We really appreciate the words of R1. We have done our best to answer R1 comments. Below, we include R1 comments (in italics) and our responses:

• I believe that all papers based on the interpretation of paleoseismic excavations should provide graphical information in the form of photographs of the main paleoseismic features used as evidence for the interpretation of earthquake history. Or at least some examples of them. It is a pity that the paper does not show photographs of the trench record or of some paleoseismic features, which could be included either in the manuscript or as supporting material. I know the quality of the research team behind this work and I am confident in the quality of the interpretations shown in the log of the trenches. However, I believe that the inclusion of field photographs would enrich the manuscript.

We agree with R1. Therefore, if the manuscript is accepted for publication in Egusphere we are adding as supplementary material high resolution images of the logs photomosaics. In these images all the palaeoseismic features will be shown.

• It would be necessary a description/interpretation of the sedimentary environment that explains the very different sedimentary records in both fault walls. I find it somewhat surprising that for such a small time interval (only a few hundred years between the ages of units A and F) there is such a different record on either side of the fault. I think this deserves some description and interpretation.

R1 emphasizes that, based on figure 7, there appear to be distinct sedimentary records in both fault walls of the RUB trench. The overall sedimentary arrangement in the trench is predominantly homogeneous, comprising silt and fine sand beds with intercalated levels of fine conglomerates. That is, there is not such a different sedimentary record in both sides of the fault. We think that the issue lies in the potentially misleading depiction of figure 7.

In figure 7, three common units (A, B, and C) are illustrated, cropping out on both sides of the fault. From that point to the top, we represent distinct units in the hanging wall and the footwall. Units A, B, and C share the same name and color on both sides of the fault because they are observable across the fault, with their bottoms and tops traceable through the fault zone. The same cannot be said for the other units. We think that this is because the fault zone developed in an area with lateral facies changes. For instance, conglomerates of unit X in the footwall laterally grade to silts, corresponding to unit D in the hangingwall (wich is made up of silts with interfingered conglomerates). The lateral transition of unit Y is not so straightforward. Probably the lower silt level of unit Y corresponds to unit E and part of unit F. We hypothesize that the conglomerate level within unit Y laterally disappeared around the present fault zone, and it was the source for colluvial wedge CW3. Therefore, under this assumption, the upper silt level of unit Y could correspond to the rest of unit F. Finally, unit Z would correspond to unit G. Due to these uncertainties, we decided to assign different names and colors to these units in the hangingwall and in the footwall. But we see now that this can lead to confusion. In response to R1's suggestion, we plan to include a paragraph in the manuscript explaining the discussed lateral transitions and the inherent uncertainties. Additionally, to prevent misunderstandings, we will modify figure 7 and revise unit labeling. It is important to emphasize that these uncertainties do not compromise our palaeoseismological interpretation of the trench.

• It is not clear to me what the objective of the seismic elastic deformation scenario carried out in Chapter 4.3.2 is. Moreover, the results of this scenario are not used in any of the discussions and conclusions of the paper. On the other hand, the complex geometry of the fault would require an in-depth discussion of the rake used in the model. A single rake for sections with such different orientations (almost 45º in some sectors) can lead to results that are difficult to interpret. I think it would have been more interesting to calculate the PGA associated with the ground shaking that would affect the archaeological sites described, taking into account the proposed maximum earthquake scenario.

The objective of the seismic elastic scenario in figure 11 is to present a glimpse of the potential use of our data in terms of seismic hazard assessment. We totally agree with R1 that would be really interesting to go deeper in further hazard assessment, for instance in the PGA derived from our data. However, we think that this is beyond the scope of our work. Actually, such an assessment could be the focus of future papers, as we mention in our conclusions chapter. Despite that, we followed R1 suggestion, we will add a further discussion about the implications of this elastic scenario in terms of the distribution of archaeological sites related to the deformation derived from our elastic model, as these sites are located close to areas of maximum expected deformation. 

As regards the complex geometry of the fault, we agree with R1 that needs to be better explained. Our model actually address this complex geometry and kinematics, as it accounts for the different orientation and kinematics of different faults sections. However, this was not clearly indicated in the original version of our manuscript. Moreover, we only presented in figure 11 the results of the horizontal displacement, adding probably more confusion. Therefore, we will extend the paragraph related to this model to better explain the geometric and kinematic complexities. Furthermore, we will add to figure 11 the results that we obtained of the along-dip displacement

• The rest of the comments are mainly suggestions to improve the understanding of the text and to correct some typing errors. All of them can be found in the attached PDF.

We will address all the comments indicated by R1 in the pdf, and we will correct all the errors.

We would like to thank R2 for his constructive comments, that will improve our work if published.

We include below R2's comments (in italics) and our response:

• The first criticism is that in the manuscript (especially in sections 2 and 3) there is a strong deficiency and/or inaccuracy in the presentation of the basic geological-structural elements, both at the regional and mesoscale scales, which are fundamental for the reader to understand the geological structure and the seismotectonic context in which the study was carried out. I am referring in particular to the presentation of the geological data in the figures, which should summarize and clarify what is expressed in the text. However, many of the figures showing the geological aspects (e.g. figs 2,3, 4, 6) are incomplete and/or imprecise, as explained in detail in the "specific comments" section relating to the figures.

We thank R2 for such a detailed analysis of the figures in our manuscript. We agree with him, and we will incorporate to the figures all his suggestions, that will improve them significantly.

•  I don't think it's just my convinced opinion that studies of paleoseismological trenches must always show, in addition to trench logs and interpretative schemes, orthophotos of the trench walls, at least for the sectors of greatest interest (e.g. fault zones). This does not mean lack of confidence in the interpretation work done, but has the fundamental meaning of sharing geological data with other researchers. This aspect is even more true for paleoseismological trenches where the studied outcrops (from tens to hundreds of cubic meters of geological record) will be lost to future researchers.

We will follow R2 (and R1) suggestion, and we will add as supplementary material high resolution images of the logs ortophotomosaics. In these images all the palaeoseismic features will be clearly displayed.

• I believe the Authors must clarify why in a structural context characterized by an oblique kinematic fault with a prevalent strike-slip component (based on the geological and geodetic studies as reported in the manuscript by the authors themselves), trenches  have not also been dug parallel to the strike of the fault. As the Authors know, the orientation of the trench is dictated by the inferred sense of fault displacement. For oblique and strike-slip faults, therefore, it is common practice to perform parallel trenching in order to define the “real displacement” by the offset of piercing points, while "fault perpendicular trenches are often used to locate an define the width of strike-slip slip fault zones" (McCalpin, 1992). I believe that carrying out statistical analyzes on slip-rates evaluated mainly on "dip-separation" data in trenches perpendicular to the fault may offers some problems, also due to the very small extent of the displacements that are measured, for the majority of cases in even very narrow Holocene time intervals.

As R2 indicates fault-parallel trenches in strike-slip contexts are the best way to determine the fault displacement. But, in these contexts, fault-parallel trenches rarely allow to recognize the number of palaeoearthquakes recorded. This history is more likely to be reconstructed in across-fault trenches. So, ideally, in order to have a complete characterization of the seismological parameters (offset and number of events), both fault-parallel and across-fault trenches should be dug. However, we could not afford this analysis, both for time and economic reasons. Therefore, we prioritized the reconstruction of the palaeoseismological history of the fault, and we try to get an approximation of offsets using the observed vertical displacements and the rake of the fault slickensides.

In any case, we now realize, following R2's comment, that we need to provide a more thorough explanation of our approach to the calculation of offsets and subsequent slip rates. We will incorporate the necessary sentences to clarify the significance of our slip rates, making it explicit that these values should be approached with caution, as they are not direct measurements, but derived from the vertical offsets.

R2's SPECIFIC COMMENTS

Figures:

• Fig. 1 - A clearer kinematic picture should be provided in this figure, indicating (with dashes, arrows...) at least the faults with normal kinematics and those with prevalent strike-slip kinematics.
  • Indicate the Betic Cordillera with an acronym or other symbol. The Padul Fault and the Ventas de Zafarraya Fault referred to in the text in Fig. 1, are missing in Fig. 1.

We will follow all the R2 suggestions, and we will correct the errors in this figure

• Fig. 2 – Fig.2 is indicated in the caption as a Geological map. I find the legend in which only the ages are reported to be at least incomplete; in a geological map before the age the geological unit should be defined by distinguishing at least what type of deposits they are (fluvial, alluvial, lacustrine... and where possible lithofacies and the range of thickness).
  • Insert kinematic indicators for the GF
  • Furthermore, the symbol in the legend for "towns" is not the one shown in the figure.

We will follow all the R2 suggestions and we will correct the errors in this figure

• Does glacis surface mean a glacis of erosion or accumulation of deposits?

A glacis is a geomorphic surface that can be both erosive or the result of sediment accumulation, depending on the considered area. This is the case of the Guadix-Baza Basin glacis. This surface is erosive near the porders of the basin and depositional in the central part.

• Fig. 3 – 2(a) same as fig. 2 (fix the legend: e.g. what do black point -dashed line and dashed black line-double point indicate? Distinguish geomorphological from geological features.... In Fig 3(c) the fault or fault zone crossed by trace I-I' ( and from the trenches) is not represented in the geological cross section. In the latter it is clear that the geomorphological feature "Surface 1" is associated with the sedimentary body colored in green which is not indicated either in the map of Fig. 3a or in the caption.

We will follow all the R2 suggestions, and we will correct the errors in this figure

• Fig. 4 – The trenches in this figure are indicated as trench A and trench B, while in the text and in Fig. 2(b) they are referred to as trench 1, trench 2 and trench 3. Uniform and perhaps put the same acronym also along the cross section of fig. 3(c).
  •  the acronyms of the samples next to the logs of the geological units are difficult to read.
  •  Bedrock instead of Bed Rock

We will follow all the R2 suggestions, and we will correct the errors in this figure

• In the wall of Las Acacias Gully in Fig. 3(b) it is clearly shown how the GF has lowered the fluvial terrace in its southern block (as also indicated by the red arrow), while in Fig. 4(a) it seems that the block lowered is the one towards the north. Clarify these kinematic aspects of the fault well.

Figure 3 and 4 offer a different scale of observation. The photograph in figure 3 shows the general kinematics of the fault strand cropping out in this area. Figure 4 shows the trench-scale detail of the fault zone. It is a common feature of strike-slip faults that the vertical displacement polarity changes according to the strike of the fault branches. That is the case of our study area. Despite the general kinematicks of the fault strand lowers the southern block, the small branches observed in the trench wall produced small displacements lowering the north block. Obviusly, our manuscript was not clear enough in that case, so, we will add to the text a sentence clarifying these kinematics aspects to our manuscript, if it is finally accepted for publication.

• Fig. 6 – For Fig. 6(a) and fig. 6(c) I suggest the same recommendations as in fig. 3(a) and 3(c).
  • While I would suggest changing the name of the profile trace which for Fig. 3 and Fig. 6 is always I-I'.

We will follow all the R2 suggestions, and we will correct the errors in this figure

• L76 – Guadix-Basa Basin: what kind of basin is it? Extensional basin? Pull apart basin?

The Guadix-Baza Basin is an intramontane basin with a complex tectonic evolution. Since the Upper Miocene it is an extensional basin. We will add sentence in the geological setting section to clarify this aspect.

• L96 – Specify from which type of studies “more than 2000 m of sediments accumulation” were estimated, geological, geophysical…?

It was estimated from geophysical studies (gravity and seismic). We will clarify this aspect in our manuscript.

• L110 – Transfer to what? It is not clear from the figures and the text on which structure(s) the GF transfers the deformation towards the east.

R2 is right, the term “transfer” is not correct. We meant that the GF presents a kinematic coherence with the Baza Fault. We will correct this error in the next version of our manuscript.

• L141 – Where is the sedimentary record associated with tectonic subsidence? Explain why under 5 m of fluvial deposits the lacustrine deposit of approximately 2.5 Ma is found directly at both the roof and the fault bed; this places constraints on tectonic subsidence associated with fault activity and its relationship with the extent of pre-Holocene erosion.

This is because our trench was excavated across a secondary strand within the fault zone, as stated in our manuscript (line 140). That is, the fault strand traversed by the trenches is not the main slip plane. Therefore, our palaeseismological site is located in the upthrown of the fault. Moreover, as we mention in our manuscript (lines 100-105), since Middle Pleistocene the Guadix-Baza Basin is dominated by erosion, and present deposition is restricted to the basin margins (alluvial fans and piedmont deposits) and the modern drainage system (fluvial terraces and valley-bottom deposits). So, the Holocene alluvial deposits unconformably lie over sediments of different ages in different parts of the basin. In the case of the alluvial terrace that we excavated, it was deposited over lower Pleistocene lacustrine sediments. We found the same sediments at both sides of the fault because the thickness of this Pleistocene unit is much higher than the vertical displacement of the fault. In any case, the comment of R2 indicates that this aspect is not clear enough in our manuscript. So, we will better explain this feature in the new version of our work.

• L187 – On what basis is it established that F2 is the fault that “presents the largest offset”?

It is because it produces the larger offset of the bedding observed in the bedrock units. We will clarify this aspect in our text.

• L392 – In this way, however, do you exclude the hypothesis that the source could have been different?
• L457-459 – It should be remembered that the liquefaction may be due to events on other nearby sources... (perhaps a check should be made with paleoseismological data on other nearby sources).

These two comments are related to the seismogenic source of the seismites observed in the RUB trench. As we stated in our manuscript (line 328-332), we know that seismites can occur far from the seismogenic source of the causative earthquake. We agree with R2 that a comparison with palaeosismological data on nearby sources will be of interest. Following his suggestion, we have compared the seismites-related events deduced from the RUB trench with the available palaeoseismological data of the nearby Baza Fault (Castro et al., 2018). Only one of the reported palaeoearthquakes of the Baza Fault overlaps in time with the events that we recognized in the RUB trench. The age of this Baza Fault event is poorly constrained (8485-785 BC), so it could correspond to any of the seismites-related events in the GF. Therefore, we cannot completely rule out that the Baza Fault was responsible for these liquefaction structures. However, as we also stated in our manuscript (line 332), because of the vicinity of the liquefaction features to the GF, we assume this latter as the seismogenic source.

In anycase, as we already indicated, we think that including the comparison between the RUB trench and the Baza Fault earthquake chronology will improve the manuscript. So, we will add this discussion to the future version of our work.

We will address all the detailed comments indicated by R2, and we will correct all the errors.