ABSTRACT

Unstable soil aggregates can easily be dispersed when mechanical energy is applied, denominated as readily dispersible clay (RDC) which is usually measured by turbidimetry. Faster methods that require minimum preparation at reduced costs are needed, such as the spectroscopy technique. Nevertheless, the technique has not yet been tested on soils. Electromagnetic radiation has strong interaction with soil physical elements (sand, silt, and clay) and with selective elements of the clay fraction, such as mineralogy. Thus, this study assessed the potential of 400–16,680 nm range (visible – Vis; near-infrared – NIR; short-wave infrared – SWIR; and medium infrared – MIR) spectroscopy to estimate RDC in some tropical soils. We also evaluated the wavelengths and regions that contribute to this prediction. We collected 68 samples from the underlying horizon of three different soils from three important states in Brazil The soil analyses were performed (physical and chemical) in the laboratory. RDC quantification was compared with the Vis-NIR-SWIR and MIR wavelengths models. The random forest (RF) algorithm was used to generate the prediction in the soil database. For RDC, the Vis-NIR-SWIR and MIR configuration presented a better relationship (coefficient of correlation (r) = 0.659 and root of the mean square error (RMSE) = 0.469%), although the results were similar when the regions were evaluated separately (Vis-NIR-SWIR – 350 to 2500 nm and MIR 2500 to 16,680 nm). RDC determination is more related to bands from 1980 to 2040, 6840 to 7360, and 7750 to 9820 nm. Spectroscopy showed its importance in RDC quantification, as it is fast and does not require soil preparations. The results show that spectral wavelengths are appropriate to assist on the estimative of RDC in the studied soils and may provide a basis for future research.

摘要

当施加机械能时，不稳定的土壤团聚体很容易分散，称为易分散粘土（RDC），通常通过比浊法测量。需要以更低的成本进行最少准备的更快方法，例如光谱技术。然而，该技术尚未在土壤上进行测试。电磁辐射与土壤物理元素（沙子、淤泥和粘土）和粘土部分的选择性元素（例如矿物学）有很强的相互作用。因此，本研究评估了 400–16,680 nm 范围（可见光 – 可见光；近红外 – NIR；短波红外 – SWIR；和中红外 – MIR）光谱在一些热带土壤中估计 RDC 的潜力。我们还评估了有助于此预测的波长和区域。我们从巴西三个重要州的三种不同土壤的底层地层收集了 68 个样品。土壤分析是在实验室进行的（物理和化学）。RDC 量化与 Vis-NIR-SWIR 和 MIR 波长模型进行了比较。随机森林 (RF) 算法用于在土壤数据库中生成预测。对于 RDC，Vis-NIR-SWIR 和 MIR 配置呈现出更好的关系（相关系数（r ) = 0.659 且均方误差根 (RMSE) = 0.469%)，尽管单独评估区域时的结果相似（Vis-NIR-SWIR – 350 至 2500 nm 和 MIR 2500 至 16,680 nm）。RDC 测定与 1980 至 2040、6840 至 7360 和 7750 至 9820 nm 的波段更相关。光谱学显示了它在 RDC 量化中的重要性，因为它速度快且不需要整地。结果表明，光谱波长有助于估计研究土壤中的 RDC，并可能为未来的研究提供基础。