Mapping And Spatial Variability Of Soil Properties Of Tea Garden With Multi-scales In Low Mountain And Hill Region

In this paper, taking Zhongfeng tea garden in low mountain and hilly regional as an experimenting site, Geostatistics and GIS were used to analyse the spatial variability of soil properties at the different soil sampling. We compared spatial variability and spatial heterogeneity at four sampling scales, in order to provide the basis of the experimental data for the construction of organic-ecology in tea plantation, and for the precision agriculture tea plantation management. According to the research, the results indicated that:(1) In the three research scales( Macroscale, Mesoscale and small scale), the average content of Med. and fine silt was the highest among the soil particle composition, and the average content of Coar. Silt was the lowest. However, the content of clay was the highest in the Macroro-scale, and the average content of silt was the lowest. The difference of pH value among the four scales was not obvious, and the soil was acid. The content of organic matter (OM) was on the middle level among the four scales, and the content of total nitrogen (TN) and the Alkali-hydrolysable nitrogen (AN) were relatively high. The content of available phosphorus (AP) was 1.51 mg/kg in the Macroscale, and that was on the middle level in the small scale and relatively rich in the Mesoscale and Macrooscale. The content of available potassium (AK) was 59.54～75.34 mg/kg in the four scales, and the content of slowly available potassium (S-avaiable K) was relatively high. So the ability of the soil potassium supplement was strong. The content of available iron (Available Fe), available manganese (Available Mn), available copper (Available Cu), available zinc (Available Zn) of the tea plantation soil were on the high level relatively, comparing to the quality tea plantation in the country.(2) The changing of scale affected the average value of soil property a little, but it affected the coefficient of variation (CV.) a lot, whereas the micro element showed obviously reacting to it. Among the micro element, the rates of nugget to sill of Fe gradually became larger but the range gradually became smaller, following the shrink of the scale. The best test index "I" value of spatial construction increased with the expanding of the scale.In large scale, the dimension of the TN was the largest. The dimension of the pH was smaller. In middle scale, the dimension of both of the Med. and fine silt and the AN were the largest. Whereas, the dimension of clay was the smallest. In small scale, the dimension of OM was larger than all of the other soil properties, and the smallest dimension was found in the Coar. silt. In Macroro scale, the dimension of AN was the largest, and that of clay was the smallest.(3) In Macroscale, the best fitting model of sand, Med. and fine silt, clay, Coar. silt, AP and Mn was spherical model. It showed middle to strong spatial correlation. pH, OM, AN, AK, Available Fe, Available Cu, Available Zn was fitting by exponential model, the range were 216 miters to 733 miters. TK was fitting by lineal model.IN mesoscale, the best fitting model of sand, clay, Coar. silt, and Med. and fine silt was spherical model, and the range was 100 to 265 miters. pH, OM, AN, AK was better adjusted by exponential model, and the range was 102 to 426 miters. The TN, AP, Available Fe, Available Mn, Available Cu and Available Zn were adjusted by spherical model, and the range were 82 miters to 139 miters. S-avaiable K was fitting by lineal model, and the distance of spatial correlation was 536 miters.In small scale, the best fitting model of sand, Med. and fine silt and AP was lineal model. The best fitting model of pH, clay, Coar.silt, OM, AN, AK, S-avaiable K and Available Fe was spherical model. They all showed middle to strong spatial correlation., and the range was 25 to 72 miters. The TN, Available Mn, Available Cu and Available Zn were fitting by exponential model, and the range was 46 miters to 162 miters.In Macroro scale, the best fitting model of sand, clay, Coar. silt, TN, AP, AK, Available Fe, Available Mn and Available Zn was spherical model. The rates of nugget to sill was 8.78% to 50.00%, and the spatial correlation distance was 8 miters to 204 miters. The best fitting model of Med. and fine silt, water, pH, OM and AN was exponential model, and the rates of nugget to sill was 4.30% to 44.44%, they showed strongly spatial correlation. The S-avaiable K and Available Cu were fitting by lineal model.(4) In Macroscale, sand, clay distributed as blocks, and the Med. and fine silt and Coar. silt distributed as straps. The pH value was high on the middle slop position. The spatial distribution characteristic of OM and TN were nearly the same. The high value area of both AN and AP was in the southeast part. The AK distributed as straps, and the S-avaiable K distributed as blocks, Available Cu and Available Zn have the same distribution. And the soil Available Fe has high value area in southeast.In mesoscale, the distribution of Med and fine silt and Coar silt was similar. The clay distributed as blocks, and the sand and pH distributed as straps. The spatial distribution characteristic of TN and OM was nearly the same, and they totally distributed as blocks. But the high value area of TN was at the northeast part. The AN decreased from southeast to northwest gradually. The AP distributed not balance. The AK which decreased from east to west following the attitude totally distributed as straps, and the S-avaiable K mainly distributed as blocks. The distribution of soil Available Fe has high value in center, and decreased in other direction. But the distribution discipline of soil Available Mn was not obvious, Available Cu and Available Zn has the same distribution discipline.In small scale, the sand and Coar silt distributed as blocks, and the clay and med. and fine silt distributed as straps obviously. And the distribution discipline was not obvious. The pH value was low on the middle and top of the valley region, and the TN, OM, AN and AK distributed as blocks. The spatial distribution characteristic of AP was that it decreased from top of the valley region to the bottom. The soil S-avaiable K was mainly distributed as blocks, and the distribution discipline was not obvious. Soil microelements distributed as straps.In microscale, the area of the highest Coar Silt content was right in the south, and that of the Med. and fine silt content was in the northwest, and that of the clay content was in the southeast. The spatial distribution characteristic of sand was that it decreased from centre to around. The spatial distribution characteristic of pH was similar, and the high value area was both in the southeast. The OM distributed as sfraps. The spatial distribution characteristic of TN was nearly the same to the OM, but its high value area was in the northwest. The AN decreased from south direction to north direction gradually, but the spatial distribution discipline of AP and S-avaiable K was not obvious. The AK, Available Fe and Available Mn distributed as straps, and increased from south to north. But the distribution of Available Cu and Available Zn like as blocks.