2.2 Experimental set up
In each forest we identified an area of 30 m × 20 m for establishing the experiment. Two factors, type of deposited compounds (control, N, Na, PAHs) and earthworms (with and without), were investigated. In each forest, each treatment (deposited compounds × earthworms) was replicated four times resulting 32 experimental units comprising of individual mesocosms, which were installed from February to April 2018 (for details see Fig. S1). For installation we first dug up an area of 0.5 m × 0.5 m to a depth of 0.2 m, i.e., a volume of about 50 L. From the excavated litter and soil material we hand-sorted earthworms. Then, we placed a 1 m × 1 m nylon bag into the pit (0.16 mm mesh size) and filled back the excavated soil and litter to fit the natural layering, doing that herb seedlings were removed. The nylon bags were then closed by zipper at the top and were covered with leaf litter. The pits were located away from main roots of trees. Individual mesocosms were established as fast as possible (< 30 min) to minimize effects of sunshine and drying. The mixing of soil broke the initial hotspots in soil surface, e.g., the worms’ burrows and the rhizosphere of fine roots; and the nylon bags blocked the exchange of macrofauna, parts of mesofauna, run-off and nutrients. Thus, the plots may slow down the rates of litter mass loss but overall the differences in micro-environment between plots were reduced. The distance between mesocosms was 3 – 5 m, and mesocosms were located at least 20 m away from the border of the forest to avoid edge effects.
After finalizing the establishment of the mesocosms in April 2018, we placed 10 litterbags (20 cm × 10 cm) on the soil surface within each mesocosm (Fig. S1d). Litterbags were either filled with litter ofQ. variabilis or P. massoniana , which were collected from December 2017 to January 2018 and dried at 40 °C for one month (for litter traits see Table 1). To disentangle effects of different fauna groups, two mesh sizes were used, fine mesh of 0.2 mm and coarse mesh of 5 mm. The fine and coarse mesh litterbags were used to evaluate the net contributions of microbe- and fauna-driven mass loss and were filled with 4 and 8 g litter (dry mass), respectively (Yin et al., 2022). The higher amount in the large mesh size litterbags was used as we assumed the litter to be decomposed faster due to access by macrofauna. The period of litter decomposition was 365 days from April 2018 to April 2019.
To manipulate deposited compounds, we added 500 mL aqueous solutions of NH4NO3, NaCl and PAH to the N, Na and PAH treatment every 35 days. Control mesocosms received 500 mL distilled water. The aqueous solutions of NH4NO3were added following Lin et al. (2017) and were equivalent to the mean annual deposited amount of N in the region of Nanjing (47 kg N ha-1 y-1). NaCl was added at a Na mass percentage of 0.5 % and was equivalent to a rate of 39.36 g Na m-2 y-1. The Na addition followed Jia et al. (2015) simulating future Na accumulation due to the input of Na by human activities e.g., via road salt (Li et al., 2016; Tiwari & Rachlin, 2018). We assumed this concentrations to detrimentally affect litter decomposition due reducing microbial activity (Ji et al., 2020). For PAHs, we included fluoranthene (Flu), pyrene (Pyr), chrysene (Chr), benzo[a]pyrene (BaP) and phenanthrene (Phe). The five PAHs accounted for 54% of the mass of sixteen prioritized PAHs in the soil of the urban region of Nanjing and Zijin Mountain (Wang et al., 2015). The PAHs added to a total of 128 mg per microcosm per year, which is equivalent to 1.813 μg g-1 dry soil y-1 and thus doubled the total amount of soil PAHs in study sites (Wang et al., 2015).
For investigating the actual effect of earthworms in field, treatments with earthworms received a total of 60 individuals of E. fetidaper mesocosm in deciduous and 20 individuals in coniferous forest resembling the density in these forests as investigated in 2018 (Table S2). E. fetida dominated 55.47% and 66.91% earthworm abundance in deciduous and coniferous forest as we investigated. Earthworms were picked by hand from the mesocosms, counted and placed back during May to August 2019 to validate the treatment. From April 2018 to April 2019, litterbags and soil samples were taken at 70, 140, 210, 280, 365 days resulting in a total of 640 litterbags (2 forests × 4 deposited compounds treatments × 2 earthworm treatment × 2 mesh sizes × 4 replicates × 5 sampling dates) and 640 soil samples (0 – 5 cm depth under litterbags). At the 140-day sampling, one nylon bag in deciduous forest was found broken and excluded from the analysis.