Study On Spatial Variability Of Soil Properties And Rice Information In Village Scale And Its Synergistic Mechanism

Combining the traditional statistics with the geo-statistical method, we investigated spatial variability and synergy mechanism of 5 soil physical properties(including soil bulk density, total porosity, capillary water holding capacity, capillary porosity, aeration porosity), 17 nutrient elements(including pH, available N,S, Si, P,K, Cu, Zn, Fe, Mn, total K, P, N, organic matter, exchangeable Ca, Mg and CEC) and rice growing information(including SPAD value, LAI, dry matter accumulation, NPK uptake at jointing stage, heading stage and mature period) in Wenxing village, Anshun city in Guizhou province. First, our results showed that the Cv of soil pH, total porosity and capillary porosity were lower than 10%, indicating their weak variation;Secondly, bulk density, capillary water holding capacity, aeration porosity, available N, P, K, S, Si, Fe, Zn, Cu, total N, P, K, organic matter, exchangeable Ca, Mg and cation exchange capacity had a variation coefficient between 10% ~100% for moderate variation.; But the available Mn have a variation coefficient of 102.81%,characterized by strong spatial variation. On the growing Information of rice, apart from the SPAD value of jointing stage and heading stage and the yield components in seed setting rate were lower than 10%, as well as for the weak variation, while others were moderate variation.Next, our semi-variogram analysis showed that the ratios of nugget to sill for soil bulk density, total porosity, pH, available N, P, S, total N, total P, exchangeable Ca,Mg,Cu, CEC, jointing stage SPAD value, heading stage SPAD value and N, P,K uptake, accumulation of dry matter and P, K uptake in mature period were lower than25%, demonstrating their strong spatial auocorrelations; Contrarily, available Si, Mn,dry matter accumulation of jointing stage and N, P uptake, Thousand-grain mass and spikelet number per panicle have a weak spatial correlation; the others of Soil capillary water holding capacity, capillary porosity, aeration porosity, organic matter,available K, Fe,Zn, total K, LAI at jointing stage, LAI and dry matter accumulation at heading stage, and SPAD value, LAI,N uptake at mature period and effective panicle,seed setting rate, theoretical yield and actual yield displayed positive significant spatial autocorrelation.While Moran’s I index analysis showed that the Z values of bulk density, total porosity, capillary water holding capacity, aeration porosity, available P, S, Si, Mn and Zn between-1.96 to 1.96 in soil properties, indicating they had no significant correlation spatial auocorrelations, the others Z value were greater than 1.96, there was a positive correlation spatial auocorrelations. And rice growing information displayed that the LAI of jointing stage and mature period, and the SPAD value, LAI at heading stage, P, K uptake of mature stage, and one of the yield components-seed setting rate is greater than 1.96, there was a positive correlation spatial auocorrelations,others Z values between-1.96 to 1.96, had no significant correlation spatial auocorrelations.From the kriging interpolation maps at the study area, From southwest to northeast of soil bulk density increased gradually to the two levels, total porosity,capillary water holding capacity and capillary porosity have a opposite trend with it;Organic matter, alkali-hydro nitrogen, rapidly-available potassium, total nitrogen,total phosphorus, effective silicon, sulfur and exchangeable calcium cation exchange capacity presents a increasing trend from northeast to southwest; The growth information of ricehave a broadly distribution trend in each growth period, as the change of growth period has a small change, the distribution of yield have a apparently patch shape, the highest in the northwest and southeast. It is very important to reasonably regulating the characteristic index of paddy soil and the growth informations of rice characteristics index. For some of the samples which measuring method were difficulty, the easily gained relationship samples could be used for co-kriging interpolation, and an estimated value of relatively small error should be got.