摘要
摸清土壤盐分的影响因素是盐渍化土壤改良利用的重要基础。以黄河三角洲濒海垦利区为研究区,选取地下水、植被、地貌、离海距离4个方面的影响因子,确定了地下水埋深、地下水矿化度、植被类型、植被覆盖度、地貌类型、相对高程、离海距离7个指标,分析了各指标与土壤含盐量之间的关系,通过灰色关联分析法筛选出土壤盐分的主要驱动因子,进而构建了不同深度土层土壤含盐量的多元线性回归预测模型。结果表明,地下水埋深与各土层含盐量为指数函数负相关,植被覆盖度、相对高程和离海距离与各土层含盐量均为幂函数负相关,地下水矿化度与各土层含盐量呈指数函数正相关,且随着土层深度的增加各指标与含盐量的相关性逐渐减弱;各土层土壤盐分与各影响因子关联度的排序均为地下水矿化度>植被覆盖度>地下水埋深>相对高程>离海距离,土壤盐分的主要驱动因子一是地下水矿化度和埋深,二是地表植被覆盖度;基于地下水矿化度、植被覆盖度、地下水埋深三因子构建的0~15 cm、15~30 cm、30~45 cm、45~60 cm土层土壤盐分的多元线性回归模型均为最佳盐分预测模型,模型的决定系数(R2)分别为0.742,0.777,0.794和0.828,均方根误差(RMSE)分别为2.079 5,2.081 9,1.868 3和1.623 6,验证集的R2分别为0.712,0.756,0.813和0.883,RMSE分别为1.952 0,1.879 7,1.728 9和1.227 3。
It is an important basis for the improvement and utilization of salinized soil to find out the influencing factors of soil salinity.Taking the coastal Kenli area of the Yellow River Delta as the study area,this paper selected four influencing factors of groundwater,vegetation,landform and distance from the sea,determined seven indicators of groundwater depth,groundwater mineralization,vegetation type,vegetation coverage,landform type,relative elevation and distance from the sea,analyzed the relationship between each indicator and soil salt content,and selected the main driving factors of soil salt through grey correlation analysis.Then the multiple linear regression prediction model of soil salt content in different depth soil layers was constructed.The results showed that groundwater depth was negatively correlated with the salinity of each soil layer as an exponential function,vegetation cover,relative elevation and distance from the sea were negatively correlated with the power function of the salt content of each soil layer,and groundwater salinity was positively correlated with the salinity of each soil layer as an exponential function,and the correlation between each index and salt content gradually weakens with the increase of soil depth.The ranking of the correlation between soil salinity and each influencing factor in each soil layer was groundwater salinity>vegetation coverage>groundwater burial depth>relative elevation>distance from the Bohai Sea.The main drivers of soil salinity were groundwater salinity and depth,and surface vegetation cover.The multiple linear regression models of soil salinity in 0-15 cm,15-30 cm,30-45 cm and 45-60 cm soil layers based on the three factors of groundwater mineralization,vegetation coverage and groundwater depth were the best salt prediction models,determination coefficient of model set(R 2)was 0.742,0.777,0.794 and 0.828 respectively,and the root mean square error(RMSE)was 2.0795,2.0819,1.8683 and 1.6236 respectively,validation of model set(R 2)was 0.712,0.756,0.813
作者
赵铭
常春艳
王卓然
赵庚星
ZHAO Ming;CHANG Chunyan;WANG Zhuoran;ZHAO Gengxing(College of Resources and Environment,Shandong Agricultural University/National Engineering Research Center for Efficient Utilization of Soil and Fertilizer,Tai′an 271000,China)
出处
《黑龙江农业科学》
2023年第2期23-34,共12页
Heilongjiang Agricultural Sciences
基金
国家自然科学基金(41877003)
山东省重大科技创新工程项目(2019JZZY010724)
山东省“双一流”奖补资金(SYL2017XTTD02)。
关键词
黄河三角洲
不同土层
土壤盐分
影响因素
预测模型
the Yellow River Delta
different soil layers
soil salinity
influencing factors
prediction model
