Yuling Hn, Wei M, Boyun Zhou, Akrm Slh, Mingjin Geng, Cougui Co,Ming Zhn,*, Ming Zho,*

aMOA Key Laboratory of Crop Physiology, Ecology and Cultivation (The Middle Reaches of Yangtze River), College of Plant Science and Technology, Huazhong Agricultural University, Wuhan 430070, Hubei, China

bKey Laboratory of Crop Physiology and Ecology, Ministry of Agriculture, Institute of Crop Sciences, Chinese Academy of Agricultural Sciences,Beijing 100081, China

cCollege of Resource and Environment, Huazhong Agricultural University, Wuhan 430070, Hubei, China

Keywords:Maize-rice system Straw return K fertilizer K-use efficiency Quantity/intensity curve

ABSTRACT Straw return is an effective way to improve crop grain yield and potassium(K)use efficiency by increasing soil K content. However, the effects of straw return on soil K supplying capacity, replacement of K fertilizer, and K-use efficiency under maize (Zea mays L.)-rice(Oryza sativa L.) cropping systems are little studied. A two-year field experiment was conducted to determine the physiological determinants of K-use efficiency under straw return with four K fertilization rates. Sr33 (straw returned plus 33% of K fertilizer applied)and Sr67 (straw returned plus 67% of K fertilizer applied) increased annual crop yields by 1.5%and 3.2%and increased agronomic K-use efficiency by respectively 2.9 and 1.3-fold on average in the two years, compared with the conventional practice S0K100 (no straw returned plus normal amounts of K fertilizer applied). The Sr33 and Sr67 treatments resulted in significantly greater equilibrium K concentration ratios () and specifically exchangeable K (KX) values according to quantity/intensity (Q/I) relationship analyses,indicating improvement of the potential soil K supply capacity. However, the Sr67 better maintained the soil exchangeable K level and K balance. The results suggested that K released from maize and rice straw can replace about half of chemical K fertilizer,depending on the available K content in maize-rice cropping system production.

1.Introduction

The middle Yangtze River basin is an important crop production region in China,in which double or triple cropping systems are dominant. In recent years, maize (Zea mays L.)-rice(Oryza sativa L.)cropping systems with high yield and high utilization of light and temperature resources have been established in this region [1,2]. High productivity of both crops in a double-cropping system requires high fertilization[3].However,the crops in these systems absorb much greater amounts of potassium(K)from the soil than phosphorus(P)or even nitrogen (N) [4]. Guided by conventional concepts such as “the more fertilization, the greater the output”, farmers apply large amounts of fertilizer to obtain high crop yields[3].With 60%of fertilizer applied to food crop species,crop yields in China increased from 4.0 t ha?1to 4.6 t ha?1from 1990 to 2008, but the amount of chemical fertilizer applied increased from 25.9 to 47.7 million tons [5]. After studying 209 sites in China,Zhan et al.[6]reported that,compared with treatment with the recommended K fertilizer amount, treatment with 150% of the recommended K fertilizer amount resulted in no yield increase and led to a marked decrease in K-use efficiency. Long-term inorganic fertilization also results in soil acidification [7]. Although potash resources and reserves are extensive globally, they are not abundant in China [8].China accordingly imports a large amount of K fertilizer each year[9],leading to high market prices for K fertilizer.It is thus desirable to replace a portion of the applied chemical K fertilizer and reduce farmer dependence on the imported product.Straw contains large amounts of carbon(C),N,P and K,and 80%of total K is stored in it[10,11].Improvement in K-use efficiency in China could alleviate K deficiency and reduce the potential negative impacts of K fertilization[12].

The ability of the soil to buffer K+and the concentration of K+in the soil solution determine the effective exchangeable K+available for crop absorption and use [13]. Beckett [14,15]developed quantity/intensity (Q/I) relationships for characterizing and evaluating the K+supply capacity of soils.The Q/I curve is established by plotting the change in exchangeable K relative to the activity ratio of K to calcium (Ca) and magnesium (Mg) to describe the exchange of K+for Ca2+and Mg2+and represent the soil K supply power[16].Bar-Yosef et al. [17]reported that Q/I experiments have great utility for precisely elucidating K dynamics in soils and described parameters that determine K fixation and release relationships.K in the tissues or cells of plants can be easily extracted by water;in turn,the plants in the subsequent season absorb K+from the soil solution [18]. These properties suggest that straw return could increase the K+in the soil solution in a given season and could promote the K-holding capacity of the soil by increasing soil organic matter (SOM) in the long term.

Studies on crop straw return have focused on SOM[19,20],N-use efficiency [21], weed control [20,22], tillage practices[23], and water-use efficiency [24]. Recently, Zhao et al. [25]reported that inorganic K fertilizer and straw return markedly increased crop yields and soil exchangeable K in wheat-maize cropping systems in north central China. Lal et al. [26]reported that rice straw increased soil exchangeable K from 50 to 66 mg kg?1by releasing K+within 10 days after the straw was returned. Buresh et al. [27]reported that maize-rice cropping systems (MRSs) to which maize straw was returned showed large decreases in net K export and increases in K deficiency when rice straw was not returned. Singh et al. [28]reported that, compared with no straw return, straw return increased K uptake by crops and soil available K content in an MRS in India. However, little is known about the effects of both maize and rice straw return on K use during the maize and rice growing seasons separately and the possibility of using straw K as an alternative source for inorganic K fertilizer in China.

The objectives of the present study were (1) to determine the effects of double-season straw return combined with application of reduced amounts of K fertilizer on maize and rice yields,K uptake,soil K content,and K-use efficiency;(2)to evaluate the soil K supply capacity by quantifying the Q/I relationship after two years of straw return; and (3) to investigate the feasibility of replacing chemical K.

2. Materials and methods

2.1. Experimental site description

The field experiment was conducted in the town of Qujialing(30°52′N, 112°50′E), Hubei province, China, from 2017 to 2018.The study area has a subtropical monsoon climate, with an annual average daily temperature of 16.2 °C and an annual precipitation of 1140 mm (mean of the last 30 years).Meteorological data during the experimental period were collected from a nearby weather station and are shown in Fig. 1. Upland crop-rice rotation systems are predominantly used in the paddies of this region. Wheat-rice rotation had been practiced by farmers for more than a decade at the experimental site before the initiation of this experiment.The experimental field sits atop a typical yellow brown paddy soil.The basic soil properties (0-20 cm) prior to the study were as follows: exchangeable K, 155.29 mg kg?1; total N, 1.40 g kg?1;organic matter,22.10 g kg?1;available P,13.49 mg kg?1;pH 6.73;bulk density, 1.26 g cm?3; sand, 5.3%; silt, 59.7%; and clay,35.0%.

2.2. Experimental design and field management

In March 2017, a maize-rice rotation was initiated in the experimental field. During 2017 and 2018, six experimental treatments involving straw return and K fertilizer amount were implemented following a randomized block design with three replications. The details of the treatments and field management practices are presented in Table 1. Each plot was 63 m2in area and was surrounded by a ridge 50 cm wide and a ditch 50 cm wide. The local maize cultivar Xingken 6 and the late rice cultivar Tianliangyou 953 were used in the study.In this cropping system,spring maize is grown usually from late March to late July,and late rice is transplanted after the maize harvest and harvested in early November. Irrigation was applied throughout the rice growing period. At harvest, the maize ears (including grain, cobs, and husks)were manually collected and removed from the plots, while the late rice grain was machine harvested,leaving nearly all of the straw in the plots.All the leftover straw from the plots without straw return treatments was removed. In the maize and rice plots to which straw was returned, the maize straw was buried in the field, and the rice straw was incorporated into the soil using a rotary cultivator. Other field management practices,such as pest,disease and weed control during the maize and rice growing seasons, followed local farming practices.

Fig. 1- Monthly rainfall(mm) and mean temperature(°C) during the maize and rice growing seasons from 2017 to 2018.

2.3. Sampling and measurements

In each plot,five uniform plants were removed at the maize silking/mature stages, and ten uniform rice plants were removed at the full heading/mature stage. The sampled plants were partitioned into straw and grain. After enzyme inactivation by heating at 115 °C for 1 h, the plant samples were oven dried at 70 °C to constant weight to determine the dry matter weight. During the spring maize harvest, 30 maize ears from each plot were removed, dried naturally in the sun, and threshed to determine grain yield. Late rice plants from three 5 m2subplots were harvested and their grain was dried to determine grain yield.The maize and rice grain samples were adjusted to 14% moisture content to determine yield. The costs of major inputs were recorded,including those of fertilizer, machinery operation, seed,pesticide, irrigation, and labor. The annual income was calculated for maize and rice yields at the local market price.

Soil samples (0-20 cm) from five various sites were collected from each experimental plot using a core sampler(5 cm diameter)at the maize and rice maturity stages in 2017 and 2018. Soil samples were mixed thoroughly, dried and passed through a 2 mm sieve, and then extracted with 1 mol L?1NH4OAc to determine exchangeable K content[29].

Plant subsamples of maize and rice were digested in 70%concentrated H2SO4and 30%H2O2to determine the K content in grain and plants[30].

?

2.4.Inorganic K-use efficiency and anthropogenically apparent soil K balance

The agronomic efficiency (AE) of inorganic K fertilizer and anthropogenically apparent soil K balance (0-20 cm) were calculated as follows[31,32].

where F is the K fertilizer application rate,YFis the maize and rice grain yields in the K fertilizer application treatments,and Y0is the maize and rice grain yields in the treatments without K fertilizer(control).

The anthropogenically apparent soil K balance (kg ha?1) =Kinput? Koutput, where Kinputrefers mainly to inorganic fertilizer K and Koutputincludes the grain K or the total K in aboveground plant parts that were removed from the plots.K losses by leaching and runoff are not considered in the equation. A positive value means a surplus and a negative value a deficiency in soil K balance.

2.5. Calculation of straw K replacement efficiency

To evaluate the replenishment of K using straw for inorganic K fertilizer, the following calculations were used to calculate replacement efficiency[33]:

Yield increase (kg ha?1year?1) = maize or rice yield in the S0K100 treatment(or Sr0 treatment)-maize or rice yield in the S0K0 treatment.

Replacement amount of inorganic K by straw (kg K2O ha?1)=maize or rice yield increase in the Sr0 treatment/maize or rice yield increase in the S0K100 treatment×amount of K in inorganic fertilizer applied in the S0K100 treatment.

Replacement efficiency of K fertilizer by straw (%) =replacement amount of inorganic K by straw/amount of K in inorganic fertilizer applied in the S0K100 treatment×100%.

2.6. Quantity/intensity (Q/I) determination

The Q/I was determined following Beckett [14], with some modifications by Wang et al.[34].A series of K solutions with concentrations of 0.00, 0.05, 0.10. 0.25, 0.60, 1.25, and 2.50 mmol L?1in 0.01 mol L?1CaCl2were prepared for an equilibration study.A total of 2.5 g of air-dried soil and 25 mL of K solution were added to a plastic tube. The suspensions were shaken for 30 min in an end shaker, equilibrated overnight, and centrifuged and filtered. The K concentration in the filtrate solution was measured with a flame photometer and the Ca and Mg concentrations were measured with an atomic absorption spectrophotometer (AAS) [35]. The Q/I curve was described from K concentration ratio (CR) and quantity of K adsorbed (-ΔK) [36-39], and the technical terminology used for Q/I are listed in Table 2.

The change in K concentration in solution was calculated using the following equation:

where ΔK is the difference between the initial(CKi)and final K concentrations(CKf)in the solution after equilibrium,v is the solution volume (mL) and w is the soil mass (g). Positive ΔK values indicate K desorption from the soil solid phase into the soil solution,while negative values indicate K desorption[40].

?

The CR was used to describe the intensity of K in the presence of Ca and Mg as follows[41]:

where Cafand Mgfare the respective Ca and Mg concentrations in the final equilibrium solutions. The activity coefficient of the diluted solution (0.01 mol L?1CaCl2) is close to unity;thus,the concentration of the equilibrium solution was assumed to represent the activity of these ions[35].

2.7. Statistical analyses

Analysis of variance(ANOVA)was performed using Duncan's new multiple-range test in IBM SPSS Statistics 17.0 (IBM Corporation, Armonk, NY, USA). Different letters are used in the tables and figures to indicate statistically significant differences when P <0.05. Two-way ANOVA was used to test the effects of K rate and year on crop yield, K accumulation,and apparent K balance. A linear model (ΔK = α1+ α2CR) was used to describe the Q/I relationships between ΔK and CR[36].

3. Results

3.1. Grain yield, dry matter accumulation, and annual net income

?

Treatment and year significantly influenced maize and rice grain yields (Table 3). In both years, maize yields of all the treatments with straw return and K fertilization (Sr33, Sr67,and Sr100) were higher than that of the S0K0 treatment. No significant differences in maize yield were found among S0K100, Sr33, Sr67, and Sr100 treatments in 2017, whereas Sr100 showed the greatest maize yield in 2018. Compared with the S0K0 treatment, the Sr0 treatment showed a significantly greater maize grain yield but showed an equivalent yield of late rice in 2017. The rice grain yield in the Sr100 treatment was markedly greater than that in the other treatments in 2017 and 2018.There were no differences in rice yield among the S0K100, Sr33,and Sr67 treatments in 2017; however, compared with the S0K100 treatment, the Sr33,and Sr67 treatments resulted in significant increases in rice yield in 2018.In total,compared with that in the S0K100 treatment,the annual grain yield in the Sr100 treatment was greater, with a significant increase of 11.3% averaged across the two experimental years (Table 3). The Sr33 and Sr67 treatments showed comparable annual yields, with both significantly greater than those in the S0K100 treatment in 2018. Interestingly, the Sr0 treatment resulted in an annual yield significantly greater than that in the S0K0 treatment in 2018. Thus, straw return increased the productivity of MRSs even when coupled with a two-thirds reduction in K fertilizer use.

Dry matter weights of maize and rice were significantly influenced by treatment,year,and their interaction(Table 3).Compared with the S0K0 and Sr0 treatments, the Sr33, Sr67,and Sr100 treatments resulted in greater rice and maize dry matter, especially in 2018, and their dry matter was greater than that in response to conventional management practices(S0K100 treatment). In 2018, the annual dry matter amounts in the Sr33, Sr67, and Sr100 treatments did not significantly differ but were significantly 11.8%, 14.0%, and 27.9% greater than that in the S0K100 treatment. In both years the annual dry matter in the Sr0 treatment was greater than that in the S0K0 treatment.

The net annual incomes of the MRS in the Sr100 treatment were significantly increased by 53.1%, 42.0%, and 35.3% in 2017 and by 27.8%,14.4%,and 12.2%in 2018(Fig.2).Compared with S0K100, Sr33, and Sr67 increased net annual income significantly (by 11.8%and 13.9%,respectively) in 2018.

3.2. K accumulation in plants

Total K accumulation in plants at harvest was significantly influenced by treatment, year and their interaction (Table 4).Straw accounted for 59.4% and 72.1% of total K accumulation in plants for maize and rice, respectively, averaged across treatments and years. On average, rice plants absorbed more K than did maize plants by 24.9% overall and by 56.4% in the straw, indicating that rice straw could return more K to the soil than could maize straw. The annual plant and straw K content in the treatments with K fertilizer application(S0K100, Sr33, Sr67, and Sr100) showed greater K uptake than did the control treatments (S0K0 and Sr0) in both years.Compared with the S0K100 treatment(conventional practice),the Sr100 treatment showed markedly greater plant K accumulation by maize and rice plants in the two years,with a mean increase of 17.5%.Although the annual K uptake in the plants and straw in the Sr33 and Sr67 treatments was comparable to that in the S0K100 treatment in 2017,these two treatments with straw return and reduced K fertilizer resulted in K uptake amounts that were significantly greater in 2018 than in 2017. K accumulation in the Sr67 treatment was the same as that in the Sr100 treatment in 2018. There was no significant difference in annual K accumulation between the Sr67 and Sr33 treatments in either year. Although K accumulation in each treatment in 2018 increased compared to that in 2017,the annual plant and straw K in the Sr33,Sr67,and Sr100 treatments were significantly increased in 2018 compared to that in the S0K100 treatment (Table 4). It was evident that reducing K fertilizer use coupled with straw return did not affect K accumulation by crop or straw K return.

Fig.2-Annual net income in the maize-rice cropping system in 2017 and 2018.Different lowercase letters above the columns indicate significant differences among different treatments within the same year according to the LSD test(P <0.05).

?

3.3. Soil exchangeable K and apparent K balance

Soil exchangeable K showed an increasing trend with straw return and K fertilizer amount in 2017 and 2018 (Fig. 3). The S0K100, Sr33, Sr67, and Sr100 treatments resulted in significantly greater exchangeable K concentrations than did the S0K0 and Sr0 treatments in both experimental years.Soil exchangeable K in the Sr33 treatment did not significantly differ from that in the S0K100 treatment in the 2017 maize and rice growing seasons and in the 2018 maize growing season, whereas it increased significantly by 12.6%in the 2018 rice growing season.However, the Sr67 treatment significantly increased soil exchangeable K by 8.7%,8.1%,and 18.4%in the 2017 rice growing season, the 2018 maize growing season and the 2018 rice growing season,respectively.The Sr100 treatment showed the greatest exchangeable K concentration in the 2017 maize and rice growing seasons and the 2018 maize growing season, but there was no significant difference between the Sr100 treatment and the Sr67 treatment in the 2018 rice growing season. In general,Sr100 showed the greatest exchangeable K concentration, followed by Sr67, Sr33, S0K100, Sr0, and S0K0. The soil exchangeable K level in the Sr0 treatment was similar to that in the S0K100 treatment even when the exchangeable K concentration was averaged across both years.

Straw return significantly influenced the apparent soil K balance in both years (Fig. 4). The Sr100 treatment showed a clear surplus in soil K, while the Sr67 treatment nearly balanced the K input and output in the soil in both years.However, the conventional treatment (S0K100) and the other treatments (Sr33, Sr0, and S0K0)showed a marked deficiency in soil K balance,which tended to decrease with time.

3.4. Q/I relationships and parameters

After four cropping seasons, each treatment showed a linear relationship between the Q/I of soil K at the 0-20 cm depth(Fig. 5). Straw return and K fertilizer application slightly influenced the slope of the Q/I plot. Several parameters were obtained from Q/I relations to reflect the potential soil available K supply. The S0K100, Sr67, and Sr100 treatments presented comparable equilibrium K concentration ratio()values,reflecting the amount of soil available K released into the soil solution.However,the S0K0,Sr0,and Sr33 treatments presented similarvalues, which were significantly lower than those of the other three treatments. Compared with the S0K100 treatment, the Sr67 and Sr100 treatments markedly increased the KL, which is the aggregate of -ΔK0and the specifically exchangeable K (KX) and represents the capacity of soil available K supply. Interestingly, compared with the S0K100 treatment, the Sr33, Sr67, and Sr100 treatments markedly increased the Kx value but had less effect on the-ΔK0level. The S0K0, S0K100, Sr0, Sr33, and Sr100 treatments presented similar potential K buffering capacity (PBCK), and the PBCKdecreased only in Sr67 to some extent.

3.5. AE of inorganic K

The AE of K fertilizer was evaluated based on the S0K0 treatment and was found to be significantly influenced by treatment and year in the maize and rice growing seasons(Table 5). The Sr33 treatment presented the greatest AE for maize, late rice and even the annual total among all the treatments; the annual AE significantly increased to a value that was 2.87 times that of the S0K100 treatment, averaged across the two years.There was no apparent difference in AE between the Sr67 and Sr100 treatments; however, compared with the S0K100 treatment, these two treatments presented significantly greater annual AE values in both years.

Fig.3-Dynamics of soil exchangeable K content in the 0-20 cm soil layer under application of straw and K fertilizer in 2017 and 2018.2017 M and 2018 M represent soil exchangeable K content measured during the maize harvest in 2017 and 2018;2017 R and 2018 R refer to soil exchangeable K content measured during the late rice harvest in 2017 and 2018;the mean is that of 2017 M, 2017 R,2018 M and 2018 R.Different lowercase letters above columns indicate significant differences among treatments within the same season according to the LSD test (P<0.05).S0K0,no straw returned; S0K100,no straw returned,with conventional K fertilizer applied;Sr0,both maize and rice straw returned,with no K fertilizer applied;Sr33,both maize and rice straw returned, with 33%of the conventional K fertilizer amount applied;Sr67,both maize and rice straw returned,with 67%of the conventional K fertilizer amount applied;Sr100,both maize and rice straw returned,with the conventional K fertilizer amount applied.

Fig.4-Apparent soil K balance during the harvest of late rice among treatments in 2017 and 2018.Different lowercase letters above columns indicate significant differences among treatments within the same year according to the LSD test (P<0.05).S0K0,no straw returned;S0K100,no straw returned,with conventional K fertilizer applied;Sr0,both maize and rice straw returned, with no K fertilizer applied;Sr33,both maize and rice straw returned,with 33%of the conventional K fertilizer amount applied;Sr67,both maize and rice straw returned,with 67%of the conventional K fertilizer amount applied;Sr100,both maize and rice straw returned,with the conventional K fertilizer amount applied.

?

3.6.Replacement efficiency of K fertilizer by straw

Estimates of replacement amount and efficiency of fertilizer K by return of straw are shown in Fig. 6. These values suggest the potential reduction in inorganic K fertilizer application under our experimental conditions.The replacement amount and efficiency were significantly greater in the maize growing season than in the rice growing season in both years. The replacement amount and efficiency in the 2018 rice growing season were markedly greater than those in the 2017 season.On average, the annual K replacement efficiency by straw return in 2017 was 40.7%, and this percentage increased significantly to 53.3%in 2018.

4.Discussion

4.1.Climate, crop yield and K uptake

Maize and rice are both temperature-sensitive and warmweather crop species [42]. The optimal temperatures for maize emergence and flowering are 10-12.8 °C and 25-26 °C.A temperature range of (30 ± 5) °C is beneficial for rice flowering and grain production [43,44]. Within these temperature ranges, the higher temperatures will benefit grain production. In our study, in March, June, September and October,which were the critical periods for maize emergence,maize flowering, rice flowering and grain production, the daily mean temperature was 1.2, 1.4, 1.5 and 1.5 °C greater in 2018 than in 2017 on average (Fig. 1). Therefore, the mean maize and rice yields increased by 38.8% and 31.6% in 2018 compared with 2017(Table 3).

In agreement with the results of Tan et al.[45]and Dong et al.[46],double-season straw return and K fertilizer application increased plant dry matter, yield and K uptake (Tables 3 and 4). Straw return can increase crop yield without the addition of K fertilizer [47,48], and our study showed similar results when the annual yield of the S0K0 and Sr0 treatments in 2018 were compared.The finding that the Sr33 and Sr67 treatments presented annual grain yields that were comparable to those of the S0K100 treatment in 2017 and greater than those of the S0K100 treatment in 2018 shows that the crop yield responses to K application were lower under straw return and suggests that the K fertilization amount can be reduced when straw return is also practiced [28]. In addition, straw return can increase the accumulation of soil organic carbon (SOC) and soil nutrients,which may favor crop growth and increase crop yields[49].Sui et al.[33]also reported that wheat straw return increased crop yield in two-year field experiments in Nanjing and Dafeng.In the present study,we applied enough N and P fertilizer in both maize and rice growing seasons,reduced the effects on crop yields. Although Sr100 presented a greater annual crop yield and net income than did Sr67, Sr100 had a lower AE than did Sr67. Moreover, the annual yields and net income difference between the Sr67 treatment and Sr100 treatment tended to decrease over time. The Sr67 and Sr100 treatments presented similar exchangeable K contents in the 2018 rice growing season,indicating that soil exchangeable K was not the limiting factor for these crops. It is possible that the short duration of the K reduction experiment accounted for the greater yield in the Sr100 treatment. In long-term (20-year) experiments in wheat-maize rotation research [25], K fertilization together with straw return during two seasons did not increase crop yields compared with those when normal K fertilization was practiced.Qiu et al.[7]also reported that the K application rate can be reduced if farmers return crop straw to the soil.

Crop types differ in their K accumulation and use[50].The accumulation of K by maize in the present study was significantly lower than the accumulation of K by rice;furthermore, the late rice had a greater proportion of K in the straw than did the maize plants(Table 4).This resulted in more K input to the soil when rice straw was returned.Maize was harvested by mid-July each year,and the subsequent soil conditions under water saturation in the late rice fields expedited maize straw decomposition. Maize was sown in March, four months after rice straw return, and the temperature was low during those four months, with little rainfall(Fig. 1). In consequence, the maize straw quickly released more K for the rice,while the K in the rice straw was relatively slower than that in maize [51]. The flooded environments of rice fields increase the availability of K by increasing the K diffusion distance [52,53]. Our results showed that the soil exchangeable K content in the rice growing season was greater than that in the maize growing season (Fig. 3). The greater accumulation of K by late rice than by maize may indicate that late rice had greater requirement for K and was sensitive to soil K supply[54],which could partly explain why late rice presented a greater AE of K than did maize(Table 5).

Fig.6-Replacement amount(A)and efficiency(B)of K fertilizer by straw in 2017 and 2018.Different lowercase letters indicate significant differences among cropping seasons and years according to the LSD test(P<0.05).*and**,F-values significant at the 0.05 and 0.01 probability levels,respectively.NS,not significant at the 0.05 probability level.

4.2. Soil exchangeable K, K balance, supplying capacity,replacement efficiency, and K use efficiency

Under no or low-K fertilizer inputs, most of the K used by crops originates from K stored in the soil [9].Thus, after two years in the present experiment, compared with the other treatments, the S0K0 treatment presented 15.18 t ha?1annual crop yield but had a significantly lower soil exchangeable K content within the 0-20 cm soil layer. Bhattacharyya et al. [55]reported that under normal N and P fertilizer conditions but with no K fertilizer, the K taken up by crops either originated in the deep soil layer or was derived from non-exchangeable K, a potential K source for plants. The continuous lower exchangeable K content under the inappropriate fertilizer K treatments, especially under S0K0 and Sr0 plots(Fig.3),indicated depletion of soil K,which required replenishment from adequate fertilization by combined K fertilizer and straw return [56]. Long-term unbalanced fertilization will lead to severe K deficiency in the soil in large areas of China [57]. In previous studies [26,52], among various fertilizer applications, K fertilizer combined with crop straw return was the least affected by a negative K balance in rice planting systems. In the present study, soil exchangeable K content also increased markedly when straw was applied along with K fertilizer from 2017 to 2018.The Sr0 and S0K100 treatments presented comparable soil exchangeable K contents averaged across both years(Fig.3),indicating that straw return could substitute for part of the K fertilizer,increase soil K supply levels and reduce K loss in the soil[25].In addition, the negative apparent soil K balance in Sr0 was approximately ?115.07 kg ha?1in 2018, which indicates that straw return without K fertilizer used mainly soil K.The Sr33 and Sr67 treatments presented comparable crop yields and soil exchangeable K contents averaged across both years,while the mean apparent K balances in the Sr33 and Sr67 treatments were ? 50.83 and 4.31 kg ha?1year?1,respectively,indicating that the Sr67 treatment had a K surplus, which would help ensure crop growth [58]. In view of the apparent K balance calculation, after two seasons of straw return, we speculate that a 33%-67% K fertilizer reduction was optimum. If the K application rate was exceeded, the AE of K fertilizer decreased, and leaching of exchangeable K occurred.

Several studies [59-61]have shown that the linear shape of the Q/I curve of a studied soil indicates the existence of nonspecific K-binding sites on the planar surface of clay minerals,and its related parameters(including,ΔK0,KX,and PBCK) are typically taken to describe the potential capacity of the soil K supply.The results of our study showed that K fertilization and straw return significantly affected these parameters by the end of the two-year field experiment(Fig.5).Relatively highvalues represent a relatively high capability of soil exchangeable K,and high values are usually associated with K fertilization, straw return, or originally high exchangeable K levels in the soil[39].In our experiment,compared with other treatments, the soil in the S0K100,Sr100, and Sr67 treatments presented greatervalues, in agreement with the results of Roux and Summer [62]and Beckett [15], who reported that K fertilization increasedvalues. Similar results were found for -ΔK0among the treatments. The -ΔK0values represent the portion of K that is retained on the planar surface of soil clay minerals and indicate the K releasing capacity of soils[38].Relatively high-ΔK0values indicate a relatively high K content in the soil solution for crops [61,63]. S0K0 presented the lowest -ΔK0value, mainly because of the continuous uptake of K by the crops and the relatively low amount of K released into the soil solution[61].Notably,compared with the S0K100 treatment,the treatments in which straw return was coupled with K fertilizer presented significantly greater KXvalues,indicating that there was abundant K+and more specific adsorption sites for K+in those soils [60]. The Sr67 treatment had lower PBCKvalue than S0K0 and Sr0 treatments, the result was supported by Rupa et al.[64],who found that the depletion of soil K increased PBCKvalue. Lower PBCKin soil may be associated with higher K saturation of soil[65].Overall,straw return markedly increased the potential capacity of soil K supply,yielding a greater KLeven under the 67%reduction in the conventional K fertilizer amount (Fig. 5). These results explained the significant increase in actual soil exchangeable K in the treatments with straw return coupled with K fertilizer(Fig.3).

Straw return affects soil K supply in two different ways.On the one hand, K in the straw is released into the soil,thereby increasing the soil K content and maintaining the apparent soil K balance. The Sr100 and Sr67 treatments presented a better balance between soil K input and output compared with other treatments (Fig. 4). The estimated replacement efficiency of straw K was 47% of that of the conventional K application amount (Fig. 6). On the other hand, straw return increases SOM via decomposition in the soil[66,67],and more SOM increases the number of sites with high affinity for K+[37].The applied K+then binds to the sites with high affinity for K+,which is desirable for crops as the K+is stored in the soil and protected from leaching[66].Multiple factors influence crop K-use efficiency, such as soil fertility[68], climatic factors [32], genotypic variation [50], and K fertilizer input rates[7].With an increase in K fertilizer input rate, the fertilizer response function to yield usually slows[7], and a similar trend was detected in our study (Table 5).This finding indicates that reducing the K fertilizer amount is can increase K-use efficiency, especially in conjunction with straw return.

5. Conclusions

In this study, compared with the S0K100 treatment, the Sr33 and Sr67 treatments significantly increased the annual crop yields, K uptake and AE in a maize-rice rotation. Q/I relationships analyses showed that straw return could improve the potential capacity of soil K supplies by increasingand KXvalues. Straw could be a potential economical K source for crops, and its replenishment efficiency was estimated to be 47% for inorganic K fertilizer under conventional management practices. Although the Sr33 treatment significantly benefited the AE of K and crop yield,it presented a negative soil K balance across the two years. In an actual maize-rice cropping system, maize and rice straw can be directly returned, and a 33% K fertilizer reduction is recommended to ensure crop yields, while a 67% K fertilizer reduction is recommended to achieve sustainable production.

Declaration of competing interest

Authors declare that there are no conflicts of interest.

Acknowledgments

This work was supported by the Special Fund for Agroscientific Research in the Public Interest of China (201503122)and the National Natural Science Foundation of China(31571622). The authors would like to thank Mr.Yong Tan for his help in managing the experiment fields.

Author contributions

Ming Zhao and Ming Zhao designed the experiments. Yuling Han performed most of the experiments and data collection and wrote the original draft. Wei Ma, Baoyuan Zhou, Akram Salah, Mingjian Geng, and Cougui Cao critically commented and revised the manuscript.

Appendix A. Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.04.003.