(International Journal of Remote Sensing, 2009) Wayumba, Gordon
Monitoring biophysical features of sugarcane to estimate productivity of growing cane using groundbased
crop cut techniques require immense time and equipment.
Crop biophysical parameters from representative Mumias Nucleus Estate sugarcane fields were used
to characterize biomass by gleaning spectral reflectance values to calculate five vegetation indices and
comparing them with ground-truthed data clipped from the fields.
Results indicated leaf area index and Red/Near infrared as the best biomass predictors with
Coefficients of correlation (r2) of 0.94 while a strong relationship existed between the spectral values
and field biomass with predictions of r2 of 0.78 and 0.82 for bands 3 and 4 respectively.
Temporal maps developed using transformed values of bands 3 and 4 suggested that yield and
biomass could be mapped from ETM+ satellite imagery. A model developed performed well returning
a coefficient of efficiency of 0.98 confirming the potential of remote sensing in providing data to
estimate crop yield.
