Seed Shadows of Northern Pigtailed Macaques within a Degraded Forest Fragment, Thailand

Round 1

Reviewer 1 Report

The presented article “Seed shadows of northern pigtailed macaques within a degraded forest fragment, Thailand” concerns the influence of macaque on the spread of tree seeds in the degraded forest landscape. The authors combine very precise data from the time consuming and laborious observation of animal troop behavior with the modeling of seed propagation using Hidden Markov models. The article provides interesting data collected in the field during the year of observation. It significantly increases knowledge about the role of animals in the dynamics of forests and is worth publishing in the journal Forests. Nevertheless, I suggest making some changes and improvements, which in my opinion will increase its value.

First of all, I propose to clearly define the objective of the study and provide a hypothesis that has been tested. In its current form, it is not entirely clear whether the main goal was to create a model and test it or whether the model was used as a tool to confirm or reject a certain hypothesis. The second issue is: why the model was used to determine the parameters already known from the data collected in the field (line 422). What was the rationale to use the model if more reliable data can be obtained from direct observation?

The study showed that macaques can spread the tree seeds on short and long distances, but there is no empirical confirmation that this influence composition of the studied stands. In my opinion, investigation on the composition and distribution of tree seedlings in PF and DEF can provide a real base for estimation the effectiveness of macaques as seed dispersers.  I suggest to include such discussion in the 4.1.3 section.  

The interpretation in line 446-454 in its current form is somewhat illogical. It says that the model provide inaccurate data (comparing to direct field data) and on this basis of that inaccuracy conclusions are drawn about the effectiveness of macaques as good seed dispersers.  According to the result, 5.5% of seeds were deposited in PF (line 325). It is low value but it can be really important for future changes of PF stands composition and structure.  In my opinion, it is “fair” to support the statement with real values specified in the paper. I think that such data should also be presented in the abstract. Line 124-129 is not a description of the material collection.

The formula for calculating FAI - how D and B and P were measured? What are the units of the  P?

Figure 4. in the legend there are two colors and the graph shows the third color

Figure 5. in the legend, there is the DDF habitat type, this abbreviation is not explained.

Table A2. It is unclear what values are placed in the Forest Types column. Data for some tree species (the rows)  is separate for PF and DEF, while for others there one line for both  (PF and DEF). Why?

Author Response

Please see the attached file. Thank you very much.

   

 

Author Response File: Author Response.pdf

Reviewer 2 Report

I read the manuscript with interest because it presents important efforts to combine field data with a modelling approach to estimate seed shadows, and it highlights the important role of primates as dispersal agents and for restoration of degraded habitats.

However, I found the manuscript rather technical because the material and methods are not always clearly explained nor justified. I’m confident that the authors can improve their manuscript by making substantial efforts to explain their methodological choices. It will guarantee that their data collection and use in their model is sound and robust (which is hard to tell in the present submission). My main comments are listed below:

# A paragraph explaining primate’s behavior regarding feeding and sleeping site choice is missing. It should help understanding the relevance of the modelling approach. It seems that each modelled movement/step is based on random choice (angle of the movement) from Hidden Markov Model. I would have expected primates to choose their movement direction (angle) not randomly but based on their knowledge of their habitat, towards interesting sites. Could you explain if each movement/step angle is based on primate behavior or not, and why? Similarly, it is stated that “the model can often reuse the same sleeping and feeding sites, which is not the case in reality”-line 418-419. Is it possible to “constrain” the model to avoid such unrealistic choices?

# The authors present an index of fruit availability (FAI) – line 153. From which data was it calculated? How did the author create their “phenology score”? And how the FAI can influence primate’s movements or feeding duration in an appropriate feeding site in the model?

# One important result of the model is that seed shadows vary importantly with the season. What are the underlying mechanisms (variation of FAI? and primate’s movement accordingly?) that is captured by the model, and that lead to these results? This point is not addressed in the discussion.

# The role of tree species during modelling is not clear. Did the authors use an overall average (from all their data, all species considered) for distance dispersal, for gut retention time? Or does the modelling consider which tree species is present in the feeding site, and use data from these specific species to estimate defecation and spat out points?

# Similarly, simulations of 4 specific species seed shadows were made (paragraph 3.5). The number of observations (regarding gut and cheek pouch retention times and dispersal distance) used for parametrization is not clear. For instance, for Listeria pierrei, it is indicated in table 3 and 4 that data for only a few seeds were available. For Afzelia xylocarpa, data for only 3 observations seems available (table 4). How can the authors perform their simulations for these species with so few number of observations?

# Line 187: “We only recorded cheek pouch retention time on individuals that left their feeding trees immediately after storing the seeds”. Why this choice? Can it bias the estimation of dispersal distance (because individuals considered for the study stayed longer in the same place) or retention time?

# The distribution of gut and cheek pouch retention time appeared to be normally distributed (line 232). However, in the model, the authors chose random value belonging the 95% confidence interval of the mean observed values (233). Why not using a random value drawn from the normal distribution (which is not the same)?

# ratio of primate’s occupation of DEF and planted forest are 85% and 15% respectively (line 252). How did the authors make the estimation? It is important since they used this ratio to compare with the results of their model.

# In all table, several species name are missing (species n°1 …). Precise if you could not identify the corresponding species in the legend.

# In the discussion: Do other approaches exist in the literature to estimate or analyze seed shadows in similar contexts? The authors did not discuss the originality of their approach. Is there a way to improve the estimations made by their model in future research?

 

Author Response

Please see the attached file. Thank you very much.

   

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The manuscript has been significantly improved and now warrants publication in Forests