shows the seasonal variation in nutrient ions, sulfate, Cl⁻, and F⁻- concentrations for the lakes. All the parameters, except chloride and sulfate, show higher values in the wet season than in the dry season, which produces a significant difference in temporal variation (p < 0.05). Concentrations of TN, TP, nitrate, and fluoride are higher in the wet season than in the dry season. TN and TP levels are significantly higher in the wet season due to increased urban runoff and sewage discharge. During this period, heavy rainfall washes nutrients from residential areas and streets into Beira Lake. Most of the pit latrines experience off flow during the rainy season, and urban runoff carries this into the lake, while sewage discharge from inadequate wastewater treatment facilities contributes additional nutrients. These sources collectively lead to elevated TN and TP levels, exacerbating water quality issues in the lake during the wet season.

Additionally, ammonia nitrogen, chloride, and sulfate concentrations are higher in the dry season than in the wet season. GFL exhibits high concentrations of ions in both seasons. East Lake receives a high amount of TN in the wet season (20.84 mg/L) compared to the dry season (1.24 mg/L) (Figure 5a). Cl⁻ and sulfate show higher concentration values in the dry season than in the wet season (Figure 5a). The tolerance limit for NO₃⁻-N is 10 mg/L, and it is 0.4 mg/L for PO₄³⁻-P, according to the Ambient Water Quality Guideline for Sri Lanka, category C [33]. Many sampling points exceeded the nitrate nitrogen limit in the wet season, and some sampling points exceeded the tolerance value in the dry season. All the sampling points significantly exceeded the tolerance limit for total phosphate in the wet season, and barely surpassed the limit in the dry season.

• TN, NH₃-N, and NO₃⁻-N

The statistical analysis reveals significant temporal variations of TN in both dry and wet seasons, while no significant temporal variation is observed for NO₃⁻-N during these periods. In terms of spatial distribution in the wet season, there are no significant variations for mean NO₃⁻-N and TN. However, a noteworthy spatial variation for TN is observed in the dry season. The NO₃^--N concentration shows significant fluctuations across different lakes during the dry season, but changes over the seasons are not significant (p > 0.05). Notably, Galle Face Lake records the highest TN mean value (29.54 mg/L) in the wet season (Figure 5a).

During the wet season, a slight increase in nitrate concentration is observed in the West Lake and Gangarama Lake areas, possibly indicating domestic sewage influence (Figure 6). The St. Sebastian Canal and the Hospital Canal openings into East Lake show low nitrate levels in the wet season. Nitrate concentrations exceed 30 mg/L in both seasons in Galle Face Lake, concentrating more on the edges of East Lake, Gangarama Lake, and Galle Face Lake (Figure 6). Nitrate and TN distributions are more similar in the wet season than in the dry season, with nitrate being the dominant nitrogen compound in the wet season.

Most sampling points exceed the ambient water quality standard value of 0.22 mg/L for NH₃-N at pH ≥ 8.5 [33] in both seasons. The dry season is dominant for NH₃-N, showing significant temporal variation within the two seasons (p < 0.05). Except for East Lake, all other lakes show a significant increase in NH₃-N during the dry season, with the highest mean value recorded in West Lake. NH₃-N can be a domestic sewage source indicator [47]. Some areas in East Lake, particularly where large hotels are situated nearby, exhibit high NH₃-N concentrations (Figure 6). During the wet season, West Lake also shows a high amount of NH₃-N, suggesting elevated domestic sewage input.

Nitrate and TN distributions are more similar in the wet season than in the dry season, with nitrate as the dominant nitrogen compound. The TN amount is higher in the wet season than in the dry season, particularly in East Lake, Galle Face Lake, and Gangarama Lake, indicating higher nitrogen compound inputs during the wet season (Figure 6).

The increasing trend regarding nitrate concentration is likely due to nutrient enrichment of the lake’s littoral zone from anthropogenic impacts in the catchment area. Surrounding urban settlements, the hotel industry, and sewage treatment plants, with their associated sewage canals, contribute to high nutrient enrichment in the lake. The lake receives multiple inflows of decomposing organic matter from these areas, and the primary contributor to nitrate contamination is predominantly biological, including human and animal waste, particularly during the rainy season.

• TP, Sulphate, and Chloride

The statistical test indicates significant temporal variations for TP within both dry and wet seasons. However, unlike for TN, there is no significant spatial variation for TP in the wet season. In contrast, significant spatial variation for TP is observed in the dry season (Figure 7). West Lake stands out, with the highest TP levels in both dry and wet seasons. Additionally, TP concentration demonstrates significant fluctuations across different lakes during dry seasons. Phosphorous compounds are more concentrated in the dry season than in the wet season, and according to Figure 7 [20,48], bird droppings can cause higher phosphate levels in urban lakes.

According to Figure 5a sulfate shows higher concentration values in the dry season than in the wet season. Rainfall and runoff might dilute the sulfate amount received by the lake during the wet season. Sulfate is relatively high in Galle Face Lake and Gangarama Lake in the dry season, as well as in some parts of the Gangarama Lake and Galle Face Lake areas, showing high values, even in the wet season (Figure 6), which might include a high amount of anthropogenic inputs.

Cl⁻ concentrations are high in the dry season and significantly reduced in the wet season (p < 0.05). The highest Cl⁻ concentration was recorded at 36.85 mg/L during the dry season, and the lowest value was 2.22 mg/L during the wet season (Figure 5a). Cl⁻ shows a significant difference between the two seasons, with high amounts of Cl⁻ occurring during the dry season. Mean concentration values of Cl⁻ show a significant difference among seasons for some areas in the Gangarama Lake and the McCullum Gate area in East Lake; all lakes show a low amount of Cl⁻ distribution (Figure 7). Saltwater intrusion may cause the high Cl⁻ concentration near the McCullum Gate area, and anthropogenic outputs may cause the high Cl⁻ in the Gangarama Lake area.

3.1.3. Spatiotemporal Variation of Heavy Metals

The standard values and the mean heavy metal concentration values are shown in Table S5 in the Supplementary Materials. Most all the heavy metals mean values are below the maximum permeable level, according to the recently published ambient water quality standards for aquatic life [33], except for Hg concentrations in the dry season. Significant spatial variations exist for Pb and Hg (p < 0.05) across the wet and dry seasons. Hg shows a significantly higher difference between the two seasons, with very high Hg concentrations in the dry season (Figure 5b).

Heavy metal pollution does not show spatial significant differences during the wet season. However, the study shows significant differences in heavy metal pollution across different basins, as higher values were recorded for EL and GL. GL experiences elevated heavy metal levels due to urban activities, which might be due to runoff from vehicular emissions, building materials, and waste disposal. Inadequate sewage and wastewater treatment also introduce metals into the lake. Recreational activities contribute to pollution through littering and waste from boats and equipment. EL might be impacted by industrial discharges and port activities. Highly industrial activities, such as vehicle and battery repair shops, discharge heavy metals like lead (Pb) and cadmium (Cd) into the lake. Port-related activities also might contribute to contamination from anti-fouling paints and other sources. GFL and WL might be affected by vehicle emissions, urban runoff, sewage, wastewater, and recreational activities. Rainfall increases runoff, leading to higher levels of heavy metals in these basins, particularly during the wet season.

• Mercury (Hg)

There is no significant spatial difference in Hg across the lakes in both wet and dry seasons (p > 0.05). Rainwater and surface runoff might dilute the Hg concentration during the wet season, but it becomes more significant in the dry season in all four lakes (Figure 8). Wastewater containing Hg from various sources, including residential, commercial, and industrial activities, may cause a high Hg level in the dry season (Figure 5b). Automobile emissions, vehicle battery repair centers and garages, and heavy traffic in urban areas can be particularly significant sources of Hg contamination (Figure 8). The Hg value exceeds 2.25 mg/L in most of the Galle Face and West Lake areas, where most of the construction fields are located. According to the distribution maps, a specific location in Gangarama Lake shows a high level of Hg in the dry season, and this gradually decreases in the broader lake area. This area might provide a high amount of Hg to Gangarama Lake during the dry season (Figure 8).

• Chromium (Cr)

Cr concentrations between lakes show significant fluctuations across the different lakes during the dry and wet seasons. Still, the changes in its temporal variation over the seasons are not significant (p > 0.05). However, there is a significant spatial variation between the lakes in the dry season (p < 0.05). The highest Cr concentration values for both wet and dry seasons were observed in Galle Face Lake and West Lake, and East Lake shows the lowest value in the wet and dry seasons, respectively (Figure 8). Nonpoint source pollution, caused by high rainwater runoff; localized industrial sources, like metal plating and vehicle painting industries around the lake, which can wash chromium-containing particles and debris from roads, construction sites, and industrial areas into nearby lakes, might be the main source of Cr contamination during the wet season. Improper disposal of waste, including electronic waste (e-waste) and hazardous materials into canals and drainage systems, can result in the leaching of chromium and other contaminants into the surrounding environment in the dry season.

Cr distribution is significantly increased during the wet season in every lake. The location where the Hospital Canal opens to East Lake shows a high amount of Cr in the wet season, which might contribute a high amount of Cr to the EL (Figure 8). Some areas in GL show a high amount of Cd distribution in the dry season (Figure S5). McCullum Gates, which open onto Colombo Port, may leach heavy metals like Pb, Hg, and Cd into East Lake (Figure 8 and Figure S5). This area shows a high amount of Hg and Pb in both seasons and high levels of Cd in the dry season.

• Lead (Pb)

Although the mean Pb concentrations during the wet and dry seasons appear to be similar, there is a significant temporal variation in Pb levels between these seasons (p < 0.05). While Pb distribution does not exhibit any significant spatial variance in the wet season, there is an observable spatial variation between lakes in the dry season (Figure 5b). The higher levels of lead in the lake may be caused by a variety of sources, both point and nonpoint. This could include waste discharge from battery storage facilities, automobile repair shops, petroleum and chemical manufacturing plants, as well as runoff from municipal sources. Other possible sources of contamination could include electronics manufacturing, cable and tire factories, steelworks, and servicing stations located near the canal, which opens to the lake [49,50]. Pb does not show a significant distribution difference between the two seasons. But Galle Face Lake and the McCullum Gate area in East Lake show a high level of Pb distribution in both dry and wet seasons. Particularly, the areas near McCullum Gate and Gangarama Lake show a high level of Pb distribution in the wet season (Figure S5).

3.2. Water Quality Assessment Based on Index Analysis

3.2.1. Water Pollution Index (WPI)

The proposed index, WPI, is based on the standard permissible limits of drinking water parameters recommended by the WHO in 2022 and the ambient water quality for aquatic life in Sri Lanka, according to the 2019 standards (Table S6). In this study, the WPI for each sample was calculated to evaluate the degree of pollution in Beira Lake using 13 water quality parameters (n = 13) (Table S6). The WPI classifies surface water into four categories, such as excellent, when WPI < 0.5; good water quality, if WPI = 0.5–0.75; moderately polluted water, when WPI = 0.75–1; and highly polluted, when WPI is >1. According to the seasonal variation in the measured parameters of Beira Lake, WPI index values range from 2.38 in the wet season to 2.53 in the dry season (

Table 1. Summary of the values of different indices of the whole Beira Lake network in the wet and dry seasons (Mean ± S.D.).

Season	WPI	TSI			HPI
		TSI (TN)	TSI (TP)	TSI (SD)
Dry	2.53 ± 1.32	88.61 ± 17.58	79.41 ± 7.61	82.89 ± 4.11	38.37 ± 5.62
Wet	2.38 ± 0.92	99.07 ± 3.95	129.09 ± 2.79	77.64 ± 1.44	133.86 ± 2.55

), which shows highly polluted water in both seasons. When it comes to the lake-wise analysis, WPI values range from 1.57–3.70 in the dry season and 1.02–3.94 in the wet season. East Lake shows the lowest WPI value in the wet season, and West Lake shows the lowest value in the dry season (Figure 9a). Gangarama Lake shows the highest values (Figure 9a) in both seasons, indicating that the most polluted water can be seen there.

The WPI values range from 2.38 in the wet season to 2.53 in the dry season, indicating significant pollution year-round. Seasonal variations are due to runoff and dilution effects. In the wet season, rainfall increases runoff, carrying pollutants into water bodies. In the dry season, less rainfall reduces dilution, raising contaminant concentrations. Human activities, such as the discharge of untreated sewage and industrial effluents, introduce nutrients, heavy metals, and organic matter, further impacting water quality. Reduced water flow and evaporation in the dry season concentrate dissolved substances, while higher water volumes in the wet season temporarily dilute pollutants, despite increased runoff.

During the wet season, nutrient runoff from urban areas increases pollution, while in the dry season, pollutant concentrations rise. Key contributors include urban runoff, industrial discharges, and untreated sewage, especially around Gangarama Lake, which shows that the highest WPI in the wet season might be due to dense human and industrial activities. Specific activities contribute to the extreme WPI values in different basins. Gangarama Lake, for instance, exhibits the highest WPI values, which might be due to high urban runoff, industrial discharges, and concentrated human activities

The GL area might receive more water from sewage and water canals than do other basins. Inadequate waste management practices, including improper disposal of solid waste and ill-treated or poorly treated sewage inputs, exacerbate the pollution problem when compared to those of other lakes. High rainfall runoff containing oils, heavy metals, and other pollutants might also cause higher pollution rates. Consequently, the high WPI values in GL are attributed to intense urbanization, industrial activities, and inadequate waste management practices, compounded by natural seasonal variations that lead to high water pollution in the wet and dry seasons.

3.2.2. Trophic State Index (TSI)

Beira Lake exhibits eutrophic and hypereutrophic states, according to the TSI Index values, in both wet and dry seasons. TSI(TP) shows the highest value of 129 in the wet season, and TSI(SD) shows the lowest value of 77.64 in the wet season (Table S9). According to the Trophic State Index calculation, a comparatively higher pollution content can be seen in the wet season than in the dry season. According to Figure 9b, most of the lakes indicate a eutrophic or hypereutrophic state, according to the TSI index. TSI(TP) is very high during the wet season (Figure 9b) in all lakes, which may be caused by a higher concentration of TP in this season.

The TSI values indicate that Beira Lake predominantly exhibits eutrophic and hypereutrophic states, especially during the wet season. The primary sources of nutrients contributing to this condition are urban runoff and sewage discharge, which introduce significant amounts of nitrogen (N) and phosphorus (P) into the lake. During the wet season, increased rainfall leads to higher runoff, carrying these pollutants from urban areas into the lake. Industrial effluents also contribute substantially, with discharges containing high levels of N and P.

The TSI(TN) in East Lake is very dynamic, as it reaches a hypereutrophic level in the wet season and an oligotrophic state during the wet season (Figure 9b). Increased rainfall in the wet season leads to substantial runoff from urban and industrial areas, which might carry high nitrogen loads into the lake. Stormwater influx during this period further elevates nitrogen levels. Conversely, in the dry season, reduced runoff and higher evaporation rates reduce the external nutrient input, while biological processes, such as vegetative uptake, absorb available nitrogen, resulting in oligotrophic conditions.

There is no significant spatial variation for TSI values across the lakes. However, there is significant temporal variation for TSI values. These inputs are heightened during the wet season due to increased runoff, exacerbating eutrophication. Additionally, seasonal variation exacerbates eutrophication, with heavy rainfall events leading to spikes in nutrient concentrations, especially phosphorus concentrations.

Spatially, the nutrient sources and their impacts vary across Beira Lake’s basins. GFL exhibits higher nutrient levels due to urban runoff and sewage discharge. WL experiences elevated phosphorus levels, especially in the dry season, likely from bird droppings and local activities. EL shows fluctuating nutrient inputs, with substantial urban runoff effects during the wet season. GL shows higher nitrogen levels in the wet season from increased runoff and higher phosphorus levels in the dry season due to reduced dilution. These variations highlight the localized contributions to the lake’s eutrophic state. Overall, this shows that the trophic pollution is higher during the wet season than in the dry season.

3.2.3. Heavy Metal Pollution Index (HPI)

HPI findings, based on WHO (2022) guidelines for drinking water, indicate that surface water bodies show different pollution states regarding heavy metals in the wet and dry seasons. The HPI value was found above 100 in the wet seasons and below 100 in the dry season (Table 1), when the whole lake is considered. This indicates that the lake is receiving increased amounts of heavy metals, revealing the unsuitability of water. The findings show a lower pollution state for heavy metals in the dry season, as the lake might receive a lesser amount of heavy metals during this season. Also, the results indicated the unsuitability of the surface water bodies for consumption in the wet season. Heavy metal pollution is high in Galle Face Lake and West Lake compared to the other lakes during the wet season (Figure 9c), but there is no significant difference. However, the results show a significant difference during the dry season, as high heavy metal pollution levels can be observed in East Lake and Gangarama Lake (Figure 9c). This HPI value shows a high amount of heavy metals received by Galle Face Lake and West Lake during the wet season, but these levels increased significantly during the dry season in East Lake and Gangarama Lake.

The specific sources of heavy metals during these seasons include industrial activities and urban runoff. Industrial facilities around Beira Lake discharge effluents containing heavy metals such as lead, cadmium, and mercury. During the wet season, increased rainfall can wash these pollutants from industrial sites into the lake, raising HPI values. In the dry season, the reduced flow of water might limit the dispersion of industrial pollutants, leading to lower HPI values, despite ongoing industrial activities. Urban runoff is a significant source of heavy metals, especially during the wet season. Rainfall washes pollutants from roads, buildings, and other surfaces into the lake. This runoff often contains metals from vehicles, construction materials, and urban infrastructure. During the dry season, there is less rainfall to generate such runoff, resulting in lower contributions of heavy metals from this source.

• Principle Component Analysis between Indices

The results of the correlation analysis (PCA) during the dry season reveal interesting correlations among various indexes (

Table 2. Correlation analysis of indices in the dry season.

	HPI	WPI	TSI-TP	TSI-TN	TSI-SDD
HPI	1
WPI	0.0491	1
TSI-TP	−0.09737	0.50556	1
TSI-TN	−0.10311	0.49968	0.99786	1
TSI-SDD	−0.11893	−0.2341	0.26651	0.26875	1

). The HPI exhibits a weak positive correlation with the WPI, suggesting a slight relationship between heavy metal pollution and overall water pollution levels. Conversely, the WPI also shows weak positive correlations with both the HPI and various components of the TSI, indicating potential connections between water pollution and environmental factors.

The TSI components, including TP, TN, and SDD, exhibit negative correlations with each other, which is consistent with the expected inverse relationship as the trophic state increases. Moreover, TSI-TP and TSI-TN demonstrate moderate positive correlations with WPI, indicating that higher levels of phosphorous and nitrogen in the water—markers of poorer water quality—are associated with higher levels of overall water pollution. Similarly, TSI-SDD, representing a shallower Secchi disk depth and indicative of a higher trophic state, shows a weaker positive correlation with WPI compared to TSI-TP and TSI-TN.

The correlation results for the indexes during the wet season provide insights into their relationships (

Table 3. Correlation analysis of indices in the wet season.

	HPI	WPI	TSI-TP	TSI-TN	TSI-SDD
HPI	1
WPI	0.16977	1
TSI-TP	−0.21585	−0.25025	1
TSI-TN	−0.04402	0.11843	0.08829	1
TSI-SDD	0.0492	0.25124	0.15042	0.54098	1

). The HPI shows a weak positive correlation with the WPI, suggesting a slight relationship between heavy metal pollution and overall water pollution levels. Conversely, WPI exhibits a moderate negative correlation with TSI-TP, implying that higher levels of total phosphorus in the water (indicating poorer water quality) coincide with higher overall water pollution levels. TSI-TP demonstrates weak negative correlations with both HPI and WPI, suggesting associations between phosphorus levels, heavy metal pollution, and overall water pollution.

TSI-TN shows weak positive correlations with WPI and TSI-SDD, indicating potential relationships with overall water pollution and trophic state. TSI-SDD exhibits weak positive correlations with WPI, and a strong positive correlation with TSI-TN, suggesting a potential link between shallower Secchi disk depth (indicating higher trophic state), total nitrogen levels, and overall water pollution.

Comparing the PCA results between wet and dry seasons offers valuable insights into the dynamics of Beira Lake’s water quality (Figure 10). During the dry season (Figure 10a), the correlations between indexes reveal nuanced relationships. The HPI and WPI exhibit weak positive correlations, indicating a modest association between heavy metal pollution and overall water pollution levels in the lake. The TSI components, particularly TP and TN, show moderate positive correlations with WPI, suggesting that diminished water flow and increased pollutant concentrations during this season may lead to lower overall well-being and water quality in Beira Lake. These results suggest that during the dry season, Beira Lake may face challenges related to heavy metal pollution and reduced water quality, necessitating interventions to mitigate these issues.

Conversely, the wet season reveals different patterns in the PCA results (Figure 10b), reflecting the impacts of increased precipitation and nutrient runoff on Beira Lake’s water quality. While the correlations between HPI and WPI remain consistent, indicating a continued relationship between heavy metal pollution and overall water pollution levels, the correlations with the TSI components shift. WPI exhibits a moderate negative correlation with TSI-TP, implying that higher levels of TP in the water coincide with increased overall water pollution levels, likely due to enhanced nutrient runoff from surrounding areas. The TSI components, particularly TN and SDD, show positive correlations with WPI, suggesting that higher trophic states and increased nitrogen levels during the wet season may contribute to elevated water pollution in Beira Lake. These findings highlight the challenges posed by nutrient runoff and increased trophic state during the wet season, underscoring the importance of implementing measures to address nutrient pollution and maintain water quality in Beira Lake throughout the year.

3.3. Water Quality Comparison in 2016 and 2023, Based on Remote Sensing Data

3.3.1. Changes in Water Transparency

The Secchi disk depth was calculated based on the Sentinel 2 images (Z_SD) to quantify the transparency of Beira Lake, which has changed significantly during the past 8 years (Figure 11a). It can be seen that the water transparency has been improved, with the maximum Z_SD increasing from 0.26 m to 0.45 m. The water transparency of East Lake is better than that of either West Lake or Gangarama Lake. The maximum Z_SD is found in East Lake, and Z_SD remains below 0.5 m in West Lake and East Lake in both 2016 and 2023 (Figure 11a). For East Lake and Gangarama Lake, Z_SD increased in the center of the lake from 2016 to 2023, which means that the water in the middle portion of the lakes is more transparent. This phenomenon is not found in West Lake because of its narrow shape.

The ZSD values indicate better water transparency in East Lake compared to West Lake and Gangarama Lake. This spatial variability can be attributed to differences in land use, pollution sources, and hydrodynamic conditions between these lake segments. East Lake likely benefits from regular cleaning efforts and floating wetlands, resulting in reduced pollutant loads entering the water. In contrast, West Lake and Gangarama Lake are more urbanized, with higher levels of industrial and residential runoff contributing to poorer water quality and lower transparency.

This could result from recent green infrastructure projects, such as floating wetlands that reduce runoff pollution, and the natural recovery of aquatic vegetation that helps filter the water. Higher chlorophyll concentrations along the edges of East Lake, compared to the middle, indicate elevated nutrient inputs from runoff, leading to increased algal growth near the shoreline. This growth might contribute to the increased transparency in the lake’s center.

3.3.2. Changes in Trophic State

The Trophic State Index (TSI) is calculated based on Z_SD, referring to N, P, and other nutrient content in the water body and water ecosystem state. As shown in Figure 11b, TSI declines in general, ignoring a few small high-value regions. The minimum TSI decreases from 85.28 to 79.27, and the area with low and medium values increases. The medium and high values are narrowed down to a small area around the lake. As a whole, the TSI in Beira Lake maintains a moderate level and shows a general decreasing trend.

Measures implemented between 2016 and 2023 that could have led to this decrease likely include both nutrient control and environmental management practices. Specific actions may involve stricter regulations regarding industrial discharges, reducing nutrient runoff, and increasing public awareness campaigns promoting pollution reduction. Additionally, broader efforts, such as the inclusion of floating wetlands and urban planning initiatives that minimize impermeable surfaces and enhance natural water filtration, could contribute to the overall reduction in TSI values.

High TSI values are concentrated in smaller areas, especially near narrow sections of GL and WL areas. These areas have higher nutrient loads due to the presence of untreated sewage, industrial effluents, and urban runoff containing fertilizers and organic waste. High TSI values are primarily caused by urban runoff, industrial discharges, and inadequate sewage treatment. Urban runoff introduces significant amounts of nitrogen and phosphorus, industrial activities add pollutants and nutrients, and insufficient sewage treatment releases untreated sewage into the lake.

To reduce nutrient loads and improve water quality, targeted actions include upgrading sewage treatment plants, implementing green infrastructure like constructed wetlands and vegetative buffers, and improving stormwater management systems. Public awareness campaigns and stronger regulations with regular monitoring are also essential to address nutrient pollution and promote sustainable water management practices in Beira Lake.

3.3.3. Accuracy Assessment

In order to assess the accuracy of the SDD model using QAA, the water transparency from remote sensing data is compared with the data from the field measurements. From the remote sensing results, the outliers are eliminated to make the sample points obey the normal distribution, and 43 values are chosen from 2023 for comparison (Figure S6). The linear regression method is employed to study the correlation between remote sensing monitoring results and the field measurements (Figure 12). The degree of fitting R2 in a linear regression model is 0.670. This shows that their correlation is at a high level. Moreover, the root-mean-square error (RMSE), mean absolute deviation (MAD) and mean absolute percentage deviation (MAPD) remain at a low level, 0.049 m, 0.042 m, and 26.1%, respectively. The maximum and minimum differences are 0.048 m and 0.003 m. The ZSD values from the remote sensing data are proved to be identical to the in-situ field measurements. The water transparency results from remote sensing monitoring are basically accurate.

These remote sensing analyses are verified by the chlorophyll-a (Chl.a) data and the SDD maps. GL exhibited the highest Chl.a concentration (Figure S9), indicating a potentially higher nutrient load and increased algal growth. WL and GFL also showed elevated Chl.a levels, suggesting possible nutrient enrichment from urban and human activities. Although it showed a high nutrient level (nitrate-17.39 mg/L), EL shows the lowest Chl.a concentration, signifying a comparatively lower nutrient load or effective management practices and thus, a lower eutrophic condition. The SDD map in the wet season (Figure 2), created using in situ data, is almost identical to the remote sensing analysis maps. This proves that remote sensing data is very useful and easy to obtain when investigating water quality in large-scale water bodies.

3.4. Recommendation for Water Quality Improvement of Beira Lake

3.4.1. Comparison of Water Quality with Different Urban Lakes in Sri Lanka

The observed variations in water quality among the sampled lakes can be attributed to a combination of natural processes and anthropogenic activities. Beira Lake and St. Sebastian Canal demonstrate significantly higher COD concentrations compared to those of Kurunegala Lake (13.3–28.3 mg/L) and Diyawanna Lake (5.2–7.4 mg/L), indicating a substantial organic pollution load (Table S12). Previous studies in the Colombo Canal system revealed that domestic waste and industrial discharges were the reasons for the high organic and inorganic pollution in the St. Sebastian Canal [50]. Beira Lake’s relatively high pH (9.3 ± 0.2 in the dry season) may be influenced by alkaline discharges from surrounding urban areas. DO variations in Beira Lake, with higher levels during the dry season (11.96 ± 0.70 mg/L) and lower levels during the wet season (8.48 ± 0.32 mg/L), could be linked to temperature changes, organic matter decomposition, and algal blooms. Kandy Lake’s significant increase in DO (262%) in 2011 (Table S12) might be attributed to artificial aeration or natural aeration processes. The EC variations in Beira Lake (296 ± 4 µs/cm in the dry season, 375 ± 22 µs/cm in the wet season) may reflect changes in ion concentrations due to the urban rainwater runoff traveling through construction fields. St. Sebastian Canal’s wider EC range (0.101–0.577 dS/m) indicates potential salinity variations resulting from natural and anthropogenic inputs.

The elevated TP levels in Beira Lake during the wet season (5.84 ± 2.31 mg/L) suggest increased nutrient inputs, potentially from anthropogenic runoff, while Kurunegala Lake and Bolgoda Lake maintain relatively higher TP levels, possibly due to agricultural activities and urban runoff. Increased NO₃⁻ concentrations in Beira Lake during the wet season (20.61 ± 2.96 mg/L) (Table S12) may result from sewage discharges. Kandy Lake’s 1.6 mg/L of NO₃⁻, with a low PO₄³⁻ (0.03 mg/L) (Table S12), suggests potential nutrient input, but limited primary production due to low PO₄³⁻ [51]. Compared to other lakes, St. Sebastian Canal shows higher PO₄³⁻ levels (3 mg/L), and the pollution in the canal has already been identified as resulting from industrial effluent discharges, domestic waste dumping, and inflows from the polluted canals and lakes [50].

According to Table S12, heavy metal concentrations are significantly low in Beira Lake, when considering other studies in Sri Lanka. The most recent survey of the St. Sebastian Canal, which acts as the main inlet to East Lake in the Beira Lake network, has reported high amounts of Pb concentrations [50], but there was no significantly high Pb value recorded in East Lake in either the wet or dry seasons during the study periods (Table S12).

3.4.2. Key Challenges of Restoring Beira Lake

Restoring Beira Lake and similar urban water bodies involves addressing key challenges, including a comprehensive evaluation of water pollution, effective source control, and lake remediation strategies, based on both management and technology. Untreated sewage and urban runoff introduce contaminants such as heavy metals and nutrients, impacting water quality and aquatic ecosystems. Sedimentation, increased as the result of construction activities and improper land use, further harms aquatic life. Urban expansion complicates restoration efforts, altering water flow patterns and increasing pollution. Balancing urban development and conservation is crucial. Hydrological changes induced by urbanization challenge the lake’s water balance. Community engagement is essential for successful restoration, requiring awareness and collaboration. Financial and infrastructure demands, influenced by climate change, underscore the need for secure funding and resilient solutions. Navigating regulatory issues, including clear regulations and enforcement mechanisms, is vital to prevent further degradation and ensure compliance with restoration efforts.

3.4.3. Recommendations for Improving the Water Quality of Beira Lake

To enhance Beira Lake’s water quality improvement, a comprehensive strategy must enforce strict pollution controls for sewage collection and treatment, limiting the influx of pollutants. Regular dredging is essential for the sediment management of Beira Lake. Urban planning should bring together regulations to balance development and preserve critical lake areas. Incorporating green infrastructure solutions, e.g., combined sewer overflows (CSOs), pollution control, low-impact development, sponge cities, and floating wetlands, can manage stormwater runoff effectively. Hydrological restoration requires evaluating and modifying drainage systems. Engaging local communities through awareness campaigns is crucial, emphasizing their role in monitoring and conservation. Infrastructure investment, including municipal wastewater collection and treatment facilities, is vital for the water pollution control of Beira Lake. A climate-resilient approach and strengthening regulatory frameworks ensure the long-term effectiveness of restoring Beira Lake. Collaboration among government agencies, communities, environmental organizations, and businesses, combining technical expertise, policy interventions, and community participation, can create a comprehensive and sustainable approach, promoting ecological resilience and community well-being around Beira Lake.

4. Conclusions

The comprehensive assessment of Beira Lake’s water quality revealed significant spatial and temporal variations. The water quality index analysis categorized Beira Lake as reflecting highly polluted eutrophic and hypereutrophic conditions in both seasons, particularly during the wet season. The HPI suggested elevated heavy metal concentrations, especially during the wet season. Conversely, the wet season brings different challenges regarding elevated trophic states and water pollution levels.

Notably, the lakes exhibited higher levels of COD, with Gangarama Lake reaching up to 306.40 mg/L during the wet and 373.41 mg/L during the dry seasons. The higher concentrations of nutrient ions and heavy metals were observed during the wet season. Nitrate levels exceeded the tolerance limit of 10 mg/L, reaching up to 20.84 mg/L in East Lake during the wet season. TP levels surpassed the tolerance limit of 0.4 mg/L in all sampling points during the wet season, with the highest concentration at 5.84 mg/L. Heavy metal concentrations were generally below the maximum permissible levels, except for Hg, which exceeded 2.25 µg/L in most of the Galle Face and West Lake areas during the dry season.

An analysis of Beira Lake’s water quality, based on remote sensing data, reveals significant improvements from 2016 to 2023. The Z_SD values indicate a noteworthy increase in water transparency, with the maximum Z_SD rising from 0.26 m to 0.45 m. East Lake consistently surpasses West Lake and Gangarama Lake in regards to transparency, while Z_SD remains below 1 m in the latter two lakes. Concurrently, the TSI, based on Z_SD, N, P, and other nutrient content, reveals a general decline, emphasizing an overall moderation in Beira Lake’s trophic state.

The study identified key challenges in restoring Beira Lake, including untreated sewage, urban runoff, sedimentation, and hydrological changes due to urban expansion. A comprehensive approach, crucial for the successful restoration of Beira Lake, is recommended, including strict pollution controls, regular dredging, urban planning, green infrastructure implementation, hydrological restoration, community engagement, and infrastructure investment.