River Network Data

Introduction

The Tsinghua University team proposed a watershed digitization theory based on true-scale river networks, establishing the basic principle of using slope-channel (or river channel, abbreviated) structures reflecting geomorphic features as the minimum macroscopic unit of watersheds. Based on the minimum cost path search algorithm, the system efficiently implemented methods for extracting river networks from digital elevation models (DEM), with algorithm complexity reaching the theoretically optimal O(nlogn). They proposed flow direction determination and obstacle crossing algorithms that avoid depression preprocessing, while ensuring the efficiency and accuracy of river network extraction. They revealed that channel heads are key points for determining the true scale of slope-channel structures. Through a geomorphic feature function reflecting channel development mechanisms, they innovatively created a channel head position identification method based on along-course change point detection using this function, achieving precise identification of each channel head position in watersheds one by one, ensuring the feasibility of extracting true-scale river networks from watersheds, fundamentally solving the question of what resolution watershed modeling should achieve, and laying the foundation for watershed water-sediment dynamics models to distinguish different erosion forces at different geomorphic locations. They proposed a river network extraction method based on convolutional neural networks, which can simultaneously complete river network structure recognition and true-scale control, providing feasibility for lossless integration of measured rivers, remote sensing water surfaces, and other multi-source data in river network extraction algorithms. The achievements built a set of efficient and high-precision algorithms for extracting digital river networks from surface elevation and other data, forming completely independent and controllable core software.

Using 30-meter resolution DEM, the global river network dataset Hydro30 was extracted, with a total of approximately 500 million river segments. The data resolution, completeness, and accuracy are significantly superior to the previous generation of global river networks HydroSHEDS released by the United States. Currently, work is underway to extract the global river network dataset Hydro12 using 12.5-meter resolution DEM. Based on the river network binary tree coding method, river entity identity coding rules have been released and authorized by the ISO/IEC 15459 code-issuing institution Zhongguancun Industry and Information Technology QR Code Technology Research Institute. The MA.1002 global river network identification root node has been built and operated, achieving unified coding and decoding services for river entities with global proprietary and global unique characteristics, significantly improving the capacity, efficiency, and standardization of river network management, and laying the foundation for basin data services and microservice interface management based on river networks.

Related Literature

[1] Mao X, Chow J K, Su Z Y, Wand Y H, Li J Y, Wu T, Li T J. Deep learning-enhanced extraction of drainage networks from digital elevation models. Environmental Modelling & Software, 2021, 144: 105135.
[2] Li J Y, Li T J, Zhang L, Bellie S, Fu X D, Huang Y F, Bai R. A D8-compatible high-efficient channel head recognition method. Environmental Modelling & Software, 2020, 125: 104624.
[3] Wu T, Li J Y, Li T J, Bellie S, Zhang G, Wang G Q. High-efficient extraction of drainage networks from digital elevation models constrained by enhanced flow enforcement from known river maps. Geomorphology, 2019, 340: 184-201.
[4] Li J Y, Li T J, Liu S N, Shi H Y. An efficient method for mapping high-resolution global river discharge based on the algorithms of drainage network extraction. Water, 2018, 10(04): 533.
[5] Bai R, Li T J, Huang Y F, Li J Y, Wang G Q. An efficient and comprehensive method for drainage network extraction from DEM with billions of pixels using a size-balanced binary search tree. Geomorphology, 2015, 238: 56-67.
[6] Bai R, Li T J, Huang Y, Li J Y, Wang G Q, Yin D Q. A hierarchical pyramid method for managing large-scale high-resolution drainage networks extracted from DEM. Computers & Geosciences, 2015, 85(PA): 234-247.