Using pedo-transfer functions to estimate dry soil layers along an 860-km long transect on China’s Loess Plateau
Introduction
Climate change and human activities are dramatically affecting traditional hydrological processes in terrestrial ecosystems (Zavaleta et al., 2003, Wang et al., 2016), resulting in more frequent and intense droughts (Sherwood and Fu, 2014, Anderegg et al., 2015). This has interrupted the balance in water input (e.g., rainfall and irrigation) and output (e.g., evaporation, root uptake and deep drainage) (Brown, 2002, Wang et al., 2010a, Wang et al., 2010b, Schulte et al., 2011), ultimately causing the formation of dry soil layers (DSLs) or soil desiccation in the soil profile (Li, 1983, Yang et al., 1999, Jia et al., 2015).
DSL is a soil layer with soil water content below stable field capacity (Chen et al., 2008, Wang et al., 2010a, Jia et al., 2020). It is mainly caused by a long period of excessive depletion of deep soil water through evapotranspiration and low precipitation recharge (Wang et al., 2010b, Yan et al., 2015). DSL is a global soil hydrological phenomenon that has been reported in southern Australia (Robinson et al., 2006), eastern Amazonia (Jipp et al., 1998), Russia (Yang and Han, 1985), southwestern United States (Querejeta et al., 2007) and China’s Loess Plateau (CLP) (Wang et al., 2010a). For example, Robinson et al. (2006) found that soil water depletion by Eucalyptus robusta caused the formation of low water potential layer to the depth of 2 m in the profile soil, limiting groundwater recharge by incident rainfall in southern Australia. In southwestern United States, Querejeta et al. (2007) observed DSL throughout the 0–2 m soil profile and the Quercus palustris plant derived water from moist soil layers below the 2 m depth. The formation of DSLs can negatively affect water cycle, crop yield and ecosystem succession (Nepstad et al., 2004, Chen et al., 2008, Mendham et al., 2011). Furthermore, if DSL persists for a long time, it could lead to soil degradation, regional vegetation die-off and aridity of local climatic environment (Yan et al., 2015, Jia et al., 2020). Understanding the formation and distribution of DSL is thus critical for sustainable management of soil hydrological and ecological processes at various spatial scales.
CLP is well known for its intense soil erosion, severe water shortage, fragile ecosystem and deep loess deposit. Thus, considerable effort in terms of labor, time and money has been invested into vegetation restoration in the region (Lü et al., 2012, Zhang et al., 2017, Jia et al., 2019). In CLP region, DSL was first detected in the 1960 s (Li, 1962). Since then, numerous studies have been conducted relating to the definition and classification (Li, 1983, Chen et al., 2008), formation and development (Wang et al., 2010b), spatial distribution pattern (Wang et al., 2010a, Zhang et al., 2017), temporal persistence (Jia et al., 2015) and sustainable recovery (Wang et al., 2012b) of DSLs in the region. Wang et al. (2010a) noted that DSL was found in most regions of CLP, but was generally more severe in the central and western regions of the plateau. A widespread formation of DSL with severe negative effects can endanger the sustainability and development of restored ecosystems and ruin the vast investments in environmental conservation in the study area (Wang et al., 2010a, Jia et al., 2015, Yan et al., 2015). Thus, studying the status and spatial distribution of DSLs at regional scale is critical for developing sustainable land management strategies with long-term ecosystem services in CLP and beyond.
According to previous studies, three indices are usually employed to characterize DSLs — DSL thickness (DSL-T), DSL initial formation depth (DSL-F) and DSL mean soil water content (DSL-SWC) (Robinson et al., 2006, Zhao et al., 2007, Wang et al., 2010a, Yan et al., 2015). The determination of DSL indices is driven by profile soil moisture survey. Such survey is often labor intensive, time consuming and costly, especially for larger-scale and deep-profile applications (Saxton and Rawls, 2006, Wang et al., 2010a, Zhao et al., 2016a). Difficulties in soil profile survey affect DSL research efficiency. As an alternative therefore, pedo-transfer function (PTF), which uses predictive functions and soil survey data to determine soil properties, is now widely considered in soil study (Minasny and McBratney, 2000, Chapuis, 2012, Duan et al., 2012, Yao et al., 2015, Zhao et al., 2016a). PTF is built on the premises that the main driving factors can be determined. Wang et al. (2010a) noted that the primary and secondary factors with significant impact on DSL formation and/or development were precipitation and soil type. Yan et al. (2015) observed that mean annual precipitation is negatively correlated with DSL-T, but positively correlated with DSL-F and DSL-SWC. Thus, based on potential relationships between environmental factors and DSL indices, PTFs of DSL indices can be built from environmental factors.
Although studies have simulated DSL formation and development (Wang et al., 2012b, Li et al., 2007), remains little known about PTFs of DSL indices at regional scale. The uniqueness of regional geographical and environmental variability on CLP makes DSL research difficult, especially in terms of cost and time (Jia et al., 2015). The determination of PTFs from dominant factors of DSL indices is therefore a theoretical and practical requirement for DSL research. On CLP, a policy-driven large-scale vegetation restoration has been going on since 1999, resulting in 10.8% reduction in cropland, 4.9% increase in forestland and 6.6% increase in grassland (Lü et al., 2012). There is widespread DSL in the restored ecosystems with significant negative effect on the local ecology and environment. This lays the need for developing robust PTFs to effectively and easily predict DSL indices that can enhance DSL evaluation and soil water management on CLP.
The specific objectives of this study were to: 1) characterize regional-scale variations in DSLs along an 860-km south-north transect of CLP; 2) determine the relationship between environmental factors and DSL in the study area; and 3) develop effective PTFs for the estimation of DSLs in the region.
Section snippets
Study area
The study was conducted in CLP region (34°0́–45°5́N, 101°33́–114°33́E), with a total area of 64 × 104 km2 (Fig. 1). The geomorphic landforms include the Yuan (a large flat surface with little or no erosion) and Mao (an oval or round loess hill), Liang (a long narrow range of hills), and hills/gullies of various forms (Wang et al., 2018). The climate in the study area is continental monsoon, with mean annual precipitation ranges from the northwest to the southeast of 150–800 mm, 55–78% of which
SDI and DSL distribution along transect
The profile distribution of the four-year average SDI is shown in Fig. 3. In the transect, DSL (with SDI < 1) was mainly distributed in the northcentral part (regions between sites 46–75) of CLP study area and was generally thicker and more severe in the area. For some sites, there was DSL right through the 0–5 m soil profile — e.g. sites 46, 47, 49, 50, 54, 56, 60, 62 and 65 (Fig. 3). Compared with the northcentral region of the plateau, DSL was scattered across the southcentral region and had
Spatial patterns of DSL
In CLP study area, DSL most widely occurred in the northcentral region (Fig. 3). Wang et al. (2010a) also noted that DSLs were mainly in western and northwestern regions of the plateau, although they were relatively thinner in interior regions such as Guanzhong Plain. Zhao et al. (2019) concluded that soil desiccation in the northern region was more severe than in the southern region of the Loess Plateau.
The distribution of DSLs can be explained in terms of the spatial distribution of
Conclusions
The study showed that DSL existed mainly in the northcentral half of the transect and had an average thickness of 2.77 m on CLP. The ANN approach improved the simulation accuracy of the model and its performance was better than the MR approach in terms of recognition of DSL and simulation of DSL indices in the study area. The indirect method was effective for the prediction of DSL indices. The use of these PTFs not only reduced the labor and time needed for field investigations, but also
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This study was supported by the National Natural Science Foundation of China (41877016 and 41907009), the China Postdoctoral Science Foundation (2018M641460) and the Youth Innovation Promotion Association of the Chinese Academy of Sciences (2017076). We thank the anonymous editors and reviewers for the value added to the work.
References (52)
- et al.
Predicting root zone soil moisture with soil properties and satellite near-surface moisture data across the conterminous United States
J. Hydrol.
(2017) - et al.
Using the soil and water assessment tool (SWAT) to assess land use impact on water resources in an east African watershed
J. Hydrol.
(2013) - et al.
Soil desiccation in the Loess Plateau of China
Geoderma
(2008) - et al.
Regional scale spatio-temporal variability of soil moisture and its relationship with meteorological factors over the Korean peninsula
J. Hydrol.
(2014) - et al.
Regional temporal persistence of dried soil layer along south-north transect of the Loess Plateau
China. J. Hydrol.
(2015) - et al.
Soil moisture decline due to afforestation across the loess plateau, china
J. Hydrol.
(2017) - et al.
Spatial variations in soil-water carrying capacity of three typical revegetation species on the Loess Plateau
China. Agr. Ecosyst. Environ.
(2019) - et al.
Hydrological projections under climate change in the near future by RegCM4 in southern Africa using a large-scale hydrological model
J. Hydrol.
(2015) - et al.
Modeling soil–water dynamics and soil–water carrying capacity for vegetation on the Loess Plateau
China. Agr. Water Manage.
(2015) - et al.
Soil water depletion and replenishment during first- and early second-rotation Eucalyptus globulus plantations with deep soil profiles
Agric. For. Meteorol.
(2011)
Hydraulic lift may buffer rhizosphere hyphae against the negative effects of severe soil drying in a California oak savanna
Soil Biol. Biochem.
Applications of stable water and carbon isotopes in watershed research: weathering, carbon cycling, and water balances
Earth Sci. Rev.
Simulating soil water regime in lowland paddy fields under different water managements using HYDRUS-1D
Agr. Water Manage.
Validation of the EPIC model using a long-term experimental data on the semi-arid loess plateau of china
Math. Comput. Model
Large-scale spatial variability of dried soil layers and related factors across the entire Loess Plateau of China
Geoderma
A preliminary investigation of the dynamic characteristics of dried soil layers on the Loess Plateau of China
J. Hydrol.
A new index to quantify dried soil layers in water-limited ecosystems: a case study on the Chinese loess plateau
Geoderma
Long-term temporal variations of soil water content under different vegetation types in the Loess Plateau, China
Catena
Using pedotransfer functions to estimate soil hydraulic conductivity in the Loess Plateau of China
Catena
Particle size distribution of soils (0–500 cm) in the Loess Plateau, China
Geoderma Regional
Tree mortality predicted from drought-induced vascular damage
Nat. Geosci.
Environmental data - water scarcity: forecasting the future with spotty data
Science
Multimodel inference: understanding AIC and BIC in model selection
Sociol. Method. Res.
Predicting the saturated hydraulic conductivity of soils: a review
B. Eng. Geol. Environ.
Comparison of methods to estimate saturated hydraulic conductivity in Texas soils with grass
J. Irrig. Drain. E-Asce.
Cited by (15)
Deep soil water use of old-aged vegetation (17- to 36-year stand age) after the formation of dried soil layers based on in situ monitoring
2023, Journal of Hydrology: Regional StudiesPrediction of deep soil water content (0–5 m) with in-situ and remote sensing data
2023, CatenaCitation Excerpt :Five vegetation zones, forest, forest steppe, typical-steppe, desert-steppe and steppe-desert, are distributed from southeast to northwest (Feng et al., 2017). More complete details on the study area have been published in past studies (Fu et al., 2011; Zhao et al., 2020). A total of 86 aluminum neutron-probe access tubes (5.0 m in length) were installed at 10 km intervals from south to north across the CLP transect (Fig. 1).
Building pedotransfer functions for estimating soil erodibility in southeastern China
2022, Ecological IndicatorsCitation Excerpt :MLR and machine-learning technologies are commonly used for building models. MLR has the advantages of simple calculation, strong explanatory power and convenient application and has been widely used to build PTF models for estimating soil properties (Qiao et al., 2019a, 2019b; Ostovari et al., 2019; Zhang et al., 2019b; Barbosa et al., 2019; Rivera and Bonilla, 2020; Zhao et al., 2020). The machine-learning technologies including adaptive neuro-fuzzy inference systems, artificial neural networks (ANNs), support vector machines and random forest techniques (Szabo et al., 2019; Rivera and Bonilla, 2020).
The coupling interaction of soil organic carbon stock and water storage after vegetation restoration on the Loess Plateau, China
2022, Journal of Environmental ManagementCitation Excerpt :However, Feng et al. (2016) noted that vegetation restoration has approached the sustainable water resource limit. Zhao et al. (2020) noted that the drastic increase in vegetation has led to severe soil water depletion and widespread dry soil layers. It is worth noting that the decreasing SWS has become the main obstacle hampering vegetation restoration, thereby threatening SOC sequestration (Zhang et al., 2016, 2020b; Feng et al., 2017; Wu et al., 2018).
Prediction of the van Genuchten model soil hydraulic parameters for the 5-m soil profile in China's Loess Plateau
2022, CatenaCitation Excerpt :Field measurements are direct and highly precise, but are time-consuming, labour-intensive, and costly, especially when performed over long-term period and at large spatial scales. Numerical simulations can alternatively be used for identification of missing hydrological processes by using input data (Romano, 2014; Zhao et al., 2020a,b). In numerical simulations, the processes of water flow are expressed by mainly using the water budget method or fundamental equations such as the Richards’ equation.
Soil enzyme activities of typical plant communities after vegetation restoration on the Loess Plateau, China
2022, Applied Soil EcologyCitation Excerpt :On the Loess Plateau, even though vegetation restoration has resulted in abundant nutrient inputs, potential water shortages have gradually emerged in recent decades (Feng et al., 2016; Feng et al., 2017). In particular, plants with exuberate root systems can contribute to the generation of dry soil layers, which substantially aggravate the root absorption capacity, causing severe vegetation degradation (Li et al., 2017; Xiao and Hu, 2017; Zhang et al., 2019; Zhao et al., 2020). Therefore, it is of great necessity to match suitable species with afforestation sites when implementing vegetation restoration (C. Zhang et al., 2020).