Modeling growth response to soil water availability simulated by HYDRUS for a mature triploid Populus tomentosa plantation located on the North China Plain

• HYDRUS can accurately simulate soil water dynamics in Populus tomentosa plantation.
• Soil water at different depths made different contributions to tree growth variation.
• Monitoring the surface soil water provides a robust means for predicting tree growth.
• rθ in the 0–30 cm soil layer should be maintained above 0.7 to minimize growth loss.

To establish the methodological basis for developing optimal irrigation strategies for increasing the productivity of triploid Populus tomentosa plantations using modelling methods, the accuracy of HYDRUS models for simulating one-dimensional (HYDRUS-1D) soil water dynamics under rainfed natural conditions (NC), and two-dimensional soil water dynamics (HYDRUS (2D/3D)) under subsurface drip irrigated (SDI) conditions was evaluated using field data. The relationship between tree growth and soil water availability (rθ) at different depths, which has not been thoroughly investigated in poplar plantations, was also examined. In general, the average soil water content (θ) in different soil layers predicted by both HYDRUS models and the θ within the two-dimensional domain around drippers predicted by HYDRUS (2D/3D) agreed well with the observed values. Under both treatments, the rθ increased with depth and was most variable in the surface 30 cm soil. The amount of variation in basal area at breast height (ABH) growth explained by rθ in various soil layers ranged widely, suggesting that soil water at different soil depths made different contributions to the variation in growth. The proportion of variation in ABH growth explained by average rθ was highest (R2 = 0.709) in the 0–30 cm layer, and decreased with increasing integrated depth of the root-zone. Tree growth was unconstrained when the rθ of the 0–30 cm layer was above 0.7. Based on these results, it can be concluded that HYDRUS-1D and HYDRUS (2D/3D) can be used as tools to accurately simulate long-term soil water dynamics in P. tomentosa plantations, at least in sites with similar characteristics to ours. HYDRUS modeling can be used to assess the impacts of rθ on productivity of mature P. tomentosa plantations. This study also shows that monitoring soil moisture of the surface soil provides a robust means for predicting tree growth of P. tomentosa plantations at sites with similar soil to ours.