Module 3

Coarse Digital Soil Mapping (Linear Model)

Objectives: Generate a baseline DSM for soil texture and SOC at multiple depths.
Requirements

Participants should have:

  1. Surface Layer Mapping (0–0.3 m)

    • Use a multiple linear regression model with covariates (NDWI, BSI, elevation, etc.) to predict:
      • Sand (%)
      • Clay (%)
      • SOC (%)
  2. Subsurface Layers Analysis of profile data shows:

    • Texture variation within profile layers < texture model error → assume same texture for deeper layers.
    • SOC variation within profiles > model error → predict SOC by depth.

Thus, SOC for deeper layers is estimated using:

\[ SOC\left( z \right) = SOC_{surface} \cdot e^{-kz}\]

where \(k\) is the fitted exponential decay coefficient.

  1. Outcome
    • Coarse DSMs for:
      • Sand (%), Clay (%), SOC (%) at 0–0.3 m
      • SOC at 0.3–0.6 m and 0.6–0.9 m
    • Field-wide initial soil parameter maps for MONICA simulations.