1. Introduction

The potassium content is an important index to evaluate the quality of flue- cured tobacco [1] [2] . Generally speaking, the potassium content in tobacco leaves with high quality should not be less than 2% [3] . The potassium content in tobacco leaves of the United States, Zimbabwe and Brazil and other countries ranges from 4% to 6% while it is only ranges from 1.5% to 2.0% in China, except for the Yunnan-Guizhou Plateau of southwest China and Sanming of Fujian province of southeast China [4] [5] . Thus, promoting the potassium content in tobacco leaves has been one of the key studies in China.

Xuancheng City (29˚57'N - 31˚19'N，117˚58'E - 119˚40'E), located in south Anhui Province, has been the dominant tobacco-planting region of this province since tobacco was initially planted in 2007 due to its well-known “burnt-sweet smelling” style [6] [7] . The tobacco-planting area is nearly 10,000 hm² with 10,000 tons of tobacco leaf yield [6] - [11] . However, according to the annual monitoring data, the average potassium content in tobacco leaves of Xuancheng City shows a decreasing trend, it alarmingly decreased from 2% in 2007 [6] to 1.5% in 2014.

The potassium in tobacco leaves mainly comes from soil and applied potassium fertilizer [12] , and the rational potassium fertilization depends upon the background of soil potassium content [13] . Therefore, in this paper, we analyzed the potassium contents in the topsoil in 2005-2007 and in 2015 of Xuancheng City, disclosed the changes and mechanism of potassium status in order to instruct the reasonable potassium fertilization for producing tobacco leaves with high content of potassium.

2. Methods and Materials

2.1. The Sources of Soil Potassium Data

The contents of potassium in topsoil were from the following sources in this paper, one is the contents of rapidly available potassium (RA-K) of 7730 topsoil samples (0 - 20 cm) in 2005-2007 obtained from the programme of Soil Testing and Formula Fertilization (STFF), and the other is the contents of slowly available potassium (SA-K) and RA-K of 124 topsoil samples collected in 2015 from the farmlands of model tobacco-farmers.

The spatial distribution of topsoil samples in the two periods are shown in Figure 1 and Table 1. The contents of SA-K and RA-K were determined by the method of flame photometer after topsoil samples were digested by HNO₃ and extracted by acetamide for the content measurements of SA-K and RA-K, re

spectively [14] .

2.2. The Grading Standards of Soil Potassium Contents

Table 2 is the grading standards of SA-K and RA-K used in this study, which is simplified on the basis of grades of potassium contents adopted in the 2^nd National Soil Survey (http://wenku.baidu.com/view/3f7f297e27284b73f242505c.html).

2.3. Methods of Data Statistics and Mapping

Data were analyzed with Microsoft Excel 2013 and IBM Statistics SPSS 20.0, the spatial distribution of topsoil samples was mapped on the platform of Esri ArcGIS 10.3.

3. Results

3.1. Rapidly Available Potassium Contents of Topsoil in 2005-2007

Table 3 shows the statistic information of RA-K contents of topsoil samples in

2005-2007. To the whole Xuancheng City, RA-K content ranged from 1 mg・kg⁻¹ to 844 mg・kg⁻¹ with an average of 68 mg・kg⁻¹, while the average RA-K contents in different administrative regions ranged from 47 mg・kg⁻¹ to 94 mg・kg⁻¹, generally in the deficient grade.

Table 4 shows the sample number distribution in different grades of RA-K contents in 2005-2007. To the whole Xuancheng City, 82.7% (6392 samples) of the total samples were in the deficient grade while 5.8% (9 samples) of the total samples were in sufficient grade, which indicate generally RA-K content of topsoil samples are insufficient. To various administrative regions, the deficiency of RA-K contents were most serious in Jingde and Xuanzhou, more than 90% of the corresponding total samples were in the deficient grade, followed by GuangDe, LangXi and Jixi, about 80% - 90% of the corresponding total samples were in the deficient grade, while 60% - 80% of the corresponding total samples in Jingxian and Ningguo were in the deficient grade.

3.2. Potassium Contents of Topsoil Samples in 2015

Table 5 shows the statistic information of the contents of SA-K and RA-K in topsoil samples from the model tobacco-farmers’ farmlands of Xuancheng City in 2015.

To the whole region, SA-K content ranged from 230 mg・kg⁻¹ to 1340 mg・kg⁻¹ with an average of 595 mg・kg⁻¹, in the sufficient grade; RA-K content ranged from 46 mg・kg⁻¹ to 352 mg・kg⁻¹ with an average of 134 mg・kg⁻¹, in the medium grade.

To various administrative regions, 1) average SA-K content in Jingxian was 485 mg・kg⁻¹, in the medium grade, average SA-K contents ranged from 545 mg・kg⁻¹ to 810 mg・kg⁻¹ in other region, in the sufficient grade. 2) average RA-K contents were 151 mg・kg⁻¹ and 185 mg・kg⁻¹ in Guangde and Ningguo, respectively, in the sufficient grade; average RA-K content ranged from 105 mg・kg⁻¹ to 141 mg・kg⁻¹ in other region, in the medium grade.

Table 6 shows the sample number distribution in various grades of SA-K contents of topsoil samples in 2015. To the whole region, 13.7% (17 samples) of

the total samples were in the deficient grade while 62.9% (78 samples) of the total samples were in sufficient grade. To various administrative regions, SA-K content in Jingde was in the sufficient grade, while 8% - 30% of the corresponding total samples in other regions were in the deficient grade.

Table 7 shows the sample number distribution in various grades of RA-K contents of topsoil samples in 2015. To the whole region, 25.8% (32 samples) of the total samples were in the deficient grade, while 29.8% (37 samples) of the total samples were in the sufficient grade. To various administrative regions, RA-K content in Langxi and Ningguo were in the sufficient grade, while 15-36% of the corresponding total samples in other regions were insufficient in RA-K.

4. Discussion

According to the data of RA-K contents in 1987 obtained during the 2^nd National Soil Survey in Xuancheng City, 89.4% of topsoils were lack of RA-K (<100 mg・kg⁻¹), however, in 2005-2007 the percentage of RA-K deficient was decreased down to 82.7%, indicating RA-K status of farmlands were slightly improved, which is attributed to potassium fertilization: before 2005-2007, most of the farmlands in Xuancheng were given priority to double rice rotation or rice- wheat rotation. Historical field survey data showed that during this period the annual average crop yield (double rice or rice + wheat) was 1000 kg/667 m² in total, the grains usually contain 0.4% of K₂O, almost all the crop straws were returned to the fields or burnt directly in the fields [15] [16] , thus about 4 kg /667 m² of K₂O was removed out from the topsoil by the harvested grains. Meanwhile, average 80 kg/667 m² of the compound fertilizer (containing 8% of K₂O) was applied annually to the fields, about 6.4 kg/667m² of K₂O brought into the topsoils, so annually the “net” input of K₂O to the topsoils is about 2.4 kg/667m².

For the surveyed model farmers’ farmlands with tobacco-rice rotation, only 25.8% of the total 124 topsoil samples are insufficient in RA-K content, while the proportion of RA-K content in medium or above grades was 74.2% in total, indicating that obvious improved in RA-K content in topsoil under the tobacco-rice rotation compared with the data under double rice cropping or rice-wheat rotation in 2005-2007, which is resulted from surplus potassium fertilizers applied into the topsoils during the tobacco-planting period. The field survey data surveyed in 2015 showed that during the tobacco-planting period, 55

kg/667 m² of tobacco specific fertilizer (containing 24% of K₂O), 5 kg/667 m² of potassium sulphate (containing 50% of K₂O) and 6.6 kg/667 m² of potassium nitrate (containing 44% of K₂O) were applied into the topsoil, while during the rice-planting period 15 kg/667 m² of rice specific fertilizer (containing 15% of K₂O) was applied to the soil, thus 20.85 kg/667 m² of K₂O in total was input into the topsoil from fertilization. Both the yields of tobacco leaves and stems are about 150 kg/667 m², which contain 2.2% and 2.0% of K₂O, respectively, but the tobacco stems usually were removed out from the tobacco fields to avoid disease spread. The yield of late rice is about 500 kg/667 m², which containing 0.4% of K₂O, also almost all the rice straws were returned to the fields or burnt directly in the fields, thus, about 5.0 kg of K₂O was removed out from the soil by the leaves and stems of tobacco and the grains of rice, so the “net” input of K₂O into the topsoil is about 15.85 kg/667m² annually. It is the reason for mean RA-K content (134 mg・kg⁻¹) in the model tobacco-farmers’ farmlands in 2015 is 49.3% higher than that under double rice rotation or rice-wheat rotation in 2005-2007 (68 mg・kg⁻¹), also higher than that of the 20 long-term monitoring farmlands (under double rice rotation or rice-wheat rotation) in Xuancheng City in the same year (ranged from 58 mg・kg⁻¹ to 188 mg・kg⁻¹ with an average of 104 mg・kg⁻¹).

Some tobacco-farmers thought that no potassium fertilizers should be applied into fields during the rice planting period, but the field survey data surveyed in 2015 showed further that for the 87.5% of ‘new tobacco-planting farmlands’ (under tobacco-rice rotation less than 3 years, 23 topsoil samples in total in 2015), the potassium accumulative effect in topsoils is not obvious, RA-K content in topsoil is still in the deficient grade, so the fields need continuous potassium fertilization during the rice planting period.

5. Conclusion

Generally, the content of rapidly available potassium content has increased gradually in topsoil farmlands in Xuancheng City, which is mainly due to potassium fertilization since 1980s, particularly due to farmlands under tobacco-rice rotation which initially started since 2007. However, more potassium fertilizers should be applied into the farmlands under double rice rotation or wheat-rice rotation, while more potassium fertilizers should be used during the rice planting period to the farmlands under tobacco-rice rotation less than 3 years.

Acknowledgements

This study was supported by the Program of Anhui Branch of China National Tobacco Co. (20150551009), the External Operation Program of BIC, Chinese Academy of Sciences (151432KYSB2015007), and the Intergovernmental Special Program of National Key R&D Plan (2016YFE0112700) and Sino-EU Cooperation Programme iSQAPER. We also thank the supports and help from the Soil and Fertilization Station of Xuancheng in STFF data supply and field sampling activity.