Abstract This study provides a comprehensive assessment of chlorophyll-a (Chl-a) dynamics in Chitgar Lake, Tehran, by integrating in-situ measurements (Jun-2013–Apr-2025) with Sentinel-2 and Landsat-8 imagery. Several Chl-a retrieval algorithms were evaluated, including Sentinel-2 red-edge indices (RESR, TREI, CRER) and Landsat-8 indices (CBAR, CGR, SRER, ECI). Sentinel-2 significantly outperformed Landsat-8, with the TREI index achieving the highest accuracy (RMSE = 1.21 mg/m³, MAE = 0.91 mg/m³, R² = 0.74). Landsat-8 exhibited moderate explanatory power (R² = 0.63–0.65) but higher error (~ 2.42–2.44 mg/m³), with ECI performing best. Seasonal validation revealed fall as the most reliable period for Sentinel-2 retrieval (R² = 0.59–0.67), while winter showed the lowest error (RMSE = 0.66–0.68 mg/m³) but limited explanatory power. Landsat-8 models performed better in spring and fall. Field observations confirmed elevated summer Chl-a (mean: 2.93 mg/m³; max: 3.53 mg/m³) coinciding with higher temperatures (up to 30.8 °C) and reduced dissolved oxygen, signaling eutrophic conditions. Mean total phosphorus (0.03 mg/L) remained near eutrophication thresholds. Spatial distribution maps from both satellites aligned with ground data, with Sentinel-2 offering superior resolution and sensitivity to fine-scale variability, while Landsat-8 enabled consistent long-term trend analysis. These findings validate Sentinel-2 red-edge bands as highly effective for Chl-a monitoring in urban lakes and endorse Landsat-8 for historical assessments. The results underscore the necessity of local algorithm calibration in optically complex waters and support the integration of multi-sensor satellite data with in-situ networks for improved eutrophication management. Highlights <UnorderedList Mark="Bullet"> <ItemContent> Long-term (Jun-2013–Apr-2025) assessment of Chl-a in Chitgar Lake using both field and satellite data. </ItemContent> <ItemContent> Summer Chl-a peaks coincide with warmer temperatures and low dissolved oxygen, indicating eutrophic stress. </ItemContent> <ItemContent> Seasonal validation shows fall is most accurate for Sentinel-2; winter has low error but weak explanatory power. </ItemContent> <ItemContent> Multi-sensor fusion of Sentinel-2 and Landsat-8 enables both high-resolution monitoring and long-term trend analysis. </ItemContent> </UnorderedList>

Integrating Sentinel-2 and Landsat-8 Remote Sensing with In-Situ Data for Chlorophyll-a Estimation in an Urban Lake: Spatiotemporal Assessment of Chitgar Lake, Tehran

期刊论文
发表日期：

2026年7月10日

查看全文

Javad Bayat,
Maryam Rouhi Kerigh,
Mehdi Karami,
HamidReza Heidari

摘要

Abstract

This study provides a comprehensive assessment of chlorophyll-a (Chl-a) dynamics in Chitgar Lake, Tehran, by integrating in-situ measurements (Jun-2013–Apr-2025) with Sentinel-2 and Landsat-8 imagery. Several Chl-a retrieval algorithms were evaluated, including Sentinel-2 red-edge indices (RESR, TREI, CRER) and Landsat-8 indices (CBAR, CGR, SRER, ECI). Sentinel-2 significantly outperformed Landsat-8, with the TREI index achieving the highest accuracy (RMSE = 1.21 mg/m³, MAE = 0.91 mg/m³, R² = 0.74). Landsat-8 exhibited moderate explanatory power (R² = 0.63–0.65) but higher error (~ 2.42–2.44 mg/m³), with ECI performing best. Seasonal validation revealed fall as the most reliable period for Sentinel-2 retrieval (R² = 0.59–0.67), while winter showed the lowest error (RMSE = 0.66–0.68 mg/m³) but limited explanatory power. Landsat-8 models performed better in spring and fall. Field observations confirmed elevated summer Chl-a (mean: 2.93 mg/m³; max: 3.53 mg/m³) coinciding with higher temperatures (up to 30.8 °C) and reduced dissolved oxygen, signaling eutrophic conditions. Mean total phosphorus (0.03 mg/L) remained near eutrophication thresholds. Spatial distribution maps from both satellites aligned with ground data, with Sentinel-2 offering superior resolution and sensitivity to fine-scale variability, while Landsat-8 enabled consistent long-term trend analysis. These findings validate Sentinel-2 red-edge bands as highly effective for Chl-a monitoring in urban lakes and endorse Landsat-8 for historical assessments. The results underscore the necessity of local algorithm calibration in optically complex waters and support the integration of multi-sensor satellite data with in-situ networks for improved eutrophication management.

Highlights

Long-term (Jun-2013–Apr-2025) assessment of Chl-a in Chitgar Lake using both field and satellite data.

Summer Chl-a peaks coincide with warmer temperatures and low dissolved oxygen, indicating eutrophic stress.

Seasonal validation shows fall is most accurate for Sentinel-2; winter has low error but weak explanatory power.

Multi-sensor fusion of Sentinel-2 and Landsat-8 enables both high-resolution monitoring and long-term trend analysis.