Mumbwa Agricultural Data

Mumbwa Agricultural Dataset: Field Descriptions and Research Overview
Field Descriptions

year: Calendar year of data collection (Integer)
month: Month of data collection (Integer, 1-12)
NIR: Near-infrared Spectroscopy reflectance (Percentage, %)

Measures crop canopy reflectance in the near-infrared spectrum
Higher values typically indicate healthier, denser vegetation

GPR: Ground Penetrating Radar measurements (Unitless)

Provides subsurface soil information including potential root zone characteristics
Related to soil structure and moisture distribution

pH: Soil acidity/alkalinity (pH scale, 0-14)

Indicates soil chemical properties affecting nutrient availability
Optimal range for maize is typically 5.5-7.0

nutrients: Composite soil nutrient index (Unitless, 0-100)

Aggregate measure of N-P-K (Nitrogen, Phosphorus, Potassium) availability
Higher values indicate better soil fertility

tmin: Minimum daily temperature (Degrees Celsius, °C)

Monthly average of daily minimum temperatures

tmax: Maximum daily temperature (Degrees Celsius, °C)

Monthly average of daily maximum temperatures

humidity: Relative humidity (Percentage, %)

Monthly average of atmospheric moisture content

pressure: Atmospheric pressure (Hectopascals, hPa)

Monthly average of barometric pressure readings

rain: Monthly rainfall (Millimeters, mm)

Total precipitation recorded per month

cumulative_rain: Cumulative rainfall for growing season (Millimeters, mm)

Running total of rainfall from season start to current month

faw_infestation: Fall Armyworm infestation level (Count per plant or % infestation)

Quantifies severity of FAW presence in fields
Based on trap counts and field scouting observations

yield_potential: Theoretical maximum yield under ideal conditions (Tons per hectare, t/ha)

Estimated from soil, climate, and cultivar characteristics
Represents yield ceiling without stress factors

yield_avg: Actual measured yield (Tons per hectare, t/ha)

Average harvested grain yield for the area
Ground truth measurement for model evaluation

Research Hypothesis
Our research tests whether environmental, soil, and pest pressure variables can effectively predict maize yield outcomes in Mumbwa, Zambia through an integrated modeling approach. Specifically, we hypothesize that incorporating Fall Armyworm (FAW) infestation data alongside traditional agronomic parameters can significantly improve yield forecasting accuracy compared to conventional models that rely solely on weather and soil data.