Improving monthly streamflow prediction in alpine regions: integrating HBV model with Bayesian neural network

Ren, Weiwei; Yang, Tao; Huang, Ching‐Sheng; Xu, Chong‐Yu; Shao, Quanxi

doi:10.1007/s00477-018-1553-x

Cited by 40 publications

(26 citation statements)

References 76 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In general, hydrological models have been proved effective for understanding hydrological process and providing physical information about the catchment. However, hydrological models rely on assumptions of mathematical equations that may fail to model the actual runoff generation process, especially when the forcing data are of low quality (e.g., sparse meteorological observations or reanalysis data) (Humphrey et al ., 2016; Ren et al ., 2018).…”

Section: Introductionmentioning

confidence: 99%

An integrated multi‐GCMs Bayesian‐neural‐network hydrological analysis method for quantifying climate change impact on runoff of the Amu Darya River basin

Liu

et al. 2021

Intl Journal of Climatology

View full text Add to dashboard Cite

As one of the most pressing issues in the world, climate change has already caused evident impacts on natural and human systems (e.g., hydrological cycle, eco‐environment and socio‐economy) in recent decades. In this study, an integrated multi‐GCMs Bayesian‐neural‐network hydrological analysis (MBHA) method is developed for quantifying climate change impacts on runoff. MBHA incorporates multiple global climate models (multi‐GCMs), hydrological model (HBV‐light), and Bayesian neural network (BNN) within a general framework. MBHA can provide the reliable prediction for runoff as well as reflect the impact of climate change on data scarcity catchments. MBHA is applied to the Amu Darya River basin in Central Asia. Climate data are derived from multiple GCMs (i.e., GFDL‐ESM2G, HadGEM2‐AO and NorESM1‐M) under RCP4.5 and RCP8.5. Several findings can be summarized: (1) during 2021–2100, both precipitation and temperature would increase, with more precipitation falling as rain instead of snow; (2) by 2100, glacier areas are predicted to reduce by 62.3% (RCP4.5) and 71.9% (RCP8.5); (3) under RCP8.5, monthly runoff would increase by 11.2% in 2021–2060 and reduce by 5.0% in 2061–2100; this is because the glaciers would rapidly disappear with the rising temperature after 2060. The findings suggest that the shrinked glacier and the reduced runoff threaten the water availability especially in summer seasons as well as affect the agricultural irrigation in the downstream of the Amu Darya River.

show abstract

Section: Introductionmentioning

confidence: 99%

An integrated multi‐GCMs Bayesian‐neural‐network hydrological analysis method for quantifying climate change impact on runoff of the Amu Darya River basin

Liu

et al. 2021

Intl Journal of Climatology

View full text Add to dashboard Cite

show abstract

“…In recent years, great effort has been made by scientists to combine machine learning hydrological models with conceptual models in which some hydrological processes are empirically described (Robertson et al 2013;Humphrey et al 2016;Ren et al 2018;Kumanlioglu & Fistikoglu 2019). For instance, Ren et al (2018) combined the HBV (Hydrologiska Byråns Vattenbalansavdelning) (Bergström 1992) model with artificial Bayesian neural networks, which enabled a better assessment of the uncertainties of hydrological predictions simulated on a monthly scale. Humphrey et al (2016) combined artificial Bayesian neural networks with the intermediate products of the GR4 J (Génie Rural à 4 paramètres Journalier) (Perrin et al 2003) model to resolve runoff routing and estimate streamflow on a daily scale.…”

mentioning

confidence: 99%

Untangling hybrid hydrological models with explainable artificial intelligence

2021

View full text Add to dashboard Cite

Hydrological models are valuable tools for developing streamflow predictions in unmonitored catchments to increase our understanding of hydrological processes. A recent effort has been made in the development of hybrid (conceptual/machine learning) models that can preserve some of the hydrological processes represented by conceptual models and can improve streamflow predictions. However, these studies have not explored how the data-driven component of hybrid models resolved runoff routing. In this study, explainable artificial intelligence (XAI) techniques are used to turn a ‘black-box’ model into a ‘glass box’ model. The hybrid models reduced the root-mean-square error of the simulated streamflow values by approximately 27, 50, and 24% for stations 17120000, 27380000, and 33680000, respectively, relative to the traditional method. XAI techniques helped unveil the importance of accounting for soil moisture in hydrological models. Differing from purely data-driven hydrological models, the inclusion of the production storage in the proposed hybrid model, which is responsible for estimating the water balance, reduced the short- and long-term dependencies of input variables for streamflow prediction. In addition, soil moisture controlled water percolation, which was the main predictor of streamflow. This finding is because soil moisture controls the underlying mechanisms of groundwater flow into river streams.

show abstract

“…Tomé esta determinación para combinar la rapidez y eficiencia del lenguaje compilado C++, con el potente ambiente interactivo que ofrece R, un lenguaje con una enorme cantidad de paquetes y funciones que permiten que todo el flujo de trabajo hidrológico (desde la descarga de datos hasta el análisis de los resultados) se pueda concretar en un solo ambiente (Slater et al, 2019). Elegí seguir ampliando este modelo (HBV) por dos motivos: primero, ha sido usado de forma continua para estudios hidrológicos en zonas de montaña alrededor del mundo y, segundo, porque requiere relativamente pocos datos de entrada (temperatura del aire, precipitación, y evapotranspiración potencial), condición que resulta apropiada en zonas con escasez de información meteorológica como los Andes del Sur (el lector podrá encontrar algunas aplicaciones para otras regiones montañosas del planeta en Finger et al, 2015;Ren et al, 2018;Seibert et al, 2018;Stahl et al, 2008). hydroToolkit es un paquete para leer, graficar, manipular y procesar series hidro-meteorológicas provenientes de Argentina y Chile dentro del lenguaje R (R Core Team, 2020).…”

Section: Desarrollounclassified

hydroToolkit y HBV.IANIGLA: dos paquetes en R para estudios hidrológicos en los Andes

Toum

Masiokas²,

Villalba³

2020

AJEA

View full text Add to dashboard Cite

En muchas regiones del planeta las montañas son las principales generadoras de agua, elemento que constituye el componente esencial para la conservación de diversos ecosistemas y el desarrollo socioeconómico de las poblaciones aledañas (Mariano H. Masiokas et al., 2006; Viviroli & Weingartner, 2004). En el oeste argentino, resulta indispensable para el regadío de cultivos, para el suministro de agua tanto municipal como industrial, para la producción de energía hidroeléctrica y para el equilibrio medioambiental de los ecosistemas que se encuentran en las cuencas. Sin embargo, a pesar de que la Argentina cuenta con más de 3000 km de cordillera y numerosas poblaciones que dependen directa o indirectamente de los recursos hídricos generados en las montañas, en la actualidad existen muy pocos avances y estudios centrados en la hidro-climatología andina (M. H. Masiokas et al., 2020; Villalba et al., 2003). Para abordar este tipo de estudios a lo largo de la cordillera, desarrollé dos paquetes que ya están disponibles en el repositorio oficial del lenguaje R: hydroToolkit y HBV.IANIGLA. El primero de ellos es un software diseñado para el manejo de las bases de datos hidro-meteorológicas tanto de Argentina como de Chile, y el segundo es un modelo hidrológico que permite, entre otras cosas, simular el balance de masa glaciar, aspecto fundamental para estudios hidrológicos en los Andes.

show abstract

Improving monthly streamflow prediction in alpine regions: integrating HBV model with Bayesian neural network

Cited by 40 publications

References 76 publications

An integrated multi‐GCMs Bayesian‐neural‐network hydrological analysis method for quantifying climate change impact on runoff of the Amu Darya River basin

An integrated multi‐GCMs Bayesian‐neural‐network hydrological analysis method for quantifying climate change impact on runoff of the Amu Darya River basin

Untangling hybrid hydrological models with explainable artificial intelligence

hydroToolkit y HBV.IANIGLA: dos paquetes en R para estudios hidrológicos en los Andes

Contact Info

Product

Resources

About