Acoustics of karst tourist caves: a case study in Guizhou Province, China

Abstract

The acoustic environment of karst tourist caves significantly impacts visitor experience and the effectiveness of information dissemination, yet systematic studies on their acoustics remain scarce. This study investigates the acoustic properties of a representative karst tourist cave in Guizhou Province, China, by combining in-situ measurements with 3D laser scanning. Impulse responses were systematically recorded at 22 positions across six caverns, and key parameters—including reverberation time (T₂₀), early decay time (EDT), definition (D₅₀), clarity (C₈₀), and speech transmission index (STI)—were analyzed in relation to spatial geometries obtained from 3D scans. The results reveal substantial spatial variations in acoustic conditions: mid-frequency T₂₀ ranges from 1.4 s to 3.1 s, largely governed by cavern volume, geometry, and entrance-related attenuation. The average EDT/T₂₀ ratio is 0.70, indicating non-uniform early decay behavior. While D₅₀ values are relatively consistent (0.49–0.74), C₈₀ shows greater variability (1.86–6.89 dB), both influenced by spatial complexity. Furthermore, significant correlations exist between STI and other parameters (P < 0.01), confirming that reverberation control and early-energy enhancement are crucial for improving speech intelligibility. These findings provide a scientific basis for acoustic optimization in show caves, supporting better guided tours and safety communications.

Introduction

The term “karst” originates from the Kras Plateau in Slovenia, referring to landscapes formed by the dissolution of soluble rocks^1,2. These landscapes are characterized by distinctive geomorphological features such as caves, sinkholes, sinking streams, and subterranean rivers³. China contains the world’s most extensive karst area, covering approximately 2.55 million km² (about 26.5% of its land area)^4,5. The exposed karst region in Guizhou Province represents the largest contiguous karst area globally⁶.

The formation of karst caves involves complex geological processes that preserve unique evolutionary landscapes and cultural-historical relics with significant archaeological value². Driven by rapid tourism development, karst caves have become prime targets for tourism resource development due to their distinctive speleothem formations and stable microclimatic conditions. Statistics indicate 708 developed tourist cave sites in China, including 177 A-rated scenic areas, most of which are karst formations^7,8. Notably, the intricate spatial geometries and irregular wall configurations of karst caves create distinctive acoustic environments. These acoustic properties critically influence visitor experience metrics (e.g., speech intelligibility and musical clarity) and determine the efficacy of emergency broadcast systems during incidents⁹. Therefore, systematic investigation of subterranean sound fields is essential for optimizing auditory environments and ensuring effective information transmission.

However, current research on natural cave acoustics remains limited. Traditional cave studies have predominantly focused on speleogenesis^10,11, tourism value assessment^12,13, typological classification¹⁴, anthropogenic impacts¹⁵, and conservation strategies^16,17. While emerging research directions include structural stability analysis using 3D modeling¹⁸ and applications of wireless mesh networks¹⁹, acoustic environment investigations are notably scarce. Furthermore, although theories in architectural acoustics are well-established, the acoustic properties of natural caves—semi-enclosed spaces with complex geometric boundaries—remain understudied.

Recent studies have begun to address this gap. For instance, research on Italian karst caves by Iannace and Trematerra revealed acceptable acoustic parameters compared to performance space standards, except in caverns exceeding 30,000 m³²⁰. Similarly, Wolski’s 2020 analysis of a glacial cave demonstrated morphology-dependent acoustic characteristics described as dry and intimate²¹. Subsequent 2022 research on Icelandic lava tubes identified very dry and intimate acoustics, suggesting potential applications for concert hall sound-absorbing installations²². Despite these contributions, cave acoustics research remains in its nascent stage.

To address this research gap, the present study investigated a representative tourist karst cave in Guizhou Province, China. Systematic acoustic measurements were conducted across six caverns, including parameters such as reverberation time (T₂₀), early decay time (EDT), definition (D₅₀), clarity (C₈₀), and speech transmission index (STI). A notable methodological aspect was the integration of 3D laser scanning to characterize spatial geometries, enabling analysis of how cave morphology influences acoustic parameters. Since geometric configuration is a dominant factor influencing physical fields, as demonstrated by the significant impact of complex hydraulic structures on flow behavior^23–25. The findings provide a theoretical basis for the acoustic optimization and sustainable development of karst cave systems, while also expanding the methodological and disciplinary scope of speleological acoustics.

The cave localization and morphology

The cave localization

This study focuses on the Shuanghe Karst Cave, situated within the Shuanghe Cave National Geopark in Zunyi City, Guizhou Province; its geographical location is presented in Fig. 1. The geopark, characterized by dolomite karst landforms, serves as an exceptional representative of this type of geological evolution in China and is classified as a 4 A-level tourist attraction. The complete documentation of the extensive Shuanghe Cave System remains an ongoing endeavor. As of October 2024, the surveyed total length has reached 437.1 km, ranking it as the third-longest cave system in the world and the longest in Asia²⁶.

Fig. 1.

Geographical location of the study site. (a) Location in Zunyi, China²⁷. (b) Location of the Shuanghe Cave National Geopark within Zunyi, China²⁸.

The specific cave branch investigated in this study is in the central section of the main system. Its entrance has an elevation of 710 m above sea level at coordinates 28°24′30.4″N, 107°27′51.07″E. The entrance, oriented to the northeast (45° azimuth), opens into a 750-meter-long passage with a height ranging from 12 to 25 m. This subterranean space exhibits significant morphological diversity, displaying characteristic underground canyon features. The cave presents alternating U-shaped, V-shaped, and fissure-type canyon morphologies, collectively forming a well-preserved underground touring canyon system that showcases comprehensive erosional features and considerable typological variety.

The cave morphology

The 3D laser scanning model of the Shuanghe Karst Cave was captured through field investigations using a Leica RTC360 3D laser scanner²⁹. Based on this model, the cave was divided into six distinct measurement zones (Caverns 1–6) to facilitate in-situ acoustic measurements, with the division guided by cavern typology, dimensions, and connectivity. The spatial distribution of these zones and the corresponding 3D model are illustrated in Fig. 2.

Fig. 2.

3D laser scanning model²⁹ of the Shuanghe Karst Cave, showing the cave outline and the location of Caverns 1–6.

The morphology and appearance of these six caverns are depicted in Fig. 3. While all caverns are composed of dolomite with relatively similar surface roughness, they exhibit substantial variations in spatial dimensions and configurations.

Fig. 3.

Morphology and diversity of the studied caves. (a)-(f) Representative photographs of Caverns 1–6, illustrating the variety in shape and scale.

Among Caverns 1 to 6, significant spatial variations are observed (Table 1). Cavern 3 is the most voluminous (80,000 m³) and has the greatest floor area (4,000 m²), featuring a complex "X"-shaped layout with one branching passage. In contrast, Cavern 4 is the smallest in both volume (9,000 m³) and area (700 m²), characterized as a narrow, fissure-type canyon. The cavern heights range from 12 m to 25 m, with Cavern 5 being the tallest. Cavern 1, located near the entrance, has a regular "I"-shaped configuration, while adjacent Cavern 2 exhibits a larger, "U"-shaped morphology. Cavern 6, situated near the exit, adopts a "U"-shaped form.

Table 1.

Spatial and morphological parameters of caverns 1 to 6 in the studied cave.

	Cavern 1	Cavern 2	Cavern 3	Cavern 4	Cavern 5	Cavern 6
Maximum width (m)	16	30	32	11	29	16
Maximum height (m)	12	22	20	13	25	23
Estimated length (m)	100	70	120	60	60	100
Estimated area of the cave bottom (m2)	1500	2000	4000	700	2500	1500
Estimated volume (m³)	20,000	45,000	80,000	9000	40,000	35,000
Configuration	"I"	"U"	"X"	"I"	"X"	"U"
Number of branch channels	0	0	1	0	1	0

Research methods and test schemes

Research methods

T₂₀, EDT, D₅₀ and C₈₀ were measured following the ISO 3382-2 standard³⁰, employing the integrated impulse response method with maximum-length sequence (MLS) signals. The measurement setup is illustrated in Fig. 4. An MLS signal generated by the Dirac 6.0 software³¹ was output through an audio interface and power amplifier to drive an omnidirectional sound source. The resultant acoustic response was captured by a measurement microphone, conditioned by a signal conditioner, and subsequently recorded via the audio interface for analysis. Additionally, STI was measured using an HBK 4720 Echo Speech Source.

Fig. 4.

Schematic diagram of the acoustic measurement system setup, showing the connections between the sound source, amplifier, audio interface, microphone, signal conditioner, and computer for impulse response acquisition.

The measurements were conducted in winter. The cave environment was characterized by low temperature and high humidity. The specific environmental conditions recorded in Caverns 1 to 6 during measurements were: temperatures of 10.7 °C, 12.3 °C, 9.5 °C, 10.1 °C, 9.7 °C, and 10.8 °C, with corresponding relative humidity levels of 81.2%, 85.1%, 91.9%, 87.5%, 87.6%, and 86.5%. Corrections for the influence of these temperature and humidity conditions were applied during the calculation of all acoustic parameters (T₂₀, EDT, D₅₀, C₈₀, and STI) to ensure accuracy.

Test schemes

Field measurements were carried out on January 30 and 31, 2024, during the tourism off-season, in the absence of tourists. Only a three-person research team was present on site during the acoustic testing. To minimize potential interference, the team maintained a safe distance from both the sound source and microphones, remaining stationary and silent throughout the measurement process. Background noise levels were measured in each cavern prior to the active acoustic testing. The recorded values in Cavern 1–6 were 48.3 dB, 40.5 dB, 32.1 dB, 33.2 dB, 41.8 dB, and 47.1 dB. These consistently low noise levels confirm that ambient noise from the natural cave environment and the measurement equipment itself did not significantly affect the quality of the acquired impulse responses.

The layout of the measurement positions in Caverns 1 through 6 is illustrated in Fig. 5, comprising a total of 22 receiver points. Source and receiver positions were strategically distributed according to the dimensions and geometry of each cavern to ensure adequate coverage of the acoustic zones. Each receiver position was subjected to a minimum of three measurement repetitions to mitigate experimental errors. All source-receiver configurations complied with the proximity-to-boundary requirements specified in ISO 3382 − 2³⁰.

Fig. 5.

Layout of acoustic measurement points. (a)-(f) Measurement points in Chambers 1 to 6; (g) location of the caverns. This schematic generated based on the 3D laser scanning model (Fig. 2), depicts the strategic arrangement of receiver and source points across the six chambers.

Results

Reverberation time

T₂₀ is a critical parameter for evaluating acoustic field characteristics, as excessively long T₂₀ values can cause auditory blurring and degrade speech intelligibility^32,33. Figure 6a presents the frequency-dependent T₂₀ distributions across all 22 measurement positions, with median values highlighted in bold. The results reveal considerable variation in T₂₀ across the cave. The mid-frequency T₂₀ (mean of 500–1000 Hz) ranges from 1.4 s to 3.1 s, while the low-frequency T₂₀ (mean of 125–250 Hz) shows greater dispersion, ranging from 1.8 s to 4.6 s.

Fig. 6.

T₂₀ in octaves. (a) all receiving positions in the cave; (b) the average value across receiving positions in each cavern.

These values indicate substantially different acoustic experiences at different locations. For instance, the longest mid-frequency T₂₀ of 3.1 s (comparable to large churches^34,35 far surpasses the generally recommended limit of < 1.2 s for speech-dominant spaces³⁶, severely compromising speech clarity for visitors. In contrast, the shortest T₂₀ of 1.4 s is like values found in some concert halls³⁷, providing relatively better conditions for speech intelligibility. Above 1000 Hz, the T₂₀ curves at all measurement points become relatively flat and exhibit a monotonic decrease, with variations primarily dominated by resonant fluctuations. This behavior shows similarities to the acoustic characteristics observed in glacial caves by Pawel’s research team²¹.

Figure 6b illustrates the variations in T₂₀ across Caverns 1 to 6. A comparative analysis shows significant numerical differences in T₂₀ values among the caverns, despite shared general trends in their frequency responses. Cavern 3 exhibits the longest mid-frequency T₂₀ (mean of 500–1000 Hz: 3.14 s), followed closely by Cavern 2 (3.10 s). This correlation aligns with their larger spatial volumes, as quantified in Fig. 5; Table 1. Such extended reverberation times (> 3 s) would produce a profoundly spacious acoustic atmosphere, which could enhance musical resonance but would significantly compromise speech intelligibility for tourist groups. Collectively, these findings confirm that variations in reverberation time are primarily governed by cavern dimensions, geometric configurations, and acoustic attenuation effects related to cave entrances.

Early decay time

EDT, defined as six times the duration required for the sound pressure level to decay by 10 dB after the sound source ceases, serves as a critical metric for assessing perceived reverberation and exhibits a strong correlation with the reverberation time³⁸. Figure 7a illustrates the frequency-dependent variations in EDT across all 22 measurement positions, revealing the greatest dispersion at low frequencies (0.8 s to 3.9 s). The mid-frequency EDT ranges from 0.6 s to 3.0 s, indicating substantially different early decay characteristics across the cave system. For visitors, this implies that the perceived reverberance varies significantly with location. The measured EDT values fall within the typical range observed for unoccupied concert halls or multi-purpose auditoriums³⁰.

Fig. 7.

EDT in octaves. (a) all receiving positions in the cave and (b) the average value across receiving positions in each cavern.

Figure 7b further details the EDT frequency characteristics for each cavern. Caverns 2 and 3 exhibit the higher mid-frequency EDT values (2.4 s and 2.0 s, respectively), suggesting that sound energy decays relatively slowly in the initial period. This creates a more persistent reverberant tail that could reduce speech clarity during guided tours. In contrast, Cavern 1 demonstrates the lowest mid-frequency EDT value (0.9 s), indicating a relatively rapid early decay that would help maintain better speech intelligibility near the entrance. The difference of 1.5 s in EDT between Caverns 2 and 1 represents a substantial variation in the initial decay characteristic that visitors would readily perceive. These variations primarily correlate with cavern volume and morphological complexity, which directly influence how sound energy dissipates during the critical early period that most affects auditory perception.

The discrepancies between EDT and T₂₀ provide insight into the uniformity of sound energy decay within the caves. A notably shorter EDT compared to T₂₀ typically indicates the presence of abundant late-arriving reflections and a prolonged sound attenuation process. Figure 8 compares the average EDT and T₂₀ values across the six caverns, illustrating their frequency-dependent characteristics.

Fig. 8.

The average EDT and T20 of all caverns in octaves. (Values at each frequency are averaged over all caverns.)

The data show that the EDT values are markedly lower than the corresponding T₂₀ values. For instance, the low-frequency averages are 2.2 s for EDT and 3.1 s for T₂₀, resulting in a difference of 0.9 s. This disparity can be attributed to the strong diffraction capacity of low-frequency sound waves, which allows them to sustain multiple reflections across the complex, irregular rock surfaces. In the high-frequency range, the divergence between EDT and T₂₀ is reduced, which is likely due to the greater absorption of high-frequency sound energy by the porous cavern walls.

Notably, the average EDT/T₂₀ ratio of 0.70 across the measured points underscores an accelerated energy decay in the early stage and indicates non-uniform sound attenuation patterns throughout the karst cave system.

Definition and clarity

Speech clarity, quantified by the Definition (D₅₀), and musical clarity, characterized by the Clarity (C₈₀) index, are critical parameters for evaluating acoustic environments³⁹. D₅₀ is defined as the ratio of the direct sound energy received within the first 50 milliseconds to the total sound energy, with a theoretical maximum value of 1.0 representing ideal speech transmission conditions. Practically, D₅₀ values exceeding 0.5 are generally considered acceptable for speech intelligibility, whereas values below 0.4 often indicate challenging listening conditions. Conversely, C₈₀ reflects musical clarity through the logarithmic ratio of early-arriving sound energy (0–80 ms) to late-arriving energy (80 ms to infinity). For musical perception, C₈₀ values around 0 dB suggest a balance between early and late energy, while positive values indicate greater clarity, which is suitable for articulated musical passages. In this analysis, the averaged values across the 500 Hz and 1000 Hz octave bands are used for both parameters, following standard analytical protocols⁴⁰.

Figure 9a presents the speech clarity measurements for Caverns 1 to 6. The results demonstrate relatively limited variation in D₅₀ across the caverns, with values ranging from 0.49 to 0.74 and a mean value of 0.57. D₅₀ results indicate generally adequate speech intelligibility conditions throughout the cave system, although the range suggests noticeable location-dependent variations in vocal clarity.

Fig. 9.

Single number averaging of 500–1000 Hz of D₅₀ and C₈₀ in Caverns 1 to 6. (a) D₅₀; (b) C₈₀.

Figure 9b illustrates the musical clarity values for the six caverns. In contrast to D₅₀, the C₈₀ data reveals substantial disparities, spanning from 1.86 dB to 6.89 dB. Cavern 1 exhibited the highest C₈₀ value of 6.89 dB. The remaining caverns displayed values distributed between 1.86 dB and 4.19 dB, specifically: Cavern 2 (1.97 dB), Cavern 3 (3.42 dB), Cavern 4 (1.86 dB), Cavern 5 (3.20 dB), and Cavern 6 (4.19 dB). Notably, Cavern 4 recorded the lowest C₈₀ value (1.86 dB), while Cavern 6 achieved the second-highest value at 4.19 dB, which is still 3.77 dB lower than that of Cavern 1.

These observations suggest that spatial configuration plays a key role in acoustic clarity. Regular, linear geometries like that of Cavern 1 facilitate optimal clarity by promoting beneficial early reflections. In contrast, complex geometries (e.g., X-shaped or U-shaped) and extreme dimensional proportions, as seen in Caverns 3 and 4, substantially reduce C₈₀ values due to acoustic energy dispersion or waveguide effects.

Speech transmission index

STI is a comprehensive metric for evaluating speech transmission quality^41,42, it incorporates the effects of reverberation time, signal-to-noise ratio, echoes, system distortions, and psychoacoustic factors such as masking⁴³. Its value ranges from 0 to 1, with higher values indicating superior speech clarity. According to the IEC 60268-16 standard⁴⁴, STI values can be classified into speech intelligibility ratings as follows: <0.30 (bad), 0.30–0.45 (poor), 0.45–0.60 (fair), 0.60–0.75 (good), and > 0.75 (excellent).

Figure 10 presents the STI values measured for Caverns 1 to 6. The data show that caverns located near the cave entrance (Caverns 1 and 6) exhibited higher STI values of 0.52 and 0.57, respectively, placing them in the “Fair” intelligibility category. In contrast, the four central caverns yielded lower STI values, with a mean of 0.32, corresponding to ratings between “Poor” and “Bad”. Specifically, applying the IEC 60268-16 standard⁴⁴ for objective assessment, Caverns 1 and 6 achieved a “Fair” rating, Caverns 3, 4, and 5 were classified as “Poor”, and Cavern 2 received a “Bad” rating. These results indicate that while verbal communication is feasible in the entrance-proximal caverns, effective guided tours in the central caverns, particularly in Cavern 2 where intelligibility is severely compromised, would require electronic voice amplification systems.

Fig. 10.

STI in Caverns 1 to 6.

Figure 11 illustrates the results of a linear correlation analysis between the STI and four key acoustic parameters: T₂₀, EDT, D₅₀, and C₈₀. The parameter values used for this analysis were averaged across the 500–1000 Hz octave bands from all receiver positions within each cavern.

Fig. 11.

Correlation of STI with (a) T₂₀, (b) EDT, (c) D₅₀, and (d) C₈₀. The black dots are all receiving positions in the cave.

A significant negative correlation was found between STI and the mid-frequency T₂₀ (r = -0.649, p < 0.01, R² = 0.421). This relationship indicates that a reduction in the overall reverberation time by approximately 1.0 s could potentially improve speech intelligibility from a “Poor” to a “Fair” quality level in these environments, which aligns with the theoretical predictions of improved STI with shorter reverberation times outlined in ISO 3382 − 2³⁰.

A stronger significant negative correlation was observed between STI and the mid-frequency EDT (r = -0.758, p < 0.01, R² = 0.575). This suggests that optimizing the early decay characteristics may have an even greater impact on enhancing speech intelligibility than reducing the overall reverberation time. Shortening the early decay time is particularly effective for improving the clarity of initial syllables, which is crucial for guide communications.

Conversely, STI showed significant positive correlations with both D₅₀ (r = 0.639, p < 0.01, R² = 0.408) and C₈₀ (r = 0.697, p < 0.01, R² = 0.486). These relationships confirm that increasing the early-to-late sound energy ratio—whether through strategic architectural modifications or targeted acoustic treatments—could substantially improve speech intelligibility for visitors.

The strong correlations identified in this study offer valuable insights for cave management. The findings suggest that strategic acoustic interventions, such as installing sound-absorbing materials in highly reverberant areas (e.g., Caverns 2 and 3) and optimizing guide positioning to leverage beneficial early reflections, could significantly enhance speech intelligibility during tours. Furthermore, these results provide a scientific basis for the design of effective public address systems in karst caves, ensuring the clear delivery of safety instructions and educational content in these challenging acoustic environments.

Discussion

This study provides a systematic analysis of the acoustic characteristics of six caverns within the Shuanghe Karst Cave, revealing distinct spatial patterns in key parameters including T₂₀, EDT, C₈₀, D₅₀, and STI. The measured mid-frequency reverberation times (1.4–3.1 s) are comparable to those reported for Italian karst caves (1.8–3.5 s)²⁰, suggesting shared acoustic features in similar geological settings. However, the broader distribution of T₂₀ observed in this study, compared to glacial caves^21,22, may be attributed to fundamental material differences. While ice-dominated cave walls promote sound absorption and rapid decay, the rough, porous surfaces of dolomite induce complex reflection-scattering interactions, thereby amplifying fluctuations in reverberation.

A key finding is that spatial parameters, particularly volume and geometric complexity, emerged as the primary determinants of acoustic behavior, despite lithological variations. The prolonged reverberation in larger caverns is consistent with acoustic principles observed across diverse environments, from glacial caves²¹ to constructed spaces like concert halls⁴⁵. This correlation suggests a potential universality of spatial-scale dominance over sound fields across both natural and architectural settings. Nevertheless, cave systems exhibit greater acoustic complexity than designed spaces due to their irregular geometries, branching channels, and openings. These features introduce additional sound attenuation and reflection paths that are not typically encountered in conventional architectural acoustics.

Our results further highlight the distinctive acoustic imprint of karst environments through pronounced low-frequency reverberation and non-uniform decay patterns (e.g., an average EDT/T₂₀ ratio of 0.70). These characteristics deviate from typical auditorium acoustics. While such features can reduce speech intelligibility in central caverns—a critical consideration for guided tours and safety announcements—they may also create unique sound qualities suitable for cultural or musical activities.

Although this study is based on a single cave system, the consistency between our findings and those from other karst regions (e.g.,²⁰) implies that the acoustic roles of volume, geometry, and surface material may be generalizable to similar dolomitic karst caves worldwide, particularly where comparable morphologies (e.g., large chambers, fissure networks, canyon passages) exist. Future studies across a wider range of karst terrains and lithologies are needed to validate and refine these extrapolations.

This research has several limitations. First, the analysis included only one cave system, which necessitates future expansion to other karst regions to enhance the generalizability of the findings. Second, the influence of specific factors, such as detailed wall roughness and cross-cavern connectivity, was not fully quantified. Future work should involve multi-cave comparisons, detailed 3D laser scanning-based geometric modeling, and in situ absorption measurements to establish more generalized spatial-acoustic relationships in karst environments.

Conclusions

This study conducted a comprehensive investigation of the acoustic environment within the Shuanghe Karst Cave, Guizhou Province, China, through systematic in-situ measurements and spatial analysis of six distinct caverns. The principal conclusions are as follows:

The reverberation time (T₂₀) exhibited considerable spatial variation, with mid-frequency values ranging from 1.4 s to 3.1 s. This variability was found to be predominantly controlled by cavern volume, geometric configuration, and sound attenuation effects associated with cave openings.

An average Early Decay Time (EDT) to T₂₀ ratio of 0.70 was observed, indicating a rapid and non-uniform decay of sound energy in the initial stages. This characteristic decay pattern is a key contributor to the distinctive acoustic signature of karst caves.

Speech clarity, evaluated by D₅₀, showed relatively consistent conditions across the cave system (0.49 to 0.74, mean: 0.57). In contrast, musical clarity (C₈₀) demonstrated significantly greater variability (1.86 dB to 6.89 dB). Both parameters were strongly influenced by cavern spatial complexity and dimensions.

The Speech Transmission Index (STI) revealed a clear spatial pattern: “Fair” intelligibility near the cave entrances (0.52–0.57) contrasted with “Poor” to “Bad” intelligibility in the central caverns (mean: 0.32). Significant correlations between STI and key acoustic parameters confirm that speech clarity can be improved by reducing overall reverberation and enhancing early arriving sound energy.

Collectively, these findings provide valuable practical insights for optimizing acoustic conditions in tourist caves. The results suggest that strategic interventions—such as installing sound-absorbing materials in large, highly reverberant caverns and deploying directional loudspeaker systems in central zones—can significantly enhance speech intelligibility for guided tours and emergency announcements. These measures are expected to improve the overall visitor experience and ensure the effectiveness of safety communications.

Author contributions

Y.W. and Q.L. proposed the research topic and designed the study plan. W.Z. and Q.L. collected and organized the research data and conducted the detailed research. W.Z. drafted the manuscript, while Y.W. and Q.L. reviewed and revised the paper. All authors participated in the acoustic experiments and reviewed the manuscript.

Funding

This research was funded by Guizhou Provincial Philosophy and Social Sciences Planning, through Grant No. 22GZYB33.

Data availability

The data will be provided by the corresponding author on request.

Declarations

Competing interests

The authors declare no competing interests.