Introduction

Soil water plays a vital role in vegetation growth, food security, and the functionality of arid and semi-arid terrestrial ecosystems1,2,3,4. However, soil water is highly variable due to factors such as precipitation, evapotranspiration, and plant characteristics. Insufficient and unevenly distributed rainfall throughout the year poses significant challenges to plant growth, leading to deep soil desiccation5,6. This desiccation significantly impacts ecohydrological processes by hindering or slowing down the movement of soil water to deeper layers7,8,9. Additionally, the long-term use of nitrogen fertilizers can result in excessive nitrogen accumulation in the soil, which may leach into deeper layers and cause environmental issues such as surface water and groundwater pollution10. Therefore, understanding and addressing environmental concerns like soil desiccation and nitrate leaching is essential to promoting the stability and sustainable development of ecosystems in arid and semi-arid regions.

As vegetation cover increases, particularly in artificial forests that depend on water extracted from deep soil layers (> 20 m) for their growth, the soil has become extremely dry11. This has led to deep soil water deficiencies and the formation of dried soil layers over time4,11,12. Consequently, there is a surge of interest in innovative methods to replenish soil moisture, such as tree felling, agroforestry, and altering land use patterns13,14. Tree felling has emerged as a widely adopted strategy to combat soil water deficits and prevent drought-induced tree mortality15,16. Research by Cui17 demonstrates that tree thinning can significantly enhance soil water infiltration rates and cumulative infiltration, effectively replenishing soil water storage in the short term. Additionally, agroforestry offers promising solutions to tree water shortages18,19. For example, aged apple trees tap into deep soil water and surrounding farmlands for their growth20,21. Moreover, transitioning from deep-rooted to shallow-rooted crops can mitigate soil water shortages. Bai22 found that soil water in the 0–4 m layers can be fully restored by converting shrublands to grasslands. However, it takes approximately seven years for soil water to recover in the 0–5 m layers after artificial forests are converted to grasslands in the same region23. Liu and Shao24 reported that soil water in the 0–10 m layers would take 6.5 to 19.5 years to recover following the conversion of a 30-year-old apple orchard to winter wheat cropland. Thus, converting apple orchards (deep-rooted) to cropland (shallow-rooted) could be a novel approach to reducing deep soil desiccation and boosting agricultural production in arid and semi-arid agroforestry systems. By exploring these innovative strategies, we can address the pressing issue of soil water deficiency, paving the way for sustainable agricultural practices and environmental conservation.

The Chinese Loess Plateau (CLP), nestled in an arid and semi-arid region, grappled with the dual challenges of scarce water resources and fragile ecosystems, further aggravated by severe soil erosion. To combat this, extensive vegetation restoration initiatives have been launched over the past decades, transforming sloping agricultural lands into verdant expanses of trees and grass4,25. Simultaneously, the region’s ideal climate for apple orchards and their lucrative economic returns have driven local farmers to convert vast tableland croplands into thriving apple orchards in the tableland region of the CLP26,27. By 2016, these orchards spanned 25.2% of the CLP’s cropland area, contributing an impressive 26.3% to China’s total apple productio28. Unlike traditional crops, trees boast deeper root systems, leading to heightened soil water consumption7. With regional precipitation averaging only 500–600 mm, insufficient for tree growth, these trees tap into deeper soil water reserves, progressively causing soil desiccation4,11. This relentless soil drying triggers drought stress, curtails the fruiting period, diminishes productivity, and fosters frequent disease outbreaks29. Coupled with a dwindling labor force, this often necessitates the clearing of apple orchards to replant more manageable crops like corn or winter wheat30. Furthermore, the drive for high yields has led to the application of excessive fertilizers in apple orchards, three to four times higher than in traditional croplands31. In agricultural production, the soil environmental pollution caused by nitrate leaching has been increasing serious, especially in orchards. Despite the fact that groundwater in CLP is deeply buried and cannot be easily replenished by rainfall22, researchers have found that an increasing number of groundwater samples in this region exceed the WHO drinking water standard for nitrate (10 mg NO3−-N)32. However, the removal of trees increases the permeability of the soil in the short term, which can easily lead to the infiltration of soil nitrate through water into deeper soils and groundwater and pose a significant environmental risk. Therefore, given that this large-scale conversion is a recent phenomenon in the loess tableland region, the effects of vegetation conversion on soil water and nutrient distribution, particularly in terms of depth and recovery time, remain largely unexplored. Delving into this research not only deepens our understanding of the interplay between deep soil water and nutrient dynamics under vegetation conversion but also bolsters the stability of regional cropland productivity and the sustainable development of ecosystems. This knowledge is pivotal for ensuring the long-term viability of agricultural practices and ecological balance in the CLP.

To bridge current knowledge gaps, this study conducted field experiments in the loess tableland region to investigate the impact of converting orchards to croplands on soil water and nutrient dynamics. Soil samples were collected from a 0–15 m soil profile to analyze various soil indicators. Initially, we assessed the soil water content deficit in a 21-year-old apple orchard and hypothesized that this deficit would be replenished within five years following orchard-to-cropland conversion. Additionally, we posited that nitrate would leach into deeper soil layers with water infiltration. The objectives of this study were to: (1) clarify the relationship between soil water recovery rates and depths over the years following orchard removal; (2) examine the relationship between the residual nitrate peak and leaching depth over the recovery years and their environmental implications; (3) identify the primary factors influencing soil water recovery and residual nitrate leaching. This research aims to provide valuable insights into the role of orchard-to-cropland conversion in soil water replenishment and the management of residual nitrate in arid and semi-arid regions.

Materials and methods

Study area

The study was carried out at the Changwu Agro-Ecological Experimental Station, situated in Changwu County, Shaanxi Province, China (35° 14′ N, 107° 40′ E, 1200 m a.s.l.). This station is positioned within the tableland region of the CLP. The region experiences a continental monsoon climate, with an average annual temperature and precipitation of 9.1 °C and 568 mm, respectively, based on data from 1992 to 2021. Precipitation exhibits significant annual variability, with over half of the total annual rain falling between July and September. The area receives roughly 2230 h of sunshine annually. The local terrain is relatively flat, with a substantial soil layer, and the primary soil type is loess, featuring a consistent soil texture across different locations and through the year4. Groundwater is about 50–80 m below the land surface in this area33, the primary water supply for vegetation is precipitation, and there is no irrigation. Croplands, apple orchards, and grasslands serve as the principal types of land use in this region.

Fig. 1
figure 1

Location of the study area on the Loess Plateau and the annual temperature and precipitation from 1992 to 2022 (the data were from the Changwu Agro-Ecological Experimental Station).

Full size image

Experimental plot selection for orchard and cropland

To determine the influence of orchards to cropland conversion on soil water and nitrate levels, we selected orchards in their prime fruiting period and nearby croplands that was formerly an orchard, now used for continuous maize cultivation, with varying periods of recovery since the orchard’s removal, as our research subjects in July 2022. We employed the space-time substitution approach to select sampling locations across different recovery durations, to assess how these durations impacted soil water content (SWC), soil organic carbon (SOC), and soil nitrate-nitrogen (NO3−-N) levels. Our study included a long-standing orchard established in the year 2000, as well as croplands that had been growing maize for 1-, 3-, 5-, and 10-years post-orchard conversion (with the orchards having been cleared in 2021, at 21 years of age). The recovery periods were identified based on discussions with local farmers. For the sake of contrast, the apple orchards were labeled as AO, while the croplands reclaimed 1-, 3-, 5-, and 10-years post-orchard removal were labeled RC1, RC3, RC5, and RC10, respectively. A control cropland (CK), which had been continuously cultivating maize since 1966, was selected to represent the long-term soil water and nutrient status of the original croplands. These sites, all within a 1 km2 area on the Loess Plateau, shared similar core characteristics (altitude, exposure, gradient, and relief) save for their recovery periods.

Sample collection and processing

For each sampling site, disturbed samples from 0 to 15 m soil depths were collected using a hollow-stem hand soil auger (4.5 cm in diameter) every 20 cm. The soil samples were divided into two parts: one portion was placed in an aluminium box and subjected drying (105 °C, 24 h) to determine the gravimetric SWC (g g−1, %)34; the other portion was air-dried, ground, and passed through 1 mm and 0.25 mm fine sieves to determine the soil particle size composition and nutrients, respectively. The soil particle size composition was analysed by Marvin laser particle size analyser (Masterizer 2000, Malvern Instruments, England) divided into clay (< 0.002 mm), silt (0.002–0.02 mm), and sand (> 0.02 mm) contents according to the American agricultural classification35. The differences in soil particle composition between orchards and croplands in the region were small, and soil bulk density (BD) varied less than 0.01 horizontally in the deep profile36. Therefore, the study cited BD of orchards and croplands from the data measured by Lu37. NO3−-N was measured using a continuous flow analyser (Auto-Analyser-AA3, Seal Analytical Norderstedt, Germany)38, and SOC was determined by the Walkley-Black method39.

To evaluate the dynamics of deep soil water deficit (SWD) and soil water storage (SWS) with recovery years of cropland, we focused on SWC in the 3–15 m soil layers, since the 0–3 m soil layers were more susceptible to single rainfall events40. The SWD was calculated as follows:

$$\begin{array}{*{20}{c}} {SWD=SW{S_i} - SW{S_F}} \end{array}$$
(1)
$$\begin{array}{*{20}{c}} {SW{S_i}=\sum {\theta _i} \times {h_i} \times {{10}^{ - 1}}} \end{array}$$
(2)

Where \({SWS}_{i}\) is the soil water storage of site (mm); \({SWS}_{F}\)is the soil water storage of the control cropland (mm); \({\theta}_{i}\)is the soil water content of site (%); and \({h}_{i}\) is the thickness of the soil layer (cm).

The soil water recovery (SWR) was calculated as follows:

$$\begin{array}{*{20}{c}} {SWR=\mathop \smallint \limits_{0}^{L} \left( {{\theta _{t2}} - {\theta _{t1}}} \right)dz} \end{array}$$
(3)

Where \(SWR\) is the soil water recovery (mm); \({\theta}_{t2}\)is the soil water content of returning cropland (%); \({\theta}_{t1}\)is the soil water content of the apple orchard (%); \(L\) is the thickness of the soil layer (cm).

Data analysis

One-way analysis of variance (ANOVA) was used to analyzed the effects of recovery years and soil depths on the SWC, SOC, and NO3−-N. Regression analysis was used to analyse the relationship between SWR, SWD, and soil water infiltration depth and recovery years. The significant differences in measured variables were determined with Duncan’s multiple range tests at P < 0.05 and P < 0.01 All statistical analyses were performed using SPSS 22.0 (SPSS Inc., Chicago, IL, USA). A map (Fig. 1) was created in ArcGIS 10.5 (Environmental Systems Resource Institute, USA), and the other charts were generated by Origin 2018 (Origin Lab, Northampton, ME, USA).

Structural equation modeling (SEM) was used to determine the standard total effects of the variables on the spatial variations of SWC using Amos 24 (IBM Corporation Software Group, NY, USA), which showed systematic perspectives and driving mechanisms across different variables by partitioning direct and indirect effects. The fitness of SEM was evaluated using the chi-square test (χ2), goodness-of-fit index (GFI), Adjusted Goodness-of-Fit Index (AGFI), and root mean square error of approximation (RMSEA)41.

Results

Spatial variations of SWC in orchard-converted croplands with different recovery years

The SWC greatly varied with the recovery years of orchards to croplands (Fig. 2). The average SWC of AO (13.91%) in the 0–15 m profile was significantly lower than that of croplands with different recovery years. In addition, no obvious variation was noted in average SWCs (16.54% and 16.23%) between the croplands with RC1 and RC3, similar variations of SWCs (17.74% and 17.92%) were also found in RC5 and RC10 with longer recovery years. However, the SWC in croplands with the longer recovering years (RC5 and RC10) was significantly higher than that of shorter recovery years (RC1 and RC3) (P < 0.05). Moreover, the study found that the soil water mainly recovered in the 6–9 m profile of cropland after returning from orchards. Specifically, the SWC of 6–9 m profile maximumly increased from 11.82% (AO) to 17.86% (RC1), 17.86% (RC3), 21.09% (RC3), and 21.39% (RC5), respectively.

Fig. 2
figure 2

Spatial distribution of the soil water content (SWC) of apple orchard (AO) and returned croplands with recovery years of 1,3,5, and 10 (RC1, RC3, RC5, and RC 10), and control cropland (CK).

Full size image

Soil depth also exerted a substantial influence on the SWC of returning orchards to cropland (Table 1). The SWCs of orchard and cropland showed an “S-shaped” distribution trend in 0–15 m profile, which increased and then decreased with the increase of soil depth (Fig. 3). Specifically, the SWCs in 0–3 m profile were significantly lower than in the 3–6 m profile in AO and cropland. In particular, the SWCs of AO, RC1 and RC3 reached the maximum value in 0–15 m profile at 3–6 m were 17.77%, 21.15% and 20.64%, respectively, while the highest SWC of RC5 and RC10 were 21.09% and 21.52% at 6–9 m, respectively. As the recovery years increased, the deep SWC of the cropland returned from orchards increased gradually, but the deep SWC of RC10 was significantly lower than that of CK.

Table 1 Mean soil water content (SWC) within 0–15 m soil layer of returning apple orchards to cropland.
Full size table

Spatial variations of SWR and SWD in cropland with different recovery years

The SWR of returned cropland was closely related to the recovery years (Fig. 3). Compared to AO, the SWR of returned cropland exhibited a ‘logarithmic’ distribution (R2 = 0.91), which showed a rapid upward trend then stabilized with the increase of recovery years. The SWR of 3–15 m in the recovered croplands increased from 45.33 mm (RC1) to 70.86 mm (RC10), the increasing rate of SWR slowed down when the recovery year was more than 5. Specifically, the SWR reached its highest value at 6–9 m, and the recovery amounts of RC1, RC3, RC5, and RC10 accounted for 41%, 43%, 37%, and 41% of the recovery amounts of 3–15 m. Overall, the annual increase in the SWR in recovered croplands was 0.59 mm m−1 a−1. While the SWD showed the opposite downwards trend then stabilized. The SWD of AO (112.7 mm) was higher than that of other croplands. Meanwhile, the SWD of the 3–15 m in the orchard decreased from 9.39 mm m−1 (AO) to 3.44 mm m−1 (RC10). In addition, the results showed that the soil water infiltration depth of recovered cropland had the same logarithmic trend of change with SWR (R2 = 0.92). With the increasing recovery years, the soil water infiltration depths rapidly deepened to 6.2 m (RC1) and 6.3 m (RC3), then slowly increased in 5 years after orchards fell, deepened to 8.2 m (RC5) and 8.6 m (RC10).

Fig. 3
figure 3

The vertical distribution of SWC in 0–15 m profile (a). AO represents the 23-year-old apple orchard, RC1, RC3, RC5 and RC10 represent the years of returning orchard to cropland for 1, 3, 5 and 10 years, respectively. The bule area represents the soil water recovery (SWR) in recovered croplands compared to orchard, the gray area represents the soil water deficit (SWD) in recovered croplands compared to CK. The soil water deficit (SWD), soil water recovery (SWR) and soil water infiltration depth of orchard and croplands with recovery years.

Full size image

Spatial variation of soil NO3 −-N in AO and croplands with different recovery years

The NO3−-N value for both AO and croplands with different recovery years in the 0–15 m soil profile is plotted in Fig. 4a, b. There were similar vertical variations in NO3−-N for the two land use types, increasing first and then decreasing with increasing soil depths and being relatively low and stable in the deeper profile. The average content of NO3−-N in AO (104.32 mg kg−1), RC1 (77.53 mg kg−1), RC3 (89.16 mg kg−1), RC5 (110.26 mg kg−1) and RC10 (91.46 mg kg−1) were significantly higher than that of in CK (5.93 mg kg−1) (P < 0.05). The peak of nitrate content in AO and recovered cropland was significantly higher than that in the CK (P < 0.05). Significantly linear relationships were found between the peak depths and recovery years (R2 = 0.92, P < 0.05), with the peak depths in AO, RC1, RC3, RC5 and RC10 were 3.4 m, 4.8 m, 4.4 m, 7.0 m and 5.6 m, respectively, and reached its highest value in CK (9.4 m), with a peak downward rate of 0.72 m a−1 after orchard removal.

Fig. 4
figure 4

Illustration of nitrate characteristics in the soil profile, including peak depths, peaks of nitrate contents, nitrate leaching depths (a). Data on AO and CK are used as an example. Vertical distribution of soil nitrate nitrogen (NO3−-N) in AO and croplands with different recovery years in 0–15 m profile (b) (The small plot represents the linear fit between the peak depth of NO3−-N and the recovery years).

Full size image

Response of SWC and NO3 −-N dynamics to environmental factors

To analyze the influence of environmental factors on SWC and NO3−-N dynamics, we separately assessed the direct and indirect path correlation coefficients using SEM modeling (Fig. 5). The results showed that the relationship between and the driving factors could be well fitted by established SEM models, with RMSEA ≤ 0.060 and GFI ≥ 0.938 for the various soil layers (Fig. 5). As shown in Fig. 5, recovery years, soil depth, SOC, and BD had both direct and indirect effects on SWC. Recovery years, clay and SOC exhibited positive standardized total effects on SWC, soil depth only had positive effects on SWC in 0–5 m soil profile (Table 2). In 0–5 m profile, soil depth directly affected the SWC with a direct effect of 0.988, and indirectly decreased the SOC by affecting SWC, with an indirect effect of 0.414 (Table 2). Therefore, the positive standardized total effect of SWC on soil depth (0–5 m) was 0.643. In contrast, the direct path relationship between recovery years and SWC was not significantly. With the increasing soil depth, recovery years were the most important environmental determinant of the SWC in the 5–15 m soil profile, with the standardized total effect was 0.57 and 0.38 in 5–10 m and 10–15 m soil profile, respectively.

Similarly, path analysis identified recovery years, soil depth, clay, and SOC as being factors having the most important influence on NO3−-N (Fig. 5), among which, soil depth, with a direct path coefficient of 0.71 and 0.65 in 0–5 m and 5–10 m soil profile, respectively, appeared to play a decisive positive role in 0–5 m but negative role in 5–10 m in determining NO3−-N. Conversely, recovery years, BD, and SWC, with direct path coefficients of 0.21, 0.22, and 0.47 in 5–10 m soil profile, respectively, were shown to have positive effects on NO3−-N.

Fig. 5
figure 5

Structural equation modeling (SEM) used to examine standard total effects on SWC and NO3−-N respectively for the 0-5 (a), 5-10 (b), 10-15 (c) soil profile. The effects of the recovery years, soil depth and soil factors (i.e., clay content - CL and sand content - SL) on SWC and NO3−-N as determined SEM for different soil layers. Red and blue arrows represent.

Full size image
Table 2 Standardized total, direct and indirect effects of the driving factors of SWC (soil water content) in the Loess Tableland study area determined using structural equation models.
Full size table

Discussions

Soil water recovery process and dominated factors in orchard-to-cropland conversion

Long-term cultivation of deep-rooted vegetation can induce or exacerbate permanent soil desiccation13. As apple trees matured, soil water content (SWC) in the apple orchard decreased with increasing soil depth. For instance, with increasing orchard age, it showed a significant decrease in SWC in deep soil and accompanied by a significant increase in soil water deficit21. In the study, the average SWC in a 21-year-old apple orchard (13.68%) within the 3–15 m soil profile was significantly lower compared to cropland (20.89%) (Table 1), resulting in severe water deficit, consistent with prior research findings4,20,32,41. Reclaiming permanent soil desiccation on the CLP is challenging due to generally sparse rainfall, deep-root vegetation consumption, and intense evaporation5,13. Studies have explored methods to mitigate soil desiccation by removing deep-rooted vegetation30,40,42. Addressing the issue of deep soil water desiccation in aging orchards on the CLP, we observed that differences in soil water recovery (SWR) were influenced by recovery years and soil depths. SWR predominantly occurred in the overlap between the depth of apple root growth and the depth unreachable by rainfall infiltration. Specifically, in the 0–3 m soil layer, there was no significant difference in SWC between the original orchard (AO) and the converted croplands (RC). However, deeper soil layers (3–9 m) exhibited rapid SWC replenishment in RC1 and RC3 following orchard-to-cropland conversion. This was attributed to the removal of apple tree roots, which opened up root channels and increased non-capillary soil space, facilitating preferential water pathways and enhancing water infiltration during rainfall events17,43,44. As recovery years increased, the rate of deep SWR decreased progressively from RC5 to RC10, consistent with previous findings45. Four years after conversion, root channels collapsed and were compacted, resulting in a decline in infiltration rate with prolonged cropping. In our study, the inherently loose nature of loess soil meant that disturbances such as orchard removal only had short-term effects on soil physical properties14. This initially impacted deep soil water within five years, as shown by the intercept of the linear relationship between SWR and recovery years post-orchard removal (Fig. 3). In conclusion, orchard-to-cropland conversion effectively replenished deep soil water primarily because mature orchard root systems no longer deplete soil water post-removal, and decayed roots form pore spaces that enhance rainfall infiltration. Our previous research indicated that apple orchards aged 10–30 years experienced an annual soil water deficit rate of 5.94 mm m−1 a−121. Conversely, the calculated soil water recovery rate post-conversion was 5.90 mm m−1 a−1, suggesting that soil desiccation caused by mature orchards can potentially recover within corresponding years following orchard removal.

Nitrate leaching and its potential impact on the environment

The nitrate peak and leaching depth in recovered croplands were significantly correlated with the years since orchard removal. Our study revealed substantial nitrate accumulation in the deep soil profiles of mature apple orchards, with both nitrate peak levels and leaching depths showing a notable upward trend as recovery years increased (P < 0.05). This trend aligned with previous findings46,47,48. In the 0–5 m soil profile, intensive nitrogen fertilizer applications and increased water infiltration contributed to significant NO3−-N accumulation49. However, following orchard removal, reduced nitrogen inputs and enhanced soil water recovery led to nitrate dilution and horizontal diffusion21, resulting in decreased NO3−-N concentrations. Persistent nitrogen surplus and water balance issues could perpetuate nitrate accumulation and migration within the soil profile10. The CLP, predominantly rain-fed agriculture, experienced concentrated high-intensity rainfall from July to September, facilitating downward water movement and nitrate leaching10,11. Furthermore, the decomposition of roots post-orchard removal created large pores where preferential and piston flows coexist, promoting rapid nitrate loss following precipitation events in recovered croplands. Despite dried and clayey soil at 5–10 m soil layers causing slow water transport and long residence time which resulted in the small nitrate leaching rate50,51, the rapid water replenishment under orchard-to-cropland conversion still promoted the leaching48,52,53,54,55. Soil water dynamics were critical in governing NO3−-N movement56,57. Orchard removal enhanced soil water infiltration and increased soil water storage in deep profiles, yet it also accelerated downward nitrate leaching, potentially exerting prolonged impacts on soil NO3−-N and posing risks to groundwater quality.

Trade-off between soil water recovery and nitrate leaching

This study demonstrates that long-term apple orchard cultivation leads to decreased soil water content (SWC) in deeper layers, resulting in soil desiccation that partially limits nitrate leaching into the deep soil. Conversely, following orchard removal, croplands rapidly replenish soil water, alleviating soil deficits initially caused by orchard-induced desiccation. Orchard-to-cropland conversion expands cultivable land and mitigates economic losses associated with declining or dead apple orchards. However, due to substantial differences in nutrient requirements between apple trees and crops, coupled with increased water inputs that exacerbate surplus nitrate leaching and accumulation in deep soil, there is a significant threat to groundwater quality, presenting a notable environmental concern for the CLP. Clearly, there is a trade-off between soil water recovery and nitrate leaching as orchard converting to cropland. Further research is imperative to understand the complex interplay between soil water dynamics and nitrate behavior. Optimizing nitrogen management post-conversion, possibly through cultivation of crops with higher nitrate uptake efficiency, represents a promising avenue for addressing these challenges.

Conclusions

In this study, the space-for-time substitution method was employed to address the effects of apple orchard (> 20 years old) to croplands conversion on soil water content (SWC) and soil nitrate dynamics. The findings revealed that converting orchards to croplands led to rapid replenishment of previously desiccated deep soil layers (6–9 m) within five years following the removal of apple trees, followed by a stable but gradual increase thereafter. With increasing years since the conversion, enhanced soil water infiltration significantly contributed to nitrate leaching and accumulation in deeper soil layers. And SWC emerged as the principal factor influencing soil nitrate leaching. The orchard to cropland conversion induced a trade-off between soil water recovery and nitrate leaching. These insights highlight that, following large-scale conversion of orchards to croplands, while soil water in the dried soil layers is restored, nitrogen leaching also occurs. This is of significant value for future water and nutrient management after orchard-to-cropland conversions on the CLP.