the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 20 Sep 2024
Soil and Biomass Carbon Storage is Much Higher in Central American than Andean Montane Forests
Abstract. Tropical montane forests (TMF) play a key role in the global carbon (C) cycle and in climate regulation by sequestering large amounts of above and belowground carbon. Elevation gradients in TMF have helped reveal the influence of environmental factors on C stocks. However, the influence of elevation and soil nutrient availability on C stocks has not been evaluated for mixed arbuscular and ectomycorrhizal (EM) associated forests in the Neotropics. We estimated aboveground biomass (AGB), coarse wood debris (CWD), and soil C based on field inventories in ten 1-ha plots along an elevational gradient from 880 to 2920 m a.s.l varying in relative abundance of EM-trees in western Panama. Trees ≥10 cm diameter at breast height (DBH) and CWD ≥10 cm diameter were measured to calculate biomass and necromass. Soil C to 1 m depth was estimated. Furthermore, climate and edaphic characteristics were described for each plot to evaluate the influence on these variables on each C pool. AGB, downed CWD and soil C were strongly positively correlated with elevation. We found exceptionally high AGB, up to 574.3 Mg ha−1, and soil C, up to 577.9 Mg ha−1 at higher elevations. Variation in total CWD within and among plots was high ranging from 14.75 to 326.5 Mg ha−1. After controlling for elevation, neither nutrient availability nor EM-dominance had an effect on AGB or soil C. Nonetheless, high AGB at high elevations was attributed to the presence of Quercus species. Remarkably high soil C at high elevations might be a consequence of reduced decomposition rates associated with lower temperature, or geological history, where repeated volcanic eruptions buried surface soil organic layers. Our results highlight large regional uncertainty in C pool estimates in Neotropical montane forests, with potentially large underestimates for Central American C stocks.
- Preprint
(1686 KB) - Metadata XML
-
Supplement
(704 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2738', Anonymous Referee #1, 15 Oct 2024
In this study, Prada et al., conduct a very thorough field sampling in Panama to evaluate C stocks accounting for different AG fractions and wood debris, understand its drivers, and compare them among other tropical forests, studies, and methodologies. I believe the methodology is sound and it has potential for a valuable scientific contribution. Nonetheless, I believe that the interpretation behind the findings and some parts of the writing need to be strongly strengthened or reformulated.
General comments:
After reading the manuscript and incorporating the data provided, seems obvious to me that the main reason for the C stock gradient is the P availability/limitation. Many previous studies suggest that tropical forests, especially in America and in lowland forests, are largely P-limited. When transitioning to higher elevations, this P limitation gradually changes into N limitation. You can check Cunha et al., 2022, Vallicrosa et al., 2023, Wright 2019, among many others. The findings displayed in this paper seem to perfectly align with that previously reported gradient as stated in line 293 “Among edaphic variables only resin P was correlated with elevation, increasing with elevation (r = 0.88, p < 0.001))”. Nonetheless, the authors' interpretation was the opposite, line 400 “After controlling for elevation, we found no effect of soil fertility (measured as NH4-N and resin P) on AGB”. Of course, if you remove the variable elevation, which explains 88% of your variability, you end up with nothing to be explained. Here you are not performing a fertilization experiment so you cannot directly report limitations. Still, I think that the story you are telling needs to be reframed to match with that evidence.
Cunha, H.F.V., Andersen, K.M., Lugli, L.F. et al. Direct evidence for phosphorus limitation on Amazon forest productivity. Nature 608, 558–562 (2022). https://doi.org/10.1038/s41586-022-05085-2
Vallicrosa, Helena, Laynara F. Lugli, Lucia Fuchslueger, Jordi Sardans, Irene Ramirez-Rojas, Erik Verbruggen, Oriol Grau, et al. 2023. “ Phosphorus Scarcity Contributes to Nitrogen Limitation in Lowland Tropical Rainforests.” Ecology 104(6): e4049. https://doi.org/10.1002/ecy.4049
Wright, S. J. 2019. Plant responses to nutrient addition experiments conducted in tropical forests. Ecological Monographs 89(4):e01382. 10.1002/ecm.1382
During the sampling and the whole paper, roots are totally neglected. I understand that root sampling has several methodological challenges such as the difficulty to identify the individual/species and root production, as well as the labor-intensive sample processing. But if we don’t sample them, we might miss a big part of the picture. Are roots directly correlating with what we are seeing aboveground? Or maybe since they have more nutrients they don’t need to invest as much in roots and therefore larger AG lower BG? I am not asking you to repeat your sampling, I still believe it is valuable, but acknowledging that fact somewhere and providing further discussion, citations and further development is necessary.
Aligning with my previous point, AM and ECM associations have been done through a database that is generalized among species and there is no mycorrhizal sampling performed in this study. Assuming that we can fully trust that assignation, are we expecting that the colonization % of among species and trees would vary significantly? I don’t think this is acknowledged or discussed in the paper. Also, Soudzilovskaia’s table for assignation sometimes provides potential open assignations, not fully committed to AM or ECM or even non-mycorrhizal. How did you deal with that uncertainty or other mycorrhizal categories? It is not disclosed in the methodology.
In your explanation, you suggest that the association of species with AM or ECM has a role in wood density (i.e., section 2.5.2). Using an ANOVA to determine that is too simplistic. Is there a phylogenetic bias behind that? For example, gymnosperms are normally associated with AM, and they also normally have lower-density wood. Are you seeing the effect of mycorrhiza or only phylogeny?
I believe it is interesting to see that Lidar underestimates AGB, in Figure 3. Still, this section of the paper seems a bit disconnected from the rest in a way that, for example, it is not mentioned in the abstract and Lidar is not presented in the introduction. Further work should be done to better incorporate this section in the context of the paper.
Minor comments:
Line 32: Please, specify what factors.
Line 38: Include a “that” as such: “a meta-analysis found that at lower elevations…”
Line 39: It would be nice to specify the reasons why of this low productivity. Do Quesada et al., 2012 provide that? Is it the wood density reasoning that you provide immediately later?
Section 2.1. It would be desirable to include a figure that would illustrate such plots, even if it is displayed as SM.
Line 117-118: I assume the selection of the 13 sampling locations has been randomly selected within the grid because if you divide 1 ha in a grid of 20 x 20 m and sample all of the subplots you do not get 13 samples. Please, specify a bit further about the process.
Line 121-122: How has this importance been assessed? Did you get this information from the bibliography? If that is the case what papers are those? Alternatively, specify if you had performed any sort of statistical test to determine such variables.
Line 146: Do you mean to infer or to generalize instead of “to compare”? I do not fully comprehend what has been done from line 146 to line 151 and why. I guess you want to use the airborne data and the measured biomass to assess how well the two values match and thus make a regional upscaling by using the airborne data?
Section 2.4.2: If the transects happened every 10m (line 157), it is possible that the same wood debris fell in several transects. For instance, a fallen tree individual that is 30m tall, could easily cross at least 2 transects. What is the protocol for repetitions?
Section 2.5.1 and 2.5.2: I assume the analysis described here has been done in R. Please name the used packages with the respective citation.
Figure 1: Why some of the vectors are greyer than others? It is not disclosed in the caption.
Figure 2a: I suggest including the initials of each plot next to the dots in Figure 2a. This way it would be easier to associate the figure a and b and to translate the % C storage to the total of each fraction.
Line 285-286: This is interesting. Based on my experience, it is common to sample until 30 cm deep, since it is assumed that those are the most nutrient-dense horizons. Could this finding be a reason to encourage the sampling deeper than that and get until 100cm?
Figure 7: In addition to the studies citation, I would like to see the locations where they were carried out.
Citation: https://doi.org/10.5194/egusphere-2024-2738-RC1 -
AC1: 'Reply on RC1', Cecilia Prada Cordero, 11 Nov 2024
General comments:
After reading the manuscript and incorporating the data provided, seems obvious to me that the main reason for the C stock gradient is the P availability/limitation. Many previous studies suggest that tropical forests, especially in America and in lowland forests, are largely P-limited. When transitioning to higher elevations, this P limitation gradually changes into N limitation. You can check Cunha et al., 2022, Vallicrosa et al., 2023, Wright 2019, among many others. The findings displayed in this paper seem to perfectly align with that previously reported gradient as stated in line 293 “Among edaphic variables only resin P was correlated with elevation, increasing with elevation (r = 0.88, p < 0.001))”. Nonetheless, the authors' interpretation was the opposite, line 400 “After controlling for elevation, we found no effect of soil fertility (measured as NH4-N and resin P) on AGB”. Of course, if you remove the variable elevation, which explains 88% of your variability, you end up with nothing to be explained. Here you are not performing a fertilization experiment so you cannot directly report limitations. Still, I think that the story you are telling needs to be reframed to match with that evidence.
Cunha, H.F.V., Andersen, K.M., Lugli, L.F. et al. Direct evidence for phosphorus limitation on Amazon forest productivity. Nature 608, 558–562 (2022). https://doi.org/10.1038/s41586-022-05085-2
Vallicrosa, Helena, Laynara F. Lugli, Lucia Fuchslueger, Jordi Sardans, Irene Ramirez-Rojas, Erik Verbruggen, Oriol Grau, et al. 2023. “ Phosphorus Scarcity Contributes to Nitrogen Limitation in Lowland Tropical Rainforests.” Ecology 104(6): e4049. https://doi.org/10.1002/ecy.4049
Wright, S. J. 2019. Plant responses to nutrient addition experiments conducted in tropical forests. Ecological Monographs 89(4):e01382. 10.1002/ecm.1382
Response: Thank you for this comment. We agree with the reviewer that carbon stocks do not respond directly to elevation, but to a combination of mostly abiotic variables that correlate strongly with elevation. Identifying causative pathways is difficult given the strong covariation of these variables in our study. One intention in this paper was to highlight the surprising, and counter-intuitive relationship showing a positive relationship between elevation and either soil carbon or AGB. As we note in the abstract and introduction, this relationship reflects the presence of EM tree species in Central American montane forests. However, we also agree that we could do more to highlight the soil variables that, in addition to temperature, vary across our elevational gradient and could contribute to this pattern. Our supplement currently includes table of correlations of climate and soils variables with elevation (Table S5), and regressions of carbon pools with environmental factors (Table S2), and a correlation matrix of variables used in the SEM (Table S3). We will add to the supplement figures similar to those in Fig. 4 showing the relationship between elevation and a) resin-extractable phosphorus, b) inorganic nitrogen, c) soil total N:P. We will also expand the results and discussion section to highlight variation in N and P across the gradient and its potential impact in driving these relationships (in addition to temperature – which has the dominant effect in our structural equation model).
During the sampling and the whole paper, roots are totally neglected. I understand that root sampling has several methodological challenges such as the difficulty to identify the individual/species and root production, as well as the labor-intensive sample processing. But if we don’t sample them, we might miss a big part of the picture. Are roots directly correlating with what we are seeing aboveground? Or maybe since they have more nutrients they don’t need to invest as much in roots and therefore larger AG lower BG? I am not asking you to repeat your sampling, I still believe it is valuable, but acknowledging that fact somewhere and providing further discussion, citations and further development is necessary.
Response: Thanks for highlighting this omission from our manuscript. Indeed, roots can provide contribute substantially to carbon storage. We unfortunately lack information from our sites on either fine root dynamics or coarse root biomass. Nonetheless, we will incorporate in the discussion a section to review the contribution of roots in montane forests and the possible effect on our results.
Aligning with my previous point, AM and ECM associations have been done through a database that is generalized among species and there is no mycorrhizal sampling performed in this study. Assuming that we can fully trust that assignation, are we expecting that the colonization % of among species and trees would vary significantly? I don’t think this is acknowledged or discussed in the paper. Also, Soudzilovskaia’s table for assignation sometimes provides potential open assignations, not fully committed to AM or ECM or even non-mycorrhizal. How did you deal with that uncertainty or other mycorrhizal categories? It is not disclosed in the methodology.
Response: We now provide more information on the classification of mycorrhizal type in the methods. ECM status has been confirmed for the most abundant taxa by Corrales et al. (2016, 2018). We note, however, that the while AM taxa predominate at our site, there are some species that are either non-mycorrhizal (e.g., Roupala) or form ericoid mycorrhizal associations (e.g., Vaccinium). We therefore now contrast ECM and non-ECM taxa.
Corrales, A., Arnold, A.E., Ferrer, A.H., Turner, B.L., Dalling, J.W. (2016) Variation in ectomycorrhizal fungal communities associated with Oreomunnea mexicana (Juglandaceae) in tropical montane forests. Mycorrhiza, doi: 10.1007/s00572-015-0641-8
Corrales, A., Henkel, T.W. and Smith, M.E. (2018). Ectomycorrhizal associations in the tropics–biogeography, diversity patterns and ecosystem roles. New Phytologist, 220(4), pp.1076-1091.
In your explanation, you suggest that the association of species with AM or ECM has a role in wood density (i.e., section 2.5.2). Using an ANOVA to determine that is too simplistic. Is there a phylogenetic bias behind that? For example, gymnosperms are normally associated with AM, and they also normally have lower-density wood. Are you seeing the effect of mycorrhiza or only phylogeny?
Response: We have very few ECM taxa in our plots – Oreomunnea, Quercus, Alfaroa. These are all in a single clade (Fagales). In fact, all the Fagales in these forests are ECM. So, we do have a phylogenetic effect and the finding that we have higher wood density in ECM species could be a consequence of their mycorrhizal association or of their phylogenetic placement. Incidentally, we have only one coniferous species in our study, which only occurs in one plot.
I believe it is interesting to see that Lidar underestimates AGB, in Figure 3. Still, this section of the paper seems a bit disconnected from the rest in a way that, for example, it is not mentioned in the abstract and Lidar is not presented in the introduction. Further work should be done to better incorporate this section in the context of the paper.
Response: We will revise the manuscript to better incorporate the published LiDAR study in the abstract, introduction and methods. The LiDAR study highlights the need to incorporate ground truthed data specific to montane forests in generating large scale carbon storage estimates.
Minor comments:
Line 32: Please, specify what factors. Response: temperature, nutrient availability and tree mycorrhizal status.
Line 38: Include a “that” as such: “a meta-analysis found that at lower elevations…” Response: Will do
Line 39: It would be nice to specify the reasons why of this low productivity. Do Quesada et al., 2012 provide that? Is it the wood density reasoning that you provide immediately later? Response: The sentence will be changed to: “However, low values of AGB have been found in the Amazon below 500 m a.s.l due to low fertility (Quesada et al., 2012).”
Section 2.1. It would be desirable to include a figure that would illustrate such plots, even if it is displayed as SM. Response: A figure with a map including the location of the plots will be included in the supplementary materials.
Line 117-118: I assume the selection of the 13 sampling locations has been randomly selected within the grid because if you divide 1 ha in a grid of 20 x 20 m and sample all of the subplots you do not get 13 samples. Please, specify a bit further about the process. Response: To make it clear the sentence will be changed to “Soil samples were collected in a regular grid within each 1-ha plot (center of every other 20 x 20 m subplot) resulting in thirteen sampling locations for each plot”.
Line 121-122: How has this importance been assessed? Did you get this information from the bibliography? If that is the case what papers are those? Alternatively, specify if you had performed any sort of statistical test to determine such variables. Response: Citations will be added to support this point.
Line 146: Do you mean to infer or to generalize instead of “to compare”? I do not fully comprehend what has been done from line 146 to line 151 and why. I guess you want to use the airborne data and the measured biomass to assess how well the two values match and thus make a regional upscaling by using the airborne data? Response: Yes, we want to compare the aboveground carbon stock estimated using our ground measurements with values extracted for the same georeferenced locations using LiDAR data. This point will be clearer once we have included the topic of LiDAR in the introduction.
Section 2.4.2: If the transects happened every 10m (line 157), it is possible that the same wood debris fell in several transects. For instance, a fallen tree individual that is 30m tall, could easily cross at least 2 transects. What is the protocol for repetitions? Response: As we stated in the methodology we followed the protocol in Larjavaara & Muller-Landau (2011): “If a given piece crosses the transect twice, it is measured at both intersections (Photo 2), with those measurements recorded on separate lines of the datasheet.”
Section 2.5.1 and 2.5.2: I assume the analysis described here has been done in R. Please name the used packages with the respective citation. Response: Yes, the R packages will be included for these sections.
Figure 1: Why some of the vectors are greyer than others? It is not disclosed in the caption. Response: The following text will be added to the caption: “Black arrows represent significant effects (p<0.05) and gray arrows non-significant effects”
Figure 2a: I suggest including the initials of each plot next to the dots in Figure 2a. This way it would be easier to associate the figure a and b and to translate the % C storage to the total of each fraction. Response: We will incorporate the plot initials into this figure.
Line 285-286: This is interesting. Based on my experience, it is common to sample until 30 cm deep, since it is assumed that those are the most nutrient-dense horizons. Could this finding be a reason to encourage the sampling deeper than that and get until 100cm? Response: For these forests with deep carbon-rich soils it is important to sample deeper in the soil profile to fully account for storage. We will add a discussion point suggesting that future studies consider sampling soil to greater depths.
Figure 7: In addition to the studies citation, I would like to see the locations where they were carried out. Response: We will add the specific site of the studies used.
Citation: https://doi.org/10.5194/egusphere-2024-2738-AC1
-
AC1: 'Reply on RC1', Cecilia Prada Cordero, 11 Nov 2024
-
RC2: 'Comment on egusphere-2024-2738', Anonymous Referee #2, 01 Nov 2024
Soil and biomass carbon storage is much higher in Central American than Andean montane forests
by Cecilia M. Prada et alThis is well-written manuscript and takes into account most of the available work on carbon storage in tropical montane forests. The questions addressed are interesting and the high C stocks of the studied montane forests are relevant for their conservation.
The comparison with carbon stocks estimated with LIDAR is not well connected to the rest of the study,
shoud be better introduced in the introduction. The result, that remote sensing underestimated AGB the importance of ground-truthing with inventories in permanent forest plots should be included in the abstract.Table 2 (it contains the same information as presented in Fig. 2) and Fig. 6 (very similar to Fig.1, only needed to explain the variables included in the SEM) both should be moved to the supplement.
In summary, the manuscript is based on extensive and valid data, but some technical details have to be improved.
Please see some detailed comments below:
l28/29: not only coarse woody debris but also the litter layer
l35/36: additional reference for soil fertility effects on AGB: Homeier & Leuschner 2021
l70-73: additional reference for decreasing nitrogen availability with elevation and effect on wood density: Homeier & Leuschner 2021
l97/98: Were tree ferns included?
l117/118: The litter layer was removed and not part of the soil analyses, do you have any idea of the depth of the litter layer? Fine wood and litter could probably add essential amounts of C to your budgets (e.g. Phillips et al. 2019a).
l136: In FigS1 it seems that in 3 plots only heights of trees >50cm dbh were available. How does this affect your height estimates?
l137: delete "(Fig. S1)"
l139: Wood volume was quantified with the water displacement method.
l142/143: Was palm AGB also calculated with the Chave (2014) equation? There are some specific allometric equations for palms: Avalos G et al.(2022), Goodman RC et al.(2013)
l174/175: Does that mean that standing dead wood was classified into the same 3 categories as DCWD?
l184-191: Was the organic layer included in the sampling?
l204-206: Fig. 6 should be moved to the supplement. Was the climate axis (PC1 from the climate PCA) not used in the SEM?
l213/214: I would suggest to use total CWD instead of only DCWD. Regressions of total CWD should be included in Table S2.
l238: "WD" in figure legend but "WC" in figure
l286-289: In the first sentence you say there is no correlation to %EM, in the second sentence you mention there is one correlation. And the C proportion in 10-20 cm depth is negatively correlated with %EM in Fig. S5
l326: Only soil parameters of the first PCA axes are included, not hte second axis and no climate parameters
l329: Shouldn't values of relative importance (standardized path coefficients) be between 0 and 1?
l392/393: Is there any explanation for the extremely high deadwood values in the HA plot?
l406/407: increasing wood density with elevation was only found in some studies, but not in your study
l408: elevation and
l427-429: What are the wood densities of these taxa?
l477/478: This sentence is speculative I suggest to delete it.
l674: Costa Rica
Figures and TablesFig. 1: What is the difference between the variables with black and grey arrows?
"WC" should be "WD" in Fig 1bFig. 4: The size of the boxes in the right-hand column should be the same. I suggest to add also SCWD (or total CWD) as additional box to the left column to show 8 parameters in total.
Fig. 6 should be moved to the supplement
Fig. 7a: For comparing with data from Ecuador the Homeier & Leuschner (2012) study with > 80 plots would be the better choice than the Moser et al (2008) study with 5 plots.
Table 2 should be moved to the supplement (it contains the same information presented in Fig. 2).
Table S4: WC?
additional references:Avalos G et al.(2022) Allometric models to estimate carbon content in Arecaceae based on seven species of Neotropical palms. Frontiers in Forests and Global Change, 5, 867912.
Goodman RC et al.(2013) Amazon palm biomass and allometry. Forest ecology and management, 310, 994-1004.
Homeier J & Leuschner C(2021) Factors controlling the productivity of tropical Andean forests: Climate and soil are more important than tree diversity. Biogeosciences 18:1525-1541
Venter M et al (2017) Optimal climate for large trees at high elevations drives patterns of biomass in remote forests of Papua New Guinea. Global Change Biology, 23(11), 4873-4883.
Citation: https://doi.org/10.5194/egusphere-2024-2738-RC2 -
AC2: 'Reply on RC2', Cecilia Prada Cordero, 11 Nov 2024
This is well-written manuscript and takes into account most of the available work on carbon storage in tropical montane forests. The questions addressed are interesting and the high C stocks of the studied montane forests are relevant for their conservation. The comparison with carbon stocks estimated with LIDAR is not well connected to the rest of the study, shoud be better introduced in the introduction. The result, that remote sensing underestimated AGB the importance of ground-truthing with inventories in permanent forest plots should be included in the abstract.
Response: We will revise the manuscript to better incorporate the published LiDAR study in the abstract, introduction and methods.
Table 2 (it contains the same information as presented in Fig. 2) and Fig. 6 (very similar to Fig.1, only needed to explain the variables included in the SEM) both should be moved to the supplement.
Response: We agree that there is overlap in the information provided in Table 2 and Figure 2. We agree to move Table 2 and Figure 6 to the supplement. However, we prefer to retain Fig 1 and 2 in the manuscript as they convey the major results of the study
In summary, the manuscript is based on extensive and valid data, but some technical details have to be improved.
Please see some detailed comments below:
28/29: not only coarse woody debris but also the litter layer. Response: We will change the sentence to “This carbon is sequestered in several pools, consisting of live biomass, necromass (i.e. coarse woody debris or litter), and soil”
35/36: additional reference for soil fertility effects on AGB: Homeier & Leuschner 2021. Response: This reference will be added to the sentence.
l70-73: additional reference for decreasing nitrogen availability with elevation and effect on wood density: Homeier & Leuschner 2021. Response: This reference will be added to the sentence.
97/98: Were tree ferns included? Response: Tree ferns were included. This sentence will be improved: “In this study to facilitate comparisons with other published datasets only trees, palms or ferns ≥ 10 cm DBH were included in the analyses.”
117/118: The litter layer was removed and not part of the soil analyses, do you have any idea of the depth of the litter layer? Fine wood and litter could probably add essential amounts of C to your budgets (e.g. Phillips et al. 2019a). Response: Not all plots have a distinct litter layer. We did record litter depth when present but did not quantify the carbon content of this pool. We will add a couple of sentences to address the litter layer in the discussion.
136: In FigS1 it seems that in 3 plots only heights of trees >50cm dbh were available. How does this affect your height estimates? Response: Correct - we lack height measurements for smaller trees specifically for these plots. These are all plots at similar elevation at the Fortuna site. To determine how this impacts our height measurements we will compare AGB estimates for these plots using the available height data versus (i) the height vs DBH relationship for nearby plots with similar soils (ie Copete data for Hornito and Alto Frio, and Honda A for Samudio; (ii) we will recalculate AGB estimates for all the plots using pooled height-DBH data for all the plots.
137: delete "(Fig. S1)". Response: We will do this.
139: Wood volume was quantified with the water displacement method. Response: We will edit this sentence as follows: “Wood volume in the field was quantified with the water displacement method (Chave, 2006).”
142/143: Was palm AGB also calculated with the Chave (2014) equation? There are some specific allometric equations for palms: Avalos G et al.(2022), Goodman RC et al.(2013). Response: We will incorporate the specific equations for palms suggested.
174/175: Does that mean that standing dead wood was classified into the same 3 categories as DCWD? Response: Yes. We will make this clearer in the manuscript.
184 -191: Was the organic layer included in the sampling? Response: Since not all the plots had an organic layer data, we just used the mineral soil.
204-206: Fig. 6 should be moved to the supplement. Was the climate axis (PC1 from the climate PCA) not used in the SEM?Response: This figure will be moved to the supplement and a clearer explanation will be added to this sentence.
213/214: I would suggest to use total CWD instead of only DCWD. Regressions of total CWD should be included in Table S2. Response: We decided to remove Standing CWD because it was not correlated with the other variables of interest, but also because the uncertainty of this pool was too large. Regressions of total CWD are now included in Table S2.
238: "WD" in figure legend but "WC" in figure. Response: This will be corrected.
286-289: In the first sentence you say there is no correlation to %EM, in the second sentence you mention there is one correlation. And the C proportion in 10-20 cm depth is negatively correlated with %EM in Fig. S5. Response: The sentence will be corrected as: “The percent of C in each depth relative to the total C% in each plot was not correlated with % EM, except in the 10-20 cm depth (Fig. S5). For this layer, percent of C was negatively correlated with the percent of basal area contributed by EM-trees (% EM; r2 = 0.36, p < 0.05,), however this effect was not significant after controlling for elevation.”
326: Only soil parameters of the first PCA axes are included, not hte second axis and no climate parameters. Response: Climate parameters were included in the original model (Fig S2), however they were not included as predictors in the best model
329: Shouldn't values of relative importance (standardized path coefficients) be between 0 and 1? Response: Yes, the values in the figure and the model will be checked and corrected.
392/393: Is there any explanation for the extremely high deadwood values in the HA plot? Response: In this plot there were two big standing dead oak trees with some of the branches in the subsamples. We will add this explanation to this sentence.
406/407: increasing wood density with elevation was only found in some studies, but not in your study. Response: “Decreasing temperature can reduce growth rates and is associated with increased wood density (Muller-Landau, 2004; Chave et al., 2009), which may increase AGB, however it was not the case in our study”
408: elevation and. Response: I will change the word.
427-429: What are the wood densities of these taxa? Response: We will add the values of wood density for these species.
477/478: This sentence is speculative I suggest to delete it. Response: We will delete the sentence.
674: Costa Rica. Response: We will correct this reference.
Figures and Tables
Fig. 1: What is the difference between the variables with black and grey arrows? Response: The following text will be added to the caption: “Black arrows represent significant effects (p<0.05) and gray arrows non-significant effects”
"WC" should be "WD" in Fig 1b. Response: We will correct this text.
Fig. 4: The size of the boxes in the right-hand column should be the same. I suggest to add also SCWD (or total CWD) as additional box to the left column to show 8 parameters in total. Response: We will rearrange the figure and a figure for SCWD and make the panels the same size.
Fig. 6 should be moved to the supplement. Response: We will do.
Fig. 7a: For comparing with data from Ecuador the Homeier & Leuschner (2012) study with > 80 plots would be the better choice than the Moser et al (2008) study with 5 plots. Response: We will evaluate the Homeier & Leuschner (2012) reference and add it to the analysis.
Table 2 should be moved to the supplement (it contains the same information presented in Fig. 2). Response: Table 2 will be moved to the supplement
Table S4: WC? Response: We will correct this text.
Citation: https://doi.org/10.5194/egusphere-2024-2738-AC2
-
AC2: 'Reply on RC2', Cecilia Prada Cordero, 11 Nov 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 158 | 57 | 11 | 226 | 16 | 3 | 3 |
- HTML: 158
- PDF: 57
- XML: 11
- Total: 226
- Supplement: 16
- BibTeX: 3
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1