Detecting spatial variability of potato canopy using various remote sensing data

Prediction of potato yield before harvest is important for making agronomic and marketing decisions but rarely done with the use of remote sensing (RS). The potential of ground-, satellite-based and UAV-carried sensors for monitoring plant growth over time was tested over a 11.1 ha potato field, located in central Poland. Independently from the method used for monitoring of plant growth, the highest correlation of the vegetation indices (VIs) with yield was obtained at the end of flowering to fruit development at 67-78 days after planting (DAP). Late monitoring of plant growth at about 100 DAP, despite the increase of coefficient of variation of VIs, showed very weak relationship with yield. The degree of senescence of the potato plants expressed by VI variability did not reflect tuber yield at this growth stage. Magnesium content, soil compaction at the layers of 0-10 and 11-20 cm and soil bulk density at 15 cm showed significant, negative correlation with yield.