Multi-Indicator Satellite-Based Monitoring of Water Quality Dynamics in Freshwater Reservoirs

Technical Summary: Multi-Indicator Satellite-Based Monitoring of Water Quality Dynamics in Freshwater Reservoirs

Problem Statement
Surface water reservoirs are critical sources of drinking water, yet their quality is increasingly vulnerable to eutrophication, sediment transport, and organic matter accumulation driven by climate variability and land-use changes. Traditional in-situ monitoring, while accurate, is constrained by discrete sampling locations, limited frequency, and high operational costs. These limitations often result in sparse datasets that fail to capture spatial gradients or short-duration events like algal blooms and sediment pulses. While satellite remote sensing offers a solution for synoptic observation, applications in data-scarce regions like Pakistan often lack integrated frameworks that combine persistent water delineation with the simultaneous assessment of multiple optical water quality indicators. There is a specific need to evaluate how multiple indicators collectively characterize contrasting reservoir conditions and seasonal behaviors within a repeatable, cloud-based framework.

Methodology
The study implemented a unified, cloud-based framework using Google Earth Engine (GEE) and Sentinel-2 MultiSpectral Instrument (MSI) imagery to monitor the Khanpur and Simly reservoirs in Pakistan during 2023–2024. The workflow consisted of four primary components:

1. Data Acquisition and Preprocessing: Sentinel-2 Level-2A surface reflectance data were filtered for cloud cover (<10%) to ensure temporal consistency.
2. Surface Water Delineation: Persistent water extents were extracted using the JRC Global Surface Water dataset. A 40% occurrence threshold was applied to generate binary masks, ensuring that water quality parameters were calculated only over stable water surfaces to minimize shoreline contamination.
3. Indicator Retrieval: Five optical indicators were derived using empirical algorithms adapted for Sentinel-2 bands:
   • Chlorophyll-a: Estimated via a red-to-near-infrared band ratio (Gitelson et al., 2008; Pahlevan et al., 2017).
   • Turbidity and Total Suspended Solids (TSS): Calculated using red-band reflectance formulations (Nechad et al., 2010; Dogliotti et al., 2015).
   • Colored Dissolved Organic Matter (CDOM): Estimated using an exponential green-to-red spectral band ratio (Chen et al., 2017).
   • Floating Algae Index (FAI): Calculated using a baseline method involving NIR, Red, and SWIR bands (Hu, 2009) to detect surface algal accumulations.
4. Analysis: Monthly mean composites were generated for 2023 and 2024. Data were aggregated into four seasonal periods (winter, pre-monsoon, monsoon, post-monsoon). To facilitate comparison across indicators with different units, standardized seasonal responses (z-scores) were calculated. Results were interpreted qualitatively against field-based reports from the Pakistan Council of Research in Water Resources (PCRWR).

Key Results
The study revealed distinct spatial, seasonal, and interannual dynamics between the two reservoirs:

• Khanpur Reservoir: Characterized by higher biological productivity and suspended matter. It exhibited consistently higher chlorophyll-a (annual means ~294–297 mg m⁻³) and turbidity (annual means >380 FNU) compared to Simly. Khanpur showed elevated responses during pre-monsoon and monsoon periods, with higher TSS variability and intermittent positive FAI values, suggesting transient floating algal or vegetation features. Spatial analysis indicated that elevated values were concentrated near tributary inflows and shallow shorelines.
• Simly Reservoir: Exhibited a clearer seasonal reversal and higher CDOM concentrations (annual mean ~3.0). Winter conditions were dominated by turbidity, TSS, and CDOM, while monsoon conditions were characterized by enhanced biological indicators (chlorophyll-a peaking >300 mg m⁻³ in late summer). FAI values remained consistently negative, indicating an absence of persistent large-scale floating algal blooms.
• Seasonal Patterns: Standardized multi-indicator profiles highlighted that Simly has a pronounced shift from a winter regime of suspended matter/CDOM to a monsoon regime of biological activity. In contrast, Khanpur maintained relatively elevated chlorophyll-a and suspended-matter responses throughout the pre-monsoon and monsoon periods with less pronounced seasonal separation.
• Interannual Consistency: Both reservoirs showed remarkable similarity in annual curves between 2023 and 2024, suggesting that seasonal processes exert a stronger influence on water quality than interannual variability over the study period.

Key Contributions
The paper's primary contribution lies in the integrated multi-indicator assessment rather than the development of new retrieval algorithms. Specifically, the study:

• Demonstrates the utility of combining persistent surface-water delineation with simultaneous spatial and temporal assessment of five optical indicators.
• Provides a comparative analysis of two contrasting drinking-water reservoirs within a common, repeatable cloud-based framework.
• Highlights the value of evaluating complementary optical indicators jointly to distinguish reservoir-specific regimes (e.g., sediment/productivity-driven vs. CDOM-rich) that might be missed by single-parameter proxies.
• Offers an operationally repeatable workflow suitable for data-scarce regions where systematic spatial water-quality observations are limited.

Significance and Claims
The authors claim that this integrated framework successfully distinguishes contrasting seasonal reservoir regimes and demonstrates the utility of satellite observations for complementing conventional drinking-water monitoring. The study emphasizes that while satellite-derived parameters cannot replace laboratory-based measurements, they provide a spatially continuous and temporally consistent perspective essential for early warning and adaptive response.

The paper maintains a modest stance regarding its claims:

• The derived values are interpreted as satellite-derived indicators of reservoir condition rather than direct substitutes for laboratory-measured parameters.
• Quantitative validation was not performed due to a lack of concurrent field measurements; instead, the results were qualitatively contextualized using existing PCRWR reports.
• The study acknowledges limitations, noting that the empirical algorithms were applied primarily for relative assessment and that future work requires synchronized field observations for local calibration and validation.

Ultimately, the study positions satellite-based monitoring as a scalable, cost-effective pathway to extend monitoring coverage across space and time, supporting routine water quality assessment and informed resource management in regions with limited field sampling infrastructure.