Introduction

Increase in crop yield due to the use of modern crop varieties has enhanced nutrient-mining from the soil: in fact, nutrient removal exceeds annual replacement. Moreover, farmers in China have mainly focused on the application of N and P fertilizers, and not K fertilizers, for crop cultivation; this is primarily because K fertilizers are often not as effective as they were in the past. Such imbalanced nutrient-management practises may impair the productivity of soil in China, which contains low levels of organic matter.

Yield declines have been reported in long-term experiments (Wang et al. 2005; Ma et al. 2007). In spite of balanced inputs of N, P, and K fertilizers, Yu et al. (2007) observed yield declines in north China, particularly in fields that were not treated with P or K fertilizers rather than in fields that received balanced treatment with fertilizers. In a long-term experiment, Cai and Qin (2006) calculated that the mean annual yield of maize on treatment with N and P fertilizers (NP treatment) was 96% of that on treatment with N, P, and K fertilizers (NPK treatment) for the first several years. However, this figure declined to 86% in subsequent years. Zhang et al. (2006) also reported significant yield declines with unbalanced treatments.

Regmi et al. (2002) found that even with recommended NPK or farmyard manure treatments, the yields of wheat and the first rice crop declined in rice-wheat cropping systems (no yield decline was observed in the second rice crop). They speculated that the yield declines in the first rice and wheat crops were due to inadequate K input. The cultivation of modern rice varieties increases the removal of P, K, S, and other plant nutrients from the soil because of the greater biomass of these varieties (Dobermann et al. 1996); eventually, more rice soils will become increasingly N, P, and K deficient (Ali et al. 1997).

It is therefore important to understand crop-yield trends under different nutrient-management conditions, so that a strategy for maintaining crop production and growth can be planned. Integrated nutrient management—i.e., the combined use of chemical fertilizers and organic amendments—can be used to maintain sustainable soil productivity. It is well recognized that organic soil matter is an important fertility parameter and that it largely determines soil quality. The application of swine waste—one of the potential sources of organic manure in China—may reduce the need for chemical fertilizers in crop fields.

The management of straw is another issue critical to sustainable agriculture in China. The burning of maize straw is common, particularly in north China. This leads to not only the loss of most organic C and large losses of N, P, and K, but also significant air pollution and the death of beneficial soil fauna and microorganisms (Liu et al. 2003). The use of maize straw is therefore an important environmental issue in north China.

The application of straw has been shown to improve soil structure and increase soil nutrient content (Liu et al. 2006; Zhang et al. 2005; Zhuang et al. 2000), but to date, few attempts have been made to assess the sustainability of rain fed soybean-maize systems that involve the application of recycled manure. The objectives of the present experiment are to examine the impact of the continuous use of inorganic fertilizers with and without recycled manure on the trends of soybean and maize yields as well as their impact on soil properties and the long-term sustainability of soybean-maize rotation under rain fed conditions in the plain of the Liaohe River.

Materials and methods

Experimental site, design, and treatments

A long-term field experiment has been conducted since 1990 at the experimental station of the Institute of Applied Ecology, Chinese Academy of Sciences. The station is located in Shilihe village of Sujiatun District, 35 km from south Shenyang (latitude 41°32′N, longitude 123°23′E); it has an average elevation of 31 m and a mean annual temperature of 7.0–8.0 °C (maximum, 39.3 °C; minimum, −33.1 °C). Its annual precipitation is about 700 mm, and the frost-free period is 147–164 days. The soil of the experimental field is an alfisol soil, which is the main soil type used for agricultural production.

The initial properties of the surface soil (depth, 0–20 cm) were as follows: soil texture, clay loam; pH, 6.7; organic C content, 22.1 g kg−1; total N, 0.8 g kg−1; available P, 10.6 mg kg−1; and soil-exchangeable K, 82.5 mg kg−1.

The experiment had 8 treatments: no fertilizer (CK), recycled manure (M), N, NM, NP, NPM, NPK, and NPKM treatments. N, P, and K fertilizers were applied at the rates of 150, 25, and 60 kg ha−1 year−1 in the form of urea, double superphosphate, and potassium chloride, respectively. All P and K fertilizers were basal-applied prior to sowing: 40 kg ha−1 N fertilizer was basal-applied prior to sowing, and 110 kg ha−1 N was top-dressed at the stem-elongation stage. Each plot area was 162 m2, with a buffer zone of 1.0 m. Initially, in 1990, the experiment was started with a soybean–maize–maize 3 year rotation. Each treatment consisted of three replications.

Through feeding-composting cycles, 80% harvested seeds, 100% soybean straw, and 50% corn stalk were returned to the original treatment. This completed a nutrient-recycling process that consisted of “fertilization-crop yield-absorption-feeding-composting-return to fields.”

Conventional tillage practises were used. Crops were manually harvested, and the above-ground crops were removed from the field. Soil was plowed approximately 20 cm deep, and stubble was incorporated into the field after harvesting.

Soil sampling and analysis

Every year since 1990, soil sampling has been conducted at depths of 0–20 cm in autumn. Five soil samples were collected from each plot and mixed to form a composite sample from that plot. The samples were air-dried, ground and passed through a 2 mm sieve, and stored for analysis.

Standard techniques were used to measure pH (1:2.5 soil-water suspension) (Anon 1986), extractable P (0.5 mol dm−3 NaHCO3 at pH 8.5) (Olsen et al. 1954), soil-exchangeable K (1 mol dm−3 NH4OAc at pH 7) (Jones 1973), non-exchangeable K (1 mol dm−3 HNO3) (Wood and DeTurk 1940), and total N and C (Vario EL, III; Elementary Co. Ltd. Germany).

Apparent K balance was estimated in the maize-soybean cropping system for an 18 year period (1990–2007) and calculated using the following equation:

$$ {\text{K balance = }}\sum { ( {\text{fertilizer K; recycled manure K; rain K; seed K) }} - {\text{ plant K}}} $$

The K content of the mineral fertilizer, recycled manure, rainwater, and seeds were determined. It was assumed that no K-leaching losses occurred in the soil system, when the cation-exchange capacity (CEC) exceeded 40 mmol kg−1 (Shen 1998).

Data analysis

Differences in each measurement were examined by ANOVA, and the mean values were separated by the least significant difference, which was set at p = 0.05.

Results

Effects of treatments on crop yields

The soybean yield varied with the different fertilizer treatments. The lowest grain yield of soybeans was obtained with the control treatment, and the highest yield was consistently obtained with the NPKM treatment. Although the grain yield of soybeans with the NP treatment was approximately 20.76% greater than that with the control treatment, the difference was not significant. NPK treatment increased soybean yield by about 22% over that obtained with the NP treatment, and the difference was significant. This indicates that K is one of the most yield-limiting macronutrients for soybeans.

Similar trends were observed in the case of maize yields: NPKM treatment resulted in the highest grain (7,790 kg ha−1) and straw (6,939 kg ha−1) yields, followed by the NPK treatment (7,476 and 6,789 kg ha−1, respectively). Fertilizer application significantly increased the grain yield of maize throughout the 18 year cropping period. The maize yields clearly differed between the N and control treatments as well as between the N and NPK treatments. However, no significant difference was detected in the maize yields between the N and NP treatments or between the NP and NPK treatments. The application of NPK fertilizer increased the maize yield by about 11.65% over that obtained with the NP treatment.

The N fertilizer had a significant yield-increasing effect on maize crops. However, the mean soybean yield was not affected by the N fertilizer (Jagadamma et al. 2008), most probably due to N fixation by soybeans.

In almost every year, soybean seeds and maize grain yields significantly increased with the application of recycled manure; this indicates that recycled manure is essential for yield increment. It was evident from the pooled data that with the M, NM, NPM, and NPKM treatments, the soybean yield increased by 33.51%, 34.13%, 23.13%, and 10.43%, respectively, and the maize yield, by 33.43%, 20.18%, 10.65%, and 4.20%, respectively. It is obvious that the yield-increasing effect would be minimized, and the fluctuation in yields, reduced with the development of fertilization strategies for both soybean and maize cultivation.

For both soybeans and maize grains, the variation coefficients were lower for the treatments that included recycled manure than for those that did not. The values of the variation coefficients were in the order of NPKM < NPM < NM < M, with a range of 0.21–0.28 for soybean yield. Thus, fluctuations in crop yield declined with the use of recycled manure. Similar trends were also seen for straw yields. In addition, the variation coefficient for maize was lower than that for soybeans (Table 1).

Table 1 Average yield of crop seeds in different treatments, from 1990 to 2007
Full size table

Soil-exchangeable K

Treatments that did not include manure or K fertilizers tended to significantly decrease the soil-exchangeable K content in the first few years, from an initial level of 80 mg kg−1 (Table 2). The soil-exchangeable K content reduced to minimum levels after 7 crop cultivations; however, the content did not continue to decrease because of homeostasis between soil-exchangeable K and non-exchangeable K. These results are in accordance with those of Yang et al. (1999) and Zhang et al. (1999).

Table 2 Soil-exchangeable K content with different treatments, from 1990 to 2007
Full size table

The soil-exchangeable K content tended to significantly increase with the NPK and NPKM treatments, and it significantly decreased with all K-omitted treatments. The greatest value thereof was 17.3%, which was obtained with the NP treatment, followed by the values obtained with the N (12.5%) and M treatments(10.6%). These findings are contrary to the general belief that most soils in north China are rich in K, and that K is rarely a limiting factor (Su 2001). Further, Liu et al. (2000) also found similar declines in soils in northwest China. The decline in soil-exchangeable K was greater with the NP treatment than with the N treatment, suggesting that N and P both limit the uptake of K by crops at this location.

In contrast, soil-exchangeable K levels increased over time with treatments that included K fertilizers (i.e., NPK and NPKM). This shows that the addition of K to organic materials resulted in a large increase in the soil-exchangeable K because manure generally contains high amounts of K, as previously mentioned. This study clearly showed that high levels of crop production resulting from the use of inorganic N and P fertilizers, as well as the removal of above-ground biomass, greatly reduce the levels of soil-exchangeable K.

K-omitted treatments, irrespective of whether they included recycled manure, resulted in a significant decline in the soil-exchangeable K content compared to the content at the beginning of the experiment. Greater reduction in soil-exchangeable K was observed with the treatments that did not include recycled manure than with those that did.

The aforementioned changes occurred over the entire experimentation period, though not continuously. Large discontinuities were observed between 2 successive sampling dates. There are 2 possible explanations for this behavior (Quemener 1986):

  • (1) The determination of soil-exchangeable K was influenced by the water content of the soil at the time of sampling; alternation of wetting and drying cycles favors fixation or release of soil-exchangeable K, depending on whether the soil was fertilized before sampling.

  • (2) The average amount of K returned to the field by residue at harvest was different for each year.

Non-exchangeable K content of soil

Changes in the non-exchangeable K content of soil between the beginning of the experiment and 2007 are shown in Table 3. During this period, K-omitted treatments significantly decreased the soil non-exchangeable K content (initial value, 543 mg kg−1). The application of K fertilizers alone resulted in a slight decrease (5 mg kg−1). In contrast, treatments that included K fertilizers and recycled manure led to a slight increase in the non-exchangeable K content (11 mg kg−1). Throughout the 18 year study period, the non-exchangeable K content was greater in soils that were subjected to treatments that included K fertilizers than in soils that were not treated. Non-exchangeable K content was lower in the soils that were treated with N and P, but not K, than in unfertilized soils. The non-exchangeable K soil content decreased significantly to approximately 15.57% and 18.04% with N and NP treatments, respectively. It is clear that non-exchangeable K content decreased with time until 1999 and since then non-exchangeable K content tended to steady in all treatments without K fertilizers except M treatment. One possible explanation for this result is experimental error.

Table 3 Non-exchangeable K content in different treatments, from 1990 to 2007
Full size table

The mean values of non-exchangeable K content were much higher with treatments that included recycled manure than with treatments that did not. The decline in non-exchangeable K was several times greater than that in soil-exchangeable K; this suggests that K associated with the latter fraction could be utilized by plants, and it thus represents a viable K reserve in this soil.

The average non-exchangeable K content values showed marked yearly variations, regardless of the type of treatment. For a given treatment, the difference in this value between 2 successive sampling dates was larger than that in the values between treatments for a given date.

In the case of the NP treatment, plant K uptake resulted in a considerable decrease in the soil-exchangeable K with time; this might enhance K release from the non-exchangeable K forms, because the balance between soil-exchangeable K and non-exchangeable K had changed.

K balance

K concentration in grain and straw

In both soybean and maize straw, K concentrations were greater in treatments with K fertilizers or recycled manure. Further, the K concentration in straw (i.e., 5.57 g kg−1) was highest after treatment with both K fertilizers and recycled manure, followed by that after NPK treatment. In the case of maize, while no remarkable differences were observed in the yields, the K concentrations clearly differed between the NPK and NPKM treatments. It has been suggested that the application of K fertilizers and recycled manure leads to the excessive absorption of K by maize straw and decreases the K-utilization efficiency. However, no significant difference was noted in the K concentrations of the grains, of either soybeans or maize, among the treatments. When the K supply is adequate, a relatively high proportion of K can be transferred from the roots to the stalk and grains (Guan et al. 1997); therefore, the mean K concentrations varied little: soybean grains, 10.96–11.91 g kg−1 and maize grains, 2.34–2.41 g kg−1. It has been reported that K concentrations in the grains are similar across various treatments and seasons, but that concentrations in the straw vary more between treatments—in accordance with K additions—and between replicates within treatments (Wihardjaka et al. 1999), (Table 4).

Table 4 Average concentration of K in grains and stalks of crops subjected to different treatments, from 1990 to 2007 (g kg−1)
Full size table

K removal

The K contained in the straw and grain is removed from the field during harvest, and before harvest, this content will decrease. It was believed that the potassium was partly washed from the leaves and straws back into the soil or simply moved to back into the soil through the root system. (Sayre 1948). In view of this fact, the calculated results according to K concentration in crops must be lower at harvest. From an agronomic point of view, the K washed from crops and transferred to root eventually returns to the soil; therefore, it is appropriate to treat K removed from the field during harvest as K removed from the soil.

The uptake of K followed the same trend as that observed with grain yield. K uptake was the highest with the NPKM treatment followed by that with the NPK treatment, and it was considerably lower with the control treatment. A better physical environment, coupled with sufficiency of water and nutrients, facilitated the uptake of nutrients and hence crop yield. Yang et al. (2004) found that the incorporation of organic residues significantly increased K uptake by rice plants, and facilitated the allocation and transfer of nutrient elements to rice grains.

The results of several long-term experiments have shown that considerable amounts of K are either leached from or have been fixed in the soil (Srinivasa et al. 1999b; Singh et al. 2002a). The greater uptake of K by crops after NP treatment as compared to that after the control treatment suggests that K was released from the soil K reserves in the plots under NP treatment in greater quantities than that released in unfertilized soils (Table 5).

Table 5 Average K output with different fertilizer treatments, from 1990 to 2007
Full size table

K budget

K fertilizers, manure, rainfall, and seeds are the sources of K, while K output occurs because of the removal of crop-based K and leaching losses. Stubble was incorporated into the soil using rotary tilling before sowing, and it was not removed from the field following harvest. The K balance, therefore, was independent of the presence of stubble.

Table 6 shows that over the 18 year study period, the total K input with the control treatment was 59 kg ha−2, which was mainly attributable to rainfall. The total K uptake in the control treatment was 487 kg ha−2; however, only 59 kg ha−2 K was recycled to the soil through seeds and rainfall. Thus, the apparent negative balance of K with the control treatment was 428 kg ha−2. NP or N treatment of the soils resulted in a greater negative K balance than that in the control treatment. This was because of higher K uptake in the NP and N treatments compared to that in the control treatment.

Table 6 Budget of K with different fertilizer treatments, from 1990 to 2007
Full size table

In both the N and NP treatments, the K input was similar, but a significant negative K balance was observed with the NP treatment because of the higher yield responses of the crops treated with P fertilizers. Analyzes of the results of the M, NM, and NPM treatments revealed that the total K removed by the crops surpassed the amount of K recycled to the soil over the 18 year period, which shows a net negative K balance ranging from 184 to 229 kg ha−2.

Compared to treatments that involved the addition of recycled manure, those that did not resulted in greater K deficits; this indicates that the addition of recycled manure can correct the K deficit to a large extent. In the case of the NPK treatment, the K input from chemical fertilizers was 1,080 kg ha−2 over the 18 year study period, and the total K removal from the crops was 793 kg ha−2. Compared to the control treatment, the NPK treatment resulted in an apparent K balance of 346 kg ha−2 because under this treatment, 60 kg year−1 K was derived from chemical K fertilizers. This result demonstrates that under these experimental conditions, an annual application of 60 kg K is adequate for maintaining the soil K concentration at its initial value, apart from the annual fluctuations that were not fully explained. NPKM treatment consistently improved the apparent K balance by about 43 kg ha−2 year−1. The positive K balance observed with NPKM treatment was greater than that with NPK treatment because the addition of recycled manure in the former resulted in greater K input.

Discussion

Nutrient limitations for crops and cropping systems

Increasing crop yields is a primary concern for farmers in the plain of the Liaohe River because land is scarce in this region. However, future increments in crop yields are likely to become smaller and will depend on changes in nutrient management (Dobermann et al. 2002). It has been shown that high crop yields can be attained through N and P management strategies that take into account plant requirements and better matching of K supply with plant demand (Mussgnug et al. 2006). However, K fertilizer application is much less common than N and P fertilizer application owing to the lack of awareness of the effects of K on crop growth. A 20 year field experiment in north China (Kong et al. 2006) showed that the content of soil-exchangeable K had decreased by 30%. K is the most yield-limiting macronutrient for soybeans in north China, and fortunately, K application has been studied in greater detail and considerable progress has recently been made in this area. Many reports are available regarding the yield-increasing effects of K fertilizers (Zhou et al. 2007; Tan et al. 2007a; Liu et al. 2007).

Exchangeable and non-exchangeable K

The results showed that imbalanced fertilization (i.e., N and NP treatments) significantly decreases the soil-exchangeable K content. This shows that nutrient depletion and deterioration of soil fertility are inevitable when only chemical fertilizers are used (Su et al. 2006). In contrast, the soil-exchangeable K content declined slightly over the experimental period in the case of the treatments that included recycled manure (M, NM, and NPM), but did not reach levels expected on the basis of nutrient-balance estimates. Even long-term application of K (60 kg ha−2) without the addition of recycled manure did not significantly increase the soil-exchangeable K content. It was reported that the annual application of 33 kg ha−2 K was adequate to maintain the soil-exchangeable K content at its initial value (Jouany et al. 1996). These results were obtained for soils with high concentrations of K.

Among all the treatments, the NPKM treatment resulted in the highest soil-exchangeable K content. Soil-exchangeable K content is usually used a primary index for determining the availability of soil K for plants in a given season, while the content and release velocity of non-exchangeable K should be taken into account to determine the long-term availability of soil K (Xie and Zhou 1999). The soil-exchangeable K content in brown soil was low in the beginning of the experiment and approached the “lowest level.” (The decomposability of K+ in soil colloids is affected by the large amounts of Ca2+ in calcareous brown soil.) Therefore, the K absorbed by the crops chiefly consisted of mineral K and K released from non-exchangeable K; their content-and-release velocity was adequate to meet the demands of the crops because the release of soil K can be hastened by treatment with only N or P fertilizers for long periods (Jouany et al. 1996). The decline in non-exchangeable K and a simultaneous increase in soil-exchangeable K concentrations, as was clearly observed in the case of the NPK treatment, showed that most of the K taken up by the crops had originated from non-exchangeable forms—i.e., via the soil solution and soil-exchangeable phase—with the equilibrium among the K forms being reestablished (Mengel and Rahmatullah 1994; Singh et al. 2002b). Greater amounts of K were released when organic manure was applied together with NPK fertilizers, which is consistent with the findings of Srinivasa et al. (1999a).

K depletion and apparent K balance

The greater removal of K (i.e., 52 kg ha−1 year−1) with the NPKM treatment than that with the NPK treatment could be associated with other benefits of the application of recycled manure (such as decreased loss of nutrients from the soil and improved soil physical properties), apart from additional nutrient supply (Bhattacharyya et al. 2006). The negative apparent K balance observed with different fertilizer and manure treatments in the long-term experiments was attributed to the uptake of larger quantities of K by crops than the amount of K added to the soil or the amount of labile K that had been either leached from the plow layer or fixed in the mineral interlayer (Swarup and Wanjari 2000). The intensification of crop production, in combination with unbalanced fertilization practises, has already resulted in a depletion of K in soils over large areas of China (Tan et al. 2007b). Our results show that recycled manure combined with K fertilizers can satisfactorily be used to attain positive K balance.