| Popis: |
Rodent pests cause extensive damage to agricultural crops worldwide. Farmers’ ability to monitor rodent activity and damage within crops is limited due to their inability to simultaneously survey vast agricultural areas for rodent activity, the inability to enter certain fields, and the difficulty of monitoring rodent numbers, as well as using traps due to trap shyness and high labor costs. Drones can potentially be used to monitor rodent numbers and damage because they can cover large areas quickly without damaging crops and carry sensors that provide high-resolution imagery. Here, we investigated whether rodent activity (Levant voles Microtus guentheri and house mice Mus musculus) is related to vegetation health and biomass in Alfalfa (Medicago sativa) fields. We used a drone to photograph one hundred and twenty 10 × 10 m plots in nine fields and calculate the plots’ normalized difference vegetation index (NDVI) and biomass. On each plot, we also trapped rodents, counted rodent burrows, and evaluated the harvested dry crop yield. The number of burrows was positively related to the number of Levant voles trapped (F1,110 = 12.08, p < 0.01) and negatively related to the number of house mice trapped (F1,110 = 5.23, p < 0.05). Biomass extracted from drone images was positively related to the yield harvested by hand (F1,83 = 3.81, p < 0.05). Farmers, therefore, can use burrow counting in place of trapping Levant voles, and biomass estimates from drones can be used in place of manual yield calculations. NDVI (F1,95 = 73.14, p < 0.001) and biomass (F1,95 = 79.58, p < 0.001) were negatively related to the number of Levant voles trapped, and the number of burrows were not related to the number of house mice trapped. We demonstrate that drones can be used to assist farmers in determining the Levant vole presence and damage within crop fields to control rodents using precision agriculture methods, such as adding rodenticides in specific areas, thus increasing efficiency and decreasing the amount of pesticides used. |
