We thank the reviewer for their review and comments on this pre-print. Our response is made in the context that the pre-print has been submitted to Geoscientific Instrumentation, Methods and Data Systems, where our understanding is that the journal aims and scope are on advances in concepts, design and description of instrumentation and data systems, and that the preprint, therefore, focuses on what we believe is an advance, using instrumentation situated in human-made and natural spaces in the vadose zone to quantify the timing of groundwater recharge events. 

As stated on lines 74-75 of the pre-print, the monitoring concept for NGROS is to (1) provide data on the timing of recharge at the event timescale and (2) determine the amount of precipitation needed to generate recharge at an event scale. We agree with the reviewer that this concept focuses on the temporal aspects of groundwater recharge. We agree that the focus is on the frequency domain, and the approach precludes the determination of recharge volume or amount of recharge, which would require further analyses that are outside the scope of this pre-print. We propose that we add some new text to the manuscript that addresses the fact that the NGROS network does not address recharge volume. We agree that the approaches of Mahmud et al (2016, 2018) could be used by future research to generate recharge volume estimations at cave sites by combining the data produced by NGROS and techniques such as in-cave lidar. The methods of Mahmud et al (2016, 2018) would not work at the mine sites, as the approaches required the counting of stalactites to quantify the number of water flow pathways into the cave.

The groundwater.unsw.edu.au website provides access to the NGROS data. It went live in mid-October and unfortunately was not available at the time of review. 

We agree that a summary table will be helpful given the large number of initial sites described in this pre-print. We will add this to Section 2.1, and at the same time check that everything is standardised between sub-sections. 

We appreciate the request to provide results from all 14 initial NGROS sites. We believe that level of data analysis and interpretation would be outside the scope of the Geoscientific Instrumentation, Methods and Data Systems (see https://www.geoscientific-instrumentation-methods-and-data-systems.net/about/aims_and_scope.html). We would like to note that the pre-print focuses on the initial fourteen sites and is intended to present the monitoring concept and research platform that is NGROS, including three examples illustrating the type of data collected.  We expect to expand the number of sites over time as more partners join to collaborate with the project. 

We checked whether the text relating to Figures 4, 5 and 6 was complete, and at this time would not propose to add any further text about these three examples.

Finally, we would like to take this opportunity to correct the description for Capricorn Caves (section 2.1.10) which incorrectly duplicates text from section 2.1.12. We apologise for that mistake. The correct text is: 

“This natural cave site is in an isolated c. 50 m high limestone hill with 390M yr old Devonian reef limestone geology. It has a native dry rainforest vegetation cover. The local climate has a mean annual temperature of 23.4 °C, mean annual rainfall of 815 mm, and modelled pan evaporation of 2525 mm. The climate is on the boundary of Cfa (temperate, no dry season, hot summer) and tropical savanna climate classifications (Aw). Loggers have been placed in the lowest cave passages, closest to the groundwater. The site is privately owned and managed.”

References

Mahmud, K., Mariethoz, G., Treble, P.C. and Baker, A., 2016. Terrestrial LiDAR survey and morphological analysis to identify infiltration properties in the Tamala Limestone, Western Australia. IEEE Journal of Selected Topics in Applied Earth Observation and Remote Sensing, 8, 4871-4881.

Mahmud, K., Mariethoz, G., Baker, A. and Treble, P.C., 2018. Hydrological characterization of cave drip waters in a porous limestone: Golgotha Cave, Western Australia. Hydrology and Earth System Sciences, 22, 977-988

Thank you very much for taking the time to make a community comment. 

We confirm, as in our response to Reviewer Comment 1, that we will clarify that this methodology focuses on the temporal occurrence of recharge, and the approach precludes the direct determination of recharge volume or amount of recharge, which would require further analyses. This data generated is predominantly of the timing of recharge at the event timescale and the determination of rainfall recharge thresholds. This can then be combined with other data e.g. of recharge volumes using techniques such as the soil moisture water balance, water table fluctuation, chloride mass balance or isotopic techniques, or comparison with other vadose zone monitoring approaches such as co-location with Critical Zone Observatories. 

We will also expand the emphasis in Section 4 to include the novelty of this approach. Namely (1) to provide data on the timing of recharge at the event timescale at sites where the source of the recharge water is known e.g. direct and focused recharge from precipitation and (2) determine the amount of precipitation needed to generate recharge at an event scale. We will include our statements provided above, that the network would complement and add value to other methods of recharge estimation and vadose zone monitoring, and that this would be suitable for applications globally (a comment made by RC2).

We thank the reviewer for their comments. We confirm that we will add to the Figure 1 caption to clarify that the dark brown layer represents an underground void (e.g. tunnel, cave or mine).

The drip sensors are located where water movement into the tunnel, mine or cave is identified (L82) and we agree that this is likely to relate to the overlying fracture network. Our existing text on L79 that states there is a focus on fractured-rock lithologies is misleading, and we thank the reviewer for alerting us to the fact. We propose that a better description to clarify that our focus is to monitor where there is observed water ingress in varied lithologies. This is likely due to the presence of fractures; however the precise nature of the overlying fracture network is unknown. Inferences might be possible from the drip hydrology time series obtained from individual loggers and through additional analyses such as lidar surveys (see related reviewer comment at https://doi.org/10.5194/egusphere-2023-2053-RC1). We propose to add some text at L79 to clarify this.

The number of sensors at each site necessary to identify recharge events is guided by our previous experience monitoring in caves and karst where we have successfully interpreted recharge data using nine loggers in one cave system (Baker et al 2021) and 14 loggers distributed across six caves in a regional analysis (Baker et al 2020).  We will add further text to explain this at L84.

We agree that the Figure 3 caption is very long, and we will modify Figure 3 to shorten the caption as suggested.

Jones et al (2009) provide analysis of the quality of the gridded precipitation data available for Australia and report a mean absolute error (MAE) for daily precipitation of 1.2 mm for Australia as a whole, with highest uncertainties in the tropical north where rainfall extremes are highest and the monitoring network sparsest. This is outside of the NGROS monitoring region. The gridded precipitation product has the lowest error in the southern half of Australia, and we would anticipate the MAE to be lower for our sites. Where local stations are available, the gridded and observational data can be compared, noting that the local station is likely to be included in the gridding process. 

Both reviewers have asked for an additional Summary Table for the sites, and we will include this in the revised manuscript.

We will clarify that the loggers used count drips over a user specified period that could range from seconds to days. We have chosen to count the number of drips over one hour, as this provides sufficient temporal resolution to identify recharge events and allows logger deployment for several years before the logger memory is full. Data output is in the form of a .csv or .xls files and data processing only involves data screening in the case that the logger is moved. Detailed analysis of logger precision is provided in Collister and Mattey (2009). We will add more detail about all of these aspects in section 2.2.  

We agree that the Australian network could be replicated elsewhere in the world and that it could benefit any country where a greater understanding of the timing of groundwater recharge is of importance. We will add this to the summary and conclusions.