Fire Regime Has a Greater Impact Than Selective Timber Harvesting on Vegetation in a Sub-Tropical Australian Eucalypt Forest
Round 1
Reviewer 1 Report
This study made efforts to compare impacts of selective logging and fire disturbance on Eucalyptus forests in south-eastern Queensland, Australia. The work seems well executed, clearly presented. I think the research is interesting. However, for a Forest Ecology and Management section of Forests journal, I find the manuscript is a bit light. This study missed in examining some important issues: what's the relations between selective harvesting and fire on tree species diversity, tree density and light levels in the Eucalyptus forest understorey?
Author Response
This study made efforts to compare impacts of selective logging and fire disturbance on Eucalyptus forests in south-eastern Queensland, Australia. The work seems well executed, clearly presented. I think the research is interesting. However, for a Forest Ecology and Management section of Forests journal, I find the manuscript is a bit light. This study missed in examining some important issues: what's the relations between selective harvesting and fire on tree species diversity, tree density and light levels in the Eucalyptus forest understorey?
Thank you for the positive review.
We disagree that the level of content is a bit light. We already included many response variables (i.e. live and dead tree basal area and density, canopy cover, CWD volume, ground-layer live plant biomass, litter debris biomass, percentage cover of perennial grasses, sedges, herbs and other herbaceous species and woody plants with DBH <10 cm, bare ground and litter, transect species richness, average understorey plant height to 7.5 m per transect, topsoil P, N, C and effective cation exchange capacity), to the point where we felt the paper was starting to get too long. Nevertheless, we added some information on tree species richness (lines 298-299) and tree density is already included (lines 292-296). Unfortunately, we did not measure light levels, but I think canopy cover provides a reasonable surrogate measure of this. A sentence was added to the discussion regarding light levels and the understorey vegetation (lines 579-581). Further studies are needed to investigate this in more detail.
Reviewer 2 Report
Reviewer Report – forests
Fire regime has a greater impact than selective timber harvesting on vegetation in a sub-tropical Australian eucalypt forest
This study describes the effects of selective timber harvesting in a sub-tropical forest that has experienced contrasting experimental fire regimes over many decades – a very impressive experimental setup, that the authors have made use of to address the questions of (i) which of selective timber harvesting or fire regime has the greatest impact, and (ii) whether the effects of harvesting would be greatest in the most frequently disturbed (burnt) part of the forest. They have made a wide range of measurements, from microclimate to vegetation structure and floristics soil nutrient concentrations, in attempting to answer this question. They conclude that fire regime has a greater impact than harvesting, and that the effect of harvesting was not greatest in the annually burnt forest. These conclusions are reasonable within the constraints of their study.
‘Selective harvesting’ is a fundamental aspect of this study- but there is no single standard way to do this. Unfortunately, the protocol used is not described in this ms – there is a single sentence stating ‘Harvesting adhered to currently accepted Department of Agriculture and Fisheries tree marking guidelines…’ These need to be explained to the reader, at least briefly. We are told that the selective harvesting resulted in an average removal of live basal area of 17, 25 and 37% in the three fire treatments – these differences among treatments are quite large, and could influence some of the results and conclusions; the logging disturbance in the single wildfire treatment was more than double that of the triennially burnt one (despite the authors stating that there were no significant differences between these values, which suggests very high variability among plots). In the discussion, we are informed that another study in Australia of selective harvesting removed 60-70% - which could have quite different effects from those reported here - it is good to see that the authors do discuss this. Nonetheless, it means that we need to be cautious in generalising the findings from this study to other intensities of harvesting and burning disturbance. The second constraint is that it could be difficult to detect interactions between fire regime and harvesting, given the inherent high variability in the system and the rather small number of replicates.
The authors describe the intensity of the fires, but I would also have liked a bit more information on their severity – I assume that the low intensity, prescribed fires are surface fires, restricted mostly to the understorey vegetation. What about the moderate intensity wildfire – was this a canopy fire? How much tree mortality was there? Was tree basal area still recovering from this at the time of harvest? It would also be worth explicitly stating that there was no long unburnt control, and that there were two distinct differences between the single wildfire treatment and the prescribed burning treatments – the much lower frequency and longer time since fire, but also the fire it did experience was more intense (and more severe?). However, it seems that the lower frequency effect predominated?
Notwithstanding the lack of power to detect interactions, it would be useful to summarise the significance of the main effects of harvest and fire regime (and interactions) on all your variables in a table – it becomes hard to keep track of all these results.
My comments and suggestions are all quite minor and should be easy to address.
Specific comments
L14 – specify the intensity of harvest – e.g. insert ‘(17 to 37% live basal area removed)’
L34-35 – awkward sentence that’s a bit difficult to interpret – perhaps ‘Selective timber harvesting (i.e. partial harvesting) has been a feature of many dry eucalypt forests, including those of sub-tropical Qld over the last century’
L90 – it would be better to give the range of maximum and minimum temperatures for both the coolest and warmest months – e.g. monthly mean maximum temperatures range from X in July to Y in January and monthly mean minimum temperatures ….
L104 – give a brief description of the severity of the fires also – e.g. prescribed fires were surface ones, the wildfire burnt some tree crowns and killed X% of mature trees
L116 – more details of the harvesting protocol required
Fig. 1 – this is very useful in helping the reader picture the trial layout
L154-156 – this doesn’t mean there aren’t differences, just that variability must be very high (a two-fold range in average harvest intensity seems like a lot)
L177-182 – ‘densiometer’ or ‘densitometer’? (I don’t think it matters which, but be consistent
L201 – ‘coarse’, not ‘course’
L210 – how did you define ground layer biomass? (to what height?)
L224 – This should be ‘near’ rather than ‘at’ ground level – or for consistency (lines 241 and 305) refer to ‘understorey temperature’
L240 – I think this should be P<0.05
L246 – note that a square root transform does not usually need the ‘+1’, given you can use sqrt(0) – the disadvantage of the +1 (for both log and square root transform) is that it will vary according to the units used e.g. will be a bit different for mm and cm.
L282-288 – why are these live BA data not presented graphically – they are among the most interesting you have collected – more interesting than dead BA
L291 – specify Fig. 2a – but see next point
Fig. 2 – the two panels don’t seem to belong naturally together, given both axes are different, the symbols are different. I would like to see the data for live standing area presented graphically – a time course showing this would fit nicely with existing Fig. 2a. Then existing Fig. 2b could become Fig. 3. Also, in Fig. 2a (or the caption) it would help to remind the reader the harvest was in 2015.
L306 – should this be average daily understorey temperature? It’s not clear what these averages are (the fact it varied among months is underwhelming – perhaps just say average monthly maxima ranged between X and Y , and minima between W and Z)
L326-327 – this sentence is grammatically incorrect, you could change ‘whereas’ to ‘by contrast’.
Fig. 3d – why are the fire treatments lumped together here? Is it because there was no fire treatment effect? If this is the case, why do you show all 6 treatment combinations for herb cover in Fig. 3b, given the only significant difference for this measure was with time? Be consistent in your approach.
L384-386 – were either of these two non-native species influenced by fire or harvest treatment?
L389 – change ‘have’ to ‘having’
Topsoil results – it would be easier for the reader to have a figure than try to follow all these values
L577 – add to the end of this sentence ‘than either alone’.
L589-593 – rewrite this sentence – it is grammatically incorrect and the meaning is unclear
L596 – I would add ‘low-intensity’ after ‘multiple’ – it’s good to keep reminding the reader of this
Author Response
This study describes the effects of selective timber harvesting in a sub-tropical forest that has experienced contrasting experimental fire regimes over many decades – a very impressive experimental setup, that the authors have made use of to address the questions of (i) which of selective timber harvesting or fire regime has the greatest impact, and (ii) whether the effects of harvesting would be greatest in the most frequently disturbed (burnt) part of the forest. They have made a wide range of measurements, from microclimate to vegetation structure and floristics soil nutrient concentrations, in attempting to answer this question. They conclude that fire regime has a greater impact than harvesting, and that the effect of harvesting was not greatest in the annually burnt forest. These conclusions are reasonable within the constraints of their study.
We appreciate this positive feedback, thanks.
‘Selective harvesting’ is a fundamental aspect of this study- but there is no single standard way to do this. Unfortunately, the protocol used is not described in this ms – there is a single sentence stating ‘Harvesting adhered to currently accepted Department of Agriculture and Fisheries tree marking guidelines…’ These need to be explained to the reader, at least briefly. We are told that the selective harvesting resulted in an average removal of live basal area of 17, 25 and 37% in the three fire treatments – these differences among treatments are quite large, and could influence some of the results and conclusions; the logging disturbance in the single wildfire treatment was more than double that of the triennially burnt one (despite the authors stating that there were no significant differences between these values, which suggests very high variability among plots). In the discussion, we are informed that another study in Australia of selective harvesting removed 60-70% - which could have quite different effects from those reported here - it is good to see that the authors do discuss this. Nonetheless, it means that we need to be cautious in generalising the findings from this study to other intensities of harvesting and burning disturbance. The second constraint is that it could be difficult to detect interactions between fire regime and harvesting, given the inherent high variability in the system and the rather small number of replicates.
We have added additional information on the tree marking guidelines (lines 126-134). We agree that there can be great differences in the intensity of selective harvesting regimes and now caution that the results reported are not widely applicable to all selective harvesting regimes (lines 549-550).
Agree, the variability in harvesting intensity within treatments could have made it difficult to detect effects and interactions with fire treatment. Unfortunately, this was unavoidable and is just a feature of this type of harvesting in heterogenous forest. For this reason, we did investigate the influence of harvest intensity (basal area removed from an individual transect) on the response variables, but very few clear relationships were detected (exceptions were for wood understorey cover, species richness, CWD and topsoil %N, all reported in the text). We have added a note of caution to the text (lines 550-52) regarding our ability to detect interactions with our design and limited replication.
The authors describe the intensity of the fires, but I would also have liked a bit more information on their severity – I assume that the low intensity, prescribed fires are surface fires, restricted mostly to the understorey vegetation. What about the moderate intensity wildfire – was this a canopy fire? How much tree mortality was there? Was tree basal area still recovering from this at the time of harvest? It would also be worth explicitly stating that there was no long unburnt control, and that there were two distinct differences between the single wildfire treatment and the prescribed burning treatments – the much lower frequency and longer time since fire, but also the fire it did experience was more intense (and more severe?). However, it seems that the lower frequency effect predominated?
Additional information has been added to the text on the fires. The wildfire was not a crown fire but did burn through the mid-storey vegetation that was abundant in this previously unburnt treatment. There was little mortality of large trees (>30 cm DBH) associated with this fire (most resprouting following crown scorch). However, this wildfire did result in mortality of smaller trees (<30 cm DBH). This was picked up in the results (line 311, pre-existing higher dead tree basal area in the wildfire treatment). We have added an additional sentence to the discussion on this also (line 515-516). Tree mortality results are also reported in [69] Lewis, T. Very frequent burning encourages tree growth in sub-tropical Australian eucalypt forest. Forest Ecology and Management, 2020, 459, 117842.
We now explicitly state the differences between the fire treatments, as suggested (lines 112-115). Unfortunately, it is difficult to disentangle the effects of frequency vs the effects of intensity/severity as the they are related (i.e. more frequent burning results in lower intensity and severity).
Notwithstanding the lack of power to detect interactions, it would be useful to summarise the significance of the main effects of harvest and fire regime (and interactions) on all your variables in a table – it becomes hard to keep track of all these results.
A summary table has been prepared for supplementary material (Appendix A).
My comments and suggestions are all quite minor and should be easy to address.
Specific comments
L14 – specify the intensity of harvest – e.g. insert ‘(17 to 37% live basal area removed)’
Added, as suggested.
L34-35 – awkward sentence that’s a bit difficult to interpret – perhaps ‘Selective timber harvesting (i.e. partial harvesting) has been a feature of many dry eucalypt forests, including those of sub-tropical Qld over the last century’
Re-worded, as suggested.
L90 – it would be better to give the range of maximum and minimum temperatures for both the coolest and warmest months – e.g. monthly mean maximum temperatures range from X in July to Y in January and monthly mean minimum temperatures ….
Detail added, as suggested.
L104 – give a brief description of the severity of the fires also – e.g. prescribed fires were surface ones, the wildfire burnt some tree crowns and killed X% of mature trees
Additional details have been added to this paragraph, as suggested.
L116 – more details of the harvesting protocol required
Additional details have been added, as suggested.
Fig. 1 – this is very useful in helping the reader picture the trial layout
L154-156 – this doesn’t mean there aren’t differences, just that variability must be very high (a two-fold range in average harvest intensity seems like a lot)
A comment on the variability was added to this sentence.
L177-182 – ‘densiometer’ or ‘densitometer’? (I don’t think it matters which, but be consistent
Thanks for picking up this typo. We have used ‘densiometer’ for consistency.
L201 – ‘coarse’, not ‘course’
Corrected, thanks.
L210 – how did you define ground layer biomass? (to what height?)
Added ‘to 1 m in height’ here.
L224 – This should be ‘near’ rather than ‘at’ ground level – or for consistency (lines 241 and 305) refer to ‘understorey temperature’
We have used ‘near-ground’ temperatures for consistency.
L240 – I think this should be P<0.05
Correct, changed, thanks for picking this up.
L246 – note that a square root transform does not usually need the ‘+1’, given you can use sqrt(0) – the disadvantage of the +1 (for both log and square root transform) is that it will vary according to the units used e.g. will be a bit different for mm and cm.
Noted, although this is not really an issue in our case.
L282-288 – why are these live BA data not presented graphically – they are among the most interesting you have collected – more interesting than dead BA
A graph has been added for live tree BA (Fig. 2a). We think the dead tree BA result is particularly interesting, given the fire treatment by harvesting interaction.
L291 – specify Fig. 2a – but see next point
Now specified.
Fig. 2 – the two panels don’t seem to belong naturally together, given both axes are different, the symbols are different. I would like to see the data for live standing area presented graphically – a time course showing this would fit nicely with existing Fig. 2a. Then existing Fig. 2b could become Fig. 3. Also, in Fig. 2a (or the caption) it would help to remind the reader the harvest was in 2015.
A graph for live tree basal area was added to this figure as suggested. We opted to keep the dead tree basal area graph with this figure (Figure 2c) to keep all ‘tree layer’ graphs in one figure. The caption was revised accordingly.
L306 – should this be average daily understorey temperature? It’s not clear what these averages are (the fact it varied among months is underwhelming – perhaps just say average monthly maxima ranged between X and Y , and minima between W and Z)
This wording has been modified. We reported temperature readings at 9 am daily, so decided not to report maxima and minima.
L326-327 – this sentence is grammatically incorrect, you could change ‘whereas’ to ‘by contrast’.
Changed, as suggested.
Fig. 3d – why are the fire treatments lumped together here? Is it because there was no fire treatment effect? If this is the case, why do you show all 6 treatment combinations for herb cover in Fig. 3b, given the only significant difference for this measure was with time? Be consistent in your approach.
This was done based on the statistical analysis output. In all cases except 3d, the time × fire treatment × harvesting treatment interaction was significant, so it is appropriate to graph the full interaction plots. In the case of 3d only the interaction between harvesting and time was significant but the interaction with fire treatment was not (hence it is not technically appropriate to plot the separate lines for the fire treatments on this graph).
L384-386 – were either of these two non-native species influenced by fire or harvest treatment?
Both tended to be associated with the wildfire treatment (Figure 4), but occurred in relatively low abundance. This has now been added to the text (line 409).
L389 – change ‘have’ to ‘having’
Corrected, thanks.
Topsoil results – it would be easier for the reader to have a figure than try to follow all these values
Our preference was to avoid adding a new figure here. Instead, we included a table as supplementary material (Appendix A) to make it easier to see where the differences were significant.
L577 – add to the end of this sentence ‘than either alone’.
Added, as suggested.
L589-593 – rewrite this sentence – it is grammatically incorrect and the meaning is unclear
This sentence was edited accordingly.
L596 – I would add ‘low-intensity’ after ‘multiple’ – it’s good to keep reminding the reader of this
Added, as suggested.
