Introduction

Levantine rock art is a prehistoric cultural expression unique to the Mediterranean coast of the Iberian Peninsula, characterised by its dynamic pictorial scenes, distinctive style and narrative quality1,2. It stands as one of the earliest known forms of visual communication, offering valuable insight into the material and socio-cultural realities of early human communities.

Its outstanding universal value was recognised in 1998 with its inclusion in the UNESCO (United Nations Educational, Scientific and Cultural Organisation) World Heritage List under Criterion III, as part of the “Rock Art of the Mediterranean Basin on the Iberian Peninsula” (ARAMPI)3, acknowledging it as exceptional testimony to a disappeared cultural tradition.

Unlike Palaeolithic art, Levantine rock art is predominantly found in open-air shelters and is closely integrated with the surrounding landscape4. The Albarracín Cultural Park (PCA) in the Lower Aragon region (Teruel) (see Fig. 1) hosts numerous sites, with the Sierra de Albarracín standing out for its naturalistic depictions of large zoomorphic figures and the predominant use of white pigment4. This is exemplified by the Toros del Prado del Navazo shelter (see Fig. 1), or the Cocinilla del Obispo (see Fig. 2) and the Cabras Blancas shelters.

Fig. 1
figure 1

Site map showing Spain, Aragón, Teruel and Albarracín in the upper image; in the lower image: The rock art shelter of Toros del Prado del Navazo.

Full size image
Fig. 2
figure 2

The rock art shelter of Cocinilla del Obispo.

Full size image

The deterioration of the PCA paintings is multifactorial, influenced by environmental and anthropogenic factors, as well as prior conservation and restoration interventions. Although initial studies began following the early discoveries5,6,7, the conservation-restoration studies did not achieve specialisation until the late 20th century8,9,10. Since then, direct conservation interventions gradually gained momentum11,12,13, ultimately leading to preventive conservation as the primary action strategy in subsequent years14.

Ensuring the preservation of this prehistoric heritage for future generations is one of the significant challenges in conservation and restoration. A primary concern in the preventive conservation of Levantine art is the development of an effective integrated management proposal15,16,17. In the absence of a standardised protocol, the proposal for PCA rock art should include at least two essential stages that assess the condition of the site from various perspectives:

  • Describe the shelter’s conservation state: establish an accurate diagnosis.

  • Identify current and potential threats through risk analysis.

At both stages, a comprehensive understanding of environmental factors is essential. Extreme climatic conditions, along with variations in temperature (T) and relative humidity (RH), can adversely affect the physico-mechanical properties of both the support and the paintings. These conditions can lead to cracking, blistering and, in severe cases, delamination of the paint film18,19. In addition, thermohygrometric variations favour the growth of biological agents within the coatings, which can invade the paintings and trigger chemical reactions that lead to corrosion phenomena, pH destabilisation and the formation of concretions, among other pathologies that compromise stability.

The primary objective of this work is to design and implement a real-time environmental monitoring system based on IoT technologies to track essential environmental parameters, including temperature (T), relative humidity (RH) and light (L), affecting Levantine rock art shelters. It also aims to reduce the frequency of site visits, which will help to decrease the associated greenhouse gas emissions (GHG). Furthermore, the initiative tries to align with the Sustainable Development Goals (SDG) by promoting responsible, innovative and sustainable practices in cultural heritage management.

This system is demonstrated at the Toros del Prado del Navazo rock shelter and aims to optimise the preventive conservation of the site by collecting and processing the data relevant to diagnosis, risks and restoration processes, facilitating informed decision-making to prevent rock art deterioration.

Methods

Archaeological site: rock art shelter of Toros del Prado del Navazo

The Toros del Prado del Navazo shelter is located in the Albarracín mountain range, approximately 4 km from the municipality of the same name, UTM coordinates (40.393158, −1.401465). The location is surrounded by an exceptional geomorphological landscape, with a predominance of reddish sandstone of the Buntsandstein facies18,20 from the Lower Triassic period21. Its mineralogical, structural and textural characteristics make it particularly vulnerable to weathering processes, including haloclastic weathering and silica dissolution20,22,23,24,25,26,27,28. This lithological type presents a macro-modelling that features rippled shapes and blocks separated by alleys affected by surface weathering that smooths edges and vertices29. The reddish landscape is complemented by a wooded area composed mainly of Pinus pinaster (see Fig. 3a) declared Protected Landscape of the Pinares del Rodeno (according to ref. 30 Decreto 91/1995, of 2 May, of the Diputación General de Aragón).

Fig. 3: The Pinares de Rodeno.
figure 3

Marked in red is the shelter of the Toros del Prado del Navazo (a) and Prado del Hostal (b).

Full size image

In terms of hydrology, the Navazo shelter has no nearby watercourses. Approximately 200 m away lies the Prado del Hostal (see Fig. 3b), a flat area that historically held water, giving the shelter its name. The availability of water favoured prehistoric human settlements, as evidenced by the abundance of cave paintings found in the Pinares de Rodeno22 (see Fig. 3a).

The Navazo painting support, comprised of a sedimentary rock, a sandstone known as “rodeno”, consists of quartz, potassium feldspars and micas23. It features sub-angular clasts, a rocky matrix, siliceous cementation and the presence of iron oxides, which gives the characteristic red colour of the support on which the large, whitish-coloured paintings are found. The site houses 18 representations, predominantly of bovids depicted with a certain naturalism, alongside a group of stylised anthropomorphs and some undefined zoomorphs.

Existing studies document a range of pathologies linked to naturally occurring alterations, including debris accumulation, crack formation, material displacement or disintegration, salt efflorescence, leaching and biological colonisation18,23,24,25. These alterations, identified by Bea and Angás26, result from the expansive–contractive behaviour of the clay minerals present in the rock, which generate internal stresses that lead to microfractures, as well as processes of disintegration or detachment.

At the Navazo shelter, significant fluctuations in relative humidity have been recorded, with levels exceeding 75% in winter and dropping below 20% in summer. Additionally, annual temperature variations range from –3 °C to 27 °C, with the daily thermal oscillation representing a low risk, remaining around +4 °C in winter and +2 °C in summer. Consequently, relative humidity has been identified as the primary risk factor (refer to Table 3).

The environmental conditions observed significantly contribute to the deterioration process in this clay-rich sandstone. The swelling phenomenon occurs when relative humidity surpasses 35%, with expansion rates reaching up to 75%—a trend consistently recorded during the study period23. This process is exacerbated by cycles of wetting and drying, which have led to visible signs of flaking and material loss in areas subjected to nocturnal condensation.

Conversely, daily hygrometric fluctuations exceeding 18%, whether in winter or summer, can induce microfractures due to the differential expansion of minerals such as quartz and clay. This issue is exacerbated by the crystallisation of soluble salts within the pores during drying phases, a process that accelerates the fracturing of the material18,23.

This situation raises a question: How can we approach preventive conservation in this context? Why not consider direct curative conservation?

Since 2020, numerous studies have investigated consolidation treatments for the Navazo shelter25. These works build upon earlier research that explored the influence of RH on nano-SiO2 binders27. They have established the connection between relative humidity and the growth of silica nanocrystals, assessing their distribution, morphology and effect on the rock’s porosity. These studies are included in the PhD thesis currently being developed by Claudia Serrano. This finding has redefined priorities, objectives and conservation needs, highlighting the importance of precise actions based on a rigorous diagnosis.

In order to make informed decisions regarding preventive measures aimed at mitigating the effects of degradation agents and threats, it is essential to understand the physical conditions of the shelter environment, both historical and current. This understanding can be achieved through the measurement of key parameters such as relative humidity, temperature and light, as suggested by Zalbidea et al.28. Analysing this data allows us to assess the site’s environmental factors and implement more effective conservation strategies.

Technological framework

The Government of Aragon, responsible for managing the rock art heritage of the PCA, has promoted interdisciplinary studies to address the conservation challenges faced by rock art. These studies involve monitoring using measuring stations installed at specific sites28,31,32,33,34. Monitoring locations include the Cultural Parks of the Vero River (Muriecho and Fuente del Trucho shelters), the Martín River (Borriquitos and Trepadores) and Albarracín (Cabras Blancas and Cerrada del Tío Jorge). Equipped with atmospheric sensors integrated into a CR10X data logger (Campbell Scientific, Inc.) and powered by lead-acid batteries and solar panels, these stations recorded various parameters such as ambient temperature, relative humidity, solar radiation, wind speed and direction, among other parameters.

In addition, between 2014 and 2015, monthly monitoring was carried out in the Navazo shelter as well as five other shelters in Lower Aragon: Plano del Pulido, Cuevetas de Poyuelo I, Los Chaparros, Cañada de Marco and Ceja de Piezarrodilla. Monthly measurements of relative humidity, rock surface and ambient temperature and daily measurements of relative humidity and temperature were collected using iButton (Maxim Integrated) data loggers32. Unfortunately, the complete datasets of the daily monitoring and the stations installed in Albarracín (Cabras Blancas and Cerrada del Tío Jorge) are unavailable, as they were never published and have since been lost. Although access to these sensors was granted in May 2023, it was not possible to retrieve the information due to depleted storage batteries. This incident highlights the need for long-term data preservation strategies.

To monitor the microclimatic conditions within the shelter of Navazo, a LOG32® TH data logger (Dostmann electronic®) was installed inside it. These devices have an accuracy of ±0,5 °C (−10 °C … +40 °C), ±3%RH (40…60%), ±3,5%rH (20…40% and 60…80%). The logger recorded temperature, relative humidity and dew point temperature data between March 3, 2020, to October 7, 2021. The data obtained from the logger was processed in Microsoft Excel® to create a climogram (see Fig. 4). In October 2021, the LOG32® TH monitoring device was replaced with LOG210® (Dostmann electronic®). During a periodic check on July 20, 2023, it was noted that the latter had ceased data collection on September 19, 2022, due to memory limitations. A subsequent check in July 2024 revealed that the device had stopped recording on July 20, 2023, because of battery issues. These failures are not isolated incidents, as another rock art complex located in the PCA, the Toros del Barranco de las Olivanas, also experienced interruptions in data collection due to datalogger failures, resulting in only seven months of recorded data12.

Fig. 4
figure 4

Climogram of Navazo shelter conditions between March 2020 and October 2023.

Full size image

To evaluate and compare preventive conservation guidelines with deterioration risk models, Table 1 presents the recommendations established by ASHRAE35, which were used in this study to analyse environmental fluctuation thresholds. The evaluation of the measurements was carried out with reference to the values corresponding to ASHRAE Class A2, since the paintings are located outdoors, where conditions are less stringent than those established for the stricter classes.

Table 1 Guidelines on temperature and relative humidity according to ASHRAE 2023
Full size table

In light of these challenges, the WiMOSA project36 aims to establish a novel system capable of monitoring parameters such as light levels, ultraviolet (UV) radiation, vibrations, temperature, relative humidity and dew point temperature. This system will be wireless and designed for real-time shelter monitoring through Internet of Things (IoT) technologies37, providing essential information for timely decision-making. Furthermore, the wireless nature of this system has the potential to significantly decrease the number of trips to Albarracín, thus helping to lower direct greenhouse gas emissions.

On October 16, 2022, a visit to Navazo was conducted to evaluate the potential for installing wireless equipment and to assess connectivity options. It was determined that mobile communication services, such as 3 G and 4 G, as well as specialised IoT-oriented services like those provided by Sigfox, Helium, or The Things Network, were unavailable.

The lack of wireless connectivity, often found in remote areas housing cultural heritage sites, requires the development of a tailored strategy for these environments. It was decided to implement LoRaWAN connectivity38 on-site by installing a solar-powered gateway nearby, which will be linked to a mobile operator for data transfer to the Internet. LoRaWAN is a low-power wide-area network (LPWAN) technology widely used in smart city scenarios, enabling sensors to penetrate, long-range transmissions, operating on battery power for several years without the need for maintenance.

Due to the complex orography (see Fig. 5) and the limited availability of suitable sites for installing the gateway, simulations were conducted to evaluate the feasibility of the proposal from a radio-electrical perspective. The simulation utilised Radio Mobile software39, which highlighted uncertainties regarding the viability of LoRaWAN technology at the potential installation points.

Fig. 5
figure 5

Orography of a part of the Cultural Park of Albarracín with highlighted shelters and the Prehistoric Rock Art Information Centre.

Full size image

Considering the inherent uncertainties and the need for caution when interpreting simulation results, a portable, battery-powered LoRaWAN system was developed to meet these specific requirements. This system is based on a Wisgate Edge Pro RAK7289 (RAK Wireless) gateway and a Barracuda OMB.868.B08F21 (Taoglas) antenna. To validate the performance of a potential fixed deployment, an ARF8123AA (Adeunis) Field Test Device (FTD) is utilised at the measurement points of interest. On May 25, 2023, the portable system was positioned at the designated evaluation points (see Fig. 6). The FTD was employed in the shelters to assess estimated performance of the wireless sensors (see Fig. 7). The findings indicate that it is feasible to install a gateway at the Prehistoric Rock Art Information Centre of Albarracín, which will also provide coverage for other nearby shelters, such as Cocinilla del Obispo and Tío Campano.

Fig. 6
figure 6

Portable LoRaWAN system installed at the Prehistoric Rock Art Information Centre of Albarracín on 25 May 2023.

Full size image
Fig. 7
figure 7

Adeunis ARF8123AA field test device is being utilised to evaluate radio signal performance in the Cocinilla del Obispo (a) and the Toros del Prado del Navazo shelters (b).

Full size image

After obtaining the necessary authorisations, the final gateway was successfully installed on March 12, 2024, in partnership with the Puyo Área Tecnológica company. The installation was simplified due to the Albarracín Town Council’s provision of Internet connection in the area of the Prehistoric Rock Art Interpretation Centre. This gateway offers an open and free LoRaWAN service via open The Things Network service for anyone wishing to deploy their standard LoRaWAN-compatible devices. To approximate the covered distances, we utilised the FTD in distant locations and plotted them in the TTN Mapper service, resulting in the beam map illustrated in Fig. 8.

Fig. 8
figure 8

Coverage test of the installed LoRaWAN gateway using the TTN Mapper service.

Full size image

Following the testing of the gateway infrastructure, three types of sensors were deployed in the Navazo: a CollectionCare Optimised sensor40, developed as part of the European CollectionCare project41, which measures T, RH, L and UV levels; a commercial sensor Milesight EM320-TH42; and a commercial sensor Dragino S31B-LB outdoor43, both of which measure T and RH. The CollectionCare Optimised sensors and the Dragino S31B-LB comply with the EN 15758:201144 and EN 16242:201445 standards, which ensure an uncertainty of 0.5 °C for temperature (T) and 2% for relative humidity (RH). In accordance with these standards, both sensors undergo annual calibration in line with ISO/IEC 17025 and the European EA-4/02 M:2013 guideline. The performance of the Milesight sensor has been verified by comparison with a calibrated Rotronic HP32 thermohygrometer (serial numbers 5220874 and 20631672), as it is not calibrated.

To ensure a seamless integration with the environment, the sensors installed inside the shelter were visually camouflaged using hydro-printing techniques (see Fig. 9b).

Fig. 9
figure 9

Location of the sensors in the shelter: Dragino (a), CollectionCare (b) and Milesight (c).

Full size image

Since their deployment, the sensor data has been transmitted and stored in the cloud in real-time. This data is readily accessible through a dashboard created in ThingsBoard46, which also calculates the dew point and sends alerts via email or SMS in case of any anomalies, such as sensor transmission failure, low battery levels, or excessive temperature. Figure 10 illustrates an example of the dashboard.

Fig. 10
figure 10

Screenshot of the dashboard showing the T, RH, L and UV values collected by the CollectionCare sensor during the first few months of 2025.

Full size image

Recognising the importance of the collected data for the scientific community, the WiMOSA project has established a Data Management Plan (DMP) that outlines the guidelines for the collection, storage, handling and accessibility of project data in alignment with open data47 principles and the FAIR48 approach. In this sense, the environmental data collected at the Toros del Prado del Navazo rock art shelter site is made available through Zenodo49 open data repository using the comma-separated value (CSV) format and the appropriate metadata. This approach ensures the long-term accessibility of the data for the scientific community.

CO2 impact of the data collection method

Prior to the installation of the wireless system, data collection required a minimum of four trips per year to retrieve information from the data loggers. To estimate the emissions associated with these activities, the parameters set by the Spanish government’s Ministry for Ecological Transition (MITECO) were utilised50. These emissions fall under Scope 1, as they represent direct emissions from sources such as combustion in boilers, furnaces and vehicles, which are owned or controlled by the entity.

Emissions are calculated using a combination of activity data and the relevant emission factor. The activity data reflects the extent or intensity of fuel-consuming actions, such as kilometres driven or litres of fuel used. Fuel consumption and resulting emissions are determined based on vehicle specifications, the type of fuel used and the distance travelled using formula (1).

The team drove a Peugeot 3008 Diesel/AdBlue with an estimated fuel consumption between 5.0 and 6.0 L/100 km, depending on driving conditions and vehicle efficiency. Therefore, an average value of 5.5 L/100 km is used. To calculate the emission factor for diesel vehicles, we utilise formula (2), incorporating the MITECO emission factor50, which indicates an output of 2.49 kg CO₂eq/L for a diesel vehicle. It is important to mention that vehicles equipped with AdBlue technology do reduce certain NOx emissions; however, the impact on CO₂ emissions is practically unaffected.

$$total\,comsuption\,\left(l\right)\,=\,\frac{vehicle\,consumption\,\left(l\right)}{100}\,\times \,distance\,(km)$$
(1)
$$C{O}_{{2}}eq\,emissions\,\left(kg\right)\,=\,total\,consumption\,(l)\,\times \,2.49$$
(2)

Table 2 presents a sensitivity analysis of annual emissions by vehicle type and number of site visits, based on a total distance of 396 km for a round trip from the monitoring laboratory to the site. As shown in the table, replacing traditional loggers with a remote monitoring system requiring only one visit per year can reduce emissions by up to ~75% for a diesel vehicle (from 231.74 to 57.93 kg CO₂eq per year). This environmental benefit is even more significant when considering low-emission vehicles, such as electric models.

Table 2 Annual CO₂ emissions (kg) by vehicle type and number of 396 km trips, calculated using MITECO’s (2022) emission factors
Full size table

An annual visit will be required for the new wireless system to replace the sensors with calibrated ones that comply with EN 15758:201144 and EN 16242:201445 standards for temperature and relative humidity measurement. However, if calibration is not required, the sensors can operate effectively for approximately five years without maintenance.

Results

Between 2020 and 2023, data collected from the LOG32® TH and LOG210® dataloggers revealed that extreme fluctuations in temperature and relative humidity have a significant impact on the sandstone rock supporting the paintings (see Fig. 4 and Table 3).

Table 3 Climatic parameters at the Toros del Prado del Navazo shelter related to damage risk metrics according ASHRAE Class A2 2023
Full size table

Specifically, the daily temperature range was found to be up to ±4.3 °C in winter and ±2.6 °C in summer. Both of these figures exceed the ±2 °C limit established by ASHRAE Class A2 for acceptable short-term fluctuations. These values indicate high thermal variability, which can lead to mechanical stress, fatigue and cracking in mineral substrates.

Relative humidity presented even more critical conditions. While the average relative humidity (RH) remained within the recommended range (40–60%) at around 45 ± 2.17%, the daily hygrometric range reached ±18.6% in winter and ±18.8% in summer. These values are nearly double the tolerable fluctuation threshold of ±10% and suggest an unstable microclimate. Such variability fosters repeated wetting–drying cycles in the clay minerals and soluble salts present in the sandstone, contributing to granular disintegration, haloclastism and salt efflorescence formation. These microclimatic dynamics directly correlate with the physical deterioration observed in situ, where cracks, powdering and salt crystal blooms were documented in the most exposed areas. These findings confirm previous studies12,23,24,28 suggesting that thermohygrometric fluctuations accelerate the deterioration of rock paintings.

A recurring challenge associated with standalone dataloggers has been the disruption of data collection due to equipment failure, which compromises the continuity of monitoring (see Fig. 4). Despite the data gaps, the gathered information has highlighted critical periods of heightened conservation risk, particularly during months with significant thermal oscillation, namely January and July. As illustrated in Fig. 4, there is a marked increase in climatic variability during the winter and summer months. In January 2021, for instance, temperatures experienced a substantial nighttime drop, recording a minimum of −3 °C and a maximum of 10 °C. Conversely, in July 2021, while daytime temperatures rose, the minima recorded were higher, hitting 15 °C, with maxima reaching 25 °C.

This pattern of heightened thermal oscillation is a result of the extreme conditions to which the paints are exposed, particularly influenced by factors such as indirect sunlight. Relative humidity exhibits an inverse relationship to temperature, with elevated humidity levels typically recorded as temperatures decrease, especially during cooler nights. Although this situation could raise the risk of damage due to condensation, data on dew point indicate that such occurrences have not taken place. Understanding this climatic behaviour is essential for planning restoration interventions at critical times and for preventing damage to the constituent materials when integrating restoration materials.

As illustrated by the Toros del Prado del Navazo shelter deployment, the adoption of IoT technologies for monitoring rock shelters has enabled the effective acquisition of real-time data on essential climatic parameters, including temperature, relative humidity and dew point. This information is vital not only for tracking environmental conditions within the shelters but also for evaluating the effects of these factors on the rock paintings. Such insights enable the development of more effective conservation strategies.

The CollectionCare sensor was configured to sample and immediately transmit data every five minutes to evaluate the battery life and the packet data transmission loss rate in this specific scenario involving extreme temperatures and frequent storms; the standard monitoring interval for conservation purposes is usually longer, and the 5-min interval was a technical stress test. The system notified us of a sensor failure on November 10, 2024, at 20:30 hours. We visited the site on November 13, 2024, to replace the sensor. A subsequent analysis of the missing data, covering the period from the initial sensor installation on March 12, 2024, to the present study date of January 17, 2025, reveals that, excluding the three days of sensor downtime, the total data availability stands at 95.4%. However, when accounting for the days missed, the data availability drops to 94.1% for this specific configuration and time lapse case. During the period from March 2020 to July 2024, measurements conducted with classical dataloggers revealed that these devices were inoperative for a total of 575 days. This results in an uptime percentage of 64%. Increased visits to the site for data downloads and operational checks could have significantly reduced this percentage.

Assuming that the wireless sensors deployed in the shelter are 100% operational, it is necessary to consider that the specific configuration of the IoT system deployed in Albarracín will result in multifactorial packet losses (i.e. environmental measurements transmitted to the receiving platform). These losses are due to three factors: the type of LoRaWAN wireless transmission used; the availability of the LoRaWAN Network Server via all internet connections (TTN in this case); and the availability of the platforms through which the messages circulate (i.e. the MQTT protocol to the Thingsboard cloud platform). This loss is inherent to this type of deployments. To evaluate these losses, the losses of packets transmitted by the CollectionCare sensor, specifically configured for transmission without confirmation every five minutes, were taken as a reference. The number of packets actually received between 13 March and 11 July 2025 was counted. To apply basic statistical measures, the period was divided into 15-day batches (4,320 theoretical measurements), ensuring that the batches were large enough to be reasonably normal and statistically independent, and allowing for sufficient degrees of freedom in a confidence analysis. The average reception rate was found to be M = 98.10%, with a standard deviation of SD = 1.16%. A one-sample t-test was conducted to evaluate whether the sample mean significantly differed from a hypothesised population mean of 0.98 (98%). The results indicated that the difference was not statistically significant, t(8) = 0.26, p = 0.804, M = 0.9810, SD = 0.0116. Thus, the null hypothesis was not rejected. There is insufficient evidence to conclude that the true population mean differs from 0.98. This positive outcome exceeded our expectations and confirms the suitability of the chosen technology. Nevertheless, we indicate lines of future work to approach 100% data recovery in the conclusions.

The process of downloading data from the initial dataloggers required a trip from Valencia to Albarracín in a Peugeot 3008 Diesel vehicle. Based on the emission factors determined using MITECO emission parameters (approximately 1.88 kg CO₂), resulting in a total of 24.65 kg for a single trip. If we include the return journey to Valencia, the emissions double to 49.30 kg CO₂. Over the course of a year, conducting at least four such journeys for datalogger checks is estimated to generate a total of 197.20 kg of CO₂ emissions.

Discussion

One of this project’s key achievements is the implementation of a real-time environmental monitoring system. This system enables continuous assessment of site conditions, reducing the need for frequent physical visits. This proposal promotes responsible production and consumption (SDG 12) and promotes an environmentally friendly approach (SDG 13). As a result, this research supports our continuous efforts to protect and preserve the world’s cultural and natural heritage (SDG 11, target 4).

IoT-based remote monitoring technologies have emerged as an essential tool for enhancing the conservation techniques applied to Levantine rock art, as evidenced by their implementation at the Toros del Prado del Navazo shelter (SDG 9). Monitoring parameters such as temperature, relative humidity and light have made it possible to immediately detect environmental fluctuations that adversely affect the stability of the paintings and the rock substrate. This data is crucial for understanding deterioration mechanisms, assessing risks and planning conservation interventions, particularly consolidation actions, where humidity plays a significant role.

Table 4 provides a comparative overview of relevant heritage conservation projects using IoT or remote monitoring technologies. While notable examples exist—such as Chauvet Cave and Herculaneum—the implementation at Prado del Navazo introduces several key innovations, including the first open LoRaWAN gateway deployed in a rock art site (based on The Things Network initiative) and effective long-range transmission in rugged natural terrain, facilitating the deployment of a wide variety of sensors by us or by third parties covering other shelters, such as the Tio Campano and the Cocinilla del Obispo ones (planned in the near future), or the natural environment in which they are located. Unlike the predominantly closed or wired systems found in heritage contexts, our open-access wireless model enables community participation, facilitates scalability and supports long-term sustainability. We have recently taken advantage of the infrastructure deployed to install a rock surface temperature sensor and a fire detection sensor, and we plan to install visitor counting sensors shortly.

Table 4 Comparative summary of monitoring projects in similar prominent locations
Full size table

Unlike most sensor deployments at other heritage sites, we have worked to integrate the sensors aesthetically through hydroimpression camouflage, reducing visual impact without creating false authenticity. This camouflage is applied exclusively to the temperature and humidity sensor housing, which is attached to the side wall using a non-invasive, easily removable bracket. This configuration fully complies with the ICOMOS principle of reversibility. The purpose of the camouflage is to minimise the stark visual contrast of the original white sensor housing. It is important to note that the sensor is not located inside the painted panel and does not interfere with the artwork itself. The intervention is limited to reducing the visual intrusion while avoiding any suggestion of authenticity. In remote and visually sensitive environments such as Prado del Navazo, features like these represent an advancement in environmental monitoring for the preservation of cultural heritage.

Remote monitoring has proven more reliable than traditional datalogger-based systems, primarily by minimising or eliminating data collection failures. However, challenges remain regarding the performance of wireless sensors, especially in improving their backup storage capacity to enable remote data retrieval following communication interruptions. Additionally, remote monitoring contributes to climate change mitigation by reducing carbon footprints, helping to maintain lower global warming levels (SDG 13).