Effects of different artificial planting schemes on invasive weeds

Abstract

Weed invasion is the main reason affecting grassland productivity, which also serves as a key difficulty encountered during technical grassland restoration. Modern agriculture relies heavily on chemical herbicides to control weeds; however, this is not applicable to the restoration of degraded grasslands. In this study, artificial establishment methods employing interspecific competition were used to suppress weed invasion while greatly increasing grassland productivity.The establishment of different combinations and densities were then evaluated. Compared to the blank control block, the high-density planting of four pastures was found to reduce the types of invasive weeds by 69.4% and the number of invasive weeds by 79.4%. In addition, the individual biomass of weeds fell by 96.3%. Moreover, 7.56 t hm−2 dry weight of forage was provided in the first year, and 6.52 t hm−2 dry weight of forage was continuously provided in the second year, which had the best input-output ratio in this experiment. This study also demonstrated that the low-density mixed sowing scheme of the three forages can effectively inhibit the invasion of weeds, reducing the types of invasive weeds by 58.3%, the number of invasive weeds by 73.9%, and the individual biomass of weeds by 89.1%, which was the least economical investment among weed control programs. At the same time, unsupervised learning, which was included in the machine learning framework Scikit-learn, was used to verify the relationship between the establishment density of pasture and the number of weed invasive species. The results indicated that the contribution rate of Avena sativa and Elymus nutans to the number of invasive weed species was 89.79%. The present study demonstrated that artificial planting methods can effectively control weeds and increase pasture yields. Accordingly, they have great potential and research value in terms of ecological and economic recovery of degraded grasslands.

Keywords

• Ecological restoration
• Artificial planting
• Weed management
• Economic recovery
• Machine learning
• Unsupervised learning