Review of “Simulating Bark Beetle Outbreak Dynamics and their Influence on

Carbon Balance Estimates with ORCHIDEE r7791”

This manuscript describes work to implement bark beetle outbreaks in ORCHIDEE, a land surface model. The authors adapt a previously published implementation of outbreaks from the LandClim model, which has different spatial and temporal characteristics.  ORCHIDEE was run in aspatial mode with conditions from various locations to describe the outbreaks’ effects on carbon cycling, and model results are presented.

This topic is very challenging, and the authors should be commended for their study.  The authors have obviously put in a lot of effort.

However, I think there are major issues with the manuscript.  First, I don’t think there was really an evaluation, qualitative or quantitative.  My suggestion is that the manuscript should add steps (my recommendation) or call their study a sensitivity study, not an evaluation (not great, though).  I acknowledge that evaluation of the model results is difficult given the multiple processes and parameters, but more could be done.  Second (and related), more presentation of model results that relate to beetle outbreaks will improve the reader’s ability to assess the study.  Third, there are major differences among bark beetle species, and this manuscript about modeling Ips typographus should ensure that readers understand that the model cannot be applied to other species without (potentially significant) modification.

Major comments

My main concern about this manuscript and study is the evaluation.  Stating that the model has been evaluated (qualitatively or quantitatively) means to me that subsequent studies can claim that the model is ready for use in assessing beetle impacts on the carbon cycle.  So the words matter, in my opinion.  The analysis that is presented consists of running the model across a number of sites and for various windthrown intensities, then evaluating cumulative wood volume and NEP and time series of NPP.  Furthermore, the comparison with other studies (Table 5) is flawed.  This study is a first step, but does not rise to the level of an evaluation.  Rather, this is a model sensitivity study, in my opinion.  At a minimum, to use the term “evaluation”, I think the following additional steps are needed:  a) presentation and analysis of the variables and metrics associated with bark beetle outbreaks (e.g., the important variables of Equations 1-14) for the model runs, including time series (similar to Figure 2, but more detailed); b) comparison of time series of drivers (climate, windthrow, substrate, etc.) and outbreak variables to assess how the drivers are affecting the outbreak variables; c) at least one run, and ideally multiple runs, for locations with observed outbreaks, and a comparison of model results with observations (or perhaps results from previous modeling studies?); d) improved selection of other studies for comparison in Table 5, deleting studies that are not relevant or address beetle species whose behavior is different and adding other modeling studies.  Unless the authors include these steps, I don’t think the authors should use the term “evaluation”, and text throughout the manuscript should be changed to avoid implying an evaluation.  See below for detailed comments.

The manuscript is too vague about which bark beetle species is simulated.  I’m assuming it is Ips typographus.  Some aspects of bark beetle biology and ecology are generalizable across species, but others are not (including what triggers and influences outbreaks). The manuscript needs to state the species of interest explicitly in the Abstract, Introduction, and Methods, and add some discussion about generalizability to other bark beetle species in the Discussion. 

Related to this, because of differences in ecology among beetle species, the model implementation described in this manuscript is not necessarily applicable to other locations/systems.  For instance, mountain pine beetle, the major beetle species in North America, is not triggered by windthrow events, has multiple climate influences, and has one, not multiple, generations per year.  And the beetle phenology model used in this study, while maybe appropriate for Ips typographus, is not appropriate for other beetle species.  Differences and lack of direct applicability should be included in the Discussion.

Introduction

In the Introduction, please include more description of past studies that modeled effects of bark beetles on ecosystem properties.  Please distinguish studies that prescribed outbreaks from those that modeled outbreaks.  Briefly describe system (beetle and host) and results.  Suggested studies include Temperli et al. (already cited, but what was the study about); Jonsson et al., AgForMet, 2012; Seidl and Rammer, Landscape Ecol., 2017.  Please do a literature search to identify others.

In the Introduction, please provide more information about Ips typographus outbreaks, including what fraction of a 25 x 25 km ORCHIDEE grid cell might be affected.  What have been observed outbreak areas or volumes, and what are these values relative to a grid cell?  (I realize in this study, the model is run in aspatial mode.)

Methods

Winter beetle mortality from very low temperatures has been modeled previously for Ips typographus (Jonsson et al.) and other beetle species.  Why wasn’t this included here?  Add to Discussion as missing processes.

Throughout the description of equations:  it would be helpful to interpret for the reader the equation, including what processes are represented and why they are represented the way they are.  The authors do some of the this, but for other equations, I am confused.

L 138: “k”, not “K”, to match Eq 3?

L 225-240:  Both G and rDD are described as number of generations in a year.  Please refine the description of one of the variables.

Equation 1: 

1. Please describe the rationale and ecology/biology behind the apparent averaging of G and Act.

1. What is Cpd?

1. What is the the rationale and ecology/biology behind multiplying the different factors?

Assuming that Si (Eq 1, 5) is the same as SI (L 265), this variable should not be in the calculation of RI twice.  It makes more sense to be to have susceptibility separate from beetle pressure, so remove from Eq 1 and clarify L 255?  Maybe I’m confused.

L 259:  I assume what is meant is susceptibility TO bark beetle (as in L 255), not OF.  If so, please change.  If not, please explain more.

Please describe the differences in Wr and Ww equations depending on beetle population stage.  What do these differences mean about the biology/ecology of the insect?

I think both drought and competition should weaken tree defenses in similar ways, so I don’t know why Wd doesn’t switch based on beetle population stage similar to Wr. 

Section 2.2.3 opens with a comparison between tree mortality from mass attacks and tree mortality from, presumably, beetles in the endemic population phase.  So I’m confused about if the killed biomass is applicable to both situations, or just one?

L 366:  Did I miss the description of the modifications to account for the differences in spatial scale?  If they are not present, please add.

Please add units where appropriate.  Examples include Binf, Bt, Bdb.  There may be other instances.

It might be easier to understand if Section 2.2.5 were moved before the equations that represent the processes.

L 421:  conspecific (not interspecific) competition

Results

I think the Results section should add a presentation about the simulated outbreaks.  Please add figures that illustrate time series of outbreak metrics (BPI, wood volume killed by beetles) for each site, similar to Figure 4, and discuss.

Also, please move Section 3.2, which describes outbreaks and comparisons with other studies, before the sections that discuss how outbreaks vary across climate and windthrow gradients.  Seems more logical in this order.

Section 3.2:  Please provide text that interprets Table 5.  What are the most important points?

Figure 3: The circle area represents “tree damages”.  Can the authors be more specific about what metric this represents?  Wood volume?  Carbon?  Area?  Units?

Figure 4:  Why do the highest wind speeds result in faster NPP recovery than some lower wind speeds?

Figure 5:  Would be helpful to separate the NPB effect of windthrow from that from bark beetles.  I realize that may be tricky, and may require two figures, but it will provide readers with a better sense of the contribution of outbreaks to the carbon dynamics.

Figure 6:  Same comment as for Figure 5 above.

The modeling approach described in this study is not necessarily applicable to outbreaks of other bark beetle species.  Text throughout the manuscript should clarify this point.

Comments/questions on Table 5:

1. Throughout this table: the Edburg et al., 2012 study is not in the reference list, but if that’s the Frontiers article, that was a conceptual paper and should not be used for model evaluation. Please select something specific to Ips typographus.  Similarly, the Hlásny et al. 2021 article is a review; please replace that with observational studies that support the statements in the table.

1. Also, many metrics are specific to a particular (and different) beetle species, which may have different dynamics and drivers.The comparisons will be strongest when studies of Ips typographus are included.

1. The entries in this table, which are intended to represent an evaluation, include processes not simulated by ORCHIDEE (grey stage and last row).Please remove from table (and therefore from the evaluation).

1. Mean climate and interannually varying climate are implicitly included in the “Climate with endemic stage” row, right column, so I’m not sure that the statements that a cold or warm year following a wind event and the effects on outbreaks is represented in Figure 3.Either demonstrate these ideas by showing time series of outbreak metrics together with interannually varying climate, or remove these statements from the table.

1. First row of “Red or epidemic stage”:Please demonstrate this idea that high populations can kill healthy trees by plotting time series of the various variables that represent tree health (susceptibility to drought), BPI, etc.

1. Second row of “Red or epidemic stage”:The cited studies do not report observations of outbreak length, that I can see.  Please use other studies that are not reviews or conceptual papers.

1. The recovery part of the Pfeifer et al. study was only modeling.Modify the text (not observational).  And comparing results from different models is okay but weaker than observations.

Discussion

Please compare your modeling to previously published studies (see my comments on the Introduction for some studies, but not all?), noting similarities and differences in model formulation, applications, and results.

Minor comments

As written, paragraphs are difficult to distinguish, making reading more challenging than needed.  Please add a carriage return between paragraphs.

Figure 1, L 295:  The authors use “green phase”  (or stage) as the phase in which trees are attacked but haven’t yet changed color.  Do these trees successfully repel attacking beetles and therefore survive?  The forest entomology literature usually refers to the stage of attacked (and killed) trees as “green-attack” because “green phase” could easily be interpreted as “unattacked”.  I haven’t heard of a stage of successfully defended trees.  Please consider changing the description.

Around L 110:  Is this a dynamic global vegetation model that allows for competition among PFTs?  Please be explicit in the description.

L 145:  The phrase “Since revision 7791” suggests that previous work implemented some aspects of beetle outbreaks.  If the authors mean that this manuscript describes revision 7791, please alter the phrase.  If there is previous work, please cite the study.

L 170 (and elsewhere if needed):  Please italicize species genus and species names following convention.

Figure 1 is a nice representation of the processes included in the modeling.  Some comments:

Figure 1 legend:  “developed” is misspelled.

Figures 1 and 3:  Typically, circular motion (like life cycles) is drawn in a clockwise direction.  The counterclockwise direction confused me momentarily. Maybe that’s just style, but it might be easier for readers to reverse the direction...?

Figure 1 caption:  It would be more accurate to discuss beetle population dynamics separately from tree phases.  They are linked, of course, but in some systems visible changes in trees occur 9 months after beetle attack.  And even during the declining population stages (if one is discussing population at the regional, not stand, level), beetles can be attacking trees, putting them into the green-attack phase (followed by the red phase).  So the caption could be improved with more attention in two ways:  a) decoupling population stage and tree stage; b) being clearer about populations at the outbreak (landscape) level versus at the stand or tree level.

Figure 2: provide y-axes labels and descriptions in caption (L 400 suggests the need).  “threshold” is misspelled in figure.

Table 4:  I assume the authors intend this to be read in landscape, not portrait, mode.  If so, please reverse the order of the columns so the reader readers, left to right, the site name, full name, etc.

Figures 3 and 4 and Table 4:  Are the temperatures in Figures 3 and 4 supposed to match, and supposed to match those in Table 4?  They don’t…

Figure 4:  Would be helpful to add a vertical line at Year=2, which shows the windthrow effect on NPP (right?), so that the reader can see the effect of outbreaks on subsequent years.  Too difficult to figure out now.

Figure 4:  Add that what is plotted is annual NPP (right?).

Several places refer to Table 6.  Is this supposed to be Table 5, or something else?

In this manuscript, Marie et al. describe the implementation of a bark beetle outbreak model into the ORCHIDEE dynamic global vegetation model. Pests are an important driver of forest dynamics but are notoriously hard to simulate, so development in this area is always welcome. The authors do not attempt to reproduce specific real-world bark beetle outbreaks in this work, focusing instead on the model description and a more general evaluation of the model’s behavior. They show that at least some patterns seen with regard to bark beetle dynamics, such as the duration of outbreaks and the effects of temperature, are reproduced well. I consider this effort a good first step and, in principle, worthy of publication.

However, there are a number of issues that prevent the current version of this manuscript from being accepted for publication. Some parts of the model description are confusing or too vague, and some analyses are insufficiently described. The paper would also benefit greatly from some reorganization, and I have a large number of more minor comments and corrections. With all that in mind, I recommend this paper be reconsidered after major revisions.

Please see the attached file for more details.