Prediction of groundwater depth in an arid region based on maximum tree height

Groundwater is the most important water resource in arid regions. However, the groundwater traditional models and field measurements are limited due to the fluctuations in hydrological cycles and the difficulty in the accurate quantification of associated parameters. Here, we hypothesized that maximum potential tree height could be used to predict groundwater depth due to the hydraulic limitation of water transportation. To address this hypothesis, we measured two proxy indicators of maximum potential tree height, i.e., the maxima of heights and volumes, of three common dominant tree species in northwest China to construct classical measurement error (CME) model for predicting groundwater depth in an arid region. Our results showed that the optimal model based on maximum tree height had the best predictive performance of groundwater depth, particularly the tallest plant species. The CME model showed that maximum tree height played a vital role in predicting groundwater depth. Mathematically the model can be expressed as: [Eln(Dw)lnTh,θ)=7.11-1.85ExlnTh,θ), where Dw and Th are respectively the theoretical values of groundwater depth and maximum tree height; x is the measured maximum tree height; θ = {7.11, −1.85, 7.19, 0.15, 1.91, 13.45}; R² = 0.82; Marginal log-Likelihood = −131.04; RMSE = 0.33]. In addition, Leave-One-Out Cross-Validation together with correlation analysis indicated that groundwater depth prediction based on maximum tree height in arid regions was an accurate and promising approach. In conclusion, our study showed that the hydraulic limitation of water transportation led to a negative relationship between maximum tree height and groundwater depth. Our developed model for predicting groundwater depth with maximum tree height has provided the important basis for the conservation and management of groundwater resources in arid regions.