The following technical concerns are identified, each pointing to potential errors or limitations in the study design, data interpretation, or methodological approach:
1. The study use 5 replicates per treatment, but all replicates are within the same vineyard under the same management. The experimental design do not appears to include true spatial replication (e.g., multiple vineyards or geographically separated blocks). This raise the risk of pseudoreplication, where environmental or management confounders (e.g., microclimate, vine age, soil heterogeneity) are not adequately controlled.
2. The authors use FUNGuild to assign ecological functions (e.g., saprotrophs, pathogens) to fungal ASVs. However, FUNGuild is a predictive tool based on literature and databases, not experimental validation. Many fungal taxa have multiple ecological roles or are poorly characterized.
3. The authors construct co-occurrence networks from the top 50 ASVs and interpret network metrics (e.g., connectivity, stability) as indicator of microbial community health or function. However, correlation-based networks are highly sensitive to Compositional bias in sequencing data, Filtering thresholds (e.g., top 50 ASVs), and Lack of causal or functional inference.
In Section 2.1, you mention that the vineyard follows a three-year alternate inter-row management: each year, only one inter-row is tilled (0–20 cm), while the other two remain untilled and grass-covered. Biochar was applied only to the inter-rows in 2009/2010 and incorporated to 30 cm depth at that time.
Since the tillage rotates annually, the physical location of the tilled inter-row, and thus the depth of recent soil disturbance, changes each year relative to the vine rows. When you sampled in 2019 (10 years later), did you account for which inter-row was tilled in the sampling year or in the years just before sampling?
If bulk and rhizosphere samples were taken without considering the tillage rotation, the immediate disturbance history and root distribution could differ substantially between plots and treatments. This might introduce uncontrolled variability in soil structure, organic matter distribution, and microbial communities independent of biochar effects.
Could you clarify 1. Whether the sampling distinguished between tilled and untilled inter-rows in 2019? 2. How you ensured that bulk soil samples were comparably influenced by recent management across all treatments? 3. If this was not controlled, it might affect the robustness of the bulk vs. rhizosphere comparisons and the attribution of observed differences solely to biochar.