LPJ-GM 1.0: simulating migration efficiently in a dynamic vegetation model

Lehsten, Veiko; Mischurow, Michael; Lindström, Erik; Lehsten, Dörte; Lischke, Heike

doi:https://doi.org/10.5194/gmd-12-893-2019

Articles | Volume 12, issue 3

https://doi.org/10.5194/gmd-12-893-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-12-893-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 12, issue 3

Model description paper

|

08 Mar 2019

Model description paper |

| 08 Mar 2019

LPJ-GM 1.0: simulating migration efficiently in a dynamic vegetation model

Veiko Lehsten, Michael Mischurow, Erik Lindström, Dörte Lehsten, and Heike Lischke

Download

Final revised paper (published on 08 Mar 2019)
Supplement to the final revised paper
Preprint (discussion started on 16 Aug 2018)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

SC1: 'Executive Editor comment on gmd-2018-161', Astrid Kerkweg, 12 Sep 2018
AC1: 'Response to comment by editor', Veiko Lehsten, 18 Sep 2018
RC1: 'Review of paper', Anonymous Referee #1, 19 Sep 2018
- AC2: 'Response to paper review', Veiko Lehsten, 20 Sep 2018
RC2: 'Review for "Simulating migration in dynamic vegetation models efficiently with LPJ-GM"', Anonymous Referee #2, 21 Sep 2018
- AC3: 'Response to review 2', Veiko Lehsten, 09 Oct 2018
RC3: 'Review of paper', Anonymous Referee #3, 25 Sep 2018
- AC4: 'Response to review 3', Veiko Lehsten, 09 Oct 2018

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Anna Wenzel on behalf of the Authors (06 Dec 2018) Author's response

ED: Referee Nomination & Report Request started (07 Dec 2018) by Christoph Müller

RR by Anonymous Referee #2 (14 Dec 2018)

Suggestions for revision or reasons for rejection

I appreciate the steps taken to clarify the applicability of LPJ-GM 1.0.
Unfortunately, it is very difficult to follow the tracked changes. In particular in section 4.3 and 4.4.3, but e.g. also in section 1 and 2, several paragraphs are marked as new/old, although it appears to be exactly the same text as before).
This makes a reassessment more difficult than necessary, and it particularly suggests much more changes than the authors have really applied!

Nevertheless, in my opinion most comments have been addressed and I am mainly left with one important aspect:
From the text provided in the manuscript I still would not be able to fully really assess when and how to apply which algorithm:

1. In 2.7 the authors refer to the performance test of the two methods SMSM and FFTM in in Supplementary 2, showing that "SMSM is still up to an order of magnitude slower than the FFTM". They pronounce that the implementation is in a different environment, but state: "However in a general sense we can see no reason why they should not reflect the performance differences between the algorithms."
Given this, when looking at the last Figure in Supp 2, I do not understand how the computation time for SMSM in Table 1 can nearly equal those of FFTM (and even perform better with transects than FFTM).

2. In the responses to the review the authors state "A simulation in which you want to allow 20 cells maximum distance require twice as much computation time for the SMSM. For long tailed species the FFTM is certainly better suited, given that this method has no such limitations." -- I would strongly recommend to make this also explicit in the discussion (e.g. add this to first paragraph of 4.2).

3. In the responses to the review the authors state "the idea is to only use the corridors in homogeneous landscapes and to speed up the simulation there. In heterogeneous landscapes this simplification is not suitable."
I would appreciate if this is also stated that explicitly in the discussion.
In addition, thinking about this statement of the authors I really wonder why the SMSM is shown with transects at all, since its only preferable for heterogeneous areas and in these it cannot use transects?
I would also appreciate if it would be explicitly stated in the text (e.g. 4.2) that SMSM with heterogeneous area as in Fig.6 cannot be simulated with transects.

4. Although not fully comparable with the other simulations, it would be interesting to have the simulation depicted in Fig.6 also listed in Table 1, or at least to have mentioned the computation time required for this simulation in 3.3.
What was the computation time required for the simulation with SMSM and the non-homogeneous dispersal area?

==
Additionally, I list a couple of minor comments/suggestions, mainly typos/issues in single sentences:

l39: "where the local dynamics is simulated" -> are?

l55: check sentence ".. replace existing vegetation – the processes gap models describe successfully – but they .." Maybe "vegetation – that/which"?

l94-96: check grammar/sentence: "If it is assumed in contrast that the simulated forest is uniformly distributed in the cell, with each time step some seeds reach the neighbour cell, leading to a resolution dependent speed up of migration."

l109: "while Snell et al. (2014)" its --> Snell (2014) not Snell et al. (2014)

l122 and
l151: I still do not agree that the temporal resolution of (tree species) seed production is necessarily annual (particularly not in the tropics), and suggest to e.g. stop after "annual basis.", or add "in temperate forest", or similar (again see e.g. Owens 1994, Brokaw, 1998))

l157: "temperatures; see; Smith et al. 2001" -> "temperatures; see Smith et al., 2001"?

l161: check sentence: "previous applications simulate a certain number of replicate patches are simulated per grid cell"

l163: "1000m2" -> "1km2" for consistency

l177: introduce LAI abbreviation when used first

l179-180: why do you cite (Lischke et al., 2006) here? there is no LAI in that paper

l218: "( 0.99" space

l250: "rarely"? ++ missing bracket ++ order of references?

l392: "...for simulations with transects"

l483: "which resulted a computing time" -> "... resulted in a ..."

l562: "TreeMig ;Lischke" -> "TreeMig; Lischke"

l578: check sentence: "where alien species a planted."

l585: still mix of plural and singular "require a relatively high amounts"

Fig.5: I would appreciate if the y-axis of the distance plots on the right would have similar scales, this would really help for comparison

Kruse, S., ..., in review, 2018.
=> published during the mean time

Table 1: still has 67% instead of 64% for FFTM with 10

Supplementary 3
- readability: sentences are cut onto two pages
- should be 931.6 not 9317

Hide

RR by Anonymous Referee #1 (18 Jan 2019)

Suggestions for revision or reasons for rejection

Overall Comments:
In general, I found the response to reviewers to be thorough and the authors comments to be informative. Once the authors address the remaining minor comments, this paper will be ready for publication.

Minor comments

Throughout – the text needs to be checked for grammar and clarity. In general, check the uses of singular versus plural (e.g., “is” versus “are”, L45), the use of “that” or “which” (e.g., L111, L112), unnecessary “the” (e.g., L78), and the use of present versus past tense. Methods and Results should be written in past tense, not present (e.g., L366, L367, should be in past tense).

Places where the text needs to be revised to improve clarity:
L56 – replace “to have a sufficient amount of seeds at a given location to successfully establish” with “to have seeds present at a given location …”

L64-67 – please rephrase. These two statements are difficult to understand.

L76-78 – this statement needs a reference.

L110 – please include … “between patches within each grid cell” to the end of this sentence.

L114 – please consider rephrasing this “allow simulating species migration of several species simultaneously”, as it is awkward.

L121 – 123 – I would delete this sentence, as it is pretty vague and doesn’t help a reader understand LPJ-GUESS. Especially as much more useful information is presented subsequently.

L140 – not at the end of every year? What is a “migration year”?

L161 – “are simulated”, needs to be deleted from this sentence.

L166 – 173 – this is an example, of where the explanation found in the response to reviewers was much clearer, than what is in the manuscript. Please rephrase this. Also, it is not clear how Figure 3 explains the comment, “LPJ-GM represents a 0.5 x 0.5 degree cells with 200 simulation cells”. Shown is an 11 x 11 box, which is 121 grid cells, of which only a subset are actually simulated.

L223 – please add in the value for loss of germination.

L225 – 227 – this statement is confusing and needs to be rephrased.

L237 – The choice of pi as a symbol in this equation is confusing. To improve clarity and understanding, please use a different symbol.

L238 – “seed number” or do you mean “number of seeds”?

Figure 1 – the coloured scales need to be labeled, and units included. Is this number of seeds? Probability of dispersal? Why does one panel go from 0 – 1000, and the other go from 0 – 14 (x 10-3)?

L292 – instead of referring to grid sizes in degrees latitude/longitude, it would be helpful to mention size (to make it easier to compare to the following statement about maximum dispersal distances of 200 m).

L318 – dispersal ability is not the correct term. This implies the ability of the tree to disperse. Perhaps “differences in available habitat”, or “differences in barriers to dispersal”?

L321 – seeds can reach those areas, however if they do reach it – they don’t germinate. Perhaps this confusion is between biology (i.e., in reality, these areas would be a city or a parking lot – seeds arrive but don’t germinate because it is not suitable), and programming (i.e., in the model, it is simulated as seeds not arriving).

All figures – Scales are missing a label. X and Y axes should be capitalized.

L325 – “local dynamics on the corridors” is unclear. Please clarify you mean that you are simulating vegetation successional dynamics.

L336 – 337 – this sentence is unclear - “neglected the points within the first 5 km”. Does this relate to the starting location? Or the ending position, if they migrated < 5 km?

L372 – 373 – this statement (and subsequent statements, e.g., L394-395, 427) about the reduction or increase in computing time by percentages is unclear. Table 1 is very helpful and clear, but these summary statements are not. Perhaps because the Table includes + and – so it is clear if the total times were increasing or decreasing? A reduction BY 88% is different than a reduction TO 88%.

L392 – please clarify here, that the migration rate is reduced when you use corridors.

There are numerous statements in the discussion which are awkward, and need to be edited for clarity. E.g.,
L448, L460 – 461, L463 – 464, L482 – 484, L516 – 518, L531-532, L568-570, L575 – 578,

L492 – The approach by Snell (2014) did not use a reduced number of patches. There were 400 patches per grid cell.

Hide

ED: Publish subject to minor revisions (review by editor) (19 Jan 2019) by Christoph Müller

AR by Veiko Lehsten on behalf of the Authors (09 Feb 2019) Author's response Manuscript

ED: Publish subject to technical corrections (18 Feb 2019) by Christoph Müller

AR by Veiko Lehsten on behalf of the Authors (25 Feb 2019) Author's response Manuscript

Short summary

To assess the effect of climate on vegetation, dynamic vegetation models simulate their response e.g. to climate change. Most currently used dynamic vegetation models ignore the fact that for colonization of a new area not only do the climatic conditions have to be suitable, but seeds also need to arrive at the site to allow the species to migrate there. In this paper we are developing a novel method which allows us to simulate migration within dynamic vegetation models even at large scale.