Elsevier

Journal of Hydrology

Volume 540, September 2016, Pages 565-573
Journal of Hydrology

Research papers
Responses of soil water percolation to dynamic interactions among rainfall, antecedent moisture and season in a forest site

https://doi.org/10.1016/j.jhydrol.2016.06.038Get rights and content

Highlights

  • The DP fluxes increased as the initial condition became wetter.

  • Antecedent moisture had the greatest impact on DP, followed by rainfall and season.

  • As antecedent moisture increasing, influence of season on DP increased.

  • The TDC influenced DP by affecting the responses of DP to other factors.

  • The DP of TDC_B (rainfall intensity linearly increased with time) was the lowest.

Abstract

Knowledge of soil water percolation below the rooting zone and its responses to the dynamic interactions of different factors are important for the control of non-point source pollution. Based on 3600 scenarios in Hydrus-1D simulation, this study revealed the integrated effects of rainfall characteristics (rainfall amount, maximum rainfall intensity or MRI, time distribution characteristics of rainfall or TDC), antecedent moisture and the season on deep percolation (DP) at a forest site in Taihu Lake Basin, China. Results showed that Hydrus-1D model can well simulate the soil water dynamics at this site. Antecedent moisture had the greatest relative contribution to DP (85.7%), followed by rainfall amount (10.9%) and MRI (3.4%). As the antecedent moisture increased, the relative contribution of the season on DP increased from 0.0% to 16.4%. In comparison, that of MRI decreased from 58.7% to 38.5% and that of rainfall amount followed a bell shape pattern (greatest when the antecedent moisture was 0.26 m3 m−3). The relative contribution of antecedent moisture to DP in summer was the greatest (87.8%), while that of the rainfall was the least. The TDC influenced DP by affecting the responses of DP to other factors. When the rainfall amount was ⩾80 mm and the antecedent moisture content was ⩾0.34 m3 m−3, effect of TDC on DP could be observed. The DP of TDC_B (rainfall intensity linearly increased with time) was the lowest, while that of TDC_E (rainfall intensity kept constant with time) was the greatest. Findings of this study have practical significance for investigating the water and pollutant transport in vadose zone.

Introduction

Soil water percolation below the rooting zone is a momentous link in the terrestrial hydrological cycle. As a primary pathway to recharge groundwater, deep percolation (DP) provides important hydrologic and ecosystem benefits in different regions (Ochoa et al., 2007, Zhang et al., 2007, He et al., 2012). In addition, DP is closely related to the leaching of non-point source pollutants (Bethune et al., 2008, J. Wang et al., 2014, Z. Wang et al., 2014). Therefore, it is important to reveal the generation, processes and quantities of DP as influenced by different factors.

The processes of DP have often been studied by field experimental methods and numerical simulation (NS). Although field experimental methods (e.g., soil water flux meters, lysimeters or soil tracers) have been widely used in past studies to assess DP (e.g., Chigira et al., 2006, Sasaki, 2006, Arbel et al., 2010), they have been also recognized to be expensive, time-consuming, laborious, and unreliable in some cases (Bond and John, 1998, Walker et al., 2002). Recently, NS combined with appropriate soil and climate datasets emerges as a popular method for DP estimation. With the superiority of low cost/benefit ratio and flexibility, NS can conveniently establish and assess the DP under a variety of potential scenarios (e.g., extreme climate conditions) (Bah et al., 2009). The Richards’ equation is an example of NS that has been broadly applied to estimate DP (Vrugt et al., 2004, Stewart et al., 2006, Min et al., 2015). However, models based on Richards’ equation require soil hydraulic parameters that are difficult to be measured (Bethune et al., 2008, Li and Shao, 2014). Thus, inverse modeling has been used to estimate these parameters from the field observation of soil water contents (Bethune et al., 2008, Selle et al., 2011).

The DP was influenced by the magnitude of water inputs (e.g., precipitation or irrigation) and their characteristics (Ochoa et al., 2007, Min et al., 2015). For example, Liu et al. (2015) indicated that DP was significantly correlated with rainfall amount and intensity in the mobile sandy lands in Inner Mongolia, China; He et al. (2012) found that larger rain events (>20 mm) played an important role in elevating soil water content and producing DP. In addition, threshold of water input for triggering DP has also been observed in previous studies. As reported by Yang et al. (2015), when the precipitation + irrigation (P + I) was above 3 mm·day−1, the P + I was linearly correlated with DP. Besides rainfall (or irrigation) amount and intensity, the time distribution characteristics (TDC) could also affecting DP. As Verma et al. (2011) proposed, the daily rainfall distributions considerably affected the estimations of vegetation water-stress, leakage and runoff occurrence, and thus the water balance. However, to our knowledge, the response of DP to TDC of rainfall has been rarely reported.

The DP has also been widely reported to be influenced by antecedent moisture and factors related to the season (e.g., evapotranspiration, temperature and plant water uptake). High antecedent moisture always means high percolation capacity in the upper layer of soil. Ochoa et al. (2007) demonstrated that less DP occurred under drier antecedent moisture conditions; Zhu et al. (2014) observed that deep soil at toe slope was recharged when rainfall amount >35 mm and antecedent soil moisture >0.30 m3 m−3. Besides, factors related to the season (e.g., evapotranspiration) influence the soil water storage, thus affect the DP (Allen et al., 1998). Jiménez-Martínez et al. (2009) indicated a high recharge rate when potential evapotranspiration was low. In the study by Li et al. (2014), air temperature and evapotranspiration were two of the most important factors in estimating monthly groundwater recharge rate. However, the extreme antecedent moisture and weather conditions may not be able to observe through the field experiment. Therefore, the mechanisms of DP responding to various combinations of antecedent moisture and weather conditions are not able to comprehensively and deeply be revealed by field experiment.

In addition, most previous studies focused on analyzing the independent effects of impact factors on DP and neglected the integrated influences of multiple dynamic factors. Only few studies assessed the interactive effects of multiple variables on DP. For example, Bah et al. (2009) evaluated how the interactions of soil permeability, vegetation rooting depth and growth duration affected drainage in northern New South Wales, Australia; Liancourt et al. (2012) investigated the interplay among climate change, local abiotic conditions (slope) and biotic factors (presence of vegetation) on soil water balance in a steppe grassland on the south exposure of a northern Mongolia valley. Nonetheless, more thorough and comprehensive information is still required to understand the response of DP to the interactions among multiple dynamic factors such as rainfall, antecedent moisture and the season.

Using the Hydrus-1D model based on the 1-D Richards’ equation, objectives of this paper were to (i) quantify the relative contribution of different dynamic factors (rainfall, antecedent moisture and the season) to DP; (ii) reveal the influence of the interactions and dynamics among rainfall, antecedent moisture and the season on DP; and (iii) analyze the effect of TDC of daily rainfall on DP in a forest site in Taihu Lake Basin, China. Scenarios were set to cover different combinations of rainfall amount, rainfall intensity, TDC of rainfall, antecedent moisture and the season.

Section snippets

Study site

This study site (31°66′N, 120°55′E) is located in Xishan District, Wuxi City, in northern Taihu Lake basin, China (Fig. 1). It is in the center of an artificially planted evergreen camphor forest (Cinnamomum camphora (L.) Presl.) with an area of approximately 10 ha. Trees in this forest are young (<15 yr) and well-aligned with a space of around 2.5-m. Bushes and grasses are rarely observed. The study site is typical of northern subtropical monsoon climate with four distinctive seasons. The annual

Calibration and validation of the Hydrus-1D model

The Richards equation-based Hydrus-1D model was capable of describing soil water movement in this study site. The NSE values were above 0.60 and 0.50 at different depths during the training and testing periods, respectively (Fig. 3). The RMSE values were below 0.029 and 0.034 m3 m−3 at four depths during the training and testing periods, respectively (Fig. 3). Therefore, the calibrated soil hydraulic parameters (Table 4) can reflect the soil hydraulic properties in this study site. These

Ranking of influencing factors

The relative contributions of different factors (rainfall, antecedent moisture and the season) to DP have been quantified by the CART analysis (Fig. 6). Result indicated that antecedent moisture has the greatest contribution (85.7%) to DP compared to other factors (Fig. 6). In addition, the correlation between antecedent moisture and DP was in positive and significant under P < 0.05. The rainfall amount and MRI also had the second and third greatest contributions to DP, respectively (Fig. 6).

Conclusions

This study used Hydrus-1D model and scenario analysis to reveal the response of DP to the dynamic interactions among rainfall, antecedent moisture and season in a forest site. Antecedent moisture had the greatest influence on DP, followed by rainfall amount and MRI. As the antecedent moisture increasing, the influence of the season on DP increased, while influences of MRI decreased and influences of rainfall amount followed a bell shape pattern. The influence of antecedent moisture on DP

Acknowledgements

This study was financially supported by the by the National Natural Science Foundation of China (41271109, 41571080 and 41301234) and Jiangsu Natural Science Foundation (BK20151061).

References (52)

  • L.L. Min et al.

    Estimating groundwater recharge using deep vadose zone data under typical irrigated cropland in the piedmont region of the North China Plain

    J. Hydrol.

    (2015)
  • J.E. Nash et al.

    River flow forecasting through conceptual models. Part I: a discussion of principles

    J. Hydrol.

    (1970)
  • B. Selle et al.

    Effective modelling of percolation at the landscape scale using data-based approaches

    Comput. Geosci.

    (2008)
  • B. Selle et al.

    Applicability of Richards’ equation models to predict deep percolation under surface irrigation

    Geoderma

    (2011)
  • L.K. Stewart et al.

    Estimating deep drainage and nitrate leaching from the root zone under sugarcane using APSIM-SWIM

    Agri. Water Manage.

    (2006)
  • P. Verma et al.

    A stochastic model describing the impact of daily rainfall depth distribution on the soil water balance

    Adv. Water Resour.

    (2011)
  • J. Wang et al.

    Nitrogen and phosphorus leaching losses from intensively managed paddy fields with straw retention

    Agri. Water Manage.

    (2014)
  • S. Wang et al.

    Responses of soil moisture in different land cover types to rainfall events in a re-vegetation catchment area of the Loess Plateau, China

    Catena

    (2013)
  • X.P. Wang et al.

    Effects of rainfall characteristics on infiltration and redistribution patterns in revegetation-stabilized desert ecosystems

    J. Hydrol.

    (2008)
  • S.L. Zhang et al.

    Simulated long-term effects of different soil management regimes on the water balance in the Loess Plateau, China

    Field Crop Res.

    (2007)
  • Q. Zhu et al.

    Soil moisture response to rainfall at different topographic positions along a mixed land-use hillslope

    Catena

    (2014)
  • N. AKrour et al.

    Simulation of yearly rainfall time series at microscale resolution with actual properties: intermittency, scale invariance, and rainfall distribution

    Water Resour. Res.

    (2015)
  • R.G. Allen et al.
    (1998)
  • Y. Arbel et al.

    Infiltration processes and flow rates in developed karst vadose zone using tracers in cave drips

    Earth Surf. Proc. Land.

    (2010)
  • N.N. Benjamin et al.

    Groundwater recharge from rainfall in the southern border of Lake Chad in Cameroon

    World Appl. Sci. J.

    (2007)
  • M.G. Bethune et al.

    Understanding and predicting deep percolation under surface irrigation

    Water Resour. Res.

    (2008)
  • Cited by (0)

    View full text