2026-10-07T12:49:57Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33960217 2026-09-21T23:14:33Z category_23857 category_23860 category_23863 item_type_3 month_year_09_2026

Spectral and biochemical soil data Gizachew Tiruneh (16388619) Agricultural land management Agricultural land planning Agricultural management of nutrients Enzymes - Analysis Soil organic carbon density reflectance data records <p dir="ltr">The dataset comprised <b>73 soil samples</b> collected from four Amazonian floodplain ecosystems, including açaí agroforestry, low-floodplain, regenerated, and high-floodplain areas, across four soil depths. It combined laboratory measurements of <b>soil carbon stocks and β-glucosidase (BG) activity</b> with spectral data acquired using <b>VIS-NIR-SWIR (350–2500 nm) and mid-infrared (MIR) spectroscopy</b>. These data were used to evaluate machine-learning approaches, particularly Random Forest and Support Vector Regression, for predicting soil carbon stocks and BG activity. Cross-validation was applied to assess model robustness and predictive performance. Overall, the dataset supports the development of rapid and cost-effective spectroscopy and machine-learning methods for monitoring <b>soil carbon and biological functioning in heterogeneous Amazonian floodplain soils</b>.</p> 2026-09-21T23:14:33Z Dataset Dataset 10.6084/m9.figshare.33960217.v1 https://figshare.com/articles/dataset/Spectral_and_biochemical_soil_data/33960217 CC BY 4.0