| The high-quality production of tobacco leaf is based on the soil. Soil nutrient, tosome extent, can influence the yield and quality of tobacco leaf. At the same time, soilnutrient have a high degree of spatial heterogeneity. In the modern tobaccoagricultural production, combined with the long-term neglected spatial mutation ofsoil nutrient and the large single use of fertilizer, this not only lead to the low rate ofutilization on fertilizer, but also bring the deterioration of soil texture and pollution ofsurrounding water, and some other negative effects, which will seriously influence thesustainable development of tobacco agriculture. Therefore, knowing the exact soilspace variation rules of soil nutrient, developing the reasonable zone management andfertilizer recommendation is the basis to achieve the precise management of farmland.In this study, the tobacco plantation field located in Guizhou province, QianXi,Bijie, was selected for exploring the site-specific nutrient management. The sampleswere taken from the topsoil (0~20cm) on an approximate20m grid sampling design(n=102). A real-time kinematic global positioning system (GPS) survey was used toidentify sampling locations. All samples were analyzed the pH, OM, AN, AP, AK, Fe,Cu, Zn, Mn. In the support of GIS and Geostatistics, we reveal the space distributionfeatures of soil nutrient and variation patterns in research area. The primaryconclusions were as follows:(1) AN had medium content with mean of95.70mg·kg-1, OM and AK belongto a high content, respectively were32.81g·kg-1and176.39mg·kg-1, while AP was ina low level of7.21mg·kg-1. For trace elements, Fe and Mn were with a high content,respectively24.85mg·kg-1and60.10mg·kg-1, while Zn and Cu were in a mediumcontent with mean of2.32mg·kg-1and1.10mg·kg-1. For the nine kinds of soilproperties, Fe had the largest coefficient of variation of60.62%, then for AP>Cu>Zn>Mn>AK>AN>OM>pH. pH had the minimum coefficient of variation of6.31%, showed a weak variability, the other soil nutrient belonged to the mediumvariability. All the variables were normally distributed and met the stationaryhypothesis.(2) Geo-statistics research showed that the coefficients of determination (R2) ofall the soil nutrients were greater than0.9, while the residual are smaller, so indicatinga good fit. Of all the soil nutrient, only AN has a strong spatial correlation, thenugget-sill ratios is22.84%, the variability of which is mainly determined by theautocorrelation part of terrain and soil type. The nugget-sill ratios of the other soilnutrient were between25%and75%, mainly hovering around50%, belong to themedium strength of spatial correlation. This showed that the spatial variability is theresult of coefficient between autocorrelation part and random part, such asfertilization, human activities. The range of all nutrients’ variation scope is35.10m to293.80m, the largest one was Cu(293.80m), and the second one was AP(100.70m),AN was minimum process(35.10m). For the trace elements, Cu had the largest range,which get to293.80m. While the minimum one was Mn(46.20m). All theautocorrelation distance of soil nutrient was larger than sampling distance, thisindicated that all the soil nutrients showed spatial correlation to some extent in20msampling scales, sampling dimension space could be used to mutation analysis.(3) Spatial interpolation showed that all the soil nutrients represented asignificant space distribution pattern. The soil pH in the survey region presented adistribution rule on a whole that the value in the eastern was higher, and showed thetrend of gradually decreasing towards the west, the regions whose pH value were lessthan5.5were mainly in the west. The content of OM were a little low only in thenorth and south regions, in the east and west regions seemed to be higher, and thedistribution in other regions, whose area is more than half of the study area, waseven(32g·kg-1~34g·kg-1). AN indicated a strong spatial correlation, the distributionin the study area was uneven, showed a spot distribution. On the whole, the content ofAN was higher in the north, relatively lower in the south, and moderate in the central.The content of AP seemed to be relatively lower externally, and higher internally. Thecontent of AK was lower in the northwestern regions, and more abundant in the east,which was above180mg·kg-1. And the northeast area was the key supply area of AP and AK. The content of Mn and Zn presented an opposite distribution rule. Thecontent of Mn was higher in the surrounding areas and lower in the central regions,while that of Zn was higher in the central area and lower in the surrounding area. Thecontent of Cu and Fe were gradually increased from west to east, they presented asimilar space distribution rule.(4) soil trace elements space-times change research shows that: tobacco growthperiod before and after the trace elements in soil has dropped, with Fe and Mn downsignificantly; Different periods of trace elements all has the spatial correlation ofmoderate intensity, but Fe and Zn spatial correlation of quick enhanced obviously,and Mn and Cu space correlation are slightly weakened; Fe from the whole spacedistribution pattern to present a low in the east, the west high levels of the trend, Cushowed more similar space distribution pattern, and Mn and Zn space distributionpattern is not obvious.(5) In this research, using the principal component analysis and fuzzy clustering,we divided the study area into three administrative zones based on the twomanagement division evaluation index (Fuzziness performance index, FPI andnormalized classification entropy, NCE). We also set up the recommendedfertilization equation, then implemented it in the field. |
PDF Full Text Request