Monitoring and Analyzing the Impacts of Climatic Variables and Human Activities on Water Extent Changes of Kajaki Dam Reservoir Using Advanced Satellite Spectral Indices

Introduction: Arid and semi-arid regions are highly vulnerable to both climatic variability and human activities, making continuous and accurate monitoring of reservoir water levels an indispensable task. This study aimed to quantitatively separate and evaluate the relative contributions of climatic and anthropogenic factors to water-level changes at Kajaki Dam, Afghanistan, and to assess the downstream impacts on the Helmand River in southern Iran over the past decade (2013–2023). Kajaki Dam, with a storage capacity of 1.4 billion cubic meters, plays a crucial role in meeting water demands and sustaining the Helmand River ecosystem.

Methods: This research employed multi-sensor satellite imagery, including 45 Landsat-8 OLI/TIRS and 62 Sentinel-2 MSI images. Images were temporally synchronized within a ±15-day window, and only images with less than 10% cloud cover were selected. Preprocessing steps, including atmospheric correction and cloud removal, were performed using SREM and Sen2Cor algorithms. Eleven spectral indices were computed using customized equations and cloud-based processing on the Google Earth Engine (GEE) platform to enable high-precision analysis. Water extent was extracted using Otsu thresholding and decision tree classification with 300 training samples. Evaluation metrics included RMSE, R², Nash–Sutcliffe efficiency (NS), MAE, and K-Fold cross-validation. Climatic data were sourced from ERA5, and in situ measurements were collected at 50 ground control points, including water–land boundaries and shallow and deep water areas, using GPS with ±3 m accuracy. Time-series analyses were employed to assess long-term trends in water surface dynamics and to quantitatively separate the impacts of climatic factors (precipitation, temperature, and actual evapotranspiration) and human activities (agricultural expansion, new dam constructions, and reservoir operation) on water level changes.

Results an con colousion: Spectral index analysis revealed that the S2WI index provided the highest accuracy in monitoring water extent, with R² = 0.96, RMSE = 1.46 km², and NS = 0.96. This index effectively distinguished water from other land cover types while reducing errors caused by vegetation and soil. Comparative analysis of Landsat-8 and Sentinel-2 sensors indicated that both were capable of monitoring water surface changes; however, Sentinel-2 outperformed Landsat-8 due to its higher spatial resolution and shorter revisit interval, particularly for capturing short-term dynamics. Multi-temporal comparative studies confirmed that indices such as S2WI and MNDWI can accurately track changes in water extent. The water surface area of Kajaki Dam declined from 32.48 km² in 2014 to 22.18 km² in 2021, representing a 31.7% total reduction and an average annual decrease of approximately 4.5%.

Multivariate regression analysis indicated that the combined effects of climatic and anthropogenic factors drove reductions in water extent and inflow. Approximately 60% of the decrease was attributed to climatic variables, including an annual precipitation averaging 196.8 mm with high interannual variability (CV = 42%) and a statistically non-significant upward trend of 3.3 mm/year; mean land surface temperature of 22.63°C with notable fluctuations and a slight decreasing trend; and actual evapotranspiration averaging 1,432 mm/year, strongly correlated with precipitation (r > 0.85), indicating a water-limited regime. The remaining 40% was associated with human activities, including the expansion of irrigated agriculture, the construction of multiple upstream control structures, and alterations in reservoir operation, demonstrating that water inflow reductions cannot be attributed solely to climatic changes. This ongoing decline is causing severe economic, environmental (destruction of the Hamoun Wetland) and social (migration) impacts.

The integration of multi-sensor satellite data with cloud-based processing on GEE enabled precise, repeatable monitoring of water dynamics. It provided a quantitative framework for separating the contributions of climatic and anthropogenic drivers. These findings offer a robust basis for future water resource management in the Helmand River basin, supporting adaptive strategies to mitigate the impacts of both environmental variability and human interventions.