2.1 Study area
The study area is located in the Tama Hills, 30 km west of Tokyo (35°35′N, 139°25′E; Fig. 1). The mean annual precipitation at the nearby Hachioji meteorological station is 1540 mm. The mean annual temperature is 14.3°C, with a mean minimum of 3.2°C (January) and mean maximum of 26.1°C (August) (Japan Meteorological Agency 2021). Monthly precipitation in 2019 and 2020 is illustrated in Appendix 1; precipitation in January and February 2019 and August 2020 was less than the monthly average.
In the Tama Hills, hill ridges are about 30 m higher than the adjacent valley bottom. The valley bottoms are used as paddy fields, and ridges and hillsides are used for coppiced woodlands, a land use pattern that is a typical feature of Japanese agricultural landscapes. The valley bottoms have good water availability for rice culture, but the shade from adjacent ridges and their standing trees in the narrow valleys is disadvantageous. To improve light availability, standing vegetation has been periodically mown on the lower slopes (so-called Susogari grassland). In the study area, the lowermost slopes have traditionally been mown two or three times a year (Okubo et al. 2005; Yamada et al. 2005).
The study grassland is on a southeast-facing slope (S 45° E) at a mean angle of 52.6° ± 7.2° (SD), and the slope was mown in a band about 7 m wide. The woodland upslope of the grassland is dominated by the deciduous broad-leaved species Quercus serrata, Quercus acutissima, and Castanea crenata, formerly coppiced but left unmanaged for about 50 years. Rice culture in the valley bottom adjacent to the study plots was abandoned in 2016, when mowing also stopped.
2.2 Reintroduction of management
The study area lies inside the 33-ha Zushi-Onoji Historical and Environmental Conservation Area, which was proclaimed a Greenery Designated Conservation Area by the Tokyo Metropolis in 1978. The government entrusted the management to the residents, who established the privately run Machida Historic Environment Conservation Management Organization. Since 1997, the organization has managed the abandoned land as was previously done, aiming to restore the former agricultural landscapes in the area. The organization clear-cut 0.4 ha of the woodland in November 2017, when the slope below was covered with tall herbaceous vegetation. In October 2018, the slope was mown after having a gap of 3 years. Since then, the site was periodically mown (i.e., on 7 June 2019, 7 August 2019, 2 June 2020, and 5 August 2020) by the organization or some authors (Yamada S, Yoshida W and Iida M).
2.3 Monitoring
We monitored vegetation in 2008, a period under continuous management; in 2018, 3 years after the management was terminated; and in 2019 and 2020, the first and second years of management reintroduction. This grassland is organized into relatively small, discrete linear elements because of the abrupt slope angle; hence, we established ten 1-m × 1-m plots (quadrats) along a transect at intervals of 1 m. In each plot, we visually estimated the percentage cover of each species on 28 June 2008, 26 June 2018, 28 June 2019, and 2 June 2020. Small individuals with primary leaves were rarely able to be identified in 2008, so we did not survey such individuals in 2018, 2019, and 2020. In 2019 and 2020, several seedlings recorded in the seedling emergence survey (see below) were not recorded in the vegetation survey. In 2019, many seedlings and sprouts (vegetative regrowth) began to grow, both of which are assumed to affect subsequent vegetation development, so in June 2019, each perennial and woody species was recorded along with its growth stage (seedling or sprout). In addition, vegetation cover was visually estimated on 10 May 2019, 5 August 2019, 4 May 2020, and 7 August 2020 to reveal the relationship between vegetation cover and seedling survival.
Seedling emergence and survival was monitored once a month from May to October in 2019 and from April to October in 2020. To distinguish the contribution of germination and subsequent survival to the overall rate of establishment, allowing seedlings to become a permanent part of the vegetation, we marked every germinated individual with a numbered plastic tag. To assess seedling recruitment, we identified recruited seedlings as those that still had their primary leaves and were not recorded in a previous census. Individual seedlings were then monitored in subsequent censuses. We were unable to find about a third of the tags owing to severe soil disturbance at the site. Because we were unsure of whether the plants were still alive nearby without tags or had died, we defined them as disappeared and eliminated them from the calculation of survival rates. Flowering of individual plants, if any, was recorded.
Surface soil was sampled to a depth of 5 cm with a 4-cm-diameter core sampler on 25 December 2020. Sampling was conducted at four points where the vegetation survey was not done (see Fig. 1); the center and four points equidistant from the center and each edge of the 1-m × 1-m unsurveyed points were sampled. The nitrate-N content in the surface soil was 0.5 mg/100 g, available P content was < 1.0 mg/100 g, and soil pH was 6.1, suggesting nutrient-poor conditions.
In early October 2020, light conditions were measured in the center of quadrats; a camera with a 180° hemispherical lens (Nikkor 8-mm f/2.8 fisheye lens) was positioned facing upward at 50 cm above the ground. Relative solar radiation was calculated in Gap Light Analyzer v. 2.0 software (Frazer et al. 1999), and the relative solar radiation was 66.6% ± 8.8%.
2.4 Analysis
Following Chibaken-shiryou-kenkyu-zaidan (2003), all species were classified according to their life histories: woody species; annual or biennial species (hereafter “annuals”); and perennial herb species, distinguished between small-statured plants (<30 cm) and larger perennials. Plant height classification was applied only to perennials, according to the reference above. Characteristic species representing grassland vegetation in Miscanthetea sinensis communities were defined as “typical grassland species” (Miyawaki 1994). Nomenclature is based on BG Plants (http://ylist.info).
About half the seedlings in the genus Cirsium were identifiedas either C. oligophyllum or C. japonicum, whereas other Cirsium seedlings died before identification. According to the flora list in the conservation area, unidentified Cirsium seedlings were likely to be either C. oligophyllum or C. japonicum; thus, the three categories were combined, and survival rates are given for Cirsium spp. The two Cirsium species represent semi-natural grasslands (characteristic species in Miscanthetea sinensis communities; Miyawaki 1994). We consider the establishment of Cirsium spp. to be informative for understanding the fate of typical grassland species. Carex spp. were assumed to be one of four taxa: C. lenta var. lenta, C. lanceolata, C. leucochlora var. candolleana, and C. tristachya. No seedlings could be identified to the species level, so all seedlings were combined and reported at the genus level.
Established vegetation communities in 2008, 2018, 2019, and 2020 were tested by principal component analysis (PCA) to confirm any differences in species composition among years. Presence/absence data were used for the PCA, which was conducted in PC-ORD for Windows v. 6 software (McCune and Mefford 1999).
We performed Student’s t-test and Holm’s post hoc tests to determine the significance of differences in the number of species. A P-value of <0.05 was considered significant. Statistical analyses were performed in R v. 2.13.1 software (R Core Team 2015).