The ability to disperse individuals and seeds among subpopulations is required for the sustainability of plant metapopulations. Understanding the mechanism of seed dispersal is essential for the conservation and restoration of plant diversity. In this study, Euphorbia adenochlora growing on floodplains was used as a research target to obtain new knowledge about hydrochory and ballochory and estimate potential dispersal range related to hydrological regimes. A wetland spreading around the Omi Maiko Inner Lake in Minami-Komatsu, Otsu City, Shiga Prefecture, Japan, was the study site. E. adenochlora bear capsules, which have a three-chambered hollow structure and which float on water up to 14 days, they can be dispersed by hydrochory. The capsule of E. adenochlora splits open when it dries, and ejects the seeds at a maximum distance of 4.1 m, indicating that the drifting capsule is capable for secondary dispersal by ballochory. At the study site, 2% of E. adenochlora subpopulations were within the range where fallen capsules could be dispersed by hydrochory, and 33% of E. adenochlora subpopulations could have been formed by ballochory from the drift line. These subpopulations are crucial for the conservation of E. adenochlora population at the study site. Furthermore, the fluctuating and maximum water levels from late May to mid-June for E. adenochlora were considered important to the formation of new subpopulations.

The buoyancy and ballochory experiment

In total, 164 capsules of E. adenochlora were randomly collected from several subpopulation s at the study site on May 23, 2023. Only ripe capsules were collected, although there were mixtures of ripe and brown capsules, unripe and green capsules, and capsules that had already fallen off. The collected capsules were sealed in plastic bags and taken to the laboratory, and floating experiments were started on the following day (May 24). Considering that some capsules had split open at the beginning of the experiment, 115 capsules were placed in a container filled with tap water and placed indoors. The air conditioning was switched off, so the temperature in the room was the same as outdoors, around 25°C. Approximately every 7 days, I counted the sunken capsules and ten capsules were randomly selected from the sunken capsules, dried on paper towels, and placed in a room (6 m × 7 m) unaffected by wind and rain. Approximately 7 days after placement, the number of capsules that split open and the distance between the dispersed seeds and the original place of the capsule was measured.

Estimation of the seed dispersal potential

The potential seed dispersal range of E. adenochlora was estimated based on the water depth of the inner lake estimated using the digital terrain model (DTM), the results of the buoyancy and ballochory experiments, and the distribution of E. adenochlora in 2023.

The distribution of E. adenochlora in 2023 was recorded using the method described by Niwa (2022); images were taken on April 5, 2023 using a DJI Matrice 300 RTK equipped with DJI P1. The images were processed by photogrammetry to produce orthomosaic images (1.1 mm/pixel), which were displayed in GIS at a scale of 1/15 to produce patch polygons of E. adenochlora via visual decipherment.

DTMs were generated from data measured by LiDAR mounted on an unmanned aerial vehicle (UAV). The measurements were taken on February 28, 2023, during the winter season when vegetation was scarce, using the DJI Matrice 300 equipped with DJI L1. The system accuracy of DJI L1 is 10 cm horizontal and 5 cm vertical. A flight route was created with a flight altitude of 60 m and a side lap of 70%, and measurements were taken at a flight speed of 5 m/s. The measured data were postprocessed in DJI Terra to obtain a 3D point cloud model, and GreenValley LiDAR360 v5.4 was used to remove outliers from the 3D point cloud model and classify the ground surface point cloud using default parameters. The point cloud of the ground surface was interpolated using the triangulated irregular network method to create a 5-cm DTM.

The water area on May 9, 2023, was estimated by locating the drift line on the orthomosaic image and extracting the elevation (water level) of the drift line from the DTM. Based on the data published by the Ministry of Land, Infrastructure, Transport, and Tourism, the water level of Lake Biwa on May 9, which was used as the reference date, and the maximum water level of Lake Biwa during the period when capsules can float, which was obtained from the buoyancy experiment, were determined. The water level difference from the base date was used to estimate the water area of the inner lake on the day when the maximum water level was reached.

The water area of the maximum water level during when capsules could float was defined as the area where the floating capsules could disperse. In addition, a buffer of the maximum seed dispersal distance, derived from the seed dispersal distance measurement, was created from the drift line on the day when the maximum water level was reached. This buffer was defined as the potential range of seed dispersal from the drift line. The area of the patch polygons of E. adenochlora in the water area and buffer on the day when the maximum water level was achieved was tabulated.