Skip to main content
ARS Home » Plains Area » El Reno, Oklahoma » Grazinglands Research Laboratory » Agroclimate and Natural Resources Research » Research » Publications at this Location » Publication #305053

Research Project: ADAPTING SOIL AND WATER CONSERVATION TO MEET THE CHALLENGES OF A CHANGING CLIMATE

Location: Agroclimate and Natural Resources Research

Title: Hydrological modeling using a multi-site stochastic weather generator

Author
item Chen, Jie - University Of Quebec
item Brissette, Francois - University Of Quebec
item Zhang, Xunchang

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/6/2015
Publication Date: 8/19/2015
Citation: Chen, J., Brissette, F., Zhang, X.J. 2015. Hydrological modeling using a multi-site stochastic weather generator. Journal of Hydrology. doi: 10.1061/(ASCE)HE.1943-5584.0001288.

Interpretive Summary: Weather data at several locations over a large watershed is usually required, especially when using spatially distributed computer models for hydrological simulations. In many applications, spatially correlated weather data can be provided by a multi-site stochastic weather generator which considers the spatial correlation of weather variables. Prior to using a multi-site weather generator for hydrological modeling, its ability to adequately represent the proper hydrological response needs to be assessed. This study assesses the effectiveness of a new multi-site weather generator for hydrological modeling over a Canadian watershed in the Province of Quebec. In this work, two weather generators (a multi-site vs. single site) and two hydrological computer models (one spatially lumped vs distributed) were compared for the generated weather series and simulated surface stream flows. The results show that the mean snowmelt peak discharge is accurately represented by the Multisite-generated precipitation and temperature, while it is considerably underestimated by the single site generated data. In particular, Multisite generator shows a significant advantage in representing the monthly stream flow variability which is underestimated by the data obtained from the single-site weather generator. A flood frequency analysis indicates that Multisite data consistently perform better than single site data at representing spring high flow and summer-autumn low flow. Additionally, the hydrological simulation of the physically-based distributed model is better than that of the conceptually lumped model for the studied watershed. Hence, a multi-site weather generator coupled with a physically-based distributed model is recommended for all studies requiring the use of randomly-generated weather data, especially for a large watershed with complex topography. This work will be useful to hydrologists, meteorologists, and watershed modelers for generating spatially correlated precipitation and temperature data and for simulating surface water hydrology in a large watershed.

Technical Abstract: Weather data is usually required at several locations over a large watershed, especially when using distributed models for hydrological simulations. In many applications, spatially correlated weather data can be provided by a multi-site stochastic weather generator which considers the spatial correlation of weather variables. Prior to using a multi-site weather generator for hydrological modeling, its ability to adequately represent the proper hydrological response needs to be assessed. This study assesses the effectiveness of a new multi-site weather generator (MulGETS) for hydrological modeling over a Canadian watershed in the Province of Quebec. Prior to hydrological modeling, MulGETS is first evaluated with respect to reproducing the spatial correlation and statistical characteristics of precipitation and temperature for the studied watershed. Hydrological simulations obtained from MulGETS-generated precipitation and temperature are then compared with those obtained from a single-site weather generator (WeaGETS) and a WeaGETS-based lumped approach (WeaGETS-lumped) that averages the climate series over all stations in a watershed before running the single-site weather generator. The hydrology is simulated using two hydrological models: the conceptually-lumped model HSAMI and the physically-based distributed model CEQUEAU. When using the conceptually-lumped model, the weather time series are first averaged over all stations in the watershed. The results show that the mean snowmelt peak discharge is accurately represented by MulGETS-generated precipitation and temperature, while it is considerably underestimated by WeaGETS-lumped and WeaGETS data. In particular, MulGETS shows a significant advantage in representing the monthly streamflow variability which is underestimated by the data obtained from the other two single-site weather generator approaches. A flood frequency analysis indicates that MulGETS data consistently perform better than WeaGETS-lumped and WeaGETS data at representing spring high flow and summer-autumn low flow for all return periods between one and 100 years. Additionally, the hydrological simulation of the physically-based distributed model is better than that of the conceptually lumped model for the studied watershed. Hence, a multi-site weather generator coupled with a physically-based distributed model is recommended for all studies requiring the use of randomly-generated weather data, especially for a large watershed with a complex topography.