I’ve now carefully read the new version of the MS by Hawkes et al., on massive autumn migration across the Pyrenees. I think the MS has greatly improved. I’m happy with the omission of the analysis based on the UK data, and also with the new classification on migration behaviour with just the two categories ‘High Altitude Migrants’ and ‘FBL Migrants’, which makes more sense. Also, the explanations on how migration has been quantified are now clearer, and Fig. 4 (and the analysis on meteorological factors influencing migration) has been improved.
My main concern is with the quantification of migrants annually traversing the Pyrenees. I still find it quite confusing and not very convincing. I found it hard to accept some of the assumptions in l. 367-384.
Assumption 1 does probably work for FBL migrants but not necessarily for High Altitude Migrants (HAM). Most episodes of massive migration combined tail winds in the northern side of the Pyrenees (i.e., the French valley) with head wind blowing from the southern side (i.e., the Spanish side) over the ridge. HAM were thus affected by head winds when they reached the ridge, thereby descending to ground level to minimise the force opposed to their movement. It was then when they became noticed by the observers. But it is likely that HAM not flying directly over the narrow pass would lower their altitude and simply continue their movement near the ridge, being not necessarily channelled through the 30 m-wide pass. This could lead to a serious underestimation of HAM.
But perhaps I find more difficult to accept that the number of insects migrating on a tailwind day equals the number of insects migrating on a headwind day (l. 376-377). On what does this assumption rely? It contradicts current knowledge indicating that HAM are quite selective on favourable winds to start their migration (there are a number of papers by several of the co-authors of the MS and other researchers showing this for several moth and butterfly species), which means that the HAM component could strongly vary in both kind of days.
Moreover, how do the number of tailwind viable days was calculated? In the equation in l. 381-382 you don’t specify that the number of viable days is different for headwind and tailwind days, but this was indeed the case (In Results: 8.75 viable days for headwind and 12.5 days for tailwind).
All in all, I find this section very confusing and in clear need of improvement.
I’ve some additional comments:
- l. 355-357. “was recorded on a sliding scale”. What does it mean?
- l. 517-518. Oscinella frit. I asked some Spanish experts on Diptera about the exact status of this species, following my earlier comment. Indeed, in GBIF there is an old record from Sierra Nevada, with little details. But it is sure that O. frit was very rare in Spain at that time. In the Catalogue of Diptera of Spain, Portugal and Andorra, published in 2002, the only confirmed records come from the Canary Islands, Madeira and Azores. On the other hand, there are some mentions in a few old publications by pest researchers (some as old as 1950), though the species has not become a real pest until quite recently (after year 2000). In particular, after 2000 the species has been recognised as a cereal pest in central Spain. In any case, I think it illustrates an interesting case of a migratory species that has expanded its distribution range, probably in response to climate warming, with some serious consequences in agriculture.
- l. 558. “Neither windspeed nor the interaction between windspeed and the degree of headwind…”. This is the first time when you mention that your model included interactions (cf. “Analysis of meteorological factors”). Also, in Table S6, no information is given regarding the interaction term nor wind speed. Even if they are not significant predictors, you need to include them in the table (and any other interaction that you may have tested, of course).
- l. 564-566. Move to Discussion.
- l. 937-948. Although I asked for discussion on this result, the sentences are a bit confusing. For instance, the Czech results that are reported do refer to autumn migration, as well? You say that your results are more similar to what was found in the Alps, but what could be the reason for a difference in sex-ratios between different regions during autumn migration? If you simply describe these differences but not propose the reason behind, this paragraph is not very informative.
- l. 1020-1021. “Insects were recorded almost exclusively during headwinds (Figure 4A).” But in the figure there are many cases with important number of insects recorded with tailwind!