MG.44 Dataset

Description

  • Target Soil Properties: SOC, pH, Clay

  • Groups of Features: vis-NIR

  • Sample size: 44

  • Number of Features: 351

  • Coordinates: With coordinates (EPSG: 32721)

  • Location: Mato Grosso, Brazil

  • Sampling Design: Simple random sampling from previous regular grid sampling

  • Study Area Size: 13 ha

  • Geological Setting: Heavily weathered soils originating from sandstone

  • Previous Data Publication: Full dataset published in Tavares et al. (2022)

  • Contact Information:

  • License: CC BY-SA 4.0

  • Publication/Modification Date (d/m/y): 28.02.25, version 1.0

  • Changelog:

    • Version 1.0 (28.02.25): Initial release

    • Version 1.1 (15.05.25): Unit bug fix for SOC

Details

Dataset

The dataset contains the following target soil properties and features:

Target Soil Properties:

SOC - Soil Organic Carbon

  • Code: SOC_target

  • Unit: g dm⁻³

  • Protocol: Measured through titration after oxidization of the organic carbon (Walkley & Black 1934)

  • Sampling Date: March 2016

  • Sampling Depth: 0 - 20 cm

pH

  • Code: pH_target

  • Unit: Unitless

  • Protocol: Measured in CaCl2 suspension with a glass electrode with a 2.5:1 liquid:soil volumetric ratio

  • Sampling Date: March 2016

  • Sampling Depth: 0 - 20 cm

Clay

  • Code: Clay_target

  • Unit: g dm⁻³

  • Protocol: Hydrometer method, measured through fractioning the soil into the sand fractions by sieving, and the silt and clay fractions by measuring suspension density using a hydrometer (Bouyoucos 1927)

  • Sampling Date: March 2016

  • Sampling Depth: 0 - 20 cm

Groups of Features:

vis-NIR – Visible and Near Infrared Spectroscopy

  • Number of Features: 351

  • Code(s): wl_431.6, wl_437.4, wl_449.1 … wl_2153.1

  • Unit: % (Reflectance)

  • Sensing: vis-NIR spectrometer of former Veris MSP3 (Veris Technologies, Salina, Kansas, USA) which is based on two spectrometers (USB4000, Ocean optics, Largo, FL, USA) for the visible region and (C9914GB, Hamamatsu Photonics, Hamamatsu, Japan) for the NIR region, on dried and sieved samples (<2 mm) in the laboratory, spectral range was 343 - 2,200 nm at ~5 nm intervals

  • Processing: Discarding noisy edges of the spectrum (343 - 431.6 nm & 2,153.1 - 2,200 nm)

  • Sampling Date: March 2016

  • Spectral Information (After Data Processing):

    • Data Representation: Wavelength (in nm)

    • Spectral Resolution: ~5 nm

    • Spectral Range: 431.6 - 2,153.1 nm

Examples

from LimeSoDa import load_dataset, split_dataset
from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score, mean_squared_error
import numpy as np

# Load and explore the dataset
data = load_dataset("MG.44")
dataset = data["Dataset"]
folds = data["Folds"]
coords = data["Coordinates"]

# Split into train/test using fold 1
X_train, X_test, y_train, y_test = split_dataset(
    data=data,
    fold=1,
    targets=["pH_target", "SOC_target", "Clay_target"]
)

# Fit model and get predictions
model = LinearRegression()
model.fit(X_train, y_train)
predictions = model.predict(X_test)

# Calculate performance metrics
r2 = r2_score(y_test, predictions)
rmse = np.sqrt(mean_squared_error(y_test, predictions))
print(f"R-squared: {r2:.7f}")
print(f"RMSE: {rmse:.7f}")

References

Bouyoucos, G. J. (1927). The hydrometer as a new method for the mechanical analysis of soils. Soil science, 23(5), 343-354.

Tavares, T. R., Molin, J. P., Nunes, L. C., Alves, E. E. N., Krug, F. J., & de Carvalho, H. W. P. (2022). Spectral data of tropical soils using dry-chemistry techniques (VNIR, XRF, and LIBS): A dataset for soil fertility prediction. Data in Brief, 41, 108004.

Walkley, A. & Black, I. A. (1934). An examination of the Degtjareff method for determining soil organic matter, and a proposed modification of the chromic acid titration method. Soil science, 37(1), 29-38.