Scholarlink

<p>This study addresses the limitations of traditional soil organic matter (SOM) analysis methods by evaluating a color-based approach for predicting SOM content in Moroccan agricultural soils. While previous research has established the technical feasibility of color-based SOM prediction, the economic viability of such approaches remains largely unexplored. We employed the Recursive Feature Elimination method to identify optimal color indices for SOM prediction and compared Random Forest (RF) and eXtreme Gradient Boosting (XGBoost) algorithms under both dry and moist soil conditions using nested leave-one-out cross-validation approach designed for small datasets (<i>n</i> = 74). Additionally, we conducted a comprehensive economic analysis comparing color-based and traditional Walkley–Black method across various sample throughput scenarios. For dry soils, the RF model achieved the highest prediction accuracy (R<sup>2</sup> = 0.60, RMSE = 0.377%), outperforming XGBoost (R<sup>2</sup> = 0.55) by 8.3%. Moisture reduced prediction accuracy by 13% for RF (R<sup>2</sup> = 0.52) and 9% for XGBoost (R<sup>2</sup> = 0.50), with hue-based color parameters (H and h) dominating SOM prediction and contributing 35% of total feature importance in dry conditions and 47% in moist conditions. The economic analysis demonstrated substantial cost advantages for the color-based approach, with break-even occurring at approximately 1261 samples. For operations processing 5000 samples annually, the color-based method demonstrated a 96% cost reduction compared to Walkley–Black, with a 5-year return on investment of 940.65%, a cost–benefit ratio of 10, and a payback period of less than 4&#xa0;months. These findings provide a novel framework for soil testing facilities to evaluate the financial and technical feasibility of adopting colorimetric approaches, potentially enabling more frequent and widespread SOM monitoring while reducing environmental risks associated with hazardous chemical waste.</p> Graphical Abstract <p></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

Predicting soil organic matter from color indices: economic and technical feasibility in semi-arid agricultural soils

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月24日

查看全文
  • Yassine Bouslihim,
  • Widad Ennaji,
  • Abdessamad Hilali

摘要

This study addresses the limitations of traditional soil organic matter (SOM) analysis methods by evaluating a color-based approach for predicting SOM content in Moroccan agricultural soils. While previous research has established the technical feasibility of color-based SOM prediction, the economic viability of such approaches remains largely unexplored. We employed the Recursive Feature Elimination method to identify optimal color indices for SOM prediction and compared Random Forest (RF) and eXtreme Gradient Boosting (XGBoost) algorithms under both dry and moist soil conditions using nested leave-one-out cross-validation approach designed for small datasets (n = 74). Additionally, we conducted a comprehensive economic analysis comparing color-based and traditional Walkley–Black method across various sample throughput scenarios. For dry soils, the RF model achieved the highest prediction accuracy (R2 = 0.60, RMSE = 0.377%), outperforming XGBoost (R2 = 0.55) by 8.3%. Moisture reduced prediction accuracy by 13% for RF (R2 = 0.52) and 9% for XGBoost (R2 = 0.50), with hue-based color parameters (H and h) dominating SOM prediction and contributing 35% of total feature importance in dry conditions and 47% in moist conditions. The economic analysis demonstrated substantial cost advantages for the color-based approach, with break-even occurring at approximately 1261 samples. For operations processing 5000 samples annually, the color-based method demonstrated a 96% cost reduction compared to Walkley–Black, with a 5-year return on investment of 940.65%, a cost–benefit ratio of 10, and a payback period of less than 4 months. These findings provide a novel framework for soil testing facilities to evaluate the financial and technical feasibility of adopting colorimetric approaches, potentially enabling more frequent and widespread SOM monitoring while reducing environmental risks associated with hazardous chemical waste.

Graphical Abstract