Estimating river water level using satellite altimetry and water surface temperature data in a regulated subtropical river

Study region: Dongjiang River in east Guangdong, China Study focus: River water level (RWL) is an important parameter for evaluating hydrologic regimes. RWL estimation based on remote sensing techniques has drawn great attention to the hydrologic community even with relatively low temporal resolution but acceptable accuracy. The techniques adapt well in wide rivers with little disturbance, while they have not been fully tested in heavily regulated rivers. In this study, we tested two methods for RWL estimation: the altimetry-based and the temperature-based methods. New hydrological insights: The Jason data with a ∼10-day revisit period were insightful for RWL simulations, and the temperature-based model reproduced well the daily RWL observations. Noticeably, both methods simulated low flows better than high flows. Underestimation of high flows can be a result of asynchrony between RWL and satellite scanning due to the low temporal resolutions. Temperature retrieved for narrow rivers may contain mixed information of water and non-water surfaces, weakening the water level-temperature relationship. Intentional discharge from the upstream reservoir for downstream ecological conservation during dry seasons could also challenge both methods for capturing the induced high flows. This study provides two means of estimating RWL continuously using readily available satellite products for small regulated subtropical rivers, which can be practically useful in any rivers with carefully selected processing algorithms.