Large-scale afforestation on China's Loess Plateau, driven by the Grain-for-Green program, has measurably depleted soil water to multi-meter depths, creating persistent dry soil layers. The founding regional assessment (Jia, Shao, Zhu and Luo, 2017) documented the depth and magnitude of this depletion but could not determine its permanence, its recovery timescale, or the process-level causes of the net decline. This problem asks for a regionally valid quantification of whether, and how fast, afforestation-induced deep soil desiccation recovers - naturally or under management such as thinning, density reduction, or species substitution - separated from multiyear climatic variability, together with the hydrological budget partitioning that produces the net decline. Site-level studies from 2020 to 2026 have produced recovery-time estimates spanning about 13 years to several centuries and partitioned budgets for individual orchards and stands, but a 2026 synthesis still describes the plateau's deep soil water as a finite resource carrying an accumulating hydrological debt, so the regional-scale question remains open.

Background

The Loess Plateau of northern China is covered by tens to hundreds of meters of wind-deposited loess, an arid to semi-arid region where the water table lies tens of meters below the surface and rainfall (roughly 200-650 mm per year) is effectively the only input of water to the soil. Since 1999 the Grain-for-Green program has converted steep cropland into forest and shrubland to control severe soil erosion, more than doubling vegetation cover; the dominant planted species are exotic, deep-rooted and fast-growing, notably black locust (Robinia pseudoacacia), Chinese pine (Pinus tabulaeformis) and pea shrub (Caragana korshinskii).

Soil water status is expressed as the volumetric soil moisture, θv\theta_{v}, the volume of water per unit volume of soil, typically measured at fixed depths by neutron probes in access tubes. Because seasonal rainfall wets only the upper 1-2 m, roots that extract water from greater depths can drive θv\theta_{v} far below the level that ordinary rainfall can replenish; such horizons are called dry soil layers (DSLs) and were documented across the plateau from the 1980s onward. A DSL is considered persistent when it survives multiple wet seasons, since recharge through the thick unsaturated (vadose) zone occurs mainly by slow piston flow, the downward displacement of old water by new infiltrating water over years to decades.

Jia, Shao, Zhu and Luo (2017) provided the first integrated regional assessment. They combined a 50-site neutron-probe survey of 0-5 m profiles across six watersheds, an 11-year (2004-2014) paired record of a C. korshinskii plantation and adjacent cropland at Liudaogou in the northern plateau, and a space-for-time substitution - substituting stands of different ages observed at one time for true time series - across 27 sites of different stand ages. They found substantially lower profile θv\theta_{v} under planted forest and shrubland than under cropland, linear growing-season declines of about 0.0080.008-0.0160.016 cm3^3 cm−3^{-3} yr−1^{-1} under Caragana, stabilization of the Caragana profile near θv≈0.086\theta_{v}\approx0.086 cm3^3 cm−3^{-3}, and drying of R. pseudoacacia stands for roughly the first 20-25 years, with only partial rewetting in older stands and the significance of that rewetting for net profile permanence left unresolved. They recommended lower-water-use species, thinning and density control, and planning within a soil water carrying capacity for vegetation (SWCCV), the maximum vegetation load a site's soil water can sustain.

The study explicitly left open what follows. Its regional picture rested on a one-time survey, a single paired watershed whose monitoring window overlapped drier-than-average years (2005-2006, 2009-2011), and space-for-time inference, so it could not (i) establish whether deep desiccation is permanent or reversible, or separate vegetation-driven trends from multiyear climatic variability; (ii) quantify recovery of deep soil water under natural succession or management (thinning, species substitution, reduced planting density); (iii) measure the components of the hydrological budget - canopy interception, infiltration and deep drainage, runoff, and evapotranspiration (ET, the sum of evaporation and transpiration) - that jointly produce the net decline; or (iv) scale these findings to landscape and regional management. Tools that later work brought to bear include stable and radioactive water isotopes (δ2\delta^{2}H, δ18\delta^{18}O, tritium) for dating and partitioning soil water, one-dimensional vadose-zone models such as Hydrus-1D, and satellite gravimetry (GRACE), which measures changes in total terrestrial water storage (TWS) at roughly hundred-kilometer resolution.

Problem Statement

For the Loess Plateau of China, determine: (1) whether the deep (multi-meter) soil moisture depletion caused by afforestation is permanent or reversible, and over what timescales it recovers - under natural succession or under management interventions such as thinning, reduced planting density, species substitution, or conversion to natural grassland or cropland - separating vegetation-driven depletion from multiyear climatic variability; and (2) the magnitudes of the hydrological budget components (canopy interception, infiltration and deep drainage, runoff, and evapotranspiration with its evaporation/transpiration partition) that jointly produce the observed net deep soil moisture decline under different afforestation types and stand ages. Because published recovery estimates depend strongly on the chosen depth window and reference state - spanning 13-19 years for peashrub-to-grass conversion in the 100-400 cm profile to several centuries for 2-10 m forest-to-natural-grass conversion - the determination must state explicitly the depth range defining "deep", the reference condition, and the recovery criterion used. The determination must extend beyond single sites to landscape and regional scales relevant to vegetation-water management, including the soil water carrying capacity for vegetation of the region's restoration program. These are the open questions left by Jia, Shao, Zhu and Luo (2017), whose regional assessment documented the depletion but could not establish its permanence, recovery, process attribution, or scaling to regional management.

References

  1. Jia, X., Shao, M., Zhu, Y., Luo, Y. (2017). Soil moisture decline due to afforestation across the Loess Plateau, China. Journal of Hydrology 546, 113-122. doi:10.1016/j.jhydrol.2017.01.011, https://doi.org/10.1016/j.jhydrol.2017.01.011
  2. Zhao, M., A, G., Zhang, J., Velicogna, I., Liang, C., Li, Z. (2020). Ecological restoration impact on total terrestrial water storage. Nature Sustainability 4, 56-62. doi:10.1038/s41893-020-00600-7, https://doi.org/10.1038/s41893-020-00600-7
  3. Ji, W., Huang, Y., Shi, P., Li, Z. (2021). Recharge mechanism of deep soil water and the response to land use change in the loess deposits. Journal of Hydrology 592, 125817. doi:10.1016/j.jhydrol.2020.125817, https://doi.org/10.1016/j.jhydrol.2020.125817
  4. Bai, X., Jia, X., Jia, Y., Shao, M., Hu, W. (2020). Modeling long-term soil water dynamics in response to land-use change in a semi-arid area. Journal of Hydrology 585, 124824. doi:10.1016/j.jhydrol.2020.124824, https://doi.org/10.1016/j.jhydrol.2020.124824
  5. Li, H., Si, B., Zhang, Z., Miao, C. (2022). Deep soil water storage and drainage following conversion of deep rooted to shallow rooted vegetation. Agricultural Water Management 261, 107359. doi:10.1016/j.agwat.2021.107359, https://doi.org/10.1016/j.agwat.2021.107359
  6. Meng, L.-C., Cui, Z., Huang, Z., Lucas-Borja, M. E., Dunkerley, D., Wu, G.-L. (2022). Thinning effects on forest production and rainfall redistribution: Reduced soil water deficit and improved sustainability of semiarid plantation forestlands. Land Degradation & Development 33(16), 3163-3173. doi:10.1002/ldr.4379, https://doi.org/10.1002/ldr.4379
  7. Shi, P., Gai, H., Li, Z. (2022). Partitioned Soil Water Balance and Its Link With Water Uptake Strategy Under Apple Trees in the Loess-Covered Region. Water Resources Research 59(1), e2022WR032670. doi:10.1029/2022WR032670, https://doi.org/10.1029/2022WR032670
  8. Liu, B., Jia, X., Shao, M., Jia, Y. (2022). Assessing soil water recovery after converting planted shrubs and grass to natural grass in the northern Loess Plateau of China. Agricultural Water Management 264, 107490. doi:10.1016/j.agwat.2022.107490, https://doi.org/10.1016/j.agwat.2022.107490
  9. Wang, Z. (2023). Estimating soil water storage recovery under former planted forest on steep slopes in the semiarid area of the Loess Plateau. EGU General Assembly 2023, abstract EGU23-1881. doi:10.5194/egusphere-egu23-1881, https://doi.org/10.5194/egusphere-egu23-1881
  10. Wang, X., Zhang, K., Li, J., Li, Q., Na, W., Gao, Y., Gao, Z. (2024). Response of soil water in deep dry soil layers to monthly precipitation, plant species, and surface mulch in a semi-arid hilly loess region of China. Agricultural Water Management 291, 108612. doi:10.1016/j.agwat.2023.108612, https://doi.org/10.1016/j.agwat.2023.108612
  11. Wu, X., Wang, X., Teng, Y., Chen, D., Wang, X., Ma, B. (2024). Response of soil moisture to rainfall following deep soil drying in China's hilly loess lands. Science of The Total Environment 950, 175145. doi:10.1016/j.scitotenv.2024.175145, https://doi.org/10.1016/j.scitotenv.2024.175145
  12. Zhou, J., Wang, Y., Li, R., He, H., Sun, H., Zhou, Z., Zhao, Y., Zhang, P., Li, Z. (2024). Response of deep soil water deficit to afforestation, soil depth, and precipitation gradient. Agricultural and Forest Meteorology 352, 110024. doi:10.1016/j.agrformet.2024.110024, https://doi.org/10.1016/j.agrformet.2024.110024
  13. Kong, W., Gao, F., Sun, S., Lei, Y., Gu, J., Luan, X. (2024). Advances and perspectives of research on soil moisture response to afforestation: a review. Hydrological Sciences Journal 69, 2090-2107. doi:10.1080/02626667.2024.2396060, https://doi.org/10.1080/02626667.2024.2396060
  14. Tian, H., Jin, M., Sohail, S., Ma, C., Bai, C., Qiao, J., Han, X., Shao, M. (2024). The trade-off between soil water recovery and nitrate leaching following the orchard-to-cropland conversion in the Chinese Loess Plateau. Scientific Reports 14, 29781. doi:10.1038/s41598-024-80192-w, https://doi.org/10.1038/s41598-024-80192-w
  15. Bai, C., Sohail, S., Tang, X., Tian, H., Han, X., Zhu, Y., Qiao, J. (2025). Crop-precipitation coupling drives deep soil desiccation-revival cycles in semiarid agroecosystems of Chinese Loess Plateau. Agriculture, Ecosystems & Environment 393, 109809. doi:10.1016/j.agee.2025.109809, https://doi.org/10.1016/j.agee.2025.109809
  16. Shen, X., Jia, X. (2025). Disentangling Ecological Restoration's Impact on Terrestrial Water Storage. EGU General Assembly 2025, abstract EGU25-2654. doi:10.5194/egusphere-egu25-2654, https://doi.org/10.5194/egusphere-egu25-2654
  17. Evaristo, J. (2026). Left High and Dry: Deep Soil Water Depletion and Ecohydrological Resilience on China's Loess Plateau. WIREs Water 13, e70053. doi:10.1002/wat2.70053, https://doi.org/10.1002/wat2.70053
  18. Yuan, Y., Tian, H., Qiao, J., Gao, J., Zhu, Y., Jia, X., Shao, M. (2026). Rainfall determines the temporal stability of soil water content following the conversion of deep rooted to shallow rooted vegetation in arid regions. CATENA 262, 109664. doi:10.1016/j.catena.2025.109664, https://doi.org/10.1016/j.catena.2025.109664