(Comment) This study quantifies carbon and nitrogen consumption from 50 fires in boreal forests of Sweden. The novel dataset is an important contribution to the scientific community and will fill data gaps associated with fuel consumption from Eurasian boreal fires. This study assesses the influence of climate and fuel availability on C and N emissions and redistribution within boreal forest ecosystems. However, minor adjustments are required to make the overall presentation well-structured and clear.
- The authors used a “space-for-time substitution” approach in which measurements performed in an unburned (control) plot adjacent to a burned plot were assumed to be representative of prefire conditions. Although this approach has been proven useful for studying fire effects, it also has limitations as the heterogeneity between matched plots can be high. The manuscript explains how the authors tried to overcome these limitations while selecting control plots. However, more data on vegetation composition, structure, and soil type are needed to show that paired plots were indeed similar.
- The methods section is missing a detailed description of the different ecosystems sampled across Sweden and fire severity in studied plots. Field pictures could be useful.
- Due to the large number of statistical analyses conducted, the results at times quite difficult to follow as currently presented. The manuscript would greatly benefit from additional figures and tables with models results (regression coefficients, metric scores). Please also consider restructuring the methods section. The description of the so-called compartment compositional variables (CCVs) is unclear to me. Including a table listing all the different variables used in the regressions would facilitate the interpretation of the results. Some statistical techniques are used in the results section without adequate prior description in the methods section (quadratic regression, path analysis).
- L40-42: “Particularly, the strongest driver of per area emissions of C in boreal wildfires appears to be total fuel (i.e. potentially combustible organic material) which is strongly controlled by long term forest moisture (Walker et al., 2018, 2020).”
(Comment) Please consider using carbon combustion (area-normalized C emissions) in the rest of the manuscript instead of “per area emissions” as it will make it easier to read. Moreover, this statement has been shown only for boreal North America. Please make it clear that these findings relate only to this part of the boreal biome. Fuel availability is also mainly controlled by stand age and local drainage conditions (Walker et al., 2020).
- L42-44: “However, in order for this fuel to be available to ignite and sustain fire, it must be both sufficiently dried and spatially arranged to be amenable to high heat and oxygen exposure during an active fire.”
(Comment) This sentence is not clear to me. Please consider rephrasing it. Fuels do not ignite anything but are ignited by humans or lightnings. What do the authors mean by “spatially arranged”?
- L61-66: “Boreal C emissions due to a single wildfire can be calculated by multiplying total area burned by estimates of C emissions per area (French et al., 2004; van der Werf et al., 2017). While total area burned may be evaluated directly through remote sensing (Giglio et al., 2018; Ruiz et al., 2012), per area C emissions are generally derived from labor intensive field sampling which is extrapolated to the larger scale either directly or through weighting by remotely sensed data (e.g. topography, vegetation cover, aerosol density) or poorly constrained free parameters such as total fuel load (French et al., 2004; Soja et al., 2004; van der Werf et al., 2017; Veraverbeke et al., 2015; Kaiser et al., 2012).”
(Comment) This paragraph needs further clarification. Summarizing the different approaches used to model C emissions from fires in a single sentence is challenging and makes it difficult to follow. Please add the reference to Seiler and Crutzen (1980) while explaining how C emissions are estimated (product of burned area, fuel consumption (fuel loads × combustion completeness) and emissions factors). Field measurements are generally used to calibrate C consumption models in response to environmental variables but also fire-severity indices (e.g. differenced Normalized Burn Ratio). What do the authors mean by “free parameters”?
- L71: “C loss due to a group of Siberian boreal forest surface fires was found to be 0.88 kg C m-2 (Ivanova et al., 2011)”
(Comment) This study estimated C combustion from experimental fires. This should be mentioned because experimental fires may not represent wildfires conditions. Please consider adding results from Kukavskaya et al. (2017) that quantified C emissions from surfaces fires in pine forests of Central Siberia (1.65 ± 0.70 kg C m-2). See also Veraverbeke et al. (2021) for a review of C combustion estimates from field measurements in boreal regions.
- L90-92: “Space-for-time substitution (De Frenne et al., 2013) along with a control-impact design provided insight into the possible future conditions of Fennoscandian forests in a changing climate and fire regime”
(Comment) Do the authors refer to the comparison between burned and unburned plots? Have the terms “control-impact design” been used in other studies? If yes, please indicate the reference(s). Otherwise, I suggest replacing it with 'paired-sample' design (Boby et al., 2010).
MATERIALS AND METHODS
- L109-111: “The first constraints on site selection were to avoid wetland or steeply sloping areas using prefire, topo-edaphic derived soil moisture data (TEM) provided by the Swedish Environmental Protection Agency (Naturvårdsverket, 2018) and elevation and slope data provided within the ArcGIS software environment.”
(Comment) What is the temporal and spatial resolution of the TEM product? Did the authors use a digital elevation model to retrieve elevation and slope? If so, please specify.
- L119-120: “Sentinel-2 infrared imagery was used to locate planned burnt plots near pixels showing the highest intensity within the mapped final burn scar perimeter.”
(Comment) Please indicate which bands and product types were used here to select burned plots (Top-Of-Atmosphere or Bottom-Of-Atmosphere reflectance products?). What does "intensity" mean here? Raw pixel value? Surface reflectance? Please specify.
- L120-121: “This gave greater certainty that the plots experienced a more developed fire effect rather than peripheral heating alone.”
(Comment) This sentence is not clear to me. What do the authors mean by "developed fire effect" and "peripherical heating"? Did they use a remotely sensed fire severity index such as dNBR (differenced Normalized Burn Ratio) to assess fire effects within burned plots? If so, please expand this section.
- L131-133: “Stand appearance and age were examined with historic, visual images provided by the Swedish National Land Survey verifying time since last disturbance had been at least 30 years for plot pairs and that stand structure between plot pairs appeared physically connected over this period.”
(Comment) What kind of images were used to assess stand structure? Aerial or satellite images? Please specify. Please consider rephrasing the last part of this sentence. It is unclear what "physically connected" means here.
- L134-135: “Due to their documented effects on C emissions (Walker et al., 2018), long and short term approximations of moisture were considered in this study. Long term moisture approximations were separated into a topo-edaphic component (TEM) and climatic component (MAP and MAT). Short term moisture was approximated over the first 6 months of 2018 using the Standardized Precipitation-Evapotranspiration Index (SPEI) with data from the SPEIBase (Beguería et al., 2019) (i.e. spei06 2018-06) to capture the extended desiccation process leading up to each fire. SPEI was also compared to summer 2018 anomalies in temperature and precipitation, i.e. the difference in the 2018 June, July, and August average of these values from those during the same months averaged over the period from 1961 to 2017.
(Comment) It is not clear why the authors defined the long- and short-term changing factors as “approximations of moisture”. I would suggest replacing these terms because they are ambiguous. Further details about the different products used in this study are needed (TEM, MAT, MAP, SPEIBase) including temporal and spatial resolutions. The authors mentioned documented effects from Walker et al. (2018), but mainly short-term fire characteristics (date of burn, fire weather indices) were assessed as potential drivers of C emissions in black spruce stands of Northwest Territories (Canada). If the authors are aware of previous studies that used the same products to evaluate their influence on fire-induced C emissions, please provide the reference(s). The last sentence is not clear. It seems that the authors compared SPEI, a drought index, with temperature and precipitation anomalies, but what does it mean by “compared”? Did the authors perform regression? Please specify.
- L151-152: “Each compartment was further sorted by weight into sets of characteristic components, here called compartment compositional variables (CCVs), which are to be specifically defined for each compartment in the following sections.”
(Comment) What exactly are the CCVs? It is not clear what this refers to from the following subsections. Please provide further details or examples. It would be worth adding a table (in the supplementary materials) to list all the CCVs used in this study.
- L164-165: “Four mineral soil samples were taken using a 3 cm diameter corer at four corners of a square each 15 m from the plot center.”
(Comment) Please indicate what type of corer was used to extract mineral soil samples. What was the purpose of collecting mineral soils since they were not likely to burn?
- L166-169: “Duff samples were taken near the mineral cores by excavating four soil volumes, trimming the mineral and moss/litter layers off the bottom and top of the volumes respectively, and then gently cutting right angles with sharp scissors to measure the 3 dimensions in millimeters (collected samples were at least 400 cm3 each).”
(Comment) Please indicate how the “soil volumes” were extracted. It is not clear from this paragraph whether the authors collected soil layers on the same profile (duff samples near the mineral cores, moss/litter layers trimmed from the soil volume).
- L178-179: “Individual tree bole diameter (sampled at 130 cm height above the forest floor) and species were recorded within each plot perimeter for all trees of at least 5 cm diameter at measurement height.”
(Comment) Why did the authors not measure trees smaller than 5 cm in diameter? These trees are more likely to burn, thus contributing to carbon emissions.
- L179-181: “If a fallen tree was charred only on its lower (in standing orientation) portions, it was deemed standing during fire ignition and its measurements were included if its base was within plot boundaries.”
(Comment) It seems that the amount of dead and downed woody debris and its consumption by fire was not quantified in the sampled plots. Yet it plays a fundamental role in C storage, influences other nutrient cycles, and controls forest fire behavior. Can the authors provide more information about this?
- L190-192: “Only bole diameters from the burnt plots were used to investigate the influence of overstory vegetation on C and N loss, while bole diameters from adjacent control plots were ignored.”
(Comment) Since tree diameters were not measure in control plots, how did the authors make sure that stand characteristics were similar between paired burned/unburned plots?
- L194-198: “To reduce sampling error due to small areal coverage of the plot, the sample patches were chosen by performing transects through the entire plot noting visual estimates of coverage and proportions of plant functional groups (i.e. graminoids, forbs, shrubs, and pteridophytes) which were applied in selecting representative patches for the portion of the plot that was vegetated, which was always all non bare rock surface. These values were applied to a visual estimate of non bare rock surface area of the burnt plots as an approximation of its prefire understory coverage.”
(Comment) It is not clear how the transects were performed. Were fractional coverages estimated in vegetation quadrants or at the plot level? Please consider rephrasing this section.
- L200: “All samples were dried at 40 °C for at least 3 days.”
(Comment) Were sampled dried to constant mass? If so, please mention it.
- L227: “When C and N stocks were described as losses their distribution was negated.”
(Comment) This statement is not clear. Please consider rephrasing it.
- L240-242: “The effects of C and N stock distribution amongst forest compartments were tested by entering the per plot ratios of the sums of different combinations of compartment C and N stocks into regression analyses both directly and to improve all models presented in the results section.”
(Comment) It is difficult here to understand what the authors wanted to test. The authors mentioned “different combinations”, but they are not clearly explained. Please provide details about the statistical tests performed in this section.
- L249-251: “The 50 burn plot overstories were largely dominated by pine with a percentage of spruce stems between 25-50% in 5 plots, between 50-75% in 3 plots and 2 plots with greater than 75%. Birch stems were less than 25% in 44 plots and between 25-50% in 6 plots, of which only 1 of the 6 was spruce dominant.”
(Comment) This could also be included in the "Materials and Methods" section in a site description paragraph with information on understory vegetation and soil types.
- L261: 3.2 C and N losses and restructuring
(Comment) Please add some numbers about C and N losses in this subsection. The quantification of C and N losses from Swedish boreal fires is supposed to be one of the main objectives of this study. However, estimates are only provided in the “Discussion” section.
- L270-272: “About three quarters of the moss/litter C was removed from burnt plots, comprising about half as much as the total amount of C that was removed from the duff layer.”
(Comment) Please provide more specific percentages of C consumption in the different soil organic layers.
- L294-295: “Fire-induced increases in bulk density of the soil layers counteracted C and N loss due to these depth changes.”
(Comment) This statement is not clear. How might post-fire changes in bulk density offset direct C emissions from fires? The term “counteracted” is not appropriate and should be replaced.
- L317-319: “Multicollinearity between the organic layer C:N ratio and CO (p = 0.003, r = -0.411, b = -1.96 kg C m-2) and NO (p < 0.001, r = -0.578, b = -92.2 kg N m-2) did not produce a high condition number in these models (1.55 for C, 1.93 for N) suggesting they are robust to these covariations (Alin, 2010).“
(Comment) The methodology used to derive condition numbers has not been described before. Please indicate how these numbers were obtained.
- L322-326: “CCVs and distribution of C and N stocks amongst control plot compartments could not improve these models explaining CO and NO losses in multiple regression with CO and NO respectively nor could they significantly explain the build up of organic layer fuel in control plots or production of char C or N. Relations either did not suitably meet the basic assumptions of regression, were deemed to be confounding or lacked supporting causal mechanism and were at a high risk of omitted-variable bias.”
(Comment) This paragraph is difficult to follow. More information is needed to describe the multiple regression and the combination of variables used here.
- L327: 3.4 Climatic drivers of fire-induced C and N loss
(Comment) There is no mention of applying a quadratic model fit to the variables in the methods. Plots and tables with regression coefficients, p and R2 values would be extremely helpful. This could be included as supplementary materials. Similarly, there is no clear description of applying a path analysis in the methods, although there are two figures showing results from this statistical approach. More details are needed to explain path analysis including the different assumptions on which this technique is based. Are all the relations described in both path diagrams (figures 4 and 5) statistically significant? If so, please specify (in figure captions for example).
Why did the authors not instead use structural equation modelling instead, which overcomes many of the limitations of path analysis, including non-linear relationships?
- L350-352: “The organic layer C:N ratio in the N model was able to replace the direct effect of MAT, however with decreased model fit and inflation of variables which is suggestive of a confounding influence of the organic layer C:N ratio on MAT and NO loss. Again, CCVs and fuel distribution could not improve either model.”
(Comment) This paragraph is difficult to follow, especially because there are no numbers to rely on. It is unclear what the authors mean by “inflation of variables”. The authors refer to many models in the results section that are not adequately described.
(Comment) The influence of short-term weather patterns on C emissions, as assessed by the SPEI, is not really discussed in this section. However, fire weather conditions dictate flammability and are often used to predict C emissions throughout the boreal forest. It might be interesting to elaborate on this point.
(Comment) Suggest adding the following references:
Kukavskaya, E. A., Buryak, L. V., Kalenskaya, O. P. and Zarubin, D. S.: Transformation of the ground cover after surface fires and estimation of pyrogenic carbon emissions in the dark-coniferous forests of Central Siberia, Contemporary Problems of Ecology, 10(1), 62–70, doi:10.1134/S1995425517010073, 2017.
Seiler, W. and Crutzen, P. J.: Estimates of gross and net fluxes of carbon between the biosphere and the atmosphere from biomass burning, Climatic Change, 2, 207–247, https://doi.org/10.1007/BF00137988, 1980.
Veraverbeke, S., Delcourt, C. J. F., Kukavskaya, E., Mack, M., Walker, X., Hessilt, T., Rogers, B. and Scholten, R. C.: Direct and longer-term carbon emissions from arctic-boreal fires: A short review of recent advances, Current Opinion in Environmental Science & Health, 23, 100277, doi:10.1016/j.coesh.2021.100277, 2021.
(Comment) Please find attached pdf with minor suggestions directly on the manuscript.