| Spatial variability of soil nutrients and quantification of the complexity of the variability using Geographic Information System (GIS) provide an effective approach to predict and simulate the procedure of soil development. Various data sets with different accuracy created by different spatial interpolation methods are critical information for precision agriculture.In the current study, we analyzed the variation of different soil properties (pH, organic matter (OM), total phosphorus (total P), available phosphorus (available P), total nitrogen (total N), available nitrogen (available N), total potassium (total K), and available potassium (available K)) in Jiangjin, Chongqing. The results were compared with the second national soil survey using classical statistics. GIS and geostatistics approaches were employed to examine the spatial variability of these nutrients. A comparison analysis of ordinary kriging (OK), inverse distance weighted (IDW), global polynomial interpolation (GPI), and local polynomial interpolation (LPI) was carried out to select the best model for each soil nutrient using mean error (ME) and root mean square error (RMSE). The results could provide important information for precision fertilization in the study area. The main results are as follows:(1) The cultivated soil had a low value of pH of 5.6. Soil pH values mainly varied between 4.5 and 5.5. Organic matter ranged from 3.3 to 34.7 g/kg with an average of 16.1g/kg. Total N was between 0.19 and 2.24 g/kg with an average of 1.01g/kg. Available N varied from 7.4 to 191.0 mg/kg with an average of 91.1mg/kg. Total P ranged between 0.04 and 1.19 g/kg with an average of 0.49 g/kg. Available P varied from 0.3 to 50.0 mg/kg with an average of 8.2 mg/kg. Total K was between 3.4 and 28.7 g/kg with an average of 15.9 g/kg. Available K ranged from 3.4 to 220.0 mg/kg with an average of 76.9 mg/kg.(2) The coefficients of variation (CV%) of the soil nutrients were between 21.56% and 95.68% indicating moderate variations. Available P had a highest CV% of 95.68% while pH had a lowest CV% of 21.56%. The order of the CV% was available P> available K> total P> organic matter> total N> available N> total K> pH.(3) An obvious acidification trend of cultivated soil was found in comparison with the results of the second national soil survey. There was 13.72% of the soil samples had pH values of 6.5-7.5 in the current study, while there was 23.62% of the soil samples had pH values of 6.5-7.5 in the second national soil survey. The percentage of soil samples with pH<5.5 was 58.09% in the current study, 30% higher than the second national soil survey. The percentage increment of soil samples with OM higher than 20.0g/kg was similar with that of less than 10.0g/kg. On average, increment trends were found in total N and available N whereas decrement trends existed in total P, available P, total K, and available K in the study area.(4) Soil pH showed negatively significant relationships with OM and available N while pH was positively significantly correlated with total P, available P, total K, and available K. Organic matter was positively significantly correlated with all other soil properties.(5) Soil pH, OM, total N, available N, total P, available P, total K, and available K showed moderate spatial autocorrelation on the basis of the results of the semivariogram model. The ratios of nugget to sill were 28.2%,33.1%,32.9%,32.5%,33.3%,33.0%,33.5%, and 33.5% for pH, OM, total N, available N, total P, available P, total K, and available K, respectively.(6) The comparison of OK, IDW, GPI, LPI showed that OK could be used to predict the spatial variability of pH and total K with the highest accuracy while LPI could be used to estimate the spatial distribution of OM, total N, available N, total P, available P, and available K with the highest accuracy. On average, OK and LPI provided higher accuracy for mapping the spatial variability of the analyzed soil nutrients while GPI presented lowest accuracy.
|
PDF Full Text Request