Differing agents of physical damage to artificial seedlings in remnant and restored forests

We tested the importance of physical damage to seedlings as a bottleneck to natural regeneration in restored tropical forest using an artificial seedling experiment. After one year more artificial seedlings had been damaged in restored forests (17.6 ± 2.6% SE) compared to remnant forests (12.7 ± 1.1% SE), primarily due to greater frequency of woody debris falling from planted trees. Vertebrate damage was more common in remnant forests.

Methods
Artificial seedling experiment

We placed 1024 artificial seedling models in the eight restored plots and five remnant forests in early May 2022. In most locations, we placed a total of 64 artificial seedlings (one replicate), except for one remnant forest area where we placed 256 seedlings in four replicates of 64 each that were separated by a minimum of 17 m. Within each replicate, artificial seedlings were distributed in four groups of 16 with ≥ 10 m separation. Each group of seedlings was distributed in an X pattern with ≥ 2 m between each artificial seedling (Figure S1) and minor location adjustments to accommodate the presence of obstacles such as large trees.

Each artificial seedling consisted of two 18-cm long × 0.5-cm diameter plastic straws following the design of Clark and Clark (1989) and Rosin et al. (2017), and painted pink or yellow to facilitate observation in the field. The straws were stapled in a "T" shape and attached to a 13-cm "root" wire using a plastic cable tie and a small piece of electrical tape, with 3 cm of overlap to allow burial 10 cm in the soil.

We monitored artificial seedlings over a 12-month period, monthly for the first seven months (June-December 2022) and subsequently at months nine and 12 (February and May 2023). We followed Clark and Clark (1989) for the classification of damaged or undamaged artificial models; the artificial model was considered as damaged if at least one of the T arms touched the ground.

 We classified the types of seedling damage (Figure 1) in the following manner when possible. Observations of animal activity, such as digging and trampling were categorized as vertebrate damage. Woody debris refers to crushing by branches or trunks, whereas leaf damage refers to when leaves damaged and covered artificial seedlings in the absence of woody debris. Water damage refers to when seedlings were excavated by water flow and erosion.