Interaction of liming and long-term fertilization increased crop yield and phosphorus use efficiency (PUE) through mediating exchangeable cations in acidic soil under wheat–maize cropping system

Low phosphorus use efficiency (PUE) is one of the main problems of acidic soil that limit the crop growth. Therefore, in the present study, we investigated the response of crop yield and PUE to the long-term application of fertilizers and quicklime (CaO) in the acidic soil under wheat–maize rotation system. Treatments included, CK (no fertilization), NP (inorganic nitrogen and P fertilization), NPK (inorganic N, P and potassium fertilization), NPKS (NPK + straw return), NPCa (NP + lime), NPKCa (NPK + lime) and NPKSCa (NPKS + lime). Results showed that, fertilizer without lime treatments, significantly (p ≤ 0.05) decreased soil pH and crop yield, compared to the fertilizer with lime treatments during the period of 2012–2018. Average among years, compared to the CK treatment, wheat grain yield increased by 138%, 213%, 198%, 547%, 688% and 626%, respectively and maize yield increased by 687%, 1887%, 1651%, 2605%, 5047% and 5077%, respectively, under the NP, NPK, NPKS, NPCa, NPKCa and NPKSCa treatments. Lime application significantly increased soil exchangeable base cations (Ca2+ and Mg2+) and decreased Al3+ cation. Compared to the NP treatment, phosphorus use efficiency (PUE) increased by 220%, 212%, 409%, 807% and 795%, respectively, under the NPK, NPKS, NPCa, NPKCa and NPKSCa treatments. Soil pH showed significant negative relationship with exchangeable Al3+ and soil total N. While, soil pH showed significant (p ≤ 0.05) positive relationship with exchangeable Ca2+, PUE and annual crop yield. PUE was highly negatively correlated with soil exchangeable Al3+. In addition, soil exchangeable Ca2+, pH, exchangeable Al3+ and available N were the most influencing factors of crop yield. Therefore, we concluded that lime application is an effective strategy to mitigate soil acidification and to increase PUE through increasing exchangeable base cations and reducing the acidic cations for high crop yield in acidic soil.


Materials and methods
Experimental site description. A long-term field trial was initiated in 1990 at the National observation and research station of farmland ecosystem, Qiyang county (26° 45′ 42″ N, 111° 52′ 32″ E) in southern region of China (Fig. 1). The climate at experimental site is subtropical monsoon that receives mean annual temperature of 17.8 °C and mean annual rainfall of 1290 mm. The duration of rainfall is from April to end of June every year. The climatic information during the experimental period is shown in Fig. S1, that were collected from the regional weather station following the National Standard of Specifications for Surface Meteorological Observations (1979). The soil type is Eutric Cambisol according to World Reference Base for soil resources (WRB) 40 , USDA classified this type of soil as Inceptisol with light loam soil texture and also classified as red soil based on Chinese soil classification system 41 . This soil contained 43.86% of clay content, 31.86% of silt and 24.28% of sand. The initial (1990) characteristics of topsoil (0-20 cm) included, soil pH 5.7, soil organic carbon (SOC) 7.9 g kg −1 , total N (TN) 1.07 g kg −1 , available N (AN) 79 mg kg −1 , total P (TP) 0.45 g kg −1 , available P (AP) 14.0 mg kg −1 , total potassium (TK) 13.7 g kg −1 and available K (AK) was 104 mg kg −1 .
Experimental design and crop management. This experiment was designed under winter wheatsummer maize rotation system and the treatments were arranged in split plot design with two replicates. Each plot (20 m × 5 m) was separated from adjacent plot by 20 cm cemented baffle plates to avoid the water and treatment contamination from nearby plot. The third replication was pseudo-replication for which samples were collected from specific area in one of the original replication of each treatment according to Hurlbert 42 . The pseudo-replication in this study can increase the type 1 error in the results 43 , although there is high spatial and temporal homogeneity in the production in this field 44 . For the present study, we selected seven treatments, including (1) CK (no fertilization, control); (2) NP (inorganic N and P fertilization); (3) NPK (inorganic N, P and K fertilization); (4) NPKS (inorganic N, P and K fertilization + straw); (5) NPCa (inorganic N and P fertilization + lime); (6) NPKCa (inorganic N, P and K fertilization + lime); (7) NPKSCa (inorganic N, P, K fertilization + straw + lime). Annually, fertilizer urea was applied at the rate of 150 kg N ha −1 , calcium superphosphate was applied at the rate of 120 kg P 2 O 5 ha −1 and potassium chloride was also applied at the rate of 120 kg K 2 O ha −1 . All fertilizers were applied before sowing, 30% and 70% of the annual inputs assigned to the wheat and maize crop, respectively. Every year, crop yield and straw were removed, while crop residues were remained in the field. The experimental field was disposed of for three years before conducting experiment to ensure the same soil physical and chemical properties. Two crops were sown each year with winter wheat (Xiangmai cultivar) cultivated at the rate of 63 kg ha −1 (160 seeds m −2 ) followed by summer maize (Yedan-13 cultivar) at the seed rate of www.nature.com/scientificreports/ 60,000 seeds ha −1 . No irrigation was applied to winter wheat and summer maize due to annual high precipitation. Pesticides Omethoate and Carbofuran were applied to control the wheat aphid during the postulation period and maize borers. Herbicide such as Glyphosate was applied to control the weeds after maize harvest. The crop was manually harvested and stubbles (about 6 cm height) and roots were remained in the soil. The collected grains and straw were air-dried and weighed separately for each crop.
Sampling and laboratory analysis. Air-dried, grain and straw samples of crop were oven-dried at 105 °C for half hour then heated at 70 °C to a constant weight for dry matter and P content determination. Oven-dried grain and straw samples of wheat and maize crop were ground and digested with H 2 SO 4 -H 2 O 2 at 270 °C. Phosphorus concentration in grain and straw was measured following the vanadomolybdate yellow method 45 . Topsoil (0-20 cm) samples were collected during 2012-2018 every year after maize crop harvest from randomly selected five points in each plot using a stainless steel sampler. Composite samples were mixed thoroughly and transferred to laboratory in the clean polythene bags for chemical analysis. To measure the soil chemical characteristics, a part from composite samples was ground and sieved through 0.25-mm sieve. SOC was estimated according to oxidation method using vitriol acid potassium dichromate oxidation 46 . Concentrations of total N, P and K were analyzed in accordance with Black 47 , Murphy and Riley 48 and Knudsen et al. 49 , respectively. Soil available N, P and K concentrations were determined according to procedures described by Lu et al. 50 Olsen (1954) and Page et al. (1982), respectively. Exchangeable Ca 2+ and Mg 2+ were extracted by 1 M ammonium acetate (pH 7) and determined by atomic absorption spectroscopy. Exchangeable Al 3+ was determined by NaOH neutralization titration after BaCl 2 (0.1 mol L −1 ) extraction. Soil pH was determined with a glass electrode using a 2.5:1 water-soil suspension.
Calculation. Based on amount of P fertilizer applied and P uptake by crop from 2012 to 2018, P use efficiency (PUE) in the term of P agronomic efficiency was determined for each plot using following equation 51 : www.nature.com/scientificreports/ where the PUE is phosphorus use efficiency (kg kg −1 ), YF is the annual crop yield (above-ground biomass) (kg ha −1 ) under the fertilization treatment and Y0 is annual crop yield (kg ha −1 ) under the control treatment. F is annual P input (kg ha −1 ).

Statistical analysis.
Significant differences among treatments were tested by one-way ANOVA and interaction between treatments and fertilization year were test by two-way ANOVA followed by Tukey's HSD test at P = 0.05 level of significance by using statistix 8.1 (window version). Relationships between soil characteristics, PUE and crop yield were quantified by linear regression equation. Boosted Regression Tree (BRT) analysis was performed using gbm package 52 in R version 3.3.3 to determine the relative influence of difference indexes on annual crop yield 36 . Since BRT models can incorporate both continuous and discrete explanatory variables, there is no need for prior data transformation or elimination of outliers, and they can fit complex nonlinear relationships 52 . The BRT fit was analyzed using a tenfold cross validation. BRT model was performed using tree complexity of 5 and learning rate of 0.005.

Relationships between soil pH, phosphorus use efficiency and crop yield. Linear regression
analysis showed that soil pH was negatively correlated with soil total N and exchangeable Al 3+ concentrations (Fig. 6). While, significant positive relationship (p ≤ 0.001; R 2 = 0.66) was observed between soil exchangeable Ca 2+ and pH.
Linear regression analysis showed that PUE significantly increased by increasing the soil pH and exchangeable base cation (Ca 2+ ) in soil (Fig. 7). Soil pH and PUE showed significant positive relationships with annual crop yield. PUE showed significant negative relationship with exchangeable Al 3+ . Furthermore, the relative contribution of predictor variables for the boosted regression tree model of crop yield showed that exchangeable Ca 2+ , pH, exchangeable Al 3+ , available N were the most influencing factors of crop yield under the long-term liming and fertilization (Fig. 8). Relative influence of soil exchangeable Ca 2+ , pH, exchangeable Al 3+ , available N and available P on annual crop yield was 33.5%, 23.9%, 11.6%, 7.7% and 6.6%, respectively. While, relative influence of Mg 2+ , soil total N, total P and SOC was < 5%.

Discussion
Soil acidification is one of the most important factors, limiting the high crop yield production in southern China 35,53 . In our study, long-term fertilization without lime application significantly decreased soil pH, exchangeable base cations (Ca 2+ and Mg 2+ ) and increased acidic cations (Al 3+ ), while addition of lime significantly increased soil pH, base cations and decreased exchangeable Al 3+ (Fig. 1). It has been reported in previous studies that, inorganic N fertilization induced soil acidification 54,55 , while, quicklime application reduced the soil acidification by decreasing exchangeable acidic cations effectively 27,56 . During the process of nitrification each mol of the ammonium belongs to each N-amidic and 2 mol of protons are released, which reduce the soil pH under inorganic N fertilization 57 . On the one hand, plants mostly release the net H + ions; on the other hand, when anions uptake exceeds that of cations, plant release net excess of OH − or HCO 3 −58 . Inorganic N fertilizer application reduces the base cations in soil, which decreases the soil pH. In previous study, it was found that inorganic N fertilization shifted the soil in to the Al 3+ buffering stage. In the acidic soil, at the soil pH below 5, hydrolysis of Al-hydroxides on the clay mineral surface release the Al 3+ into soil solution, which decreases the base saturation cations and accelerate the soil acidification 59 . The positive effects of quicklime application on soil pH were also due to its flocculating and cementing actions 60 . Increase in soil pH might be due to precipitation of exchangeable Al and Fe as insoluble hydroxides of Al and Fe, consequently decreasing the concentrations of Al and Fe in soil solution and acidity 61 . In present study, among fertilization treatments, highest soil pH was under NPKSCa treatment that might be due to addition of lime and straw incorporation to the field. Previous studies, observed the positive effect of straw incorporation on soil pH 24 . Positive effects of straw on soil pH might be due to addition of base nutrients through straw incorporation such as Ca and K which increases the soil pH 56 . www.nature.com/scientificreports/ In present study, available P in soil was higher under the NPK and NPKS treatments compared to the NPKCa and NPKSCa treatment. Soil P availability is very sensitive to soil pH 18 . In acidic soil, lower P availability could be due to P fixation with oxides of Fe and Al 62 . Application of lime may reduce the exchangeable and soluble acidic cations in soil solution and release the P in to soil solution, through changes in cation exchange capacity (CEC) and shifting phosphate adsorption-desorption equilibrium 63,64 . Therefore, in our study, soil pH showed significant negative relationship with exchangeable Al 3+ and highly positive relationship with Ca 2+ cation concentrations (Fig. 2). Lime application increases the microbial activities and accelerate the decomposition of organic matter which can release the inorganic P and can increase the P uptake 65,66 . In previous studies, Holland et al. 27  www.nature.com/scientificreports/ observed that lime application significantly increased the soil available P in acidic soil, which was in contrast with our results (Fig. 1). Some other studies have also found that high lime application can have negative impact on soil available P due to inorganic P fixation with Ca 67 . Acidification of soil directly or indirectly affects the soil biochemical characteristics and plant growth 7,68 . In our study, fertilization treatments with lime application significantly increased P uptake, PUE and crop yield, compared to the fertilization treatments without liming (Figs. 3, 4). These results were consistent with previous studies 27 . Kostic et al. 69 reported that lime application to the acidic soil increased P uptake and plant available P in soil through release of root exudation of citrate in P deficient soil, which in the turn increased PUE. In another study, Shahin et al. 70 described that effective liming of acidic soil improve plant root structure and growth, which positively influence the nutrient uptake. Poor soil fertility, nutrient losses through leaching, lower nutrient availability and accumulation of non-essential heavy metals are common characteristics of acidic soils 71,72 , which all negatively influence the plant growth and nutrients uptake. Therefore, in our study, wheat and maize crop yields under long-term fertilization without liming were very low, compared to fertilization with lime addition (Fig. 3). The highest increase in crop yield and PUE was observed under the NPKCa and NPKSCa treatment (Figs. 3,  4), that could be due to addition of lime and straw incorporation. In previous study, we found that combined application of wheat straw and inorganic fertilization significantly increased PUE by increasing P-cycling enzyme activities and P availability 24 . Increasing the soil pH through liming enhances the microbial activities 73 , which can regulate the soil P content and enhancing the P uptake. Furthermore, incorporation of crop straw improves the soil quality by increasing soil pH, improving soil organic matter (SOM) content, soil structure, aeration and retention of the high moisture content 74 , these all positive effects on soil of straw incorporation increase the crop yield. Therefore, in our study, soil pH showed significant positive relationship with PUE and crop yield (Fig. 3). Furthermore, Boosted Regression Tree (BRT) analysis showed that in acidic soil under long-term fertilization and liming, exchangeable Ca 2+ , soil pH, exchangeable Al 3+ and available N were the most influencing factors of crop yield (Fig. 7), indicating that soil acidification highly affect the crop yield by affecting PUE. Therefore, mitigation of acidification through liming is a better approach to enhance the PUE for high crop production under long-term fertilization.

Conclusion
We concluded that long-term fertilization without liming decreased the crop yield and PUE, because of high acidification of soil. Quicklime application significantly increased PUE and crop yield by increasing soil pH and base cations (Ca 2+ and Mg 2+ ), and reducing the exchangeable Al 3+ . Highest increase of crop yield and PUE were under the NPKCa and NPKSCa treatment, due to retention of SOC by straw and mitigation of acidification through liming. While, liming decreased soil available P in NPKCa and NPKSCa, compared to NPK and NPKS treatments, respectively. Moreover, exchangeable Ca 2+ , soil pH, exchangeable Al 3+ and available N were the most influencing factors of annual crop yield in acidic soil. Therefore, combined fertilizer, straw and lime application could be an effective strategy to achieve high crop yield and PUE in the acidic soil under wheat-maize rotation system.