Design of soil preservation conditions
We did not test full-factorial combinations of different preservation periods and conditions, and we did not consider freezing, which is unrealistic when dealing with large numbers and / or volumes of samples, as is the case for more and more metabarcoding studies. Furthermore, freezing is generally impossible when sampling remote areas, where maintaining a cold chain cannot be ensured given the logistical challenges and, in the best cases, it is replaced by preservation in a cool box (i.e. 4°C or more).
In the design of treatments, we considered approaches allowing preservation at different temperatures and for different periods. For preservation condition 2, we accounted for a certain delay (six hours) between sampling and extraction, that could correspond to local transportation from the sampling area to the nearest base station. Sometimes, even in the case of an in situ extraction, samples remain at ambient temperature for hours prior to extraction, especially when monitoring a large area or a tricky ground, but this can have an impact on the final results (Delavaux et al., 2020). Preservation at 4°C is among the most frequent approaches to soil preservation (Dickie et al., 2018). This can be attained through portable refrigerators and requires the cold chain not to be interrupted at any point during transportation, which is only possible when lab facilities are accessible in a relatively short time (Hoffmann et al., 2016; Huerlimann et al., 2020). However, one of the most attractive characteristics of eDNA metabarcoding is its capacity to provide biodiversity data for understudied areas (e.g. tropical and arctic areas; mountain chains). Preservation conditions 4 and 5 of this study refer to the situations where samples are located far from lab facilities, in areas for which traditional biomonitoring is particularly challenging. When sampling these areas, the time lag between sample collection and subsequent molecular analysis can be particularly relevant. The use of silica gel to preserve soil samples is then a good solution (Chase & Hills, 1991; Guo, Yang, Chen, Li, & Guo, 2018): silica gel allows removing up to 26% of its weight in water in a relatively short time (overnight; Taberlet, pers. communication), is cost-effective, easy to transport, and is not an issue for aircraft transportation (no flammable or dangerous preservatives). Silica gel can be added in situ during sampling (protocol 4) or within a short time lag, allowing for local transportation from the sampling area to the nearest facilities (protocol 5). Afterwards, samples can be stored at room temperature, protected from sunlight to avoid photo-degradation of DNA.