the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Jul 2026
Crop diversification shapes microbial communities and alters carbon-use efficiency across European agricultural soils
Abstract. Crop diversification practices are expected to influence soil carbon (C) dynamics by altering microbial communities, yet their effects across environmental contexts and on carbon use efficiency (CUE) remain unclear. We tested whether diversified cropping systems modify microbial alpha diversity and community composition, and whether they alter diversity-function relationships. To do so, we analysed fungal and bacterial communities and CUE in soils from six countries along a pan-European gradient. Microbial composition was strongly structured by environmental context, particularly for fungi. After accounting for this context, crop diversification explained a small but significant fraction of variation in both bacterial and fungal communities. However, responses differed between microbial groups, with bacterial communities showing consistent but minor shifts, whereas fungal responses were highly context dependent. Crop diversification did not affect bacterial alpha diversity, measured as richness and Shannon diversity, but had a marginal, context-dependent effect on fungal richness. Although crop diversification did not directly affect CUE, it modified the relationship between fungal Shannon diversity and CUE, with a weak negative association under control conditions which disappeared under crop diversification. This interaction was driven by microbial respiration rather than microbial growth, indicating shifts in C loss pathways. Overall, our findings show that crop diversification subtly reshapes microbial communities within strong environmental constraints and can alter diversity-function relationships, highlighting the context dependency of soil C stabilization.
- Preprint
(906 KB) - Metadata XML
-
Supplement
(254 KB) - BibTeX
- EndNote
Status: open (extended)
-
RC1: 'Comment on egusphere-2026-3526', Anonymous Referee #1, 23 Sep 2026
reply
-
AC1: 'Reply on RC1', Beatriz Andreo-Jimenez, 06 Oct 2026
reply
We appreciate the constructive comments. Please find below our responses to them:
1. The study pools three distinct practices into one diversified category. This is biologically unjustified and likely masks practice-specific effects. Authors should further improve the relevant content in the analysis and discussion.We agree that the three practices differ biologically and that pooling them may obscure practice-specific responses. Because practices were associated with particular sites rather than independently replicated across sites, their effects cannot be separated from site-specific environmental and management conditions. We therefore used the pooled analysis to estimate the average within-site response to locally relevant diversification treatments relative to their corresponding controls, without assuming equivalent effects across practices.
We have clarified this rationale in the Materials and Methods and revised the Discussion to acknowledge potential differences among practices and limit interpretation to the practices and contexts represented in this study.
2. Experiment design details are not reported in this manuscript. Soil cores were composited per plot, making each plot a single observation. Could the authors clearly report plot details, or discuss the associated limitations?
We have clarified the sampling design in the Materials and Methods and added Table 1 summarizing the treatments, sampling seasons and sample sizes at each site. Within each experimental plot, soil cores were collected following a W-shaped pattern and combined into one homogenized composite sample. Each plot therefore represented one independent experimental unit; individual cores were sub-samples, not independent replicates.
We also acknowledge that compositing provides a plot-level estimate but prevents assessment of within-plot spatial variability.
3. The same issue as mentioned above, CUE measurement method is omitted, provide a dedicated CUE methods subsection.
We agree that the CUE methodology should be summarized in the manuscript. Microbial C growth, C respiration, and CUE were not measured specifically for the present study; plot-level data were obtained from Schroeder et al. (2025). We have therefore added a dedicated subsection briefly describing the original measurement procedure and referring readers to that publication for the complete protocol.
4. The Introduction fails to review environmental controls on soil microbes and CUE, yet the study’s conclusion is that diversification effects are strongly context-dependent. This disconnect leaves the research questions without a solid foundation. The authors must rewrite the Introduction. Without this, the rationale for the multi-site design and the significance of subtle, context-dependent results remain weak.
We have revised the Introduction to explain how environmental conditions shape soil microbial communities and regulate CUE through their effects on microbial growth and respiration. Drawing on the perspective by He et al. (2024), we now discuss temperature, moisture, nutrient availability, substrate quality, pH and soil structure, together with microbial acclimation and interactions among these controls. This framework strengthens the rationale for the multi-site design and explains why diversification responses may be modest and context-dependent. We have also reformulated the objectives and hypotheses accordingly.
5. Fig. 1 shows that country-season explains 25.1 % (bacteria) and 72.1 % (fungi) of community variation, whereas crop diversification accounts for only 1.6 % and 1.7 %, respectively. Despite this stark contrast, the manuscript continues to centre its narrative on diversification effects? The authors further argue that high residual variation in bacteria suggests greater within-context heterogeneity or stronger influence of unmeasured local soil and management factors. Invoking unmeasured variables merely highlights the study’s failure to capture dominant drivers and further undermines the relevance of diversification. If local factors exert stronger effects, the current design cannot support claims about diversification.
We agree that country-season accounted for substantially more community variation than diversification. We have revised the manuscript to emphasize this contrast and interpret the diversification-associated differences in proportion to their small effect sizes. We have also revised the title to better reflect the context-dependent nature of these associations.
Country-season and diversification represent different components of the design: the former captures differences among contexts, whereas the latter is evaluated through replicated comparisons between diversified treatments and their corresponding controls within each context. The dominance of contextual variation therefore does not invalidate these comparisons, although accounting for country-season cannot remove all potential confounding or separate practice-specific effects from site conditions.
We now describe the diversification associations as modest and context-dependent. The fungal association remained detectable in leave-one-context-out analyses, indicating the stability of a small association rather than a large ecological effect. The bacterial association was more sensitive to context removal, becoming non-significant when either Spain autumn or Slovenia autumn was excluded. We have also removed the speculative attribution of bacterial residual variation to unmeasured local drivers.
6. Fig. 1B shows that leave-one-context-out analyses indicate a robust fungal diversification effect, yet the explained variance remains extremely low (about 1.2 %–2.1 %). While robustness to single-context removal is statistically useful, it does not compensate for the trivial effect size. The manuscript currently uses this robustness to imply biological relevance, which is misleading.
We agree that robustness and effect magnitude are distinct properties of the result. The leave-one-context-out analysis shows that the pooled fungal diversification effect is not dependent on any single country-season combination. However, it neither increases the variance explained nor demonstrates that the effect is biologically important. We have therefore revised the manuscript to avoid using statistical robustness as evidence of biological relevance.
The fungal result is now described as a small but statistically stable pooled association, with diversification explaining only approximately 1.2–2.1% of community variation across the leave-one-context-out analyses. We explicitly acknowledge that its ecological importance cannot be established from the explained variance or statistical significance alone and that country-season remains the dominant source of variation.
7. The study spans a large environmental gradient, and while linear mixed-effects models were used, they cannot fully account for unmeasured site-specific factors (soil properties, land-use history, detailed management). These factors strongly influence both microbial composition and CUE (Fig. 3 shows CUE is highly structured by country-season). Consequently, the detected crop diversification effects, though statistically significant, may be biased or inflated by residual confounding. The authors must exercise extreme caution in interpreting causality and should explicitly discuss these limitations.
We agree that mixed-effects models cannot account for all potentially relevant environmental and management variables. Diversification responses were estimated from replicated comparisons between control and diversified treatments within each country-season context, rather than from comparisons among countries. Including country-season captures broad contextual variation but does not explicitly control for individual environmental drivers or unmeasured factors that may covary with treatment. Residual confounding therefore cannot be excluded.
We have revised the Discussion to acknowledge these limitations and avoid interpreting statistical significance as evidence of causality. The estimated responses are now described as modest within-context associations observed across the field trials. Because diversification practices were associated with particular sites, their contributions cannot be separated from local conditions, and the pooled results should not be generalized to all practices or environmental contexts.
8. While the Discussion compares findings with existing literature, it avoids a deep analysis of the cross-regional results, particularly why crop diversification had negligible effects on microbial structure and CUE across diverse environments. This omission leaves the core context-dependency unexplained.
We agree that the cross-regional patterns required a more explicit interpretation. We have expanded the Discussion to consider why diversification was associated with small compositional differences and limited CUE responses. Changes in plant inputs may be moderated by soil properties, climate, season, management history, and the type, duration and intensity of the locally implemented practice. Community resistance may limit compositional shifts, whereas functional redundancy may buffer their consequences for microbial processes. Proportional changes in microbial growth and respiration could also leave CUE largely unchanged.
We now discuss these potential explanations in relation to the observed country-season patterns. Because the relevant contextual factors were not independently manipulated and practices were associated with particular sites, we present these explanations as hypotheses consistent with the results, rather than demonstrated mechanisms.
9. The Methods section fails to detail site selection, experimental design, and soil sampling procedures, merely directing readers to another publication. This is inappropriate. Key methodological details must be self-contained. The authors must provide a comprehensive description of the sampling design, including site characteristics, replication, and justification for representativeness, directly in this paper. Omitting such details undermines transparency and reproducibility, and weakens confidence in the results.
We agree that essential design and sampling details should be reported directly in this manuscript. We have expanded the Methods to describe soil sampling, compositing and the experimental unit, and added Table 1 summarizing the countries, diversification practices, corresponding controls, sampling seasons and sample sizes. Schroeder et al. (2025) provides supplementary site-specific details.
10. In Fig. 1b, the color scheme for significance is too subtle; use more distinct colors or patterns to improve discriminability. In Fig. 2b, a line plot incorrectly connects two distinct indicators, implying continuity where none exists; replace with bar charts or point plots without connecting lines to accurately represent independent metrics.
We agree with the comment and have adapted the figures accordingly.
-
AC1: 'Reply on RC1', Beatriz Andreo-Jimenez, 06 Oct 2026
reply
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 281 | 114 | 36 | 431 | 56 | 35 | 34 |
- HTML: 281
- PDF: 114
- XML: 36
- Total: 431
- Supplement: 56
- BibTeX: 35
- EndNote: 34
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This study examines the effects of crop diversification on soil microbial communities and carbon use efficiency (CUE) across a multi‑site European gradient. Based on seasonal sampling and integrated analyses of bacterial and fungal composition alongside CUE, the authors conclude that diversification subtly reshapes microbial communities and alters diversity‑function relationships, without directly enhancing CUE. While the large‑scale design and joint assessment of compositional and functional responses are strengths, several methodological and analytical limitations compromise the robustness of these conclusions, as detailed in the specific comments below.
1. The study pools three distinct practices into one diversified category. This is biologically unjustified and likely masks practice-specific effects. Authors should further improve the relevant content in the analysis and discussion.
2. Experiment design details are not reported in this manuscript. Soil cores were composited per plot, making each plot a single observation. Could the authors clearly report plot details, or discuss the associated limitations?
3. The same issue as mentioned above, CUE measurement method is omitted, provide a dedicated CUE methods subsection.
4. The Introduction fails to review environmental controls on soil microbes and CUE, yet the study’s conclusion is that diversification effects are strongly context‑dependent. This disconnect leaves the research questions without a solid foundation. The authors must rewrite the Introduction. Without this, the rationale for the multi‑site design and the significance of subtle, context‑dependent results remain weak.
5. Fig. 1 shows that country‑season explains 25.1 % (bacteria) and 72.1 % (fungi) of community variation, whereas crop diversification accounts for only 1.6 % and 1.7 %, respectively. Despite this stark contrast, the manuscript continues to centre its narrative on diversification effects? The authors further argue that high residual variation in bacteria suggests greater within‑context heterogeneity or stronger influence of unmeasured local soil and management factors. Invoking unmeasured variables merely highlights the study’s failure to capture dominant drivers and further undermines the relevance of diversification. If local factors exert stronger effects, the current design cannot support claims about diversification.
6. Fig. 1B shows that leave‑one‑context‑out analyses indicate a robust fungal diversification effect, yet the explained variance remains extremely low (about 1.2 %–2.1 %). While robustness to single‑context removal is statistically useful, it does not compensate for the trivial effect size. The manuscript currently uses this robustness to imply biological relevance, which is misleading.
7. The study spans a large environmental gradient, and while linear mixed-effects models were used, they cannot fully account for unmeasured site-specific factors (soil properties, land-use history, detailed management). These factors strongly influence both microbial composition and CUE (Fig. 3 shows CUE is highly structured by country-season). Consequently, the detected crop diversification effects, though statistically significant, may be biased or inflated by residual confounding. The authors must exercise extreme caution in interpreting causality and should explicitly discuss these limitations.
8. While the Discussion compares findings with existing literature, it avoids a deep analysis of the cross‑regional results, particularly why crop diversification had negligible effects on microbial structure and CUE across diverse environments. This omission leaves the core context‑dependency unexplained.
9. The Methods section fails to detail site selection, experimental design, and soil sampling procedures, merely directing readers to another publication. This is inappropriate. Key methodological details must be self‑contained. The authors must provide a comprehensive description of the sampling design, including site characteristics, replication, and justification for representativeness, directly in this paper. Omitting such details undermines transparency and reproducibility, and weakens confidence in the results.
10. In Fig. 1b, the color scheme for significance is too subtle; use more distinct colors or patterns to improve discriminability. In Fig. 2b, a line plot incorrectly connects two distinct indicators, implying continuity where none exists; replace with bar charts or point plots without connecting lines to accurately represent independent metrics.