1 Introduction

For centuries, fireworks have delighted millions of people worldwide. Nowadays, they are widely used to celebrate special events, such as religious festivities, national holidays, sporting events, large concerts, and shopping mall inaugurations, among others. However, while fireworks bring joy, their use also presents various problems that can cause significant harm to both health (such as burns, eye injuries, auditory damage, and poisoning) and the environment (such as contamination of the environment by inorganic components, anxiety and fear in animals) [1,2,3,4,5]. These issues underscore the urgent need to explore concrete solutions, either through the implementation of prohibitive measures regarding this form of entertainment, or by actively promoting celebration practices that are more respectful and mindful of their impact on the health and well-being of the aforementioned groups. In this sense, many countries have established regulations to govern firework displays, and in some cases, they have even been completely banned at the municipal or national level. However, in several countries, including Argentina, fireworks are widely used by the general population, including young people and children (fireworks can be legally purchased from the age of 16 years), without specific restrictions on their use in certain locations and periods. Fireworks are typically launched in residential areas, public spaces, and private properties, creating a collective experience of visual and auditory spectacles. This situation leads to unpredictable domestic use of fireworks, making it difficult for governments to implement effective preventive measures. The lack of studies in this area exacerbates the situation, as it hinders a specific understanding of the issues caused by fireworks and, therefore, hampers the development of potential solutions to the problem.

Many disorders caused by fireworks are related to noise pollution. Quantitative assessment of the acoustic impact generated by pyrotechnics on ambient noise levels has been the subject of numerous investigations. Many of these studies measured the impact of fireworks on popular festivities by comparing the variation in ambient noise levels during festivities with ambient noise levels under normal conditions. For example, Deželak et al. [6] measured ambient noise during New Year's Eve (from 11:00 PM to 01:00 AM) in the surroundings of the capital city of Ljubljana, Slovenia. The recorded levels increased the background noise level on a normal day by 50 dB owing to the use of pyrotechnics. In Krakow, Kukulski et al. [7] conducted measurements in urban areas during New Year's Eve (from 11:30 PM to 12:30 AM). In two of the three evaluated points, all the analyzed descriptors presented excessive levels, with differences in the range of 50–60 dB for typical levels during that timeframe. Similarly, two studies measured the acoustic impact on noise levels during the Deepawali festival in Haridwar City in Uttarakhand [8] and Jaisalmer City [9], India. The authors reported an increase in the average noise level during festivities of 29.6% and 30% (in residential areas) and 18.1% and 13.4% (in commercial areas), higher than the standards set by the Central Pollution Control Board, New Delhi for residential and commercial areas, respectively. Similar results were obtained in a study conducted during the Nawruz festival in Duhok City (Iraq), where the average noise levels on a holiday showed an increase of over 20% compared to a non-holiday [10]. The levels of exposure to which the population and workers are exposed when participating in festivals or pilgrimages produced by the use of fireworks were measured in the Portuguese cities of Vila Nova de Gaia, Porto, and Matosinhos [11]. The results showed that the limits proposed by the EU Directive 2003/10/EC were exceeded in the case of police officers, sound and light operators, and those responsible for the detonation of fireworks (blasters). A similar study was conducted in Japan during firework displays [12]. In this case, the authors concluded that pyrotechnics technicians were exposed to levels higher than 133 dB at the launch site (152 dB considering the range of 1 Hz to 80 Hz), 46 dB above the recommendation by the Japanese Society for Occupational Health based on Occupational Exposure Limits (OEL).

Regulations for fireworks use vary by country and jurisdiction. Some common considerations include age regulations that establish a minimum age for handling fireworks, the need to obtain licenses and permits for manufacturing or selling fireworks, restrictions on their use in specific locations and periods, detailed safety standards for design and storage, and inspections and quality control to ensure compliance with the regulations. However, in countries such as Argentina, legislation regarding fireworks use is limited and challenging to implement. Furthermore, scientific studies addressing the various issues generated by pyrotechnics, both in health and the environment, are scarce or even nonexistent. Hence, it is crucial to delve into the study of this issue to develop possible legislations and strategies that enable the mitigation of problems caused by pyrotechnics.

In this study, we present, for the first time, measurements of the impact of fireworks on noise levels during the Christmas and New Year celebrations (2020/2021) in the Autonomous City of Buenos Aires (CABA), Argentina. In addition, we provide measurements conducted during the same festivities in the city of Santiago, Chile (2019/2020). Unlike Argentina, where fireworks are widely used by the general population, in Chile, the government organizes and carries out celebrations with professionals in specific locations and periods. By comparing these two datasets, we can conduct an analysis of the acoustic impact of fireworks during festivities under varying conditions, including mass usage and specific usage scenarios.

2 Methods

2.1 Measurement points on festive days and non-festive days in CABA, Argentina

The measurement criteria adopted for the present investigation were based on the aforementioned previous studies. The measurement interval was from 23:00 to 01:00 h on December 24th, 2020, and December 31st, 2020, for Christmas and New Year's Day, respectively. The selection of measurement points was planned to obtain the largest possible number of samples distributed throughout the Autonomous City of Buenos Aires (CABA).

Nine measurement points were evaluated, with slight variations in their locations between the two festive dates. Six points (1 to 6) were located in the same position during both festive events, while the remaining points (7 to 9) changed their locations between the two festive dates. In all cases, the systems were installed in private homes located in residential areas where the maximum level allowed at night according to Law 1540 (2004) of the Autonomous City of Buenos Aires is: Leq = 50 dB(A).

Figure 1 depicts the distribution of measurement points during celebratory days, represented by varying shades of red, as well as normal days indicated in green. The measurement points captured on both holidays (1 to 6) are displayed in red, while those exclusively recorded on Christmas Day are depicted in dark red, and those exclusive to New Year's Day are shown in light red.

Fig. 1
figure 1

Geographical distribution of the measurement points evaluated during holidays and non-holidays. The icons marked in red represent the positions of measurements taken on both holidays. The dark red icons represent the measurement points for Christmas, while the light red icons represent the measurement points for New Year's Day. The icons in green represent the measurement points provided by Environmental Protection Agency (CABA—Argentina)

Full size image

The positioning of the measurement systems was determined based on the available locations and the building characteristics of the residences, whether they were houses or apartments. The systems were placed in the courtyards whenever possible. However, in the case of multi-family residential buildings, where outdoor areas are commonly shared among all residents, these points were located on the building's terraces. If terraces were not available, the measurement systems were placed on the balconies of the properties. In all cases, the measurement microphones were positioned at a sufficient distance (greater than 10 m) from the nearest noise source to avoid its influence on the recorded noise levels. On Christmas night, 67% of the systems were located on terraces, 33% on balconies, and none in courtyards. On New Year's Eve, the percentages were 55% and 33% for the first two locations, respectively, and 12% for the last one.

To compare the environmental noise levels between festive and non-festive days (usual conditions), an information request was submitted to the responsible agency, the Environmental Protection Agency of CABA (APrA), to obtain noise levels between 23:00 and 01:00 h from the nearest geographic points to the evaluated measurement points.

The information request was based on the type of roadway (street or avenue) and its proximity to the measurement point. For example, if the measurement point was situated on a street and the closest equivalent point was on an avenue, noise levels from the nearest points located on streets were requested to account for the differences in vehicular traffic circulation. In such cases, the value used for comparison corresponds to the average energy of the set of points. Each point provided by the APrA had measurements ranging from 21 to 30 days, which were averaged according to the analyzed time interval.

2.2 Measurement points on festive days in Santiago de Chile city, Chile

For this research, we collaborated with the Department of Noise, Light, and Odor of the Ministry of the Environment of Chile. During the New Year's celebration, the government organized a 20-min fireworks display at the Entel Tower, a 127-m-high building. Approximately 16 tons of pyrotechnic devices were used, resulting in around 16,000 detonations reaching an altitude of approximately 300 m. The data analyzed corresponds to the time period from 11:00 p.m. to 01:00 a.m. on December 31, 2019. This data was obtained from the Environmental Noise Monitoring Network maintained by the aforementioned agency in Santiago de Chile.

Figure 2 illustrates the geographical distribution of the evaluation points provided by the agency (depicted in light red) and provides a view of the Entel Tower (depicted in blue). It is important to note that the distance between the transducers and the noise source is not consistent across all four points. Points located closer to the event site are expected to exhibit higher noise levels. Therefore, we took this variation into account when conducting the analysis. Based on the geolocation of the monitoring stations, it was determined that the distance from points 1–2 (close proximity) and 3–4 (far proximity) to the event site is approximately 0.6 km and 1.1 km, respectively. Consequently, in addition to calculating the energy average across all points, we also calculated separate averages based on the distance to the noise source, labeled as "Nearby points" and "Far points".

Fig. 2
figure 2

Geographical distribution of the measurement points evaluated during New Years Day in Santiago de Chile, Chile. The icons marked in light red represent the measurement points for New Year's Day provided by the Department of Noise, Light, and Odor of the Ministry of the Environment of Chile. The icon marked in blue dot represents the location of the Entel tower where the pyrotechnic event took place

Full size image

2.3 Measurement protocol, equipment and data processing

The ISO 1996 measurement guidelines were adopted. At all locations, the systems were mounted using tripods to ensure a height between 1.6 and 2.5 m from the ground and a minimum distance of 1.5 m from the reflective surfaces to minimize reflections. The height range was determined according to the available location to comply with the maximum possible value. In terraces and backyards, this condition was met in all the cases. On balconies, if the dimensions allowed it, the same criterion was used; otherwise, a 1.6 m was used to prevent the transducer from being at a distance of less than 1.5 m from the ceiling. In all cases, a wind screen was used to minimize the effects of wind.

The measurement equipment consisted of an audio interface, calibrated omnidirectional measurement microphone (94 dB(A) - 1 kHz) and computer for data acquisition.

A total of 18 recordings of 2 h duration were obtained, 9 for the night of December 24 and 9 on December 31, both from 23:00 to 01:00 h. All recordings files were in Wav format (48 kHz, 24 bits).

An exploratory analysis was performed to identify anomalous events that could affect the noise levels of each measurement, considering anomalies such as electrical noise due to cable movement, voices or shouts close to the microphone capsule (generally coming from conversations with the operator or his environment), and wind. The adopted process involved adjusting the level of this event to the mean level of an audio segment with ambient noise. From this analysis, one measurement point was discarded (Point 9 in Crhismas day), where the recorded levels corresponded mainly to loud music from a nearby neighbor.

The acoustic parameters were calculated and obtained by developing the MATLAB R2019a code. Before processing the set of collected signals, script validation was performed using a pink noise signal recorded with a Class 1 sound level meter. The following environmental noise descriptors established by the ISO1996:2007 standard were calculated: Equivalent continuous sound pressure level (Leq,A) and percent exceedance level (LN: L01, L05, L10, L50, L90). Three time intervals were considered to evaluate the differences between the measurements. The intervals according to the festive day, conditions and time were (h:mm):

  • Christmas or New Year's Day

    • Without fireworks

      • 60 min (23:00–23:59 h)

      • 30 min (23:30–23:59 h)

      • 15 min (23:45–23:59 h)

    • With fireworks

      • 60 min (00:00–01:00 h)

      • 30 min (00:00–00:30 h)

      • 15 min (00:00–00:15 h).

2.4 Outliers analysis

To detect and discard extreme estimates (outliers), we analyzed the collected data employing the median absolute deviation (MAD) method with a threshold of 3 MADs, using the Routliers package for R [13]. To this end, we calculate the Leq and LN parameters (L01, L05, L10, L50 and L90) in collapsed dB for each measurement point in each condition and time interval according to festive day (Christmas/New Year's Day). As a result, no outliers were found.

2.5 Statistical analysis

We calculate the averages parameters LeqA and LN (L01, L05, L10, L50 and L90) averages in collapsed dB for each measurement point in each condition and time interval according to festive day using the Seewave [14] package in R [15]. We fitted linear models (LMs) to estimate the associations of the magnitudes (LeqA and LN) with the condition, interval time and festive day using the Stats [16] package in R. The parameter estimates and confidence intervals were calculated using the ggstatsplot package in R [17]. We tested the fixed effects with an F-test (Satterthwaite's method for degrees of freedom) using the R’s function anova. The detailed syntax and formal definition of all the fitted models are also presented in the OSF repository.

3 Results

Table 1 presents all parameters results for each integration time interval, condition type, and celebration day. Figure 3a presents the energetic mean [dB(A)] of all the points considered for the analysis, corresponding to the 2020 Christmas and New Year celebrations (position 9 was excluded for Christmas day). For the Christmas celebration, the LeqA averaged over the complete measurement interval (23:00–00:00 h) was equivalent to 70.4 ± 2.6 dB(A), while for New Year's it was equivalent to 68 ± 2.7 dB(A).

Table 1 Acoustical parameters by integration interval time, type of condition and festive day [dB(A)]
Full size table
Fig. 3
figure 3

A Sound pressure level mean [dB(A)] as a function of the time, corresponding to the 2020 Christmas (right) and New Year (left) celebrations. The points in blue color represent the condition without fireworks and in red the condition with fireworks. The lines represent the linear model adjusted according to the time interval and condition. B Equivalent continuous sound pressure level (Leq,A) and C percent exceedance level (L01) collapsed across time and festive day (mean ± sd) for the intervals time. In green, the measurement data of a normal day in CABA is represented. Each point represents an individual measurement location

Full size image

Before 00:00 h, some events can be observed that correspond to isolated explosions recorded in the vicinity of the measurement points, resulting in maximum levels that are clearly detectable from an individual analysis of each position. In particular, in the case of Christmas, this condition was observed, for example, at 23:01 and 23:03 (point 2, see supplementary material: FS01), 23:31, 23:43, 23:44, and 23:52 (points 6 and 7, see supplementary material: FS01). After 00:00, it is also possible to detect isolated peaks, although the multiplicity and simultaneity of explosions makes it difficult to identify them clearly, which can be exemplified by observing the peaks that occur at 00:02 and 00:12, attributable to explosions close to point 7 (point 7, see supplementary material: FS01). In this case, these are not isolated events because they predominate only over the rest of the explosions that occur during the celebration. After 00:15 h, all the maximum levels observed were attributable to the use of fireworks, for example, the peaks present in the intervals 00:21–00:31, 00:36–00:42, and 00:46–00:53 corresponding to points 5, 2, and 1, respectively (see supplementary material: FS01). Because the energetic curve continues with a decreasing trend, it is not possible to define whether these events are isolated from the rest, as explosions still occur with minor periodicity.

During the New Year, the evolution of the curve is similar to that of the Christmas celebration. Although in lower quantities than that during Christmas, isolated events corresponding to explosions near the evaluation points also occur. In particular, the peaks at 23:08, 23:24, and 23:26 correspond to point 7, whereas the peak at 23:40 belongs to point 1 (see supplementary material: FS01). After 00:00 h, the set of peaks observed in the interval 00:03–00:13 also corresponds to point 1, those in the interval 00:16–00:22 to point 6, and the last ones to point 7 (see supplementary material: FS01). In all cases, these events were attributed to fireworks. After 00:15 hs, all the maximum levels observed are attributable to the use of fireworks,for example, the peaks present in the intervals 00:21–00:31, 00:36–00:42 and 00:46–00:53 correspond to points 5, 2 and 1, respectively (see supplementary material: FS01). Because the behavior of the curve continues with a decreasing trend, it is not possible to define with certainty whether these events are isolated from the rest, since explosions still occur at a lower frequency.

After fitting a linear model with condition (with fireworks/without fireworks), interval time (15 m;30 m and 60 m) and festive day (Christmas/New Year's Day) as a fixed effect, we confirmed that the condition with fireworks generated a significantly higher noise level than without fireworks: LeqA: β = 12.3 dB(A), 95% CI [4, 20.6] dB(A), fixed effects test: F(1.88) = 70.36, p = 7.43e−13; L01: β = 14.2 dB(A), 95% CI [5.35, 23] dB(A), fixed effects test: F(1.88) = 87.21, p = 8.22e−15; L10: β = 12.2 dB(A), 95% CI [3.86, 20.6] dB(A), fixed effects test: F(1.88) = 68.1, p = 1.41e−12 and L50: β = 8.8 dB(A), 95% CI [0.25, 17.3] dB(A), fixed effects test: F(1.88) = 31.91, p = 1.95e−7. No significant differences were found in the interval time factor. This may indicate that the first 15 min had a significant influence on the overall average value. Furthermore, no significant differences were found in the festive day factor, which suggests a similar use in the execution of fireworks on both festive dates. The model found no significant differences in any interaction (second or third order).

To evaluate the significance of the acoustic impact generated by pyrotechnics during festivities on environmental noise levels, we fitted linear models (LMs) to estimate the associations of the magnitudes (LeqA and LN) with the condition (with fireworks/without fireworks/Environmental noise of a normal day) and interval time (15 m; 30 m and 60 m). The results are presented in Fig. 3B and C. Our results show significant differences in the condition factor in all the measured parameters, however, no significant differences were found either in the interval time factor or in any interaction (second or third order). However, when making multiple comparisons between conditions (using the Holm–Bonferroni adjustment), we only found significant differences between—with fireworks—and—Environmental noise of a normal day—when interval time is—15 min—(L1: ΔM = 6.68 dB, 95% CI [0.71, 12.65] dB, t(23) = − 2.31, p = 0.029) and − 30 min—(L1: ΔM = 6.52 dB, 95% CI [0.48, 12.54]%, t (23) = − 2.23, p = 0.035). This result can be observed in Fig. 3C. When we study the significant differences between—without fireworks—and—Ambient noise on a normal day, we find that all studied noise parameters are significantly lower when there is a festive day than on a normal day at any time interval.

3.1 Comparison with Santiago de Chile's New Year celebration

Unlike Argentina, where New Year's fireworks are widely used by the general population after 12:00 a.m., in Chile the government organizes and carries out celebrations with professionals in specific places and periods. Figure 4A shows the mean energy [dB(A)] of all the points, corresponding to the 2020 New Year celebration in the city of Santiago de Chile, Chile.

Fig. 4
figure 4

A Sound pressure level mean [dB(A)] as a function of the time, corresponding to the 2020 New Year celebrations from Santiago de Chile, Chile. B Average sound pressure level [dB(A)] as a function of collapsed time according to the measurement points near and far from the sound source in Santiago de Chile, Chile. The points in blue color represent the condition without fireworks and in red the condition with fireworks. C Equivalent continuous sound pressure level (Leq,A) collapsed across time and country (mean ± sd). D Equivalent continuous sound pressure level (Leq,A) collapsed across time and country (mean ± sd) according to the measurement points near and far from the sound source. Each point represents an individual measurement location

Full size image

For Santiago de Chile, noise levels remain relatively constant, resulting in a LeqA of 69.4 ± 1.1 dB(A) during the first hour of observation. A considerable variation in noise levels is observed at the beginning of the celebration, resulting in a LeqA of 82.8 ± 2.4 dB(A) for that interval, which corresponds to an energetic increase of 13.4 dB compared to the initial LeqA. After the event ended, the noise levels decreased and tended towards the initial mean level. The high sound energy concentrated during the fireworks show results in a global LeqA for the entire measurement interval of 80.1 ± 2.2 dB(A), a difference of 10.7 dB higher than the pre-event mean noise level.

Figure 4B shows the same result as Fig. 4A but divided according to the proximity to the sound source (Nearby points in solid line and Far points in dotted line). The behavior of the curve is similar to the average of both, however, they maintain an average difference of 11.5 dB(A) throughout the time.

After fitting a linear model with condition (with fireworks/without fireworks) and location to source (Nearby points/Far points) as a fixed effect, we confirmed that the condition with fireworks generated a significantly higher noise level than without fireworks in Chile: LeqA: β = 8.41 dB(A), 95% CI [2.94, 19.8] dB(A), fixed effects test: F(1.4) = 19.6, p = 0.01; L5: β = 9.41 dB(A), 95% CI [4.01, 22.8] dB(A), fixed effects test: F(1.4) = 18.1, p = 0.01; L10: β = 10.3 dB(A), 95% CI [2.88, 23.5] dB(A), fixed effects test: F(1.4) = 19.41, p = 0.01 and L50: β = 8.58 dB(A), 95% CI [2.35, 19.5] dB(A), fixed effects test: F(1.4) = 18.01, p = 0.01. In addition, we confirmed that the more distant sources have lower noise levels than those closer to the source: LeqA: β = 11.5 dB(A), 95% CI [0.14, 22.8] dB(A), fixed effects test: F(1.4) = 6.01, p = 0.04. The model found no significant differences in any interaction.

To evaluate the significance of the acoustic impact generated by pyrotechnics during New Year on different government policies, we fitted linear models (LMs) to estimate the associations of the magnitudes (LeqA and LN) with the condition (with fireworks/without fireworks) and country (Chile/Argentina). The results are presented in Fig. 3B. As expected, we confirmed that the condition with fireworks generated a significantly higher noise level than without fireworks: LeqA: β = 13.4 dB(A), 95% CI [5.7, 21.1] dB(A), fixed effects test: F(1.22) = 18.26, p = 0.0003; L5: β = 13.4 dB(A), 95% CI [5.6, 21.2] dB(A), fixed effects test: F(1.22) = 19.4, p = 0.0002; L10: β = 13.4 dB(A), 95% CI [5.6, 21.2] dB(A), fixed effects test: F(1.22) = 19.65, p = 0.0002; L50: β = 13.4 dB(A), 95% CI [5.7, 21.1] dB(A), fixed effects test: F(1.22) = 17.35, p = 0.0004 and L90: β = 13.4 dB(A), 95% CI [5.7, 21.1] dB(A), fixed effects test: F(1.22) = 13.68, p = 0.001. We found significant differences between the country factor, which confirms a higher level of noise between the measurement points in Chile and those observed in Argentina: LeqA: β = 20.8 dB(A), 95% CI [11, 30.6] dB(A), fixed effects test: F(1.22) = 39.6, p = 2.46e−6; L5: β = 26.1 dB(A), 95% CI [16.2, 36] dB(A), fixed effects test: F(1.22) = 57.1, p = 1.5e−7; L10: β = 24.6 dB(A), 95% CI [14.7, 34.6] dB(A), fixed effects test: F(1.22) = 50.16, p = 4.19e−7; L50: β = 17.8 dB(A), 95% CI [7.92, 27.6] dB(A), fixed effects test: F(1.22) = 30.53, p = 1.48e−5 and L90: β = 10.8 dB(A), 95% CI [1, 20.7] dB(A), fixed effects test: F(1.22) = 17.4, p = 0.0003.

However, when making multiple comparisons between location to source (Chile: Nearby points/Far points; Argentina: Random), we only found significant differences between—Nearby point—and—Random—when condition is—with fireworks—(Leq: ΔM = 21.27 dB, 95% CI = [20.47, 32.62] dB, t(8) = − 9.97, p = 5.72e−6—are presented in Fig. 3C; L5: ΔM = 26.62 dB, 95% CI = [25.76, 38] dB, t(8) = − 11.84, p = 1.14e−6; L10: ΔM = 25.12 dB, 95% CI = [24.34, 36.47] dB, t(8) = − 11.46, p = 2.07e−6 and L50: ΔM = 18.14 dB, 95% CI = [17.2, 29.55] dB, t(8) = − 8.59, p = 1.35e−5).

4 Discussion

In this study we present acoustic measurements carried out in various places in the city of CABA during the Christmas and New Year festivities of the year 2020. In Argentina, these events are characterized by the widespread use of fireworks, which contribute significantly to the overall ambient noise levels. In addition, we also present acoustic measurements conducted during the New Year celebration in the city of Santiago, Chile, where the government regulates the use of fireworks through a specific event held at a designated time and location in the city. Comparing the widespread use of fireworks to a singular event allows us to understand the differences between these scenarios and analyze their respective implications, problems, and benefits.

Our results in CABA showed a clear increase in noise levels during the Christmas and New Year's festivities. Measurements conducted at various locations throughout the city revealed a significant rise in ambient noise during the midnight hours, where a peak in noise levels was observed, coinciding with the time when numerous fireworks were being set off in different areas of the city. These noise peaks remained elevated for a considerable period (approximately 1 h). This result suggests that, despite the absence of legal restrictions on the use of fireworks by the public, the majority of individuals primarily use fireworks at midnight during the celebration of Christmas and New Year's. The question that arises from these results is what the appropriate temporal window should be to measure the impact of fireworks in such situations. On average, the duration of the events was approximately 1 h for both Christmas and New Years, as observed in the figure where the noise values take that time to return to their original levels. Therefore, it seems reasonable to use a 60-min time window to measure this parameter, as has been utilized in previous studies [6]. However, our findings demonstrated similar values regardless of the temporal window used (windows ranging from 15 min to 1 h). This may indicate that the first 15 min had a significant influence on the overall average value. Thus, our results suggest that measuring the first 15 min after midnight on festive days such as Christmas and/or New Year's is sufficient to capture the energy of these fireworks events. Reducing the measurement time would benefit the economic, human, and logistical costs required by a government entity to control this type of noise.

Our results demonstrate a significant increase in all measured parameters before and after 12:00 AM for both Christmas and New Year. We did not find significant differences between Christmas and New Year's, or in the hours leading up to 12:00 AM during the festivities after that time. These results indicate that in Argentina, both major festivities (the largest ones of the year) are celebrated in similar ways by the population. The comparison of environmental noise during Christmas and New Year with noise levels measured on a non-festive day yielded interesting results. Firstly, we observed that before 12:00 AM, the noise levels during Christmas and New Year's were significantly lower than on a regular day. While this result was unexpected, it is logical considering that on a festive day, families gather in their homes and there are fewer people on the streets than on a normal day. Indeed, noise levels in the city of Buenos Aires on a non-holiday day were measured at 15 dB above the limits established by city regulations. These results are similar to previous studies [18, 19] and should be an alert for governments in order to raise awareness and generate policies that mitigate this problem. Another unexpected result was that when comparing the noise levels of a regular day obtained using the overall parameter LeqA, there were no significant differences with the maximum noise levels measured after 12:00 AM. Again, this result highlights the high noise levels present in the city of Buenos Aires on non-festive days. Additionally, it emphasizes the inadequacy of the LeqA parameter in measuring the acoustic impact of fireworks, as it underestimates the noise levels caused by impulsive sounds. Therefore, we decided to reanalyze the data using other parameters, such as percentiles, as suggested by previous studies for similar events [20, 21]. Our results show significant differences with the ambient noise level using the L01 percentile when comparing a regular day to the noise levels measured after 12:00 AM. This finding demonstrates that this parameter is more effective than the LeqA parameter in characterizing impulsive events.

Another relevant finding from our study is that we did not observe differences in the noise levels caused by fireworks on Christmas and New Year's between the measurement points. This indicates that regardless of the location in the city, individuals will be exposed to high levels of noise generated by fireworks. This is one of the main disadvantages of the widespread use of fireworks, as the noise levels impact the entire population, in contrast to cities where fireworks celebrations are held in controlled events. This can be clearly observed when comparing the results obtained in the city of CABA with the noise levels measured in the city of Santiago, Chile during the New Year celebration, where fireworks are organized as a state-sponsored fireworks display at a specific location. Although in Santiago, Chile, noise levels near the event are considerably higher compared to measurements in Argentina (due to the magnitude of the event), these values decrease significantly as the measuring points move away from the noise source. The measurements in Santiago clearly show how the ambient noise levels caused by fireworks decrease as the measurement point moves away from the focal point of the display. Our results demonstrate that at 11 blocks away from the fireworks display (1.1 km), noise values similar to those measured in Argentina were obtained, indicating that beyond that distance, the population begins to be exposed to lower noise levels. However, it is crucial to acknowledge that the propagation of noise generated by fireworks exhibits substantial variability, strongly influenced by climatic factors such as wind, which were not accounted for in our study. Furthermore, it is essential to consider the diverse nature of the fireworks used in Argentina. In some instances, mortars are employed to launch projectiles that detonate in the upper atmosphere, with the projectile's power primarily dependent on the mortar's diameter. Conversely, ground-based fireworks may also be used, exhibiting varying power levels associated with their distinctive structures. Consequently, we exercise caution when generalizing the results of this study to other settings with different climates and fireworks that may differ from those in Argentina and Chile. To the best of our knowledge, there have been no systematic acoustic studies addressing the wide variety of fireworks in Argentina. Therefore, future research should prioritize conducting precise measurements and detailed analyses to comprehensively understand and characterize the diverse types of fireworks and their noise emissions.

It is interesting to compare both strategies regarding the use of fireworks in relation to the potential health and environmental harm they may cause to humans and animals. One of the primary health issues associated with fireworks pertains to injuries incurred by the general public, including burns, eye damage, and mutilations. In this context, children, especially those over the age of 10, emerge as the most vulnerable group to firework-related injuries [22]. Bagri et al. [22] suggest that this problem could be effectively addressed by enacting laws that prohibit the public sale of all fireworks and restrict their private use, while promoting public displays conducted by professional pyrotechnicians. It would be interesting to compare the number of injuries during festive dates between Santiago, Chile, and Buenos Aires, where regulations concerning fireworks differ significantly. Another issue stemming from fireworks is the adverse effect observed in individuals with Autism Spectrum Disorder (ASD), commonly known as autism. It has been reported that 70% of children with ASD experience hypersensitivity to auditory stimuli, particularly those with high-functioning ASD (individuals with better social adaptation and lower impairment in daily life) [23, 24]. Once again, significant disparities are evident between widespread use and organized events. As observed, in Argentina, fireworks noise was present at all measurement points, whereas in Santiago, Chile, there was only one source of fireworks-related noise. This means that the latter strategy affects a considerably smaller number of residents. Furthermore, such events follow a predictable schedule, allowing families with vulnerable individuals to take necessary precautions. Finally, fireworks have detrimental effects on animal health. The most widely recognized case is the hypersensitivity of dogs to fireworks [2, 25, 26]. In the study by Sulser et al. [27], it is noted that approximately half of urban dogs exhibit some of the aforementioned symptoms. While the majority of studies focus on domestic animals like dogs and cats, they are not the only ones experiencing adverse effects from fireworks. Reports indicate harmful impacts on various species of birds and aquatic birds [28], marine animals inhabiting coastal areas of Chile [29], horses [30], and anxiety or panic effects have also been observed in rhinoceroses, cheetahs, elephants, and rodents [31], among others. Once again, these issues are much more severe in scenarios of widespread fireworks use compared to their occasional, regulated use. In summary, of all the issues related to fireworks, they appear to be less significant in organized, regulated usage compared to their widespread use by the general population. Therefore, we would recommend the former of the two options to mitigate many of the health and environmental problems caused by fireworks.

5 Conclusions

Our results demonstrate a clear increase in ambient noise caused by fireworks after midnight, both during Christmas and New Year's Eve. The data collected from various locations clearly indicate the widespread use of fireworks by the population in these festivities in CABA. It is important to note that the recorded noise levels during the festivities exceeded the limits established by environmental regulations by a significant margin. This raises concerns about the potential negative effects on auditory health and the well-being of the population exposed to such high noise levels. This mass practice of pyrotechnics poses various health problems, particularly in terms of burns and injuries that may occur, particularly in sensitive areas such as the eyes and hands. Furthermore, the extensive use of fireworks during these festivities also leads to significant drawbacks in terms of noise pollution, which significantly affects vulnerable groups, such as individuals with autism spectrum disorder and animals, particularly dogs. These findings highlight the need for stricter control and regulation measures regarding the use of fireworks during these celebrations. Additionally, they emphasize the importance of raising awareness among the population about the harmful effects of excessive noise and promoting safer and more environmentally-friendly alternatives for celebrating these festive occasions.

On the other hand, organized fireworks displays by authorized entities have, in our understanding, several important advantages. While they do contribute to air and noise pollution, the fact that these displays are programmed and managed by professionals mitigates several issues associated with the widespread use of fireworks by the general public. One notable advantage is the significant reduction in accidents such as burns and eye injuries, which are common with unregulated fireworks usage. These organized displays are conducted under controlled conditions, ensuring the safety of both the professionals handling the fireworks and the spectators in attendance. Moreover, the noise generated by these displays tends to affect a smaller number of people, as they are typically concentrated in a specific location rather than dispersed throughout the entire city. Additionally, since these displays are scheduled for a specific time, individuals who may be sensitive to noise or have concerns can take measures to minimize the potential negative effects. We consider organized fireworks displays to be a safer alternative for enjoying fireworks and pyrotechnics. Another alternative would be a complete ban on their use. However, implementing such a ban raises other issues related to the loss of livelihood for many families who depend on the fireworks industry. In light of these considerations, finding a balanced approach that takes into account safety, environmental concerns, and the socio-economic impact on the community becomes crucial. This may involve implementing stricter regulations, promoting public awareness campaigns about responsible fireworks usage, and exploring alternative forms of entertainment that can provide similar enjoyment while minimizing the negative consequences associated with unregulated fireworks usage.