Assessment of Soil Injection Treatments in Street Tree Beds: Impacts on Soil Microclimate and Foliar NDVI

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The title emphasizes treatment effects through the phrase “Impacts on Soil Microclimate and Foliar NDVI”; however, soil moisture and temperature sensors were actually installed on only one tree per treatment. It is therefore questionable whether the treatment effects are supported by sufficient independent experimental replication.

 

The Abstract states that the combined Mycorrhiza + Hydrogel treatment produced the greatest improvement in soil moisture. However, at the Brno site, the Aeration treatment showed the highest soil moisture in both soil layers, and the mean values under the Mycorrhiza treatment alone were also higher than those under the combined treatment. Thus, the summary in the Abstract is not fully consistent with the reported results.

 

The Abstract states that reductions in soil temperature occurred mainly under the Mycorrhiza + Hydrogel treatment. However, at the Brno site, Aeration significantly reduced soil temperature at both monitored depths, with a reduction of up to 1.98 °C in the deeper layer. Therefore, this statement does not fully reflect the results.

 

The Introduction hypothesizes that Hydrogel reduces soil moisture variability and that the combined treatment improves “moisture and thermal stability”. However, the Results mainly compare mean soil moisture and mean soil temperature, without statistical results for coefficients of variation, fluctuation amplitudes, or other indicators of stability. Thus, some of the stated hypotheses do not fully correspond to the variables actually evaluated.

 

The differences between Brno and Znojmo are mainly used to explain how contrasting soil structural conditions affect treatment performance. However, the two sites also differ in city, roadside environment, soil type, and management background. Consequently, site effects and soil structural effects cannot be readily separated.

 

Brno is characterized as having heavy, compaction-prone soil, whereas Znojmo is described as having lighter, skeletal soil. However, relatively few field measurements are provided to substantiate these contrasts. Key soil physical properties, including bulk density, particle-size distribution, porosity, penetration resistance, field water-holding capacity, and hydraulic conductivity, are not reported.

 

Section 2.4 states that each treatment was replicated twice at each site, resulting in a total of 20 experimental trees, whereas Section 2.6 explicitly states that soil moisture and temperature were monitored for “one tree per treatment”. This represents a clear inconsistency between the nominal experimental design and the actual experimental units used for the core soil microclimate measurements.

 

The Mycorrhiza, Hydrogel, and Mycorrhiza + Hydrogel treatments were all subjected to compressed-air treatment at 6–7 bar before amendment injection. These treatments therefore also included an Aeration effect. Relative to the completely untreated Control, the independent effects of the amendments cannot be fully separated from those of pneumatic soil loosening.

 

The Mycorrhiza, Hydrogel, and combined treatments were all injected using 30 L of water as the liquid carrier, whereas the Control and Aeration treatments did not receive an equivalent 30 L liquid input. Because soil moisture is a central response variable in this study, this additional water input itself represents an important treatment difference.

 

The product used in the so-called “Mycorrhiza” treatment, Endomyk PROF + TRI, contains both AMF and Trichoderma spp. and therefore is not a single mycorrhizal fungal inoculant. In addition, no evidence is provided regarding root mycorrhizal colonization rates or successful establishment of the inoculum. Consequently, the biological source of the observed effects cannot be clearly distinguished.

 

Irrigation, fertilization, mulching, and other municipal maintenance practices at the two urban sites during the three-year experiment are not clearly described. These management factors could directly influence root-zone moisture and foliar NDVI of young street trees, particularly during years of extreme drought.

 

No pretreatment baseline data are provided for soil moisture, soil temperature, or NDVI. The first NDVI measurements were conducted 15–16 weeks after treatment application; therefore, it is difficult to determine whether subsequent differences among treatments were entirely attributable to the treatments without pretreatment data.

 

In the NDVI methodology, NDVI is described as “Relative chlorophyll content”. However, NDVI is a spectral index based on red and near-infrared reflectance and is not equivalent to direct measurement of chlorophyll content. No independent chlorophyll measurement was conducted to validate this interpretation.

 

Five leaves were measured per tree for NDVI, whereas Tables 5 and 6 state “n = 5”. If n refers to five leaves, these are subsamples within the same experimental unit rather than independent replicates. If n refers to five trees, this is inconsistent with the experimental design of two trees per treatment. The actual unit of statistical replication for the NDVI analysis is therefore unclear.

 

Soil moisture and temperature represent repeated measurements collected over multiple years from the same tree and the same sensor, yet the statistical analysis is described only as conventional ANOVA. The treatment of tree-level random effects, repeated-measures structure, and temporal autocorrelation is not specified.

 

The significance thresholds used in the Methods are inconsistent: the Shapiro–Wilk test uses α = 0.001, Levene’s test uses α = 0.05, ANOVA uses α = 0.01, whereas the multiple comparisons in Tables 3–6 use α = 0.05. Thus, the statistical decision criteria are not applied consistently throughout the manuscript.

 

Tables 3 and 4 both report “n = 10”, whereas the Methods indicate that soil moisture and temperature were monitored on only one tree per treatment. It is therefore unclear whether n = 10 refers to trees, dates, months, or another statistical unit, and this value cannot be readily reconciled with the experimental design.

 

The monitoring depth for soil temperature at Znojmo is inconsistent throughout the manuscript: the Methods and the note to Table 4 indicate 0.35 m, the Figure 4 caption indicates 0.5 m, the Results refer to 0.2–0.4 m, and the Discussion again refers to 0.5 m. The actual monitoring depth therefore cannot be determined consistently from the manuscript.

 

For the combined treatment at Brno, Moisture I is reported as 23.63% in the main text but as 23.36% in Table 3, indicating a numerical inconsistency between the Results text and the table.

 

For the 2025 NDVI results at both Brno and Znojmo, the Control is labeled “ab” and the Mycorrhiza + Hydrogel treatment is labeled “b”. Because the two treatments share the letter “b”, they cannot be considered significantly different. However, the main text and Discussion contain relatively definitive statements that the combined treatment was significantly higher than the Control.

 

The Discussion attributes treatment responses to specific mechanisms, including mycorrhiza-induced changes in hydraulic properties and soil aggregation, improved infiltration or enhanced drainage following Aeration, and Hydrogel-induced changes in soil pore structure and gas exchange. However, the study did not directly measure mycorrhizal colonization, porosity, infiltration rate, hydraulic conductivity, soil water fluxes, gas diffusion, or soil thermal properties. These mechanistic interpretations therefore extend well beyond the variables actually measured.

 

The Conclusions state that a significant reduction in soil temperature occurred “only” under the Mycorrhiza + Hydrogel treatment. However, at Brno, Aeration significantly reduced soil temperature at both monitored depths, and Mycorrhiza also reduced temperature in the deeper layer. In addition, the Conclusions extend the findings to tree vitality, urban green infrastructure resilience, and urban adaptation strategies, whereas NDVI was the only direct plant-level response measured. Therefore, some statements in the Conclusions are not fully consistent with the reported results, and the ecological and management-level extrapolations extend considerably beyond the direct measurements of the study.

Author Response

The responses to the reviewers’ comments are provided in the attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Please find below my recommendations for "Assessment of Soil Injection Treatments in Street Tree Beds: Impacts on Soil Microclimate and Foliar NDVI":

1. L61-74: Instead of generic statements and enumerative reference I recommend to better evidence for readers what was done and achieved by these studies. Do not forget to mention conditions/species/etc
2. Also, based on these the manuscript should better highlight the specific gap in the injection-treatment literature which this study fills beyond the "site-specific" effects mentioned
3. L95-104: I recommend a better formulation for the hypothesis as the current version are looks more as clear statements that transmit something well known
4. Why just like this? I saw that only two sites are used, each with different soil types, different measurement depths (Brno vs Znojmo) and different sensor placement according to those stated in L178-183 which for me means that the site and soil type are perfectly confounded (pseudoreplicated at the site level, n=1 per soil type), therefore site-dependent effects attributed to soil texture in Discussion cannot be statistically distinguished from any other site-specific factor (microclimate, management, traffic exposure)
5. L168-173: I have doubt here, in my opinion n = 2 trees per treatment per site (total 20 trees) is too low for ANOVA-based inference with 5 treatment levels; I think that pseudoreplication is likely because sensors were installed on one tree per treatment, meaning environmental time-series measurements are not independent biological replicates but repeated measures on a single tree, yet they are treated as if generating independent daily/seasonal replicates. I recommend careful consideration
6. In section "2.5. Soil amendments and applications" the manuscript should much clearly evidence the control description. Looking at table 1 and statement from L211-212 "Soil aeration was performed for all treatments, including the Aeration treatment, using compressed air at 6–7 bar" I understand that every treatment (Mycorrhiza, Hydrogel, Mycorrhiza+Hydrogel) is confounded with an aeration effect; therefore there is no factorial design isolating amendment effect from aeration effect. Further this invalidates any additive/synergy interpretation later stated
7. L224-226: This won't conflates the treatment zone with the measurement zone? I think that without a buffer there is a risk of sensor artifacts from physical disturbance during injection rather than genuine treatment effect. Please include a motivation statement as support in the manuscript
8. L246: Please justify this "α = 0.001"
9. Please mention also the corrections considered for multiple comparisons across the many factors/interactions tested
10. Please better clarify for Tables 3,4,5,6 the ANOVA sample sizes because I saw that they are reported as n = 10 (Tables 3–4) and as n = 5 (Tables 5–6) which appear to derive from repeated daily/seasonal measurements on the same trees and not independent trees which means that this pseudoreplication artificially inflates degrees of freedom and produces artificially small p-values (most p < 0.001)
11. Table 3/4: In my opinion the manuscript should also consider formal site × treatment interaction test instead of analyzing the two sites entirely separate ANOVA tables in order to could support the latter formulated "site-dependent" effect by a tested statistical interaction (L422-427)
12. Same in case of "synergistic effect" of Mycorrhiza+Hydrogel at Znojmo
13. L305-310: This need to be better considered here as I saw that a physical explanation is deferred entirely to Discussion rather than examined against data precision (SE values in Table 3 are implausibly small, like ±0.05–0.06 °C from what is effectively n=1 tree per treatment measured repeatedly). The small SE across treatment means, given only one instrumented tree per treatment, indicates SEs were computed across daily/temporal pseudoreplicates, not biological replicates which in turn materially misrepresents the true uncertainty of treatment comparisons...
14. I recommend to carefully consider the discussions section and avoid causal overinterpretation from correlational and confounded design. For example I understand based on those wrtitten that the site-dependent magnitude differences are attributed to soil texture, but in my mind because site and texture are perfectly confounded with unmeasured covariates like for example microclimate, traffic load, tree row orientation, canopy shading, etc. the texture-based causal attribution is unsupported by the design
15. L527-543: Which data support these statements? For me these looks more as a speculation as the unexpected hydrogel-induced surface warming is explained by a hypothesized reduction in gas exchange and pore restriction when this mechanism is asserted post hoc with no soil-gas or porosity measurements collected in this study to support it
16. I recommend also special attention when consider pot/container-scale literature to interpret field result. Same recommendation in case of NDVI mechanism inferred by analogy to unrelated stress physiology
17. The manuscript should better consider the potential confounds affecting NDVI (canopy age, phenology, inter-annual climate) as alternative explanations before attributing the M+H NDVI advantage to soil treatment
18. The manuscript should also allocate a subsection dedicated for limitations
19. Considering all those stated I recommend that after improvements implementation to revise manuscript conclusions in accordance

Author Response

The responses to the reviewers’ comments are provided in the attached file.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have clarified that soil moisture and temperature were monitored on two trees per treatment and that the continuous daily measurements were aggregated into monthly means for each tree. However, the statistical analysis still relies on a GLM with Treatment, Year, and Month as fixed factors. Because the same trees were repeatedly measured across different months and years, it remains unclear whether Tree was incorporated into the model as a repeated-measures subject, random effect, or other within-subject correlation structure. Aggregating the data into monthly means may reduce short-term temporal autocorrelation, but it does not eliminate the non-independence of repeated observations from the same experimental unit. Therefore, concerns remain regarding the statistical independence of the soil microclimate treatment effects.

For the NDVI analysis, the authors have clarified that each treatment included two trees and that five leaves were measured per tree, resulting in 10 leaf measurements per treatment. However, Tables 5 and 6 still appear to use the 10 leaf measurements as the basis for statistical analysis. The five leaves measured from the same tree represent subsamples rather than independent experimental replicates, whereas the tree itself is the true independent experimental unit. It remains unclear whether the leaf measurements were first averaged at the tree level or whether the nested structure of leaves within trees was accounted for statistically. Therefore, the issue regarding the independent unit of replication in the NDVI analysis has not been fully resolved.

The authors explain that the Aeration treatment was included to distinguish the mechanical effect of compressed-air soil loosening from the additional effects of Mycorrhiza, Hydrogel, and their combination, which is a reasonable experimental rationale. However, the main results and the Site × Treatment comparisons in the revised manuscript are still largely interpreted relative to the untreated Control. Because the Mycorrhiza, Hydrogel, and combined treatments all included an Aeration step, some statements attributing the observed responses specifically to the amendments cannot fully exclude the contribution of compressed-air treatment.

The authors acknowledge in their response that the Mycorrhiza, Hydrogel, and combined treatments each received 30 L of water as a one-time injection carrier and state that this issue has been addressed in the limitations of the study. However, the revised Limitations section mainly discusses the small number of experimental replicates, the lack of direct measurements of soil physical properties, field spatial heterogeneity, and the absence of independent chlorophyll measurements for NDVI interpretation. A clear discussion of the potential short-term confounding effect associated with the 30 L water input is still not evident.

Author Response

The responses to the reviewer’s comments are provided in the attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for considering to improve "Assessment of Soil Injection Treatments in Street Tree Beds: Impacts on Soil Microclimate and Foliar NDVI". I do not have new remarks for this version of the manuscript.

Author Response

Thank you very much for your careful review of our manuscript and for your valuable and constructive comments throughout the review process. We greatly appreciate your time and thoughtful feedback.