Interactive comment on “ Climate changes since the mid-Holocene in the Middle Atlas , Morocco ”

A detailed review of the available literature must be done. This will strongly improve the present manuscript which doesn’t present the most relevant or recent studies from the western Mediterranean, Morocco and Middle Atlas: Reille 1976; Barker et al 1994, Lamb et al 1989, 1991, 1994, 1995, 1999; Cheddadi et al. 1998, 2009; CombourieuNebout et al., 2009 ; Detriche et al 2008, 2009, 2013; Rhoujjati et al. 2010; 2012; Amami et al. 2013; Fletcher et al., 2008 and 2013; Zapata et al., 2013; Reddad et al, 2013; Giraudi et al. 2013, Wassenburg et al., 2014; El Bait et al. 2014; Muller

c-Presentation of the calibration results in Table 1 does not follow recommendations given in the CALIB manual (in particular indication of 2 sigma cal age ranges and relative area under distribution).
== Table 1 now follows the Calib recommendations and each column is properly titled.1, what are SD1 and SD2 == These were the standard deviations.We have corrected this in table 1. e-and does "Calibrated age" stand for median probability?== It corresponds to the median probability provided by Calib.This has been corrected in Table 1 3) Methods: C3004

d-In Table
a-References for all methods used should be given.
== References are now provided in this version of the ms.
== This sentence is now replaced by "delta 13C isotopic ratio".
== This has been corrected.
-l. 23: Justify why Cyperaceae pollen from this site is included in aquatic plants.
== Cyperaceae pollen percentages are included in the aquatic plants because they originate mainly from the genus Cyperus which often grows around the lakes.
== We agree.This sentence has been re-written.
-l. 14-15: the method used for delta 13C measurements should be described.
== We have provided a full description of the method.
== The database of georefenced plant species has been set up by R. Cheddadi from the following two plant distribution atlases: Hulten and Fries (1986) and Flora Europaea (Jalas et al. 1972, 1973, 1976, 1979, 1980, 1983, 1986, 1989, 1991, 1994).These references are now cited in the manuscript.The database is a local one and its coverage corresponds to the two cited atlases.

4) Pollen data:
*The description of the pollen diagram is very short and uncomplete.In contrast to the author's statement, other taxa than Pinus and Cedrus, such as Poaceae and Quercus, present strong variations which are not described or interpreted.The authors do not give a clear description of the pollen diagram and do not present a comprehensive history of the vegetation changes at the site.Only the replacement of pine by cedar at 3750 cal yr BP is pointed out while for instance, two important forest opening episodes are clearly recorded, the first one preceding the cedar expansion and the second one between 2500 and 2500 cal yr BP marked by a drastic cedar reduction.
== Thank you for this remark.In order to make an efficient and useful presentation of our data, we now provide a detailed description in a new table.This makes the main features of the pollen diagram easily accessible to the reader without overwhelming the manuscript with unnecessary details. *In the discussion, the vegetation changes recorded in the Hachlaf sequence are not compared to the other records from the Middle Atlas (only cited in the introduction) and the Alboran Sea.
== Thank you for this comment.We have added a new section that compares the aridity trend after 6ka in both the Mediterranean area and within other records outside the study area as well.This, indeed, provides a wider view of our work.

5) Pollen-based climatic reconstructions:
* The reconstructed climate changes should be discussed in the context of the vegetation variations as recorded by the pollen diagram.Is the increase in precipitation seasonality in the second half of the record supported by the vegetation changes?
== Thank you for this point.The precipitation seasonality increase coincides with substantial changes in the ecosystems around the studied site.The main observed changes are the increase of the taxonomic diversity along with an expansion of the cedar populations and the retreat of the pine forest.Mediterranean areas are known for their high species diversity and for their sclerophyllous tree species that may withstand pronounced annual rainfall seasonality.Since both the annual amount of precipitation and the winter temperature have not changed substantially, we believe that the observed vegetation changes in Hachlaf may be explained by increase of the wintersummer contrast in precipitation.
-Is Quercus taxon including evergreen or deciduous oak pollen, or both?
== Yes it does because the distinction between the two types was rather random during the pollen analyses.We decided not to use dubious assignments of the pollen identifications.
-And which pine and oak species are taken into account in the database used for the climatic reconstructions?
== For the climate reconstruction we have assigned Pinus pollen grains to P. halepensis and those of Quercus to Q. coccifera.
-Are cedar ecological requirements supporting higher rainfall seasonality?C3007 == A study of drought thresholds influencing the growth and photosynthesis was performed on different cedar stands and species (C.atlantica, C. libani, C. brevifolia & C. deodora) of different origins (Aussenac & Finkelstein, 1983).This study shows that among many conifers, cedar trees may keep a sustained photosynthesis activity even when drought is very high.Thus, strong precipitation contrast between summer and winter may not affect the cedar's overall growth as long as the total amount of rainfall is sufficient (higher than 600 mm/year) and the winter temperature is low enough for the bud burst (Larcher, 2000).-Which pine species are developing in the Middle Atlas and to which climatic characteristics are they associated with?
== Currently two pine species are developing in the Middle Atlas; Pinus pinaster and Pinus halepensis up to 2600m asl.Although these two species are rarely associated (Benabid, 2000).
[Figure 1] This plot, based on plant species distributions and their related climate variables archived in our database, shows the climatic niche (annual precipitation versus January temperature) of the two species.The median annual precipitation for P. halepensis and P. pinaster is 927 mm/yr and 991 mm/yr, respectively.The median January temperature for these two species is 4.2 • C and 4.5 • C, respectively.There is a clear overlap of their climatic niches.Therefore, assigning P. halepensis or P. pinaster to Pinus pollen grains will not affect the climatic reconstruction of neither their annual precipitation nor the January temperature.
-p. 4105, l.25: "water persisted in the lake during the summer season which allowed the presence of aquatic plants (which flower during late spring and summer)": Could the authors relate this assertion to the pollen diagram?
== The occurrence of aquatic plants (>20%)in the pollen record (figure below) requires a sustained presence of water.All the identified taxa in the fossil record (Potamogetonaceae, Cyperaceae, Typha angustifolia,Typha latifolia, Ranunculaceae, Myriophyllum, Juncaceae) flower during the summer season in a lake (or marshy) environment.
[Figure 2] -p.4016: the authors state "After 3750 cal BP, Atlas cedar spread noticeably around the site while the pine forest strongly regressed.During this ecosystem transition we do not observe any major change in the reconstructed amount of annual rainfall or in winter temperature."However, we can see in Fig. 3 that Pann become noticeably lower and Tjan more regularly higher when cedar percentages increase.
== Actually, the overall change that we observe in the amount of annual precipitation over the last 6000 years is about 50 mm and the amplitude of change in January temperature is about 2 • C.These changes may not be considered as high.However, as the precipitation index shows it, if the 50 mm/yr represent an additional difference between summer and winter seasons then that may affect the ecosystems because it increases the seasonal contrast which is already quite pronounced (3 to 4 times more precipitation in winter than in summer, please see figure 4).Thus, both summer and winter amount of precipitation decrease (since Pann decreases) but the decrease affects more the summer than winter.Such pronounced winter-summer contrast may explain the regression of pines at the elevation of our site (1700m asl) and the expansion of Atlas cedars.
*The obtained climatic reconstructions are not discussed in comparison with previous pollen-based climatic reconstructions available in this region (Cheddadi et al., 2009;Combourieu Nebout et al., 2009).This would have been interesting to note that recon-C3009 structions based on the MAT from Cheddadi et al. ( 2009) are not consistent to those presented in this manuscript and try to explain this contrasting pattern.Is it related to the method or to an actual different climate pattern?
== Thank for this remark and we will make such necessary comparison in the manuscript.
As a matter of fact, the climatic reconstruction presented in Cheddadi et al. ( 2009) are based on the pollen records from 3 sites.Two of them are from the Middle Atlas in Morocco (lake Ifrah and lake Tigalmamine) and the third site is from Algeria (Chataigneraie).The record from lake Ifrah does not cover the last 5000 years which prevents from a direct comparison with our data.The climatic data from lake Tigalmamine have already been published by Cheddadi et al. (1998).These data (with an age model based on uncalibrated 14C dates) which cover the Holocene do not show any major conflict with the data of this manuscript (based on calibrated 14C dates).Indeed, they both show a rather decreasing trend of the annual amount of precipitation (figure 4A in Cheddadi et al., 1998) after 6ka.The main "discrepancy" concerns Tjan (figure 4C in Cheddadi et al., 1998).However, if we focus only on the period of interest (the last 6ka) the reconstructed amplitude of change from both pollen records (lakes Hachlaf and Tigalmamine) is about 2 • C and in both cases the reconstructed absolute Tjan is between 3 and 5 • C.
-The authors only state p. 4104: "Besides the low range, both Pann and Tjan show consistent trends.Pann decreases progressively since 6000 cal BP which is in line with the aridity trend that has been observed in other records from the Mediterranean borderlands (Risacher and Fritz, 1992;Brooks, 2006;Hastenrath, 1991;Anderson and Leng, 2004;Umbanhowar et al., 2006)."None of these works deals with records of the Mediterranean region but to Bolivia, tropics, Greenland and USA records.
*The reconstructed increasing trend in precipitation seasonality is also opposite to conclusions of other papers such as Fletcher et al. ( 2008), dealing with Western Mediterranean pollen records.It is also counterintuitive regarding the orbital forcing which favors a decrease in seasonality along the Holocene.Can hypotheses be proposed for explaining that contrast?
== If we may, the reference proposed above, is probably Fletcher and Sanchez-Goni (Quaternary Research, 2008) rather than Fletcher et al. ( 2008) as we didn't find any Fletcher et al. (2008).If this is correct then, Fletcher & Sanchez-Goni ( 2008) have not presented any climate reconstruction in their work and ultimately haven't stated any seasonality trend for the mid-Holocene.In their conclusions they state the following: "Changes in the seasonality of insolation (precession) also had an important pervasive influence on vegetation development over the last 48,000 yr, influencing both the amplitude and composition of forest development in this Mediterranean region." While we do agree with this overall statement, we still are unable to compare our data for the last 6000 years with those of Fletcher & Sanchez-Goni (2008).
We must add that the relationship between precipitation trends and the orbital forcing is more complex than that of temperature.Reconstructed annual precipitation at 9ka, 6ka and 3ka show major spatial differences in Europe (Guiot et al. 1993) which makes their direct relation with the insolation, probably difficult.While some areas recorded more than 400 mm/yr, others received 400 mm/yr less than today (please see figure 8 in Guiot et al. 1993).The temperature trends seem to provide more coherent patterns C3011 with the long-term insolation changes (Davis et al., 2003).
Having made these statements, we would like to add that concerning our data one should analyse the total amount of annual rainfall (Pann) and the seasonality index (SI) at the same time.Pann shows a decreasing trend (of about 50 mm/yr over the studied period).SI shows that the winter amount is 3 times higher at 6ka and 6 to 7 times higher today than the summer amount.These results tend to indicate that the overall amount of annual rainfall decreases which makes the summer season even drier today than at 6ka.Davis, B. A. S., Brewer, S., Stevenson, A. C., Guiot, J., & Contributors, D. (2003) * Rapid climatic changes are presented in the introduction as a characteristic of the Holocene.Could the event at ∼ 5 ka be discussed in this context?
== Thank you for this remark.Although the thermal amplitude during the Holocene is lower than 4 • C in the temperate latitudes (including the Mediterranean), there are indeed short and abrupt changes between 2 and 4 • c which took places in a few decades (or even less).We will discuss these issues in the manuscript.
== Indeed, the first sentence was repeated twice.We will delete it.
-The last paragraph of the discussion has to be rewritten.It is repetitive and unclear.
== The last paragraph has been rewritten.
-p.4117: TN curve is not shown in Fig. 3 while it is described in the text and mentioned in the figure caption; == TOC and TN are highly correlated (0.99).The shape of the two curves is very similar, that is why we plotted only the TOC curve.We have added it to Fig. 3 (as below) and changed the legend accordingly.
[Figure 3] -what is the red line?
== The red line will be deleted as it concerned a previous interpretation which no longer appears in the manuscript -p.4105, l.4: "a SI".
-p.4118: in Figure 4, a curve showing the taxonomic diversity and a hierarchical clustering tree are displayed but not mentioned in the caption.In addition, the method and software used for implementing the hierarchical clustering should be described in the main text or indicated in the figure caption.
== Here's a new caption for Fig. 4: "Diagram showing the percentages of the main pollen taxa identified in the Hach-I core.Cyperaceae and Juncaceae were grouped within aquatic taxa.The dashed black curves show an exaggeration (×7) of the percentages of some taxa.On the right, pollen zones with their boundaries are set up using a constrained hierarchical clustering (R Development Core Team, 2013).The C3015 taxonomic diversity is computed using a rarefaction analysis." -What is the red line?
== Here it's the same; the red line will be deleted as it concerned a previous interpretation which no longer appears in the manuscript.
-Is the taxa Quercus corresponds to deciduous or evergreen oak, or both?== Quercus pollen grains include both deciduous and evergreen oak as their assignment may often be dubious and/or not reproducible by another pollen analyst.