武夷山丘陵茶园红壤土壤可蚀性近似计算初探

土壤可蚀性是土壤侵蚀内动力因素,反映土壤对外营力作用的敏感性,但其定量化计算普遍存在参数复杂、计算过程繁琐的问题。为寻找近似计算K值的途径,分别使用沉降法及丘林法对武夷山市丘陵地区32组茶园红壤机械组成与有机质进行测定,对比M层(耕作层)、N层(非耕作层)机械组成差异,以EPIC方程计算茶园红壤K值并分析变化规律,建立了并优化单个粒径数据与K值的近似方程。结果发现:(1)丘陵茶园红壤以砂粒及黏粒为主,M层pH值更低,N层K值高于M层,更易被侵蚀;(2)砂粒与K值有强负相关关系,粉粒与黏粒与K值有强正相关关系,有机碳与K值相关关系弱;(3)通过线性、幂函数、二次多项式拟合形式,砂粒、粉粒与K值可建立拟合效果良好的近似方程,二次多项式对砂粒拟合效果较好,幂函数对粉粒拟合效果较好,黏粒拟合效果差。将两种方程联立,结合多元线性回归分析赋予权重,并通过平均绝对偏差多次迭代修正的方法,可降低拟合误差。

Preliminary Study on Approximate Calculation of Erodibility of Red Soil in Hilly Tea Garden of Wuyi Mountains

Soil erodibility is an internal dynamic factor of soil erosion,which reflects the sensitivity of soil to external forces.However,the quantitative calculation of soil erodibility generally has the problems on complex parameters and complicated calculation process.In order to find a way to approximately calculate K value and let it convenient to calculate,the mechanical composition and organic matter of 32 groups of tea garden red soils in hilly area of Wuyishan City were determined by settlement method and Tyurin′s method,respectively.The differences of soil mechanical composition between M layer(cultivated layer)and N layer(non cultivated layer)were compared.The K value of red soil in tea garden was calculated by EPIC equation and the variation pattern was analyzed.The approximate equation between single particle size data and K value was established and optimized.The results showed that:(1)the red soil in hilly tea garden was mainly composed of sand and clay,the pH value of M layer was lower,and the K value of N layer was higher than that of M layer,so N layer was easier to be eroded;(2)there was a strong negative correlation between sand content and K value,a strong positive correlation between contents of silt and clay and K value,and a weak correlation between organic carbon content and K value;(3)through linear,power function and quadratic polynomial fitting,the approximation equation of sand,silt and K value could be established with good fitting effect,quadratic polynomial had the better fitting effect for sand,power function had the better fitting effect for silt,and the poor fitting effect for clay particle.The fitting error could be reduced by combining the two equations,giving weight with multiple linear regression analysis,and iteratively correcting the mean absolute deviation.