Scholarlink

Knowing soil moisture is critical to comprehend the hydrology of the vadose zone because it regulates the interchange of water-related energy fluxes at the land surface. Irrigation and agriculture in arid and semiarid regions also depend heavily on accurate and timely estimations of soil moisture dynamics. The relationship between soil moisture and physical qualities is well documented. With an increasing variety in physical characteristics of soil, understanding the subsurface hydrology requires addressing such variations at both temporal and spatial scales. Therefore, the present study aims to analyse the spatiotemporal variability of soil physical properties. The current study was carried out in the Suketi watershed of district Mandi, Himachal Pradesh, India. Samples from 104 locations were collected from the agricultural, grassland, and forest landforms. This study gives a comparative analysis of different interpolation techniques in a GIS environment for quantifying the spatial variation of two soil physical properties (namely soil moisture content and soil organic content). Various interpolation techniques (Inverse Distance Weighting (IDW), Radial Basis Function (RBF), Local Polynomial Interpolation (LPI), Simple Kriging (SK), Empirical Bayes Kriging (EBK), Universal Cokriging (UCK), Ordinary Kriging (OK) and Universal Kriging (UK)) were employed to determine and compare the spatial variability of soil physical properties in the region. Cross-validation is applied to evaluate the accuracy of interpolation techniques based on error estimates such as coefficient of correlation (R), root mean square error (RMSE), bias, and unbiased root mean square error (ubRMSE). Comparing the Geostatistical techniques, UCK and SK were the best-performing for soil moisture (RMSE ~ 2.9 and ~ 0.40 correlation values), while the UK is best performing for organic matter interpolation (RMSE ~ 2.25 and 2.29 correlation value). The LPI was identified as the best deterministic approach based on the (RMSE ~ 2.9 for soil moisture and ~ 2.24 for organic content) computation; however, LPI has a continuous negative bias (~ 0.48) for soil moisture and positive bias (~ 0.1) for soil moisture. In contrast, RBF for soil moisture provides a balanced approach which depicts a smaller bias (~ − 0.19) but the error value increases (~ 2.98). The methods compared identify better functionality of the Kriging methods to IDW methods. The Kriging with external drift (KED) must be applied for better prediction results.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Estimation and Comparison of Spatio-temporal Variability of Soil Physical Properties Based on Interpolation Techniques

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sahil Sharma,
  • Vinay Meena,
  • Shankar Yadav,
  • Deepak Swami

摘要

Knowing soil moisture is critical to comprehend the hydrology of the vadose zone because it regulates the interchange of water-related energy fluxes at the land surface. Irrigation and agriculture in arid and semiarid regions also depend heavily on accurate and timely estimations of soil moisture dynamics. The relationship between soil moisture and physical qualities is well documented. With an increasing variety in physical characteristics of soil, understanding the subsurface hydrology requires addressing such variations at both temporal and spatial scales. Therefore, the present study aims to analyse the spatiotemporal variability of soil physical properties. The current study was carried out in the Suketi watershed of district Mandi, Himachal Pradesh, India. Samples from 104 locations were collected from the agricultural, grassland, and forest landforms. This study gives a comparative analysis of different interpolation techniques in a GIS environment for quantifying the spatial variation of two soil physical properties (namely soil moisture content and soil organic content). Various interpolation techniques (Inverse Distance Weighting (IDW), Radial Basis Function (RBF), Local Polynomial Interpolation (LPI), Simple Kriging (SK), Empirical Bayes Kriging (EBK), Universal Cokriging (UCK), Ordinary Kriging (OK) and Universal Kriging (UK)) were employed to determine and compare the spatial variability of soil physical properties in the region. Cross-validation is applied to evaluate the accuracy of interpolation techniques based on error estimates such as coefficient of correlation (R), root mean square error (RMSE), bias, and unbiased root mean square error (ubRMSE). Comparing the Geostatistical techniques, UCK and SK were the best-performing for soil moisture (RMSE ~ 2.9 and ~ 0.40 correlation values), while the UK is best performing for organic matter interpolation (RMSE ~ 2.25 and 2.29 correlation value). The LPI was identified as the best deterministic approach based on the (RMSE ~ 2.9 for soil moisture and ~ 2.24 for organic content) computation; however, LPI has a continuous negative bias (~ 0.48) for soil moisture and positive bias (~ 0.1) for soil moisture. In contrast, RBF for soil moisture provides a balanced approach which depicts a smaller bias (~ − 0.19) but the error value increases (~ 2.98). The methods compared identify better functionality of the Kriging methods to IDW methods. The Kriging with external drift (KED) must be applied for better prediction results.