South-to-North Water Diversion stabilizing Beijing’s groundwater levels

Groundwater (GW) overexploitation is a critical issue in North China with large GW level declines resulting in urban water scarcity, unsustainable agricultural production, and adverse ecological impacts. One approach to addressing GW depletion was to transport water from the humid south. However, impacts of water diversion on GW remained largely unknown. Here, we show impacts of the central South-to-North Water Diversion on GW storage recovery in Beijing within the context of climate variability and other policies. Water diverted to Beijing reduces cumulative GW depletion by ~3.6 km3, accounting for 40% of total GW storage recovery during 2006–2018. Increased precipitation contributes similar volumes to GW storage recovery of ~2.7 km3 (30%) along with policies on reduced irrigation (~2.8 km3, 30%). This recovery is projected to continue in the coming decade. Engineering approaches, such as water diversions, will increasingly be required to move towards sustainable water management.


Supplementary Notes Supplementary Note Details on the South-to-North Water Diversion Project
The South-to-North Water Diversion (SNWD) Project was launched in 2002 by the Chinese government to transfer water from the Yangtze River in the south of China to the drier north through three canal and pipeline systems. The eastern route, designed to transfer 14.8 km 3 of water annually, started to transfer water from Yangzhou City (i.e., downstream of the Yangtze River) through Jiangsu, Anhui, Shandong, and Hebei Provinces and Tianjin Municipality since 2013. The central route, designed to transport 9.5 km 3 of water annually from the Danjiangkou Reservoir in the middle reaches of the Hanjiang River (i.e., the largest tributary of the Yangtze River) through Henan and Hebei Provinces, and Beijing and Tianjin Municipalities began to operate since Dec 2014. The western route, planned to divert 20 km 3 from the upper reaches of the Yangtze River (i.e., Tongtian, Yalong, and Dadu Rivers) to the Yellow River in northwest China is still being planned and has yet started to transport water.

Supplementary Note 2. Details on water diversion to Beijing
There are two phases of water diversion to Beijing discussed in this study. During phase I (2008During phase I ( -2014, major reservoirs (i.e., Gangnan, Huangbizhuang, and Wangkuai) in Hebei Province supplied 1.6 km 3 of water to Beijing, which was not included in the initial plan of the central SNWD route. During phase II (Dec 2014-Dec 2019) when the principal part of the central SNWD route operated, a total of ~ 26 km 3 of water has been transported to North China, of which 5.2 km 3 (~ 1 km 3 yr -1 ) reached Beijing (the amount of 1 km 3 yr -1 accounting for 26% of annual total water use of 3.8 km 3 in Beijing). This has profoundly altered the water supply structure and indirectly impacted groundwater storage (GWS) there.

Supplementary Note 3. Impacts of interannual variability in precipitation on GWS projection
We also projected GWS changes by incorporating interannual variability in precipitation. Two new precipitation scenarios during 20192030 were generated based on the original RCM precipitation: (P1N) original RCM annual precipitation during 20192030 (mean annual precipitation of 750 mm yr -1 ) multiplied by a factor of 0.72 (540/750), to represent the climatology (i.e., mean annual precipitation during 20002018, 540 mm yr -1 ), and (P2N) original RCM annual precipitation during 20192030 multiplied by a factor of 0.77 (580/750), to represent a wet climate (i.e., mean annual precipitation during 20082018, 580 mm yr -1 ). Groundwater use scenarios are the same as the initial projection.
New projections of GWS during 20192030 clearly show the impacts of interannual variability in precipitation on GWS change and higher rates of the recovery than the initial projections (Supplementary Figure 5). The lowest recovery rate was estimated to be 28 mm yr -1 (groundwater depth of ~ 13 m in 2030) under scenario I, and the highest recovery rate was found to be 50 mm yr -1 (groundwater depth of ~ 5 m in 2030) under scenario IV. However, these trends may have been overestimated, because increased groundwater withdrawal during drought was not incorporated into these scenarios. For instance, annual precipitation in 2026 was estimated to be 282 mm yr -1 (P1N) after bias correction, which is much less than the historical precipitation level and would likely result in substantial groundwater pumping to alleviate the drought. However, the finding of overall increasing trends in GWS in Beijing in the coming decade did not change.

Supplementary Tables
Supplementary

Supplementary Figures
Supplementary Figure