The authors thank the reviewers for the detailed comments, which have greatly improved the paper.
Authors’ response to Dr Lars Gunnarsen
 The paper analyses data from continuous monitoring of air temperature, relative humidity, and velocity behind books in bookshelves and in the space in front of the shelves at three locations.
 The driving forces behind air movements behind the books are seen as the result of differences in air temperatures influencing air density and the activities in the buildings housing the historic bookshelves including movements by visitors, renovation works and opening of doors.
 The paper presents a simple model assumed to predict air velocities that is demonstrated to have some prediction capability but lacks precision and inclusion of important parameters such as surface temperatures of walls behind bookshelves and more precise information about openings between space and hidden environments.
Response: We agree with the reviewer that the model proposed is in fact very simple, and indeed should not be seen as an attempt to describe mathematically the physics of the system. As the reviewer points out, a physically-based model should include surface temperatures and the dimensions of the channel. It should possibly be based on natural convection correlations between flat planes, which have been developed for years for engineering applications. However, in this case, the model is used only as an attempt to extract the maximum possible value from the available data. These limitations were mentioned in the previous version of the paper. In the revised version, more detail has been added, including what additional parameters would be needed for a more accurate model.
 The analysis would benefit from more precise modelling of boundary layers and buoyancy taken from more comprehensive fluid dynamic and heat transfer theory. The analysis would also benefit from more precise modeling of wall surface temperatures and information about outdoor temperatures during measurements.
Response: The main purpose of our paper is to report experimental data that displays interesting trends. The use of the model is merely a means to explore this data further, and give an indication that there are probably subjacent physical mechanisms that require further exploration. The reviewer is right in saying that more advanced modelling will be helpful to answer these questions, but we would like to argue that this falls outside the remit of our publication. In fact, we have begun CFD simulations of these microenvironments, that we plan to include in a future publication.
In the revised paper, we have added further clarification of the remit of this work, explaining that it focuses exclusively in extracting interesting conclusions from a unique, if imperfect, dataset. This is visible in changes in the abstract, which now includes a mention to further research, as well as this new paragraph in the introduction:
“This study aimed to understand which physical mechanisms are responsible for the air movement within bookshelves, using data collected in-situ. This study focused on historic bookshelves of three UK National Trust properties. The selection of properties was made based on different scenarios of mould development incidence; a relatively constant presence of mould (Charlecote Park), some previous mould development (Blickling Hall), no records of mould development (Ham House). This analysis resulted in an interesting dataset of monitored data that offers an insight in the behaviour of this unique microenvironment. The purpose of this paper is to present and analyse this dataset, as well as indicating which additional data future research should collect in order to reach a good understanding of the dynamic behaviour of these environments.”
 Mold will normally start growing in the range 70-75 %RH. Absolute humidity measured as water weight per weight of dry air will without sinks and sources be constant in air flows being heated and cooled. Analyses of water content based on temperature and relative humidity would also be relevant in the available data. Moisture content of the cold exterior walls could be at equilibrium with somewhat higher relative humidity than measured in the air and be at higher risk of mold growth.
My thinking is that increased air flows behind the books in bookshelves may not be sufficient to prevent mold growth on cold outer walls but may increase the temperature of the books and their margin to mold formation.
Paper would benefit from more precise modeling of air flows, assessment of wall temperatures and inclusion of absolute humidity in the analysis.
Response: We would like to thank the reviewer for this insightful comment. We have incorporated this reasoning in the paper, and we believe it frames very strongly the needs for future data collection and the limitations of the current data. A new paragraph in the paper states the following:
“The question remains whether this phenomenon is sufficient to make a difference on mould growth. To explore this question, the conditions of the walls and back of the books should be examined separately. In the absence of sources or sinks of water, the absolute humidity next to these two surfaces should be the same. Exterior walls will tend to be colder and therefore have higher relative humidifies. The back of books may be colder than the rest of the room, due to the external wall acting as a heat sink. In this scenario, the mild air movement detected in this research is far mor likely to increase the temperature of the books than to increase the temperature of the external wall. This hypothesis should be verified by monitoring surface temperatures. “
Authors’ response to Dr Fernando Sarce
 At the abstract, it has been stated that "the use of ventilation holes in small micro-environments has been proposed as a mechanism to improve the environmental conditions of moisture and temperature within bookshelves". However, this has not been discussed on the results nor on the conclusions.
Response: In response to this comment, we have made it more clear in the abstract that "the use of ventilation holes in small micro-environments has been proposed” by the National Trust and not by this study.
 The stack effect is a natural thermodynamic phenomena, but how this related to the improvement of the "moisture excess" that you would like to partially eliminate? This would be a valuable research statement for your research.
Response: We have revised the introduction to make this research statement more clear. In the revised paper, we clarify that there are two ways in which air movement behind a shelf can contribute to mitigating mould growth: by equilibrating the temperature of the microenvironment with that of the room (thus reducing the RH), or by removing moisture if the room has a lower absolute humidity than the microenvironment.
In this case, as demonstrated by the absolute humidities reported in section X, the absolute humidity behind the shelf and in the room are very similar. This indicates, firstly, that there are no sources of additional humidity behind the shelf and secondly, that any mitigation effect will be through thermal equilibration. This does not exclude the possibility, which should be explored in future research, that ventilation behind shelves reduces the risk of mould by driving moisture away from wet walls in drier rooms.
 Please refer as indoor/ outdoor environmental conditions. Indoor air relative humidity or indoor air pressure, etc.
Response: In response to this comment, we have made changes in the manuscript to clarify when environmental conditions refer to indoor or outdoor conditions.
 Are you interested on the air motion to then determine an effective way to estimate a mass air flow rate (to dilute the presence of excessive moisture) depending on the air psychrometric conditions? In this case, is your contribution aimed to complement the current understanding on temperature and humidity modelling simulations?
Response: The main interest of this paper is to report an interesting dataset on a microenvironment for which no previous empirical information exists. We do not aim at making a new contribution to hygrothermal modelling, which would be outside of the scope of this work. Instead, a simple model is used to maximise the value we can extract from the available data, showing that there is a possibility that observed small differences in temperature drive measurable air velocities. The revised version of the paper clarifies that the use of the model is exploratory and that alternative modelling strategies should be followed for a more accurate prediction.
 Please explain why have you selected those heights for the anemometers locations. What is the rationale behind? It should be clear at the introduction of section 2.1.
Response: In response to this comment, we have made changes in section 2.2, first paragraph, to clarify the reason behind the height selected for the anemometers.
 Improve the quality of Figure 3. The data lines are difficult to see.
Response: In response to this comment, we have added labels (T Shelf, RH Shelf, T Room, RH Room) to plot lines in Figure 3 to improve legibility.
 Please refers as air pressure in your hypothesis stated in Section 3.2.
Response: “Pressure” has been revised by “air pressure” for clarity as suggested.
 Is it possible that air leakage on the window's frame could affect air speed together with some of the indoor/ outdoor openings in the buildings? If yes, can this be introduced in your methodology or for further research?
Response: This is a possibility because historic buildings are not always well sealed. We have added a paragraph suggesting future work to explore the impact of air leakage from windows on air movement present in bookshelves.
The use of ventilation holes in small micro-environments has been proposed by the National Trust as a mechanism to improve the environmental conditions of moisture and temperature within bookshelves. At one National Trust historic property, this mechanism has been used to encourage air movement behind books as a possible strategy to reduce the risk of mould growth. It is believed that including ventilation holes as a passive design solution to promote airflow within micro-environments could prevent decay from occurring in archives of historic buildings. This paper investigates the mechanisms that cause airflow behind bookshelves using field measurements in three National Trust historical libraries. The measurements indicate that small but measurable velocities, up to 4 cm/s, can be generated passively behind bookshelves. Air movement in such confined micro-environments is probably caused by a combination of natural convection, caused by temperature differences between the walls and the interior and the exterior of the bookshelf, and forced convection due to drafts in the surrounding environment. While in some cases one mechanism prevailed, both mechanisms may be present simultaneously in most cases. Further research is needed to clarify how surface temperature drives air motion behind shelves.