Li, T., Wang, G. & Chen, J. A modified binary tree codification of drainage networks to support complex hydrological models. Computers & Geosciences 36, 1427–1435 (2010).

Madsen, T., Franz, K. & Hogue, T. Evaluation of a distributed streamflow forecast model at multiple watershed scales. Water. 12(5), 1279 (2020).The scale of the water resources challenge means we need to take a strategic approach to planning our future water supply. We’ve been working closely with other water companies to look at options that could provide a large volume of water for more than one water company to use. These options are called strategic resource options (SROs). Stein, J. L., Hutchinson, M. F. & Stein, J. A. A new stream and nested catchment framework for Australia. Hydrology and Earth System Sciences 18, 1917–1933 (2014).

On registering their interest, third parties will be registered on Thames Water’s IASTA SmartSource portal within two weeks and provided access to complete an initial pre-qualification (PQQ1) survey. NASA/METI/AIST/Japan Spacesystems, U.S./Japan ASTER Science Team. ASTER Global Digital Elevation Model. NASA EOSDIS Land Processes DAAC https://doi.org/10.5067/ASTER/ASTGTM.002 (2009). Although the spatial resolution of existing hydrological data sets has been improved dramatically, which helps to enhance the results accuracy and scales refinement in water resources evaluation, there are some deficiencies. Firstly, it is difficult to automatically generate correct digital rivers using original DEM data, thus affecting the regional flow concentration relationship. This is due to the low vertical resolution 18 or some observation gaps in high-relief mountains and over water bodies of the original data 14, which makes it impossible to judge the correct flow direction of natural rivers on the land surface without or with very little auxiliary data. Secondly, the basins of many data sets are divided and coded using the Pfafstetter coding system 19, which limits the number of sub-basins to no more than 9 20, 21. It fails to highlight the details of the flow concentration relationship among sub-basins. Finally, the coding method of reaches and their zones is complex and does not directly correspond to the river spatial network structure. It also requires a correlation or calculation between fields to obtain the upstream and downstream relationships among basins.Falorni, G., Teles, V., Vivoni, E. R., Bras, R. L. & Amaratunga, K. S. Analysis and characterization of the vertical accuracy of digital elevation models from the Shuttle Radar Topography Mission. Journal of Geophysical Research: Earth Surface. 110, F2 (2005). Oki, T. & Kana, S. Global hydrological cycles and world water resources. Science. 313, 1068–1072 (2006).

Ministry of Water Resources of the People’s Republic of China, Code for china river name (SL 249-2012). We chose HydroSHED 9 and HDMA 11 databases for accuracy comparison. They are two widely used global hydrological data sets in which the secondary products (HydroRIVERS 34 and Stream layers) can provide us with an objective comparison. We refer to the comparison method in GRNWRZ V1.0 to compare the accuracy of our RN with HydroRIVERS and Stream layers. It should be noted that few rivers in Stream layers can meet the criteria for rivers at level 7 in GRNWRZ V2.0, so the rivers in Stream layers cannot completely cover the comparison points. Therefore, we only compare with HydroRIVERS at level 7. The comparison is made in the following way: Step 1. We randomly selected 400 rivers from L2-L7 RN of each continent, and three randomly points were generated at each river by ArcGIS (because the Level 1 RN consists of all Level 2 RN that flow into the same ocean, they are essentially the same rivers but divided by different criteria). Step 2. We imported these random points into Google Earth and manually marked the center point of natural rivers nearest to the random point. Step 3. We calculate the deviation distances from the center point obtained in step 2 to rivers in the other two data sets by ArcGIS. Then, we performed statistics in Excel to derive the comparison results. Li, X. et al. Hydrological cycle in the Heihe River Basin and its implication for water resource management in endorheic basins. Journal of Geophysical Research: Atmospheres. 123(2), 890–914 (2018). USGS-United States Geological Survey. HYDRO1K elevation derivative database, https://lta.cr.usgs.gov/HYDRO1K (2001).For Essex & Suffolk Water, our supply demand forecasts are significantly different to our WRMP19 forecasts because of: Major changes from our WRMP19 tables are in the columns for 2020/21 and 2021/22 data in Tables 5, 6 and 7. The values from WRMP19 have been overwritten with figures from our annual reporting. Create the original river. The original rivers were firstly automatically generated based on DEM by the hydrological module of ArcGIS, through several calculation process of filling depression, flowing direction, flowing accumulation, and crating river network, similar to most studies 2. The area threshold of rivers was judged by the National River Code of China, which stipulates that the catchment area of major rivers is larger than 1000 km 2. The ASTER GDEM at resolution of 30 m was resampled into 90 m as same as the SRTM data before creating the river. The determination of capacity is based on a standardised national review of the available information. Cohen, S., Wan, T., Islam, M. T. & Syvitski, J. P. M. Global river slope: A new geospatial dataset and global-scale analysis. Journal of Hydrology 563, 1057–1067 (2018).