the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Apr 2024
Where to start with climate-smart forest management? Climatic risk for forest-based mitigation
Abstract. Natural disturbances like windthrows or forest fires alter the provision of forest ecosystem services like timber production, protection from natural hazards or carbon sequestration. After a disturbance, forests release large amounts of carbon and therefore change their status from carbon sinks to carbon source for some time. Climate-smart forest management may decrease forest vulnerability to disturbances and thus reduce carbon emissions as a consequence of future disturbances. But how to prioritize stands most in need of climate-smart management? In this study we adopted a risk mapping framework (hazard × vulnerability) to assess the risk to climate-related forest ecosystem services (carbon stock and sink) in forests prone to windthrow (in the Julian Alps, Italy) and forest fires (in the Apennines, Italy). We calculated hazard by using forest fire and windthrow simulation tools, and examined the most important drivers of the respective hazards. We then assessed vulnerability by calculating current carbon stocks and sinks in each forest stands. We used these values together with the calculated hazard to estimate “carbon risk”, and prioritized high-risk stands for climate-smart management. We show that combining disturbance simulation tools and forest carbon measurements may help in risk-related decision making in forests, and taking planning decisions for climate-smart forestry. This approach may be replicated in other mountain forests to help understanding their actual carbon vulnerability to forest disturbances.
- Preprint
(1740 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-758', Anonymous Referee #1, 05 Jun 2024
General comments
I think this manuscript is a good example of how to leverage field-based information and disturbance simulation tools, to inform decision-making in forest management under risks connected to global change, and represents a promising way forward in climate-smart forestry, though its actual implementation might still find obstacles related to the socio-economic context.
Though I overall positively assess the study, I think the manuscript would greatly benefit from a thorough revision of the language by a proficient English speaker. What is pointed out in the “technical corrections” does not represent an exhaustive list, as I’m not a native speaker myself. Furthermore, a series of clarifications on the methodology is required before the publication.
Specific comments
- In the section “Study area” you could more explicitly describe the typical disturbance regime of the two areas. Furthermore, add info on the climate station from which you got temperature and precipitation data, and specify to which interval of years these data refer to.
- In the field sampling section (lines 99-100), could you specify how the 3 trees per species and size were chosen? E.g. as the closest 3 to the plot centre? Furthermore, in cases when there were less than 3 trees of a species and size class, how did you proceed? You just sampled what you could find or got info from trees outside the plot?
- Are management unit (section 2.2) and stand (2.3.1) equivalent? If so, why using different terminology?
- You mention “unmeasured forest units” at line 159, but earlier you wrote “For every unit, we randomly established a circular sampling plot with 10 m radius.” (lines 97-8). Could you clarify how the unmeasured forest units came up?
- Could you elaborate on why you applied different statistical models (multiple linear regression with stepwise backwards selection, and regression trees) to identify the main hazard drivers for wind and fire, respectively, and not the same type of model in both cases?
- Could you further elaborate on how the estimation of “belowground, deadwood, litter and soil carbon using empirical equations from the Italian National Forest Inventory” was done (line 181-2)? Since they represent an important part of the carbon stock together with the aboveground tree biomass, I think it’s important to be more detailed about the estimation methodology and its reliability.
- I think the role of microbial respiration, besides that of disturbances, in limiting forest’s ability to be a carbon sink should not be overlooked, especially if this methodology would be used, as suggested, “to calculate the amount of carbon credits” (line 308). I think you should elaborate on this in the discussion.
- Concerning forest management, I appreciated the identification of the main drivers of wind and fire hazards, to easily inform forest practitioners, and the part of the discussion including thorough management recommendations. I think what could be a good adding here are some considerations about the feasibility of implementing these decision support tools and related management strategies, given the scarcity of planning and ownership fragmentation in the Italian context.
Technical corrections
26 would use “furthermore” rather than “on the other hand”, since the two sentences (about fire and wind respectively) are not in opposition, you’re rather adding something to the previous sentence
29 a better reference than Runkle 1985 here would be: Thom, D., & Seidl, R. (2016). Natural disturbance impacts on ecosystem services and biodiversity in temperate and boreal forests. Biological Reviews, 91(3), 760-781.
34 Would remove “Indeed, ”
43 Would use the word “strategy” instead of “goal”
50 it would be good to add a reference at the end of this sentence
69, 83 Use the verb “offer” instead of “have” with (provisioning or regulating) ecosystem services
71 can be rewritten as “Fusine area features high forests managed with shelterwood system …”
111 no reference is reported for Brown et al., 1982
159 typo: tree height
161 typo: then
162 would use “resolution” instead of “grain”
215 would use “varied among” instead of “differed in”
254 use “or” instead of “/” between C stock and CO2 sink to avoid potential misinterpretation of that being a ratio
261-2 rewrite as “… forest ecosystem services (ES), such as climate change mitigation, hydrogeologic protection and recreational value, should also be considered as vulnerable assets.” I would argue that ES are, by definition, “of interest to society”, so it’s superfluous to repeat that.
295 add “found to be” as “tree height and tree species were found to be the most important predictors”
299 use “greater solar radiation” instead of “more sun rays”
Citation: https://doi.org/10.5194/egusphere-2024-758-RC1 -
RC2: 'Comment on egusphere-2024-758', Anonymous Referee #2, 12 Jun 2024
The manuscript describes an attempt to operationalize the CSF concept for two areas in Italy, using state of the art disturbance models to identify high-risk areas.
Although I support the idea of the paper very much, I think there are several critical issues with the manuscript.
- The presented risk maps are relative, showing the areas where risks are higher than elsewhere. However, it would be very important to know the absolute risk to judge if action is really needed. If the absolute risk is lower than your tolerance level, no need to do anything. ForestGALES provides the critical windspeed, which is commonly used in combination with wind climate characteristics to judge what is the probability of exceedance of this critical windspeed.
- It is unclear how the combination of windspeed/fire risk and carbon stocks/sinks are made, what values result, and how they are grouped into classes. Only in the caption of figure 7 there is a remark that classes are “equally distributed by quantiles”. This should be part of the methods. And again, this makes the classes relative, always showing 20% of your area at very high risk. Even if you would take action based on your analysis, there will always be 20% area at very high risk.
- A critical windspeed of 5 m/s as presented in the manuscript is extremely low, and will likely be exceeded every year. In other words, such a stand shouldn’t be able to exist. Although the CWS is only used for a relative judgement, it raises questions on how trustworthy the results are. A validation would be nice, or at least a reflection on the CWS that are found.
- I have difficulties with the approach of averaging the critical windspeed in case of multiple cohorts within a stand. Low CWS will have a high likelihood of occurring while a high CWS will have a low likelihood, in a very non-linear way. In fact I think you should take the lowest CWS and not the average, as the lowest CWS will determine your risk.
- The authors present carbon stock as well as carbon sink as a potential asset at risk, but don’t give any judgement or guideline which of the two are more important. Personally I think the carbon stocks are more important. If you lose them, it takes a long time to rebuild. You may harvest high-risk stands with high stocks and use the wood for products etc. A high carbon sink basically means a good growing forest. I don’t think it is a good option to harvest those and put trees back in that grow less. Your risk map will show a reduced risk, but I wouldn’t call this a climate-smart option.
- After the detailed risk models are applied, the authors apply simplified regression models to determine what are the underlying risk factors. Not surprisingly, they show exactly the same factors that are entered into the risk models, and that are already described as being important in risk modelling. I think the value of these regression models is that they could be used in other areas, without the need to run the detailed risk models. Showing such an application (plus perhaps validation) would greatly enhance the study.
- At the same time, I wonder what this study in practice brings for the forest managers of the study area. It tells them that stands with high stocks, tall trees and a high share of spruce are more vulnerable. Isn’t this common knowledge? Do they need such an approach to locate such stands? Or would the approach be more suitable to map risk at larger scale?
- On the same line, the authors only give very generic advice on what to do if you have a high-risk stand. Does that really apply to all stands? Line 320 includes the advice to avoid unstable edges (which is not included in the modelling approach at all I think), and selective thinning for better anchorage. However, thinning an overstocked tall spruce stand is extremely risky, as it opens up the canopy to the wind, and a total collapse may follow. In practice you may rather clearfell the whole stand, or leave it until something happens.
- For among others the above reasons, I think there is absolutely no ground for the statement in line 283/284 that “we showed that hazard modelling can be a valuable support…”.
- To me it is unclear how much of the characteristics of the forest/the stands is really measured, and how much is averaged/interpolated and how much is taken from the forest management plans. One single random plot (line 97) in a stand is by no means enough to characterise the stand. Maybe a map could be included that shows what part of the area was measured, what was interpolated and what was taken from management plans or otherwise.
- The protocol as described in line 102-124 gives rise to many questions:
- Line 105 states that all pieces of deadwood of 2.5-7.5 cm are counted. Line 110 applies a correction factor of 1.13 for diameters smaller than 7.5 cm, but these have been counted already? I thought perhaps the authors mean 2.5 cm, but line 121 shows that the fuel load for deadwood smaller than 2.5 cm is measured separately. Why the correction factor?
- Line 106 states that the amount of deadwood is calculated in t/ha, while equation 1 results in the volume (V) deadwood in m3/ha.
- Line 115 indicates that the number of shrubs intersecting the transect was used, while in line 104 only the height of shrubs was measured at 1 m intervals. When is a shrub considered to intersect with a transect? Only at the 1 m intervals or always? What happens if the same shrub has a canopy that extends over multiple 1 m intervals? How do you estimate shrub density per ha from the number intersects?
Citation: https://doi.org/10.5194/egusphere-2024-758-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 179 | 49 | 19 | 247 | 12 | 12 |
- HTML: 179
- PDF: 49
- XML: 19
- Total: 247
- BibTeX: 12
- EndNote: 12
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1