Can merged gridded rainfall products replace sparse gauge data in hydrological modelling? A case study with the HYPE model in Ethiopia

Abstract

AIM: We investigate if merged gridded satellite rainfall products (CHIRPS, MSWEP, or SUPER), used with the HYPE hydrological model, can replace limited rain gauge measurements when modelling streamflow in poorly gauged basins. Three research questions frame this aim: (1) How does the HYPE hydrological model perform in a poorly gauged basin? (2) Which rainfall product performs better (CHIRPS, MSWEP, SUPER, or in-situ gauge measurements) in simulating streamflow? (3) Can merged gridded rainfall data replace limited in-situ measurements in simulating streamflow? The Upper Gilgel Abay Basin in Ethiopia was chosen to represent a typically poorly gauged basin.
METHODS: First, each gridded precipitation data set was compared to the best available in-situ gauge measurements. We then ran HYPE four times, once per precipitation input, holding all other inputs constant. Eight influential parameters were identified through a Morris sensitivity analysis. These eight parameters were calibrated via 10 000 random parameter simulations. The top performing parameter sets for each HYPE run were used to evaluate model skill against daily and monthly measured streamflow at the outlet of the basin.
RESULTS: (1) All four precipitation data sets resulted in good performance in simulating daily streamflow and very good performance in simulating monthly streamflow according to the performance categories proposed by Moriasi et al (2007). This shows that even when using a computationally inexpensive sensitivity analysis and a single general calibration round of 10 000 simulations, HYPE can perform well in a poorly gauged basin. (2) All three gridded precipitation products performed equally well. They resulted in the same sensitive parameters but different
optimal parameter values for each top performing parameter set. This adds some uncertainty to the results as different hydrological components in the model compensated for the differences in precipitation. (3) Gridded precipitation products can perform as well as in-situ gauge precipitation when the in-situ measurements are sparse. CHIRPS, MSWEP, and SUPER all emerge as viable alternatives to in-situ measurements when simulating either daily or monthly streamflow.

Popular Abstract

Understanding how water flows across the landscape is essential for securing clean water, avoiding floods, and planning for renewable energy. Using information on the amount of rainfall in an area, we can estimate how much water will reach the river farther down the landscape. Unfortunately, in most areas of the world we have very limited physical rainfall measurements. So, how can we estimate the flow of water in these areas when no physical rainfall measurements exist?
To answer this, we tested whether rainfall estimates from satellites could be used to predict the amount of water that flows in a river per day and per month. We set up a hydrological model, called HYPE, and then we ran the model using four different rainfall data sets for the same area. Three of the rainfall data sets used satellites, and one of the data sets used only physical measurements. This allowed us to compare how well the model could predict the flow of water depending on which rainfall data set was used.
The results were clear: all four rainfall data sets gave equally good estimates of daily and monthly river flow. In other words, the rainfall data based on satellite measurements can be used in areas where no or few physical rainfall measurements exist. However, in this study, we compared the estimated river flow at only one location. This means that although the models could accurately predict the amount of water flowing at that specific point, we could not gain any significant insights into what components of the water cycle were the most important in leading the water to that point. Future research can build on the methods and results we presented to gain better understanding of the water cycle components that affect the flow of water depending on the topography and climate of the area.

• author: Black, Steven
• supervisor: Zheng Duan; Renkui Guo
• organization: Dept of Physical Geography and Ecosystem Science
• course: NGEM01 20251
• year: 2025
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: Physical Geography and Ecosystem analysis, hydrology, streamflow, CHIRPS, MSWEP, SUPER, HYPE, Upper Gilgel Abay Basin, Ethiopia, gridded precipitation products
• publication/series: Student thesis series INES
• report number: 738
• language: English
• id: 9212664
• date added to LUP: 2025-09-18 11:16:14
• date last changed: 2025-09-18 11:16:14