Funded by: Afeyan Foundation
PI: Habet Madoyan
Researchers: Vahe Movsisyan, Anna Zatikyan, Gayane Shahnazaryan
This project aims to build a multidisciplinary research team capable of analyzing environmental and water quality datasets collected by the Armenian Hydrometeorological Center (HayHydromet). Initially, the work focused on water chemistry data from Lake Sevan and was later expanded to cover the broader river basin across Armenia. The overarching goal is to strengthen HayHydromet’s capacity for monitoring lake and river water quality, introduce more efficient and AI-driven monitoring approaches, and establish a sustainable research team capable of securing and executing future projects.
Understanding spatial and temporal patterns in river water chemistry is essential for designing effective monitoring systems in lake basins. This study analyzes long-term hydrochemical data (2010–2024) from 16 monitoring sites across 10 major tributaries of the Lake Sevan basin. The dataset includes major ions, nutrients, heavy and trace metals, and key physicochemical parameters.
Using data science techniques, the study identifies dominant chemical gradients, classifies monitoring sites, and evaluates temporal stability. Principal Component Analysis reveals that calcium, magnesium, sodium, and nitrate are the main drivers of spatial variability, reflecting both geological factors and downstream anthropogenic influences. Temporal variability is primarily associated with zinc, iron, and barium.
Hierarchical clustering identifies four distinct hydrochemical regimes: mining-impacted rivers, upstream sites with predominantly natural composition, river mouths affected by wastewater and agricultural runoff, and sites with mixed anthropogenic and geogenic influences. Trend analysis shows that 84% of site-parameter combinations exhibit no significant long-term trends, indicating overall hydrochemical stability. However, localized increases in iron, silicon, and barium, along with decreasing zinc levels, highlight site-specific environmental pressures.
This integrated spatiotemporal analysis provides a foundation for optimizing monitoring networks and supports evidence-based strategies for protecting the ecological integrity of Lake Sevan.