Impacts of vegetation restoration on soil physicochemical properties, bacterial communities, and metabolites in newly reclaimed croplands
Abstract. To find one suitable vegetation restoration type as a good means of restoring newly reclaimed croplands in subtropical of China. This study investigated the effect of vegetables, corn, and peach in soil properties, bacterial communities, and metabolites of newly reclaimed lands after three years restoration. Results from this study indicated that soil physicochemical properties were differentially affected by vegetation restoration of three different plants, while the effect depends on both the vegetation types and the kind of soil parameters. Indeed, the pH, soil bulk density (SBD), soil organic matter (SOM) and total nitrogen (TN) were generally unaffected except a significant reduction in SBD (13.97 %) and SOM (35.41 %) by vegetable and peach, respectively. However, three different plants significantly increased the available phosphorus (AP) (75.03–143.02 %), available potassium (AK) (154.90 % and 103.93 %) and microbial biomass carbon (MBC) (37.71–144.93 %), with the greatest increase by vegetable relative to the control except a significant reduction in the AK (41.73 %) by peach. Furthermore, the analysis of 16S rRNA gene high-throughput sequencing revealed that the vegetation of three plants increased the relative abundances (RAs) of soil bacterial phyla and genera with 6.21–10.54 % increase in operational taxonomic units (OTUs), 6.22–10.53 % increase in Chao1 and 2.30–3.11 % increase in Shannon indices, while redundancy discriminant analysis (RDA) revealed that the change of soil properties were highly related to the variation in bacterial community composition. In addition, 130 significantly differential metabolites (SDMs) that belong to organic acid, amino acid, heterocyclic compounds between vegetable and the control were identified based on liquid chromatography-mass spectrometry (LC-MS) analysis, while the top 20 SDMs were highly correlated with the 7 enriched bacterial genera. Overall, the results showed that the vegetation of three plants, in particular vegetable can ameliorate soil quality of newly reclaimed croplands by improving soil chemical properties, and increasing the richness and complexity of bacterial community structure, as well as specific bacterial genus and metabolites.
Review MS egusphere-2025-4266
While the paper provides an extensive analysis of the consequences of planting different crops/plants species as a soil restoration measure, the resulting understanding remains rather descriptive, describing observed changes, but not translating the results into a conceptual advance in understanding the contexts, interactions and trade-offs involved. And I mean this particularly from the perspective soil health, in a multifunctional perspective. Many microbial groups and metabolite groups are mentioned, and reference is made to biochemical pathways, but I expect an explicit link from them to soil functioning and restoration. As a reader of SOIL I think the current discussion falls below my level of expectation, although I do support the publication of this data and results in another place.
It is a complicated study in terms of techniques used, I am familiar with most, but metabolomics I only have a cursory understanding of, cannot judge the quality here. I suggest to get a reviewer who can.
The starting condition is poorly described, what were the sites reclaimed from? What was the nature of the degradation, the disturbance? Without this we cannot judge the meaning of the reported effects. The organic matter levels seem to be unexpectedly high for my understanding (5-10%!) on what is described as a sandy loam in the first 20 cm.
The experimental setup is one of a practical applied experiment. In the present ms all effects are subscribed to the different plant species put in place in the treatments. But at the same time the treatments vary in planting density, but also type, timing and amount of fertilizer used. The control, received no plant, but also no fertilizer. In addition, the planting types are not described in very much detail, what variety was used? For vegetables we even completely do not know which species/crops were used. In places a plural form is used suggesting multiple species were used, but did they grow together in the same year, or was it a rotation? What type of management was done in addition to the fertilization, was there any irrigation, and pesticides were they used? How was the land prepared for the crop planting? Any tillage or other site preparation measures?
Regarding analysis, why were OTUs used to describe microbial diversity and not ASVs? How many samples were run together on a Miseq plate? What was the read count per sample, before and after bioinformatics? Also you say, ‘Principal component analysis (PCA) was conducted using Bray-Curtis metrics’, but PCA exclusively works with Euclidean distances, so what did you do? PCoA? Or PCA? What do you mean with ‘ structural variation’ (L201), why not simply variation? Also, analyses and calculations are done at phylum, family and genus level, but the choice comes across as quite arbitrary – what is the rationale for doing one analysis at genus level and another at family level? How was the heatmap constructed, what correlation matrix is lying underneath? When was a bacterium considered dominant (L203)? Finally, an analysis is made of cross-correlations among microbial taxa and metabolite groups – but I wonder, what do these correlations mean? Quite some abiotic things change across planted vs unplanted controls, to what extend are these relationship confounded? How does fertilization yes/no and rhizodeposition from plant roots yes/no influence microbiomes and associated metabolite profiles? Also for the metabolites, they are extracted from soil, but to what extend do they originate from microbes or plants? And since the treated plots have more plants then the control, what does that tell us?
In many places the taxonomy of bacteria is not correctly represented. Often all taxa are referred to as genera, but they actually represent a mixture of different taxonomic levels, that also makes me wonder about the analyses, did you merge them, if so how? And is that valid? I think in a regression trees approach where taxonomy is the predictor you could do it meaningfully in one go. This makes me worried about the identification of the metabolites as well. I also don’t appreciate the use of the word vegetation type here, in ecology vegetation type (phytosociological entities) have a very specific meaning of a naturally shaped community. Here it is not even a community, it is one crop species, artificially put in place, and maintained that way. I would talk about plant or crop species effects, or cropping types.
One thing that worries me is how the measured properties are evaluated as an improvement or not, why is it better to have more OTUs? Often these things are done intuitively, which I understand, but we need to be critical, why is that really better? In intensive arable system usually bacterial diversity is higher then in more extensive systems, is that good? I don’t think so, I think the soil health is often better in the extensive systems, that have less bacteria, but more fungi, and as a consequence improved organic matter (quality), water infiltrating and holding capacity, and less nutrient leaching.
The conclusion largely repeats the results. No need to mention RDA and other experiment technical aspects – tell me 1) what is the message the data told you, 2) what are the wider implications for the research field, for practice, and perhaps beyond.
Minor
Title: can you include what the croplands were reclaimed from?
L28: why is lowering of OM by peach considered a soil improvement? This needs an explanation to be understood.
L30: metabolites extracted from what?
L30: do these gains have a cost somewhere to? Many of the processes involved are governed by (e.g. stoichiometric) trade-offs, that requires some consideration.
L31: what do you mean with ‘ a gain in soil’ ? also for microbes and metabolites – what increased? Abundance, richness, diversity, particular components (species, compounds)?
L32-33 this sentence is not complete. Also, why focus on one vegetation type? Why not a mixture, or several types that is spread over the landscape in a mosaic?
L34-35: reclaimed from what? What were they before? What is the degradation like?
L40-42, you need to tell something before about the situation in the control so readers can put these effects in context.
L45-47: you start with talking about relative abundances, but the evidence you give is about overall diversity levels. Please make consistent. And add some comma’s in the right places to help your reader.
L47: redundancy discriminant analysis (RDA), RDA in the literature stands for Redundancy Analysis, what the discriminant is doing in your term is unclear to me. You have linear discriminant analysis (LDA) or redundancy analysis (RDA) – mathematically related approaches, but with a different objective. From the figure I think you did RDA.
L51 I don’t think the ‘ while’ is appropriate here, it suggests something else is coming. While we observed X, we also saw Y.
L122 what do you mean with ‘site’ here? How many sites were there? How far apart? Differences in slope, rainfall, etc?
L123-125: I want to see what the variation is in these base variables.
L127: change the word restoration here to planting/sowing or so (what fits), because you don’t know if the treatment will not restore anything, so while it is the goal, you cannot be sure that it does what you think until you tested it.
L127-8 describe this newly reclaimed cropland
L130-131 okay so you have a compound treatment, combining plant species and fertilizer application in one combined measure. Can you rescale the amounts of fertilizer to equivalent per hectare? Also, can you give a sense of the N-input thus realized?
L133 what is in the compound fertilizer? Please give NPK amouns, and micronutrients if included.
L136: the control was sampled to the same depth? Now it reads a bit confusing, I know in the control there is no root zone, but we need to be sure you used the same setup.
L138: what do you mean with quickly? 1h, 1day, 1 week? Make precise.
L165: what do you mean with ‘clean read’.
L429: also here the sentence after ‘ while’ seems unfinished, the greater what was observed?
L431: what is ACE?
L447: please write out Ras here and elsewhere.
L461: “The improvement of microbes in soil quality” I don’t get what you mean here. Please clarify.
L463-471 how likely is it that the strains you found in your plots, do the same as other members of their genus/family in another location?
L498: what is FA and GP
L502-504: but do they have a function in the soil itself?
L518: write out AK too. Okay, and is that study relevant in this case? I don’t know it, but context matters hugely, so be careful with extrapolations.
L505: essential is too strong in my opinion, plants can also grow on mineral nitrogen
L523: what do you mean by connection here, I assume correlation? Or do you infer a functional dependence? Does microbe A produce compound X? if the latter, you need stronger evidence then an RDA or heatmap.
L525: what do you mean with CAR and FFA? Write out.
L550: Vicinamibacterales is a family, not a genus. Check the others too and correct mistakes. Is A4b a complete name?
L552: identified as what?
L526: there is not one organic acid, nor one nucleotide – including derivatives these seem like large heterogenous groups to talk about, is it meaningful to lump them?
Fig 1. Please add a picture of the control. When were photos taken and the data, how many years into the study is this?
Table 2. it is unclear what the values represent? Community composition quantified how, in what units? Is the post-hoc analysis done across the table, or can I only look across columns, or rows?
Fig. 3. In b that’s not only genus level, it’s a mix of taxonomic levels. Please correct this.
Fig 4. In b the rendering is way to small to make any sense of whats in the data.
Fig. 5. OPLS-DA is an abbreviation for orthogonal partial least squares discriminant analysis, not what you call it. Is it the same technique or not? If so, please use the common term. How do you read that donut plot? Is it a relative abundance visualization or something else? If the former, can you switch to stacked barchart, its easier to see the pattern. If the latter, more guidance on how to read the result is needed.
Fig. 6. What is VIP here? The labels and plots are too small to read, Improve the readability.
Fig. 7. Typically RDA stands for Redundancy analysis. I don’t know what Redundancy discriminant analysis is, I don’t think it exists, unless you mean Linear Discriminant Analysis? But the plot suggests classic RDA was used. Please fix for clarity. Please spell out the abbreviations for soil abiotic variables at first mention. What is SDM, please write out?
E.R. Jasper Wubs