Evaluación de la disponibilidad hídrica en la cuenca del Río Maule, considerando la evolución espacio temporal del uso de suelo.
No Thumbnail Available
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
La gestión eficiente de los recursos hídricos constituye un desafío prioritario frente a la disminución de la disponibilidad de agua y el incremento sostenido de su demanda. La modelación hidrológica se presenta como una herramienta clave para estimar y analizar caudales futuros bajo distintos escenarios, aportando insumos relevantes para la toma de decisiones. En este contexto, el presente estudio evalúa la disponibilidad hídrica en la cuenca del río Maule (Chile), considerando la evolución espacio temporal del uso de suelo, mediante la aplicación del modelo hidrológico semi-distribuido SWAT, calibrado y validado con la herramienta SWAT-CUP Premium.
La investigación se desarrolló a escala diaria, considerando un periodo de calibración comprendido entre 2010 y 2016, y un periodo de validación entre 2017 y 2019. Los parámetros físicos e hidráulicos de los suelos provinieron de la base de datos CLSoilMaps y fueron sometidos a análisis de correlación, componentes principales (PCA) y métodos de agrupamiento (K-Means y Silhouette) para optimizar su representación espacial. Se utilizaron mapas de uso de suelo para los años 1998, 2010 y 2020 y luego se modeló la evolución del uso bajo dos enfoques: un método estacionario, que mantiene fijos los datos en los años sin información, y un método dinámico, que interpola la información disponible para representar cambios interanuales de los usos de suelo.
El desempeño del modelo fue evaluado mediante indicadores estadísticos como NSE, KGE, R² y PBIAS, evidenciando resultados moderados en términos de ajuste, pero con mejoras consistentes al emplear el método dinámico respecto al estacionario. Si bien la magnitud de la mejora fue acotada, la representación espacial y temporal de los usos de suelo, junto con la calidad de los tipos de suelos generados, permitió una caracterización más realista de los procesos hidrológicos en la cuenca.
Los hallazgos obtenidos resaltan la importancia de incorporar variabilidad temporal en la información de usos de suelo y de utilizar bases de datos de suelos robustas para fortalecer la modelación hidrológica. Aun con desempeños estadísticos no sobresalientes, este trabajo aporta insumos relevantes para la planificación y gestión de los recursos hídricos en la cuenca del Maule, constituyendo una referencia para futuros estudios orientados a mejorar la precisión y aplicabilidad de modelos hidrológicos bajo escenarios de cambio de uso de suelo.
The efficient management of water resources constitutes a priority challenge in the face of declining water availability and the sustained increase in demand. Hydrological modeling emerges as a key tool to estimate and analyze future streamflows under different scenarios, providing valuable inputs for decision-making. In this context, the present study evaluates water availability in the Maule River basin (Chile), considering the spatio-temporal evolution of land use through the application of the semi-distributed hydrological model SWAT, calibrated and validated using the SWAT-CUP Premium tool. The research was conducted on a daily scale, considering a calibration period between 2010 and 2016, and a validation period between 2017 and 2019. Soil physical and hydraulic parameters were obtained from the CLSoilMaps database and subjected to correlation analysis, principal component analysis (PCA), and clustering methods (K-Means and Silhouette) to optimize their spatial representation. Land use maps for the years 1998, 2010, and 2020 were employed, and land use evolution was then modeled under two approaches: a stationary method, which keeps data fixed in years without information, and a dynamic method, which interpolates available data to represent interannual land use changes. Model performance was evaluated using statistical indicators such as NSE, KGE, R², and PBIAS, showing moderate results in terms of fit, but consistent improvements when using the dynamic method compared to the stationary one. Although the magnitude of improvement was limited, the spatial and temporal representation of land uses, along with the quality of the generated soil types, enabled a more realistic characterization of hydrological processes in the basin. The findings highlight the importance of incorporating temporal variability in land use information and employing robust soil databases to strengthen hydrological modeling. Even with non-outstanding statistical performances, this study provides valuable inputs for water resource planning and management in the Maule River basin, serving as a reference for future studies aimed at improving the accuracy and applicability of hydrological models under land use change scenarios.
The efficient management of water resources constitutes a priority challenge in the face of declining water availability and the sustained increase in demand. Hydrological modeling emerges as a key tool to estimate and analyze future streamflows under different scenarios, providing valuable inputs for decision-making. In this context, the present study evaluates water availability in the Maule River basin (Chile), considering the spatio-temporal evolution of land use through the application of the semi-distributed hydrological model SWAT, calibrated and validated using the SWAT-CUP Premium tool. The research was conducted on a daily scale, considering a calibration period between 2010 and 2016, and a validation period between 2017 and 2019. Soil physical and hydraulic parameters were obtained from the CLSoilMaps database and subjected to correlation analysis, principal component analysis (PCA), and clustering methods (K-Means and Silhouette) to optimize their spatial representation. Land use maps for the years 1998, 2010, and 2020 were employed, and land use evolution was then modeled under two approaches: a stationary method, which keeps data fixed in years without information, and a dynamic method, which interpolates available data to represent interannual land use changes. Model performance was evaluated using statistical indicators such as NSE, KGE, R², and PBIAS, showing moderate results in terms of fit, but consistent improvements when using the dynamic method compared to the stationary one. Although the magnitude of improvement was limited, the spatial and temporal representation of land uses, along with the quality of the generated soil types, enabled a more realistic characterization of hydrological processes in the basin. The findings highlight the importance of incorporating temporal variability in land use information and employing robust soil databases to strengthen hydrological modeling. Even with non-outstanding statistical performances, this study provides valuable inputs for water resource planning and management in the Maule River basin, serving as a reference for future studies aimed at improving the accuracy and applicability of hydrological models under land use change scenarios.
Description
Tesis presentada para optar al título de Ingeniero/a Civil.
Keywords
Recursos hidrológicos, Modelos hidrológicos, Cuencas hidrográficas, Uso de la tierra