A leaf reflectance-based crop yield modeling in Northwest Ethiopia

Area of Science:

• Agricultural Science
• Remote Sensing
• Agronomy

Background:

• Crop yield prediction is vital for agricultural policy and food security.
• Maize (Zea mays L.) is a critical crop, and accurate yield estimation is essential.

Purpose of the Study:

• To model maize grain yield (GY) and aboveground biomass yield (ABY) using leaf spectral reflectance.
• To assess the effectiveness of various spectral vegetation indices for yield prediction in the Aba Gerima catchment, Ethiopia.

Main Methods:

• Leaf spectral reflectance was measured using a FieldSpec IV spectroradiometer (350-2,500 nm).
• Spectral vegetation indices (e.g., NDVI, EVI, GNDVI) were calculated from reflectance data.
• Regression analyses were employed to model GY and ABY, evaluating performance with R², RMSE, and RI.

Main Results:

• NDVI showed a strong correlation with GY (R² = 0.70), while GNDVI best predicted ABY (R² = 0.71).
• Combined spectral indices improved prediction accuracy for GY (R² = 0.83) and ABY (R² = 0.78).
• Highest yields were recorded on soil bunded plots with gentle slopes.

Conclusions:

• Spectral reflectance indices derived from spectroradiometers offer a reliable method for predicting maize GY and ABY.
• Yield estimation models can assist policymakers in identifying limiting factors and implementing strategies for enhanced crop production and food security in Ethiopia.