Unmanned aerial vehicles (UAVs) can obtain time-series spatial data of plant communities and are useful for the monitoring weeds. In this study, we extract kudzu (Pueraria lobata (Willd.) Ohwi) from the UAV images by supervised learning, and we evaluate the expansion of the community in the summer. High-resolution RGB images of the riverbank (40 m × 30 m) are obtained from a 50-m altitude five times from June to October 2018. The orthomosaic images are generated and divided into 3221 grids (50 cm × 50 cm). These grids are classified into two classes (presence/absence of kudzu) using a support vector machine (SVM) based on the brightness of each band of the RGB images. The classifiers, which were trained by 5% of all grids, achieved an accuracy of over 0.9 and an F-measure of greater than 0.9, except for the image obtained on June 4. To evaluate the expansion rate of the community’s edge, a logistic model representing the relationship between the colonization probability and the distance is applied. During the period of accumulating the photosynthetic products in the storage organs (August 18, 2018–October 2, 2018), the community expanded quickly in the area in which the kudzu was mowed at the end of July. It appears that the shoots that newly emerged from the residual overwintering stems and the crowns grew quickly and compensated for the damage. These results indicate that the analysis of spatial data based on the classification of UAV imagery is useful for evaluating weed community dynamics.

Four populations of wild oats (Avena fatua L.) collected from fields in western Ibaraki, Japan, were tested between 1999 and 2004 to determine the effect of seed burial depth on seedling emergence. The persistence of seeds buried at a depth of 25 cm in the soil was also investigated. Plants were grown in the same field at NARO-ARC in Tsukuba, Ibaraki, Japan, and mature seeds were harvested. After 3 months in storage at room temperature, 100 seeds were buried in pots at 0, 5, 10, 15, and 20 cm depths with three replications for each depth. There were distinct differences in seedling emergence patterns between populations; most of the seeds in two populations emerged uniformly in the fall after the sowing year, whereas the third population emerged within a year. In the fourth population, there was a delay in the seedling emergence pattern with the first flush occurring in early spring. The relationship between burial depth and emergence patterns differed between populations. In the three early-emerging populations, seeds buried at shallower depths emerged earlier, whereas in one late-emerging population, seeds buried at depths of 15 cm and 20 cm emerged earlier than those buried at 10 cm or less. For seeds buried at a depth of 25 cm, survival was associated with the emergence patterns. Most of the seeds in the early emergence population had germinated in the soil within 6 months, whereas more than 50% of the seeds in the late emergence population did not germinate. These results indicate that the emergence uniformity of wild oats is related to the depth of seed burial. Additionally, interpopulation variation in A. fatua germination requirements were found within a narrow regional scale in Japan.