Abstract

In this study, multispectral images were used to detect toxic blooms in Villarrica Lake in Chile, using a time series of water quality data from 1989 to 2024, based on the extraction of spectral information from Landsat 8 and 9 satellite imagery. To explore the predictive capacity of these variables, we constructed 255 multiple linear regression models using different combinations of spectral bands and indices as independent variables, with phycocyanin concentration as the dependent variable. The most effective model, selected through a stepwise regression procedure, incorporated seven statistically significant predictors (p < 0.05) and took the following form: FCA = N/G + NDVI + B + GNDVI + EVI + SABI + CCI. This model achieved a strong fit to the validation data, with an R2 of 0.85 and an RMSE of 0.10 ?g/L, indicating high explanatory power and relatively low error in phycocyanin estimation. When applied to the complete weekly time series of satellite observations, the model successfully captured both seasonal dynamics and interannual variability in phycocyanin concentrations (R2 = 0.92; RMSE = 0.05 ?g/L). These results demonstrate the robustness and practical utility for long-term monitoring of harmful algal blooms in Lake Villarrica. © © 2025 Rodríguez-López, Usta, Duran-Llacer, Alvarez, Bourrel, Frappart and Urrutia.