Scholarlink

Abstract <p>For a key site in the Pre‑Salair region, the digital mapping of soil organic carbon (SOC) content in the arable layer (0–30 cm) has been conducted using the random forest algorithm implemented on the Google Earth Engine (GEE) online cloud platform. The following are used as predictors in the random forest model: 19 bioclimatic variables from WorldClim, 5 climate variables calculated based on WorldClim and the soil and climate atlas, 8 vegetation indices calculated based on Landsat 8, 26 morphometric characteristics of the relief calculated based on the ALOS digital elevation model, and 2 variables characterizing the spatial position (longitude and latitude). Correlation coefficients (<i>R</i>) between the SOC content and the predictor values are taken into account when forming the following sets of predictors: (1) BIO11+RVI, (2) Longitude+CNBL, (3) SАТ10+CC+Texture, (4) 60 predictors, (5) 42 (excluding relief curvatures, vegetation indices, and predictors with zero values), (6) 37 (all with <i>R</i> &gt; ± 0.5), (7) 32 (all with <i>R</i> &gt; ± 0.3, excluding vegetation indices), (8) 27 (all with <i>R</i> &gt; ± 0.5, excluding vegetation indices), and (9) 23 (excluding BIO1–19, relief curvatures, vegetation indices, and predictors with zero values). The SOC content modeling result based on 32 predictors and the training dataset (<i>n</i> = 42) with lower RMSE (0.72) is selected as the best. Soil bulk density modeling is performed using the pedotransfer function, which, together with the SOC content map, is used to compile the SOC stock map. SOC content in the arable layer (0–30 cm) varies from 1.3 to 6.1%, according to actual data, and SOC reserves range from 84 to 203 t/ha. The highest SOC content and reserves are found in the soils of the slope upper part, while a gradual decrease in these values is noted downslope. The soil bulk density, according to calculated data, varies in the range from 1.20 to 1.36 g/cm<sup>3</sup> and increases downslope; i.e., it has an inverse distribution trend compared to the SOC content and reserves. The total SOC reserves in the arable layer (0–30 cm) of soils of the study area of 225 ha amount to 28.7 kt.</p>

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

Digital Mapping of Organic Carbon Content and Stocks in Pre‑Salair Soils Using the Google Earth Engine Online Platform and the Random Forest Algorithm

  • 期刊论文

  • 发表日期：

    2025年12月1日

查看全文
  • N. V. Gopp,
  • T. V. Nechaeva

摘要

Abstract

For a key site in the Pre‑Salair region, the digital mapping of soil organic carbon (SOC) content in the arable layer (0–30 cm) has been conducted using the random forest algorithm implemented on the Google Earth Engine (GEE) online cloud platform. The following are used as predictors in the random forest model: 19 bioclimatic variables from WorldClim, 5 climate variables calculated based on WorldClim and the soil and climate atlas, 8 vegetation indices calculated based on Landsat 8, 26 morphometric characteristics of the relief calculated based on the ALOS digital elevation model, and 2 variables characterizing the spatial position (longitude and latitude). Correlation coefficients (R) between the SOC content and the predictor values are taken into account when forming the following sets of predictors: (1) BIO11+RVI, (2) Longitude+CNBL, (3) SАТ10+CC+Texture, (4) 60 predictors, (5) 42 (excluding relief curvatures, vegetation indices, and predictors with zero values), (6) 37 (all with R > ± 0.5), (7) 32 (all with R > ± 0.3, excluding vegetation indices), (8) 27 (all with R > ± 0.5, excluding vegetation indices), and (9) 23 (excluding BIO1–19, relief curvatures, vegetation indices, and predictors with zero values). The SOC content modeling result based on 32 predictors and the training dataset (n = 42) with lower RMSE (0.72) is selected as the best. Soil bulk density modeling is performed using the pedotransfer function, which, together with the SOC content map, is used to compile the SOC stock map. SOC content in the arable layer (0–30 cm) varies from 1.3 to 6.1%, according to actual data, and SOC reserves range from 84 to 203 t/ha. The highest SOC content and reserves are found in the soils of the slope upper part, while a gradual decrease in these values is noted downslope. The soil bulk density, according to calculated data, varies in the range from 1.20 to 1.36 g/cm3 and increases downslope; i.e., it has an inverse distribution trend compared to the SOC content and reserves. The total SOC reserves in the arable layer (0–30 cm) of soils of the study area of 225 ha amount to 28.7 kt.