Significant soil degradation is associated with intensive vegetable cropping in a subtropical area: a case study in southwestern China

Within the context of sustainable development, soil degradation driven by land use change is considered a serious global problem, but the conversion from growing cereals to vegetables is a change that has received limited attention, especially in subtropical regions. Here, we studied the effects of the conversion from paddy rice to an oilseed rape rotation to vegetable production in southwestern China on soil organic carbon (SOC), total nitrogen (TN), the C/N ratio, pH, phosphorus (P), potassium (K), calcium (Ca), and magnesium (Mg) based on face-to-face farmer surveys and soil analysis. In the vegetable cropping system, fertilizer application often exceeds the crop demand or levels recommended by the local extension service several times over. Thus, the crop use efficiency of N, P, K, Ca, and Mg was only 26 %, 8 %, 56 %, 23 %, and 28 %, respectively. In the vegetable cropping system studied, SOC, C stock, TN, and N stock were decreased significantly due to low organic inputs from crop residues and high tillage frequency. Furthermore, the soil C/N ratio decreased slightly; available P (AP) in the topsoil increased by 1.92 mg kg−1 for every 100 kg ha−1 of P surplus, and the critical levels of AP and CaCl2-soluble P in P leaching were 104 and 0.80 mg P kg−1. Besides, compared to the current paddy–rape rotation system, a clear trend of soil acidification was observed in the vegetable fields. However, increasing the contents of soil Ca and Mg significantly alleviated topsoil acidification, with the effect increasing over time. Given our findings, the potential benefits of conservation agricultural practices, integrated soil–crop system management strategies, and agricultural technology services for recovering the degraded soil and improving the vegetable productivity are discussed here. Published by Copernicus Publications on behalf of the European Geosciences Union. 334 M. Lu et al.: Significant soil degradation is associated with intensive vegetable cropping

. The terrestrial expansion of vegetable production is particularly significant in the tropics and subtropics because these regions have a longer growing season compared to temperate regions (Fischer et al., 2012). In contrast to cereal crops, vegetable production requires more intensive management, including higher tillage frequencies and fertilization rates, and higher nutrient and water inputs due to the shallow root systems of most vegetable species (Wang et al, 2018a, b;Yan et al., 35 2013;Zhang et al., 2017). Thus, vegetable production is associated with a higher risk of environmental pollution. Zhang et al. (2012) found that the per unit nutrient inputs in vegetable production are 81-120% greater than those for cereal production system in China and 1.5-12 times greater than those for vegetable cropping in the rest of the world. In particular, the overuse or misuse of fertilizers in vegetable fields (VF) is a long-standing problem (Conley et al., 2009).
Soil organic carbon (SOC) is widely regarded to be the most reliable indicator for monitoring soil degradation (Lal, 2015). 40 Bünemann et al. (2018) reviewed soil quality assessment approaches and summarized the indicators used; after SOC, the most frequently used chemical indicators were soil pH, available phosphorus (P), available potassium (K), total nitrogen (TN) and bulk density.
The impacts of land conversion from cereal to vegetable production may be more adverse in tropical and subtropical regions compared to temperate regions due to higher temperatures and rainfall combined with low organic inputs and high frequency 45 of tillage practices (NAPCC, 2016;Powlson et al., 2016;Sarker et al., 2018). For instance, Wang et al. (2014) found that SOC concentration and the C/N ratio in open VF converted from paddy rice fields decreased by 19.7% and 27.8%, respectively, in southern China.
High inputs of N fertilizers to vegetable crops and the crop removal of base cations such as calcium (Ca) and magnesium (Mg) are expected to contribute to soil acidification (Guo et al., 2010;Zhang et al., 2016). However, in southwest China, 50 fields receive large inputs of Ca and Mg from various fertilizers used for supplying micronutrients. To date, the combined impacts of these inputs and the removal of base cations on soil pH and other properties in vegetable systems have not been studied.
Systematic surveys to elucidate the effects of different cropping systems on soil properties are needed for a clear understanding of the impacts of vegetable production and the proper development of sustainable high-yield vegetable 55 systems that do not cause environmental damage. In the present study, paddy rice-oilseed rape rotation fields (PRF) were used as a reference, because all VF in the region were converted from PRF. The objectives of this study were to (i) summarize the seasonal input-output apparent balance of N, P, K, Ca and Mg in PRF and VF, and (ii) assess the impacts of land-use conversion (from PRF to VF) on soil pH, SOC, TN, soil C/N ratio, available P (AP), CaCl2-soluble P (CaCl2-P), and available base cations (K, Ca, and Mg).

Site descriptions and cropping systems
The field study sites were located in Jinping County in eastern Guizhou Province at 26° 23´ -26° 46´ N and 108° 48´ -109° 24´ E ( Figure 1A). In this area, VF covered an area of 4,814 ha, representing an increase of 24% between 2009 and 2015, and production increased by 132% between 2011 and 2018 (84,886 t). The area covered by cereal crops (14,408 ha at the 65 time of our study) decreased by 9% from 2009 to 2018. The region is characterized by a typical subtropical humid monsoon climate with an annual average precipitation of 1160 mm for the 2007-2017 period, of which 85% occurs between April and October, and a mean annual air temperature of 15.9°C ( Figure 1B). The highest temperatures and rainfall occur in the same season. The main soil type is yellow earth (Ultisol), and the main crops grown are open-field vegetables, paddy rice, oilseed rape, and fruit trees. 70 Open-field vegetable production accounts for 70-80% of the total area of dryland agricultural production, and the most common planting pattern (71% of the total VF area) is pepper-Chinese cabbage rotation system. In this cropping system, peppers are transplanted in April and harvested in mid-August, then non-heading Chinese cabbage is grown continually for three seasons, from August to January of the following year, followed by a fallow period until April ( Figure 1B). During the annual rain-fed cropping period, tillage is performed four times, and fertilizers are applied 10 times (four times for pepper, 75 and twice for each Chinese cabbage crop). Conventional farming practices commonly used throughout the region were used for all crops. In the paddy rice-oilseed rape double-cropping system, paddy rice is grown from late May to September, and then oilseed rape is grown from mid-October to early May of the following year.

Farmer surveys and soil sampling
In this study, the focus was only on the typical vegetable production (pepper -Chinese cabbage -Chinese cabbage -Chinese 80 cabbage -fallow rotation) and the paddy rice-oilseed rape rotation. Based on farmers' survey methodology (Jia et al., 2013), to obtain representative results, two of the most important vegetable-production townships in Jinping County-Xinhua and Dunzha-were selected for farmer surveys. In each township, four villages were randomly selected, and 15-20 farmers from each village who had managed PRF before switching to VF were randomly surveyed. Survey questions were related to crop varieties and the management of the two cropping systems, including sowing rate, fertilizer (chemical or organic) application 85 rate, tillage frequency, crop-residue management, crop growth cycle, cropping duration, and yield. A total of 24 surveys were related to PRF management and 133 surveys to VF management.
To investigate the effects of long-term fertilization on nutrient surpluses, enrichment/depletion, and downward transport in the soil profile in the VF, we collected soil samples randomly from 36 VF that were converted from PRF 1-3 years (n=12), 5-10 years (n=12), and ≥15 years (n=12) ago. Twelve adjacent PRF were also sampled for comparison; the soil type at all 90 sites was very similar. Soil samples were collected at six spots at 20-cm depth intervals to a depth of 60 cm within each plot and pooled by depth. Samples were collected at all selected VF and PRF from September 25 to 31, 2018, after the crops https://doi.org/10.5194/soil-2021-17 Preprint. Discussion started: 8 March 2021 c Author(s) 2021. CC BY 4.0 License.
were harvested. In addition, three other soil cores were collected randomly from each field type at the same depths for bulk density measurements.

Soil analysis 95
In this study, the moist composite fields samples were air dried and passed through a 1-mm sieve for the analysis of soil pH; CaCl2-P; AP (determined using the Bray method); water-soluble K, Ca, and Mg; and exchangeable K, Ca, and Mg. The samples were then passed through a 0.15-mm sieve for SOC and TN analysis. Soil pH in a 1:2.5 soil-to-water suspension was measured using a potentiometer, SOC was measured using dichromate digestion, and TN was measured using the semiautomatic Kjeldahl method. CaCl2-P was extracted from a 1:5 soil-to-water suspension using 0.01 mol L -1 CaCl2, AP was 100 extracted using 0.03 mol L -1 NH4F and 0.025 mol L -1 HCl, water-soluble K, Ca, and Mg were extracted in the water portion of a 1:5 soil-to-water suspension, and exchangeable K, Ca, and Mg were extracted with 1 mol L -1 neutral NH4OAC (Lu, 2000).

Calculation of nutrient inputs and balance
Detailed records were made of the farm management practices employed in each field from 2017 to 2018, and we confirmed 105 that the management history of each field was similar to the management practices employed during the survey year. The main mineral fertilizers used in this region were urea, potassium sulfate, compound or mixed fertilizers, calcium superphosphate and fused calcium-magnesium phosphate. Organic fertilizers included crop residues and commercial organic fertilizer, animal manure (from chickens, cows, and pigs), and rapeseed meal. In the PRF, rice straw was directly returned after the grains were harvested, whereas oilseed rape straw was burned in the fields. Most above-and below-ground residues 110 in the VF were removed during harvest.
Nutrient removal due to grain harvest or the removal of aboveground crop parts was calculated using yields reported by farmers and published nutrient concentrations typical of the region. Nutrient surpluses were calculated by subtracting nutrient removal due to grain harvest or aboveground removal from total fertilizer inputs.

Calculation of SOC and TN stock, accumulation of P surplus and changes in soil AP 115
The SOC and TN stocks in the PRF and VF soils were calculated by multiplying SOC and TN concentrations with bulk density and depth. Accumulation of surplus P was calculated as the annual P surplus multiplied by planting duration, and changes in soil AP in the VF were calculated as soil AP in the PRF subtracted from that in the VF.

Statistical analysis
SPSS software (ver. 21.0 for Windows; IBM Corporation, Armonk, NY, USA) was used for statistical analysis. Differences 120 in soil nutrient concentrations and stock among cropping systems were assessed using one-way analysis of variance, and mean values were compared using the least significant difference test with a significance level of 5%. Correlations were https://doi.org/10.5194/soil-2021-17 Preprint. Discussion started: 8 March 2021 c Author(s) 2021. CC BY 4.0 License. analyzed using a two-tailed Pearson test with a significance level of 5%. The linear-plateau model in SAS (ver. 9.2 for Windows; SAS Institute, Cary, NC, USA) was used to analyse the relationship between SOC and tillage frequency in the 0-20-cm soil layer in the VF, and the two-segment linear model in SigmaPlot (ver. 12.5 for Windows; Systat Software, San 125 Jose, CA, USA) was used to analyse the relationship between soil AP and CaCl2-P.

Yields, fertilizer use and nutrient balance in two intensive cropping systems
Total annual inputs of N, P, Ca, and Mg in the VF were several times higher than those in the PRF (2.38, 2.97, 4.40 and 7.14 times, respectively), but the total annual K inputs were similar (Table 1). K, Ca, and Mg inputs in the PRF were mainly from 130 organic sources, i.e., a combination of organic fertilizers and crop residues, representing 65%, 80% and 65% of the total inputs, respectively, while chemical fertilizer supplied 61% and 72% of total N and P inputs. In contrast, in the VF, these nutrients were mainly supplied by chemical fertilizers, which comprised 75%, 86%, 71%, 75%, and 79% of the total inputs for N, P, K, Ca, and Mg, respectively. Fertilizer application rates did not differ significantly among VF according to the time since conversion from PRF (Table 1). Vegetable yields also did not differ significantly. 135 In the PRF, straw was returned to the field after grain harvest, whereas in the VF, all plant biomass was completely removed during harvest. Consequently, 2.46, 9.05, 14.0, and 4.52 times more N, K, Ca, and Mg were removed, respectively, from VF than from PRF. However, the P removal rate in the VF was slightly lower. Large surpluses of N (790 kg N ha -1 yr -1 ), P (343 kg P ha -1 yr -1 ), and K (235 kg K ha -1 yr -1 ) were found in the VF. Unexpectedly, there were also large surpluses of Ca (800 kg Ca ha -1 yr -1 ) and Mg (161 kg Mg ha -1 yr -1 ) in the VF compared to the PRF (Table 1). 140

Changes in SOC and TN in VF
The concentration of SOC, TN, C/N ratio, C stock and N stock were compared between VF and PRF (Table 2). Conversion from PRF to VF led to a significant decrease in both TN concentration and stock in the topsoil (0-20 cm), with mean decreases of 15.8% and 17.6% (P＜0.05), respectively. However, conversion to VF did not result in significant changes in TN in the two deeper soil layers. Conversion from PRF to VF led to significant decrease in topsoil SOC concentration and 145 stock, even after only 1-3 years. The C stock in the 0-20 cm soil layer decreased by 10.3% after 1-3 years and increased to 24.3% after 5-10 years. Similar trends were observed in the 20-40 cm and 40-60 cm soil layers. Because proportional decreases in SOC were slightly greater than those for TN, the soil C/N ratio became slightly smaller following conversion to VF, with this trend most apparent in the 40-60 cm soil layer and decreased by 24.3% after 5-10 years.
The average bulk density in the topsoil tended to decrease upon conversion from PRF to VF but did not change significantly 150 according to years of vegetable production. Averaged over all periods of VF, C stock in the 0-60 cm soil layer was 15.9% (P One major reason SOC might have decreased in the VF was that the organic C input in the PRF was 1.35-fold higher than that in the VF (Figure 2A). Tillage practices were also performed more frequently in the VF, which supported four crops per year instead of two in the PRF. Tillage frequency, which was as high as 4 times per year, clearly affected topsoil SOC (0-20 155 cm) but not that in deeper soil layers ( Figure 2B). Moreover, despite the large N surplus (790 kg N ha -1 , assumed to be predominantly nitrate) in VF, soil N stock generally decreased in line with SOC stock, with a slight variation to this trend observed in the 40-60cm layer ( Figure 2C).

Available P changes in VF
Soil AP and CaCl2-P in the 0-60 cm soil layer were significantly higher in the VF than in the PRF, with levels increasing 160 according to time since VF conversion ( Figure 3). For example, in VF converted >15 years ago, soil AP concentration was 22-fold those in PRF and CaCl2-P in the topsoil was 9-fold higher. The increases in deeper soil layers are indicative of downward movement of P. In the 0-20 cm soil layer in the VF, the amount of surplus P accumulated (P inputs minus crop offtakes) was significantly and positively correlated with differences in soil AP between VF and PRF at the end of the growing season (VF, pepper season; PRF, paddy rice season) ( Figure 4A). For every 100 kg ha -1 of surplus P in the VF, AP 165 concentration in the 0-20 cm soil layer increased by 1.92 mg kg -1 . The phenomenon of a "change-point" has previously been observed in the relationship between soil AP and CaCl2-P (Hesketh and Brookes, 2000), and used as an indicator of the risk of P loss. In VF topsoil(0-20 cm layer), AP and CaCl2-P values at the change point as calculated using a two-segment linear model were 104 and 0.80 mg P kg -1 , respectively ( Figure 4B). Above the change-point, soil CaCl2-P increased rapidly with increasing AP. 170

Soil available K, Ca, and Mg in VF
The K nutrient removed via harvested crops in the PRF (33.8 kg K ha -1 yr -1 ) was only 11% of that removed from the VF (average, 306 kg K ha -1 yr -1 ) (Table 1). Thus, the K surplus in the PRF was, on average, 2.36 times that in the VF. However, soil available K levels (water-soluble and exchangeable fractions) were considerably higher in the VF in all three soil layers ( Figure 5A and 5D). The depth distribution of soil available K indicated that there was substantial downward movement of K. 175 The accumulation of available K, but not exchangeable K, in each soil layer increased with the number of planting years since VF conversion. Water-soluble K concentrations in the 0-20, 20-40, 40-60 cm soil layers in VF that had been continuously cultivated for over 15 years were 26%, 40%, and 275% higher, respectively, than those in VF cultivated for 1-3 years, and 193%, 180% and 650% higher, respectively, than those in PRF.
Water-soluble Ca was significantly higher in the VF (0-60 cm) than in the PRF soil ( Figure 5B). Levels increased in the first 180

Changes in soil pH in VF and effect factors
In the PRF, the soil was slightly acidic, especially in the topsoil (0-20 cm), with an average pH of 5.34 across the fields 185 sampled, and initially decreased to approximately pH 4.7 after conversion to VF but recovered eventually ( Figure 6). The initial pH was higher in the deeper soils (~pH 6.5) than in the 0-20 cm layer in the PRF, but it was significantly lower in the VF. The concentration of soil exchangeable Ca and soil pH were significantly and positively correlated in both the VF and PRF, particularly in the VF (Table 3). Soil exchangeable Mg exhibited the same trend in the VF but not in the PRF. In addition, excessive N and K surpluses were significantly correlated with decreased soil pH in the 0-20 cm soil layer. 190 Conversely, the enrichment of soil P had a negative effect on soil pH.

Nutrient balance
Through the farmer surveys, we obtained a considerable amount of data on the fertilizer application rates employed in PRF and VF, the two main cropping systems used in the southwest China. The annual application rates of fertilizer N, P and K in 195 the VF (Table 1) were much higher than those recommended by the extension service (N, 600-765, P, 79-144, K, 398-498 kg ha -1 ) (Zhang et al., 2009). In the PRF, fertilizer N and K inputs were also much higher than the recommended (N 255-330, K 112-174 kg ha -1 ), but the annual fertilizer P rate was close to the local recommended rate (P 33-79 kg ha -1 ) (Zhang et al., 2009). Consequently, nutrient surpluses developed under both cropping systems but were especially large in VF. Thus, nutrient use efficiency (nutrients removed in the harvested crop expressed as a percentage of nutrient input) in the VF was 200 very low. On average, the N, P, K, Ca, and Mg use efficiencies were 26%, 8%, 56%, 23%, and 28%, respectively. Studies assessing vegetable crops at the national scale have reported that the N use efficiency in China is 26 ± 13%, consistent with our finding (Ti et al., 2015). The low P use efficiency found for vegetable crops in the present study is even lower than the values reported in previous studies (15-20%) (Conley et al., 2009;Yan et al., 2013;Zhang et al., 2008).
Crop yields and nutrient use efficiency in VF in southwest China can be limited by a lack of support from local extension 205 services (Ju et al., 2007), including a lack of fertilization recommendations or publicity materials, remoteness of advisers from the production fields, and a lack of training for most farmers. The average farm size in this region (calculated as the weighted mean) is 0.07 to 0.1 ha, and limitations can also be due to small farm size (Zhang, 2017). Farm size is often used as a proxy for the level of professionalism exhibited by farmers and awareness of efficient nutrient-management practices (Ju et al., 2016;Wu et al., 2018). Therefore, smallholder farmers in China generally have a low awareness of fertilizer-saving 210 practices, and farm size significantly influences the usage intensity and efficiency of agricultural chemicals.
Previous work by our group has shown that there is considerable scope for increasing yields and decreasing fertilizer use in vegetable production in this region (Wang, 2018b). Integrated soil-crop system management approaches (ISSM) based on advanced crop and nutrient management have been successfully developed for cereal cropping in China (Chen et al., 2014). The grain yields of maize, rice and wheat can be increased by 35% (n=5,406), 18% (n=6,592) and 24% (n=6,940), 215 respectively, with no increase in fertilizer inputs and often, a decrease instead, using an ISSM approach (Chen et al., 2014). The integrated management system has been developed for greenhouse pepper production in the Yangtze River Basin. After 2 years, pepper yield increased by 16% and N application rates decreased by 54%, due to the optimization of soil management, plant population, and nutrient management (Wang et al., 2020). Thus, local extension services could provide support and resources for applying integrated soil-crop system management methods to increase vegetable production, 220 minimise nutrients surpluses, and mitigate soil degradation in VF in the region.

SOC and N stocks
This study provides unequivocal evidence that converting land from long-established PRF to VF sharply decreased SOC and TN concentrations and stocks. The trend was clearest in the topsoil where initial levels of SOC and TN were the highest, but it was also apparent in the 20-40 and 40-60 cm soil layers (Table 2). SOC tended to be depleted further with an increasing 225 period of vegetable production, particularly in the first 10 years (Table 2), but decreases were generally small thereafter. Low input of organic C from crop residues ( Figure 2A) and a high tillage frequency ( Figure 2B) were two responsible factors caused that. The results agreed with the findings of Wang et al. (2014), who reported that SOC concentration decreased by 25% following conversion from paddy rice to vegetable production in southern China. We estimated that the average annual C input in VF was less than 75% of that in PRF (Figure 2A). A higher frequency of tillage practice results in the break-up of 230 soil macro-aggregates, soil structure damage (Chivenge et al., 2007;Pires et al., 2016;Six et al., 2000), and an increase in soil aeration (Sarker et al., 2018), which promotes microbial decomposition of soil organic matter. Decomposition is further accelerated by high air temperatures and soil moisture levels in sub-tropical regions compared to temperate regions .
Under conventional VF management, at least four tillage operations per year are required. In contrast, an alternative strategy 235 is conservation agriculture (CA), which can create a positive soil/ecosystem C budget that often increases agricultural productivity while sequestering additional C into the C stock, or at least, minimises SOC depletion (Chivenge et  Our results also indicated that soil TN content declined proportionally less than SOC content after conversion to VF. This was accompanied by marked changes in the soil C/N ratio (Table 2). For instance, in the 20-40 cm layer depth, the C/N ratio decreased from 10.5 in PRF to 9.01 in VF where vegetable production had been practiced for >15 years, and a larger change was observed in the 40-60 cm soil layer. Although the changes were significant, the absolute quantities of C and N in the 250 deeper soil layers were small (around 7-8 g C kg -1 soil and 0.9 g N kg -1 soil; Table 2); thus, experimental error from sampling and analysis could be easily magnified when calculating ratios. However, the C/N ratio did seem to exhibit a decreasing trend, which has been noted in other studies of land-use conversion from cereal to vegetable cropping (Ju et al., 2007;Yan et al., 2012). Because of the very large N surplus in the VF (mean over three vegetable cropping periods, 789 kg N ha -1 yr -1 , Table 1), it is perhaps conceivable that a fraction of the inorganic N in the soil profile was immobilized into soil 255 organic matter under the conditions prevailing in the VF, thus counteracting the loss of N that accompanied the observed SOC loss. The slightly lower pH in subsoil in the VF compared to the PRF might have slowed nitrification such that part of the inorganic N remained in the ammonium form sufficiently long enough for some immobilization to occur ( Figure 6).
We did not directly measure N losses at the study sites, but the large N surpluses strongly imply that the losses would have been substantial. Zhang et al (2017) had found that 36 % of the N applied (at a rate of 139 kg N ha -1 ) during the growing 260 season of bitter gourd was lost by leaching. In another intensive vegetable-cropping system, Wang et al (2018a) and Wang et al (2019) measured a loss of 3.91 kg N2O-N ha -1 season -1 and a runoff of 16.5 kg N ha -1 yr -1 . Therefore, losses of this order are to be expected in our study region. Hence, to minimize N loss, which leads to undesirable environmental impacts such as water pollution, eutrophication, and greenhouse gas emissions, rational N fertilizer management for intensive vegetable production in high-rainfall, sub-tropical environments is urgently needed. 265

P enrichment
P enrichment in farmland soils in China has been a serious problem for many years, as the movement of phosphate to surface waters causes environmental pollution (Bai et al., 2013;Cao et al., 2012;Li et al., 2011;Yan et al., 2013). On average, the annual P input (373 kg P ha -1 ; Table 1) in the present study was 2.6~4.7 -fold higher than those recommended for VF (Zhang et al., 2009). Soil AP varied widely and ranged from 14.9 to 229 mg P kg -1 (mean, 120 mg P kg -1 ) in the 0-20 cm soil layer 270 with a large variation, which is several times higher than the critical AP based on yield for fruit vegetables (58 mg P kg -1 ) and leafy vegetables (46 mg P kg -1 ) (Yan et al., 2013). The AP concentration in the topsoil (0-20 cm) increased by 1.92 mg kg -1 for every 100 kg ha -1 of surplus P in the VF. In other soils and cropping systems, the rate of increase ranges from 1.44 to 5.74 mg kg -1 (Cao et al., 2012). A high level of P enrichment in the soil is detrimental to plant growth because it inhibits the rhizosphere manipulation processes employed by plants to efficiently acquire P, including the colonization of roots by 275 arbuscular mycorrhiza fungi and the exudation of organic acids or phosphatase enzymes (Deng et al, 2017;Dong et al., 2004;Ryan and Graham, 2002). In addition, CaCl2-P serves as an indicator of the risk of P loss from leaching (Hesketh and Brookes, 2000). Previous studies have reported a "change-point" in the graph of soil available P against CaCl2-P, above which the risk of P leaching increases greatly (Bai et al., 2013;Shen et al., 2018;Yan et al., 2013). Here, this value was https://doi.org/10.5194/soil-2021-17 Preprint. Discussion started: 8 March 2021 c Author(s) 2021. CC BY 4.0 License. exceeded in the 0-20 cm soil layer in many VF. As such, excessive P fertilizer application should be avoided to reduce the P 280 surplus to zero and P movement to water bodies.

Soil pH respond to land use change
At the national scale, long-term intensive farming has been shown to cause soil acidification, with an average decline of 0.50 in pH in the topsoil of major Chinese croplands from 1980 to 2000 (Guo et al., 2010). Soil acidification is caused by the excessive application of N fertilizers and the loss of base cations via leaching (Guo et al., 2010;Zhang et al., 2016), and 285 leads to decreased crop production and quality (Cakmak and Yazici, 2010). After the conversion from PRF to VF, soil pH at all depths decreased, especially at 1-3 years after conversion ( Figure 6). After this period, soil pH at all depths decreased, especially at 1-3 years after conversion. After this period, the pH in the 0-20 cm soil layer increased slightly and either stayed constant or decreased very slightly in the deeper soil layers. In the VF, exchangeable Mg and, to a lesser extent, Ca, accumulated due to the types of fertilizers used. This apparently counteracted the trend towards acidification, as shown by 290 the correlation between Mg and Ca accumulation and soil pH (Table 3). In this sense, the current fertilization practice does confer benefits in terms of improving soil quality. However, the greater accumulation rate of Mg compared to Ca is likely to cause physiological damage to crops due to the antagonism between the two elements. An imbalance in mineral nutrients in VF soil has been observed in other fields in southwest China.
The mean soil exchangeable Mg concentration in VF converted 1-3 years (73.2 mg kg -1 ) is considered deficient, as it is much 295 lower than the critical threshold (Bai et al., 2004). Cationic antagonism is also an important influencing factor that influences mineral nutrient bioavailability (Tisdale et al., 1993;Marschner, 2012). With the increase of planting duration (1-10 yrs), soil exchangeable K, Ca and Mg concentration gradually increased ( Figure 5); but the growth ratio of each nutrient was not consistent (K > Mg > Ca), and the ratio of Ex-K/Ex-Mg decreased from 1.92 to 1.37 can be observed, while the ratio Ex-Ca/Ex-Mg of remained at 6.14-6.29, with little change. However, it has been reported that plant Mg absorption would be 300 inhibited by soil K and Ca when the value of soil Ex-K/Ex-Mg or Ex-Ca/Ex-Mg greater than 0.6 or 7.0, respectively (Tisdale et al., 1993;Morton et al., 2008). Thus, current vegetable-cropping practices can easily result in Mg physiological deficiencies (Yan et al., 2016). Nowadays, more and more attention has been paid to Mg issues in soil-plant-human continuum, and this intensive vegetable rotation in southwest China urgently needs optimal Mg fertilization strategy.

Conclusions
Open-field vegetable-cropping practices in southwest China can result in significant soil degradation and potential environmental pollution after conversion from a paddy rice-oilseed rape rotation system. Fertilizer application rates in VF are often several times higher than those needed to meet crop requirements or recommended by the local extension service.
Excessive fertilization has led to the substantial soil enrichment of P, K, Ca, and Mg. We also observed a clear trend of soil acidification at 0-60 cm soil layer depths, although this was partially alleviated by the accumulation of Mg and Ca. SOC decreased rapidly after conversion to VF, and levels continued to decrease over time. This was attributed to decreased inputs from crop residues compared to PRF and a high tillage frequency, as well as the high temperatures and moisture levels prevalent under subtropical climatic conditions. The loss of SOC inevitably causes a decline in soil physical and biological properties, with potential long-term impacts on the sustainability of the vegetable-cropping system. The C/N ratio tended to 315 decrease in deeper soils layers in VF. Management practices likely to slow or reverse soil degradation and prevent and environmental pollution in the vegetable-cropping system include increasing organic inputs (from manure, organic fertilizers, or crop residues; or including cereals in the crop rotation so as to incorporate straw), decreasing fertilizer application rates to better match crop requirements, decreasing tillage frequency, and changing the types of chemical fertilizers used to avoid an imbalance between Mg and Ca. These may be achieved by applying ISSM concepts and CA practices as well as through the 320 development of effective strategies to deliver information to farmers.

Data availability
The data that support the findings of this study are available by request from the corresponding author (D. Liu).

Author contributions
DL and XC designed the experimental setup. ML, YL and YZ did soil sampling, led the lab analysis procedure and the 325 farmer survey with the input of SL. ML, DP and DL also did the statistics, prepared the manuscript with valuable contributions of DC and XC, and undertook the revisions during the review process.

Competing interests
The authors declare that they have no conflict of interest.  Organic fertilizers included crop residues, commercial organic fertilizers, animal manure (from chickens, cows, and pigs), and rapeseed 480 meal. In PRF, rice straw was directly returned after the grains were harvested, whereas oilseed rape straw was burned in the fields. Most above-and below-ground residues in the VF were removed during harvest. b Vegetable yields are expressed as the annual sum of fresh weights per season, and yields in the PRF are expressed as the annual sum of air-dried weights per season. c Only grains were harvested in the PRF, and most of the above-and below-ground residues were returned to the fields. All vegetable parts 485 were removed from the fields during harvest.   Values were averaged over twelve replicates. The least significant difference test was performed to compared values in each soil layer, and different lowercase letters indicate significant differences at P＜0.05 among cropping systems. 495 Soil depth (cm) Cropping system SOC (g C kg -1 ) C stock (Mg C ha -1 ) TN (g N kg -1 ) N stock (Mg N ha -1 ) Soil C/N ratio