Bat Activity in Organic Rice Fields Is Higher Than in Conventional Fields in Landscapes of Intermediate Complexity
Round 1
Reviewer 1 Report
This study considers effects of organic rice cultivation as well as landscape complexity on insect communities and activity of their bat predators in a tropical agricultural area. This topic is very important and also very poorly studied, and this study is a welcome addition to the literature. The manuscript is well written and the methods appear valid.
My main suggestion is that there appear to be a small number of outliers in the data. In Figure 3, I can’t figure out if the data point just above the statistical result in the upper left of each plot is a data point or somehow part of the legend. But Figures 4 and 6 each suggest that the results may have been influence by the presence of 1-2 outliers. There is some suggestion in the discussion of what this might be related to water or proximity to limestone outcrops. I believe that a direct mention of these outliers in the discussion would be helpful, including what they represent and why you decided to retain them. If you already examined the effect of discarding them, briefly mention the results.
Author Response
> In Figure 3, it is the data point that is above statistical result. In all Figures (3, 4, 6) we use "DHARMa" package to visualise, and we did not detect extreme outliers or data over dispersion. Outlier test show p > 0.05 which means that we do not have enough evidence to conclude that an outlier exists. We also examine by removing outliers from our models, and we did not find significant changes of the results .
Ex. Total activity model (Figure 4)
There is some suggestion in the discussion of what this might be related to water or proximity to limestone outcrops. I believe that a direct mention of these outliers in the discussion would be helpful, including what they represent and why you decided to retain them.
> In discussion, we have mentioned the finding that more bat activity in conventional paddies in clear landscape could be from the presence of water body (50-100m width) there. We are not sure in this case but the above package did not showed that outlier exist. So we did not mention much on outliers.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Please see both attached files in the zipped folder.
Comments for author File: Comments.zip
Author Response
please see attached file.
Author Response File: Author Response.pdf
Reviewer 3 Report
Excellent work, well designed and presented. My only concer regarding the design is about the time-frame used, which includes a single year and no replicates per block. I understand this is not possible to address, and since the data analysis, results obtained and conclusions drawn from them are sound and well supported by the data the authors gathered, I recommend the publication of the work.
Besides, a few minor comments and typos:
Line 61: grasslands as such have not been the primary focus. Consider referring to anual crops instead.
Lines 210-212: why using correlations instead of adding prey availability as predictor in the 'Total bat activity' model?
Line 227: coud the authors indicate the threshold used?
Lines 205-210: were the predictors scaled previous to running the models? It is recommendable to ensure optimal model convergence and comparable effect sizes \(please see e.g.: Zuur, A.F., Ieno, E.N., Elphick, C.S., 2010. A protocol for data exploration to avoid common statistical problems. Methods Ecol. Evol. 1, 3–14. doi:10.1111/j.2041-210X.2009.00001.x\ Zuur, A.F., Leno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M., 2009. Mixed effects models and extensions in ecology with R, Public Health. doi:10.1016/B978-0-12-387667-6.00013-0\).
Line 240: 'There were twelve taxa', should perhaps declare that 'There were a minimum of twelve..'. Unless the authors suspect that all 'unknown bats' belong to some of the identified species.
Line 244: 112 insect sampling events are reported, though according to M&M the number should match that of nigth/detectors \(56\). Please clarify.
Line 302: Table 4. It would help reading the table to more clearly separate the rows corresponding to each response variable. For instance it is unclear from the table whether Tropical storm disturbance is and independent v\ ariable variable used in the 'Total bat activity' or in the 'Feeding buzzes' model, or in both.
Line 334: Figure 8 uses superscripts to denote significant differences, which seems unnecessarily unconventional. A more conventional horizontal line with the appropiate number of asterisks would make the interpretation mor\ e straightfoward.
Line 418: 'lavartus' for 'larvatus'
Line 434: 'the tropics' for 'tropical islands'?
Line 482: typo 'scenaios' for 'scenarios'
Author Response
Excellent work, well designed and presented. My only concern regarding the design is about the time-frame used, which includes a single year and no replicates per block. I understand this is not possible to address, and since the data analysis, results obtained and conclusions drawn from them are sound and well supported by the data the authors gathered, I recommend the publication of the work.
Besides, a few minor comments and typos:
Line 61: grasslands as such have not been the primary focus. Consider referring to annual crops instead.
> We have changed this term as you suggest (grasslands to annual crops).
Lines 210-212: why using correlations instead of adding prey availability as predictor in the 'Total bat activity' model?
> We have added insect abundance and insect biomass as predictors in the 'Total bat activity' model and they did not show significant effect on bat activity.
Line 227: could the authors indicate the threshold used?
> Yes, so we have indicated the threshold used in M&M.
Lines 205-210: were the predictors scaled previous to running the models? It is recommendable to ensure optimal model convergence and comparable effect sizes \(please see e.g.: Zuur, A.F., Ieno, E.N., Elphick, C.S., 2010. A protocol for data exploration to avoid common statistical problems. Methods Ecol. Evol. 1, 3–14. doi:10.1111/j.2041-210X.2009.00001.x\ Zuur, A.F., Leno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M., 2009. Mixed effects models and extensions in ecology with R, Public Health. doi:10.1016/B978-0-12-387667-6.00013-0\).
> We have verified that every predictor is appropriate before running every model.
Line 240: 'There were twelve taxa', should perhaps declare that 'There were a minimum of twelve...'. Unless the authors suspect that all 'unknown bats' belong to some of the identified species
> we change to 'There were a minimum of twelve...'
Line 244: 112 insect sampling events are reported, though according to M&M the number should match that of night/detectors \(56\). Please clarify.
> Because each field had two insect traps. We took samples from two rice fields for one night (1 pair; organic and conventional). So we used a total of four insect traps for one night. We took samples for 28 nights in this study. As a result, the total number of traps repeated is 4 x 28 = 112.
Line 302: Table 4. It would help reading the table to more clearly separate the rows corresponding to each response variable. For instance it is unclear from the table whether Tropical storm disturbance is and independent v\ ariable variable used in the 'Total bat activity' or in the 'Feeding buzzes' model, or in both.
> We have modified the table 4 as your suggestion.
Table 4. Results of statistical analyses showing the effects of landscape characteristics on bat species richness, total bat activity, total number of feeding buzzes, activity of each bat species and activity of each foraging guild. [abbreviations: *, p < 0.05; **, p < 0.01; ***, p < 0.001; SE, standard error].
|
Response Variable |
Independent Variable |
Estimate (±SE) |
z-Value |
p-Value |
|
Bat species richness |
% of agricultural area within 2-km radius |
-0.02 (± 0.01) |
-2.19 |
* |
|
|
|
|
|
|
|
Total bat activity |
% of forest and semi-natural area within 1-km radius |
0.09 (± 0.03) |
3.07 |
** |
|
|
Tropical storm disturbance |
1.65 (± 0.39) |
4.23 |
*** |
|
|
|
|
|
|
|
Feeding buzzes |
Distance to nearest limestone outcrop |
-0.14 (± 0.04) |
-3.66 |
*** |
|
|
Tropical storm disturbance |
1.07 (± 0.37) |
2.86 |
** |
|
|
|
|
|
|
|
Buzz ratio |
Distance to nearest limestone outcrop |
-0.002 (± 0.001) |
-2.02 |
* |
|
|
|
|
|
|
|
Edge-space foragers |
% of forest and semi-natural area within 1-km radius |
0.13 (± 0.04) |
2.99 |
** |
|
|
Tropical storm disturbance |
1.93 (± 0.56) |
3.46 |
*** |
|
|
|
|
|
|
|
M. horsefieldii activity |
% of agricultural area within 2-km radius |
-0.16 (± 0.06) |
-2.72 |
** |
|
|
Tropical storm disturbance |
1.96 (± 0.80) |
2.44 |
* |
|
|
|
|
|
|
|
M. muricola activity |
Distance to nearest water body |
-0.16 (± 0.05) |
-2.93 |
** |
|
|
|
|
|
|
|
H. larvatus activity |
% of forest and semi-natural area within 1-km radius |
0.17 (± 0.05) |
3.43 |
*** |
|
|
|
|
|
|
|
Open-space foragers |
Distance to nearest water body |
-0.14 (± 0.05) |
-2.86 |
** |
|
|
|
|
|
|
|
S. kuhlii activity |
Distance to nearest water body |
-0.15 (± 0.05) |
-2.97 |
** |
|
|
|
|
|
|
|
T. melanopogon activity |
Distance to nearest water body |
-0.14 (± 0.06) |
-2.29 |
* |
|
|
Distance to linear feature |
-12.49 (± 5.64) |
-2.22 |
* |
|
|
|
|
|
|
|
Narrow-space foragers |
% of forest and semi-natural area within 1-km radius |
0.13 (± 0.04) |
3.17 |
** |
|
|
|
|
|
|
|
Rhinolophus spp. activity |
% of forest and semi-natural area within 1-km radius |
0.13 (± 0.04) |
3.03 |
** |
Line 334: Figure 8 uses superscripts to denote significant differences, which seems unnecessarily unconventional. A more conventional horizontal line with the appropriate number of asterisks would make the interpretation mor\ e straightforward.
> We have modified the Figure 8 as your suggestion by removing superscripts and add asterisks instead.
Figure 8. Mean (± SE) values of (A) total bat activity, (B) total number of feeding buzzes, (C) activity of edge-space foraging bats; and (D) activity of M. horsefieldii in matched organic and conventional rice fields before (pre) and after (post) tropical storm disturbance. Asterisks indicate significant differences between pre and post tropical storm disturbance (p < 0.05).
Line 418: 'lavartus' for 'larvatus'
> We have fixed the typo you mention.
Line 434: 'the tropics' for 'tropical islands'?
> We have changed 'the tropics' for 'tropical islands' as you suggest
Line 482: typo 'scenaios' for 'scenarios
> We have fixed the typo you mention.
