Answers are grounded in this paper and cite their sources.
The underwater terrain data of lakes serves as a crucial foundational dataset for estimating water storage capacity. Although the traditional measurement approaches are capable of precisely acquiring underwater terrain data, they are costly and time-consuming, making them impractical for widespread application to all lake water bodies. Consequently, only a limited number of lakes worldwide possess underwater terrain measurement data. To address this issue, the urgent need to develop a low-cost, user-friendly observation and simulation methods is needed. Hence, this study proposes a three-dimensional underwater terrain simulation method by using the machine learning technique Extreme Gradient Boosting (XGBoost). The method leverages the similarity between underwater and surface topography by utilizing a digital elevation model (DEM) as input data. Precision evaluation and analysis were conducted using eight representative lakes from the Tibetan Plateau. The results demonstrate that the algorithm achieves a relative error of −11.83% in average depth, −26.94% in maximum depth, and 19.36% in water storage capacity, all estimated based on the simulated underwater terrain. In regions where direct measurements are challenging or data is unavailable, this algorithm provides relatively accurate simulations of lake underwater terrain and water storage capacity, offering critical foundational data to support lake water resource management and scientific research.
No takes yet. Share an insight, caveat, or question.
Hu et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: