the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Mar 2026
Beyond the laboratory: performance and agreement of rapid methodologies for soil health assessment
Abstract. Soil health assessment increasingly relies on biological indicators because of their sensitivity and direct links to ecosystem functioning. However, conventional laboratory methods are time-consuming, require specialized infrastructure, and are often incompatible with rapid decision-making in applied contexts. Several rapid or field-deployable tools have recently been developed to address this limitation, but their comparability with standard laboratory methods remains insufficiently evaluated. Here, we compared four rapid approaches with their corresponding laboratory reference methods in a long-term grassland experiment: aggregate stability (SLAKES), soil respiration (portable CO2 analyzer), microbial biomass carbon and fungal-to-bacterial ratio (microBIOMETER®), and enzyme activities (Soil Enzymatic Activity Reader, SEAR). Agreement between methods was assessed using Spearman correlations, redundancy analyses and Procrustes analysis. Aggregate stability showed strong correspondence between rapid and laboratory measurements (R = 0.64), whereas soil respiration exhibited weak agreement, likely reflecting that in situ and laboratory approaches capture different aspects of respiratory activity. Microbial biomass carbon displayed moderate comparability between methods (R = 0.51), while fungal-to-bacterial ratios did not. Enzyme activities measured with SEAR were generally consistent with laboratory assays. Multivariate analyses indicated that overall, rapid methods captured ecological patterns similar to those revealed by laboratory protocols. These findings support the use of selected rapid tools as complementary or alternative options when laboratory facilities are unavailable or timely soil health information is required to inform management decisions.
Competing interests: Author JF was employed by Digit Soil.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(892 KB) - Metadata XML
-
Supplement
(30 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-1255', Anonymous Referee #1, 22 May 2026
-
AC1: 'Reply on RC1', Lur Epelde, 03 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-1255/egusphere-2026-1255-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Lur Epelde, 03 Jun 2026
-
RC2: 'Comment on egusphere-2026-1255', Anonymous Referee #2, 26 May 2026
Major comments:
The central research question is based around whether rapid or field-deployable tools are interchangeable with standard laboratory methods. The scientific question is interesting and important, although ultimately unanswerable using the experimental design and statistical analyses.
The methods show a sampling design of 6 subplots (3 plots × 2 management treatments) with a single point per subplot sampled at two depths sampled at 11 time intervals. This design has three nested sources of dependence: subplot (n = 6), time point (n = 11), and depth (n = 2). The real replication for the management treatment contrast is n = 3 subplots per treatment. The Spearman rank correlations and RDA appear to treat all 132 observations as independent, which is pseudoreplication. The consequences are that the Spearman rho confidence intervals are narrower than they should be, and p-values are inflated, possibly by orders of magnitude; RDA permutation tests are biased because permutations assume independence among the observations; and rank concordance between on-site and laboratory methods may largely reflect a combination of the depth gradient and within-subplot temporal autocorrelation, which are not accounted for. For the method-comparison Spearman correlations, a more defensible approach would be to aggregate observations to the subplot level (or to compute within-subplot rank correlations and test the distribution across subplots) before computing inter-method correlation statistics. For the RDA, time should be included as an adjusted covariate and permutations stratified by subplot. Without restructured analyses that account for the nested and repeated-measures design, the manuscript cannot support the asserted rank concordance between on-site and laboratory methods.
Throughout the manuscript the authors describe their analyses as testing the comparability and agreement between on-site and laboratory methods, but the statistical framework applied (Spearman’s rank correlation) does not test comparability, agreement, or sensitivity, but instead whether two methods order samples consistently. It is uninformative about whether one method could substitute for another. The visual presentation compounds this problem, as linear regression lines with confidence intervals are plotted on scatter plots showing one method against the other, while reporting Spearman’s rho on the same plots. The authors should restrict their claims to rank concordance, removing all “agreement” and “comparability” to “the methods rank samples consistently” or emphasizing the relative ordering nature of the comparison. The rank correlation comparison is still useful and important to the literature.
Minor comments:
28-30 One of the citations for this assertion is a “viewpoint” and should be removed
40-41 Slakes is not a field-based measurement, as aggregates must be air-dried prior to analysis
47-53 Assessing the performance of rapid laboratory measurements relative to traditional laboratory measurements through numerical comparison alone makes two assumptions: 1. That the methods are measuring similar properties using similar methodologies and thus have a basis for numerical comparison; 2. That traditional laboratory measurements are “better”, i.e. more sensitive or precise. I am not convinced that either assumption is met, and the manuscript would benefit greatly from addressing these assumptions. There are other studies that quantitatively evaluate these methods based on their sensitivity to management (e.g., Rieke et al., 2023)
64-67 The LUCAS framework was designed for landscape-scale soil chemistry and physical property monitoring, such as total C/N, pH, and soil texture, and are not appropriate for biological assessments, since typically the 20-50 cm depth is not very biologically active.
76-83 Were aggregates air-dried prior to running Slakes as required by the methods? This would seem to mean it is not a field-based test. Additionally, only ~9 soil aggregates 3-10 mm are typically required for Slakes as implemented commercially, which would translate to 1-5 g, maybe enough for one Eijkelkamp run. What was the actual mass of soil used to for Eijkelkamp, and how many replicates were run for each test? The aggregate sizes are likely different for both tests, with Eijkelkamp measuring aggregate stability on 1-2 mm aggregates and Slakes measuring aggregate stability on 3-10 mm aggregates. This holds implications for comparability. Finally, Slakes was validated on aggregates collected from 0-6 cm in depth, and has not been tested for other depths, please acknowledge or address
84-89 The comparison between ISO 16072 bench respiration and in-situ IRGA flux is not methodologically defensible, as the two methods differ in soil disturbance (air-drying, sieving, rewetting), environmental conditions, and what type of respiration is measured. Due to the loss of microbial biomass through air-drying, Birch-effect flush of CO2, presence or absence of autotrophic root respiration, and difference in ambient soil moisture, the alkali trap method will likely overestimate low fluxes and underestimate high fluxes (Yim et al. 2002; Jensen et al. 1996). So many sources of method bias cannot support the inferences the authors seek to draw; I recommend either removing this comparison or replacing the disturbed soil protocol with a minimally disturbed intact-core incubation
90-96 This section compares a biomass-based ratio with a relative abundance ratio (DNA copy count), which makes comparison difficult; additionally, ITS is more accepted for fungal community profiling, please justify
102-112 The fact that Spearman’s rank correlation was used to compare methods needs to be highlighted much more in the introduction; the goal of this paper is not numerical comparison, but comparison of the ordering of treatment pairs. Still, potential confounding remains and much more caution should be used in interpreting results
120 The linear regression line is not appropriate, as it implies that y has measurement error but x does not, that the relationship is linear, and that errors are normally distributed; the confidence interval is for the conditional mean of y given x, not the prediction or agreement interval; a 1:1 line should be plotted if the goal is method comparison; and the inclusion of the Spearman values is inconsistent with the linear regression
134 There is a substantial body of literature that addresses all of the differences between field-measured respiration and laboratory-measured respiration, please research, consult and amend
140-142 This is a good point and should be expanded
167-175 These should be amended to address the feedback provided in major comments
Citation: https://doi.org/10.5194/egusphere-2026-1255-RC2 -
AC2: 'Reply on RC2', Lur Epelde, 05 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-1255/egusphere-2026-1255-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Lur Epelde, 05 Jun 2026
-
RC3: 'Comment on egusphere-2026-1255', Anonymous Referee #3, 05 Aug 2026
Epelde et al. compare several rapid soil health measurements with conventional laboratory indicators. The manuscript is generally well written and easy to follow. However, I have several concerns regarding the experimental design, statistical analyses, and interpretation that should be addressed before the manuscript can be considered for publication.
My first general comment concerns the experimental design. The study is based on 132 soil samples collected from a single grassland experiment (L67). However, these samples are not independent because they follow a nested sampling design, with multiple sampling times and two soil depths. As a result, the subsequent correlation analyses treat pseudoreplicates as independent observations (Fig. 1), which may bias the reported relationships. This limitation should be explicitly acknowledged, and the authors should consider revising the analyses to account for the hierarchical structure of the data.
Related to this point, the reported correlations are valuable for understanding relationships within this experiment but should not be interpreted as broadly representative across soil types or management systems. Validation across independent sites would still be necessary before drawing more general conclusions. I encourage the authors to acknowledge this limitation. The authors could refer to Culman et al. (2012) for an example of using both within-site and cross-site analyses to evaluate the performance of a new indicator of soil health.
My second comment concerns the comparison between field-based and laboratory-based soil respiration measurements. In my view, these measurements represent fundamentally different processes. Field measurements quantify soil CO₂ flux, which is strongly influenced by environmental conditions such as temperature, soil moisture, and substrate availability. In contrast, the laboratory assay appears to quantify the potentially mineralizable carbon pool under controlled conditions. I therefore question the value of this comparison and would recommend either removing it or substantially revising the discussion.
Finally, I am not convinced that the redundancy analysis (RDA) provides more insight than the correlation analyses. An alternative and potentially more informative comparison would be to evaluate whether the rapid indicators and conventional laboratory measurements respond similarly to management or temporal variation. For example, comparing the effect sizes of the grazing treatments or temporal trends would provide a more direct assessment of whether the rapid methods capture the same ecological responses as the conventional indicators. See Culman et al. (2012) for an example.
Specific comments:
L21: The phrase “ecological patterns” is too vague. Please describe more specifically what patterns are being investigated.
Materials and Methods: The Introduction emphasizes the time-saving advantages of the rapid methods. It would therefore be helpful to briefly report the approximate time required for the rapid methods compared with the conventional laboratory analyses.
L87: Please provide a brief description of the ISO procedure used for the respiration measurements. If I understand correctly, the standard introduces the methods for basal respiration and substrate-induced respiration. A concise description would improve reproducibility without requiring readers to consult the standard.
L103: Please justify the use of Spearman's rank correlation rather than alternative correlation methods.
Culman, S. W., Snapp, S. S., Freeman, M. A., Schipanski, M. E., Beniston, J., Lal, R., Drinkwater, L. E., Franzluebbers, A. J., Glover, J. D., Grandy, A. S., Lee, J., Six, J., Maul, J. E., Mirksy, S. B., Spargo, J. T., & Wander, M. M. (2012). Permanganate Oxidizable Carbon Reflects a Processed Soil Fraction that is Sensitive to Management. Soil Science Society of America Journal, 76(2), 494–504. https://doi.org/10.2136/sssaj2011.0286
Citation: https://doi.org/10.5194/egusphere-2026-1255-RC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 785 | 294 | 88 | 1,167 | 107 | 66 | 81 |
- HTML: 785
- PDF: 294
- XML: 88
- Total: 1,167
- Supplement: 107
- BibTeX: 66
- EndNote: 81
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Please see the comments in the supplement.