Noise Pollution from Diesel Generator Use During the 2024–2025 Electricity Crisis in Ecuador

Abstract

Hydropower is the primary source of electricity in several countries in Latin America. Hydropower provides approximately 80% of Ecuador’s electricity; however, it remains highly vulnerable to climate change, resulting in uncertainties in power generation due to altered precipitation patterns, runoff, and systematic failures. Consequently, Ecuadorians are becoming increasingly reliant on diesel generators during crises, resulting in public health, safety, and economic impacts, as well as social and political disruptions. This study evaluated noise pollution in the central urban area of the city of Loja for the first time during the 2024–2025 electricity crisis in Ecuador. A Type 1 integrating sound-level meter was used to monitor noise pollution (LAeq, 10min) at 20 locations during periods of generator operation and non-operation. At each location, the number of generators, the density of commercial activities along the streets, as well as traffic and other urban characteristics, were recorded. Results revealed that the presence of generators, street width, and the number of generators significantly increased the LAeq, 10min, often exceeding the limits set by the World Health Organization and Ecuador’s environmental regulations. Frequency spectrum analysis revealed that medium frequencies increased with A-weighting, while low frequencies rose with C-weighting, suggesting potential health risks to the local population. The thematic noise map during generator inactivity showed lower noise levels, averaging around 71.5 dBA. Conversely, when the generators were operational, noise levels exceeded 79.6 dBA, indicating a significant increase in environmental noise exposure associated with their use. This highlights an urgent need to implement and expand renewable energy sources, as existing options like wind power, photovoltaic energy, and biomass are insufficient to meet community demands.

1. Introduction

Urban noise constitutes a significant environmental stressor, with documented implications for auditory health, physiological function, and cognitive performance [1,2,3,4,5]. The World Health Organization (WHO) identifies noise as a significant public health concern, linking chronic exposure to sleep disturbances, cardiovascular disease, hearing loss, and cognitive impairment [1,6,7,8].

In Latin American cities, average sound pressure levels (SPLs) often exceed recommended limits due to rapid and unregulated urbanization, limited enforcement of environmental norms, and insufficient policy integration [9]. Although most research in the region has centered on vehicular traffic as the primary source of urban noise pollution [10,11,12,13], unusual circumstances can give rise to alternative sources. In the last 30 years, Ecuador has experienced at least seven major electricity shortage events. Between 1995 and 1996, a severe drought linked to El Niño drastically reduced the water level of reservoirs [14]. Similarly, in 2002 and 2009–2010, low reservoir levels during the dry season, coupled with Ecuador’s worst drought in 40 years, attributed to El Niño, resulted in a 40% drop in the power supply, leading to daily blackouts of 2–6 h [15,16,17]. Consequently, emergency electricity imports from Colombia and Peru, combined with a nationwide economic downturn, resulted in tens of millions of dollars in losses. On the other hand, there are construction hydro plant deficits, aging infrastructure, and inadequate maintenance planning, resulting in operational failures that reduce generation capacity and the enforcement of emergency measures. In 2023, Ecuador experienced another severe drought caused by El Niño, resulting in daily power blackouts of up to 8 h. Finally, between September and December 2024, Ecuador experienced a nationwide energy crisis caused by the worst drought in 61 years, which was exacerbated by climate change and the return of El Niño. The prolonged drought severely reduced hydroelectric power generation, which accounts for over 80% of the country’s electricity supply [12,13,18]. The resulting blackouts, lasting up to 14 h per day for over 70 consecutive days, are estimated to have caused an economic loss of $2 billion, with severe public health and environmental impacts. During this period, the urban, commercial, and industrial sectors led to widespread reliance on portable electric generators, which were often installed without acoustic control or regulatory oversight [17,19].

Studies on noise pollution caused by generators at the global scale and in Latin America are scarce; however, among backup and decentralized energy systems, diesel generators are by far the worst sources of noise pollution. In general, attention has been given to the impacts on air quality, including the emission of CO, SO₂, NO₂, smog, and particulate matter produced by vehicles and the industrial sector, which also represent the primary sources of noise pollution [18,19,20,21]. Only during the last decade has the impact of noise created by diesel generators been primarily assessed within the occupational health context, including auditory impairment, stress, and health problems. In Latin America and Ecuador, the effects of electricity shortages have also been focused on air pollution and energy security. Hence, there is an evident research gap in the impacts of generators’ noise produced during electricity crises on health and the environment at the regional and national levels. This places the present study as the first comprehensive noise pollution assessment conducted in Loja, Ecuador, during the country’s most significant energy crisis. This research highlights and quantifies the acoustic impact of noise pollution caused by portable and stationary generators used as an emergency energy source across residential, commercial, and institutional zones. A real-time thematic sound map illustrates noise spatial variability in generators’ use over time, as well as the relationship between urban acoustic environment factors, features previously absent from regional assessments. Portable generators powered by internal combustion engines emit high-intensity noise through multiple mechanical elements, including the engine block, exhaust system, alternator, and vibration-prone components. Generally, to meet the ever-increasing energy demand, power generation is often replaced with fossil fuel-powered generators [22]. In this context, previous studies have documented the negative impacts of generator use on environmental quality (noise and air pollution) due to outdated technologies and low-quality diesel fuel [23,24,25]. Several studies in developing countries have focused on noise pollution in urban environments, particularly in areas sensitive to noise, such as hospitals and commercial buildings. Reported SPLs have ranged from 72.6 to 115.6 dBA [24], 70.8 dBA [26], 100–120 dBA [27], 70–90 dBA [28], which exceeded the recommended Permissible Noise Exposure Limits (PNELs) set by the WHO, USEPA, and EN [28].

Ecuadorian regulations prescribe that the maximum SPLs in commercial and industrial zones range between 55 and 75 dBA, as outlined in the Ministerial Agreement 097-A [29]. Municipal ordinances from neighboring countries and cities with similar population sizes, cultural heritage, colonial architecture, and altitudes are comparable. The maximum SPLs in commercial and industrial zones range from 65 to 68 dBA in Sucre, Bolivia [30]; 60 to 80 dBA in Cusco, Peru [31]; 60 to 75 dBA in Popayán, Colombia [32]; and 55 to 75 dBA in San Juan, Argentina [33]. These periods cover the day and nighttime. However, there is an evident gap in specific regulations governing noise emissions from portable generators within densely populated urban environments. In contrast, EU Regulation 2024/1208 stipulates maximum permissible sound levels for outdoor machinery, including generator sets, based on ISO 8528-10:2022 and referenced regarding noise in ISO 8528-13:2016 norm [34]. Several variables, such as street morphology, building configuration, traffic intensity, and source overlap, shape urban noise propagation [35,36,37,38].

These factors converge critically in intermediate cities like Loja in southern Ecuador. The combination of narrow streets, mid-rise buildings, and concentrated commercial activity facilitates acoustic confinement. In this context, the simultaneous operation of multiple generators without spatial planning results in localized noise saturation, posing increased health risks for vulnerable populations such as older adults, individuals with preexisting conditions, and workers exposed to extended periods of elevated noise levels [39,40]. By identifying zones and quantifying exposure levels that exceed acceptable limits, effective mitigation strategies can be developed to protect vulnerable populations. At the national level, few studies have evaluated the effects of infrastructure and dynamic noise contributors. For example, in cities such as Latacunga [41], Cuenca [42], Machala [43], Quito [44,45], and in Loja, the effects of noise caused by urban structure and transport have been assessed in residential, commercial, industrial, and service areas [46]. The primary objective has been to highlight the lack of public awareness and the inadequate enforcement and monitoring. Thus, this study provides empirical data to guide regulatory updates on permissible noise levels during non-standard operating conditions, such as energy shortages. It also points out the environmental trade-offs associated with fossil-fuel-based backup systems. The contrast of acoustic profiles observed during generator use and ambient conditions, identifying statistical correlations with urban context variables, supports evidence-based noise control in future crisis scenarios. This study aims to quantify generator-induced noise levels across different urban, commercial, and service areas during a severe electricity crisis. It also evaluates the compliance of noise pollution during periods of generator operation and non-operation with national and WHO noise regulations under emergency conditions. Finally, it identifies noise intensity in critical exposure zones and assesses temporal patterns of generator usage, considering urban, institutional, and infrastructural challenges.

2. Materials and Methods

2.1. Study Area

The study was conducted in the central urban area of Loja, located in southern Ecuador, with an estimated population of 250,028 [47]. Observational surveys were conducted in various urban sectors to identify potential sampling sites. The central district was selected because it concentrates most of the city’s commercial, administrative, and service activities, resulting in a high flow of pedestrians and vehicles throughout the day. Consequently, this area represents the area with the greatest potential exposure to urban noise, particularly from the operation of electric generators during power outages. In contrast, the peripheral sectors of the city are predominantly residential, with little or no presence of generators.

Figure 1 illustrates a total of 20 monitoring points strategically placed within an area of influence of 0.71 km². These locations were selected along street segments with varying generator densities and commercial characteristics to capture the variability of acoustic conditions associated with the different urban environments in the central area of Loja.

Figure 1. Study area indicating the spatial distribution of 20 noise monitoring points (orange circles) in central area of the city of Loja.

2.2. Data Collection

Noise measurements were performed under favorable meteorological conditions, avoiding rainfall events using a Delta OHM HD2010UC/A Type 1 integrating sound level meter, in accordance with Annex 5 of Ministerial Agreement No. 097-A and IEC 61672-1:2013 [23]. The instrument used has a dynamic measurement range of 20 to 140 dB, with an accuracy of ±1.0 dB, and supports A, C, and Z weighting filters, provides octave band analysis (31.5 Hz–8 kHz), statistical outputs (L₁–L₉₉), and third octave band analysis (25 Hz–12.5 kHz). In addition, the sound level meter incorporates a ½ inch prepolarized free-field condenser microphone, which was placed on a tripod at 45 and 90 degrees, depending on the sampling scenario to capture sound from multiple directions at a height of 1.5 m above ground level and 1 m from vertical surfaces to minimize reflective interference, and taking into account a minimum distance of 3 m from the nearest generator. The Delegated Regulation (EU) 2024/1208 recommends a measurement distance of 7 m, including multilateral positioning for directional assessment. However, urban constraints and the density of generators made it impractical to evaluate each unit individually. Therefore, environmental noise was assessed by encompassing as many generators as possible within a 10 m buffer area, defined and calculated using ArcMap 10.4.1. Figure 2 shows the installation of the sound level meter at one of the points considered during the study.

Figure 2. Installation of the sound level meter in monitoring points in central area of the city of Loja.

To evaluate noise levels, the sound level meter was calibrated using a Delta OHM HD9102 acoustic calibrator (94 dB at 1 kHz, ±0.3 dB) and subsequently configured to record data every 10 s during continuous 10 min measurement periods (Table 1). The primary acoustic parameter established was the equivalent continuous sound pressure level in the A-weighting filter (LAeq, 10min).

Table 1. Full specifications of the sound level meter and calibrator used.

Instrument Model	Measurement Range	Accuracy	Weighting Filters	Sampling Intervals
Delta OHM HD2010UC/A	20–140 dBA	±1.0 dB	A, C, Z	10 s
Delta OHM HD9102	94 dB (1 kHz)	±0.3 dB	-	-

Noise measurements were taken under two comparable experimental conditions: (a) during generator operation, corresponding to power outage periods, and (b) without generator operation, corresponding to periods with normal power supply once service was fully restored. Both campaigns were carried out under similar conditions at the same 20 monitoring points, during the daytime period from 9:00 a.m. to 6:00 p.m., Monday through Friday, coinciding with peak commercial activity and similar traffic and pedestrian flow to ensure environmental equivalence. Given the variability and unpredictability of power outage schedules, priority was initially given to collecting measurements during generator operation. Outage times often fluctuated from day to day and could extend for four to six consecutive hours in each sector, creating uncertainty and increasing the risk of missing data collection opportunities. Therefore, sampling during blackout periods was flexibly scheduled according to the official outage calendar, allowing the collection of six replicates per monitoring point under generator operation.

Once these measurements were obtained, control measurements (without generator) were scheduled at the same time of day, using the same monitoring points and identical configurations, to ensure temporal equivalence. Identical protocols were applied in both experimental conditions: the same sound-level meter configuration, same calibration procedure, same microphone height (1.5 m), same distance to reflective surfaces (1 m), and same traffic flow count. Therefore, the only experimental difference between the two sets of measurements was the presence or absence of generator operation.

In addition to sound data, several acoustic and contextual variables were collected (Table 2) following the methodology proposed by Torija et al. [48] and Acosta et al. [49].

Table 2. Parameters and considerations for data collection.

Parameters	Measurement Method	Classification	Configurations
Average building height	shapefile analysis
Street width and Sidewalk width	BOSCH Blaze GLM165-25G laser distance meter
Street typology		0.2 for U-shaped	buildings on both sides
		0.4 for L-shaped	buildings on one side
		0.6 for J-shaped	tall buildings on both sides
		0.8	Streets with no adjacent buildings
Commercial activity intensity	systematic street surveys to define commercial density areas	0.2 for high	>31 establishments
		0.4 for medium-high	19–30 establishments
		0.6 for medium-low	10–18 establishments
		0.8 for low-density areas	<9 establishments
Number of generators		1–11
Vehicular traffic	Manual counting	Heavy vehicles	Urban buses, trucks, and trailers
		Light vehicles	Cars
		Motorcycles	Motorcycles

2.3. Data Analysis

Raw data were downloaded using Noise Studio software 2.0 and processed using Microsoft Excel. LAeq, 10min values were calculated using Equation (1), where T represents the total duration of the measurement, Leq₁, Leq₂, …, and Leqn correspond to the equivalent sound pressure levels for each 10 s integration interval, and t represents the fractional time of each interval, such that the sum of t equals T:

$$\mathrm{L}\mathrm{e}\mathrm{q}=10\ast \mathrm{l}\mathrm{o}\mathrm{g}({\displaystyle \frac{1}{\mathrm{T}}}({10}^{{\displaystyle \frac{\mathrm{L}\mathrm{e}\mathrm{q}1}{10}}}\ast \mathrm{t}1+{10}^{{\displaystyle \frac{\mathrm{L}\mathrm{e}\mathrm{q}1}{10}}}\ast \mathrm{t}2+\cdots +{10}^{{\displaystyle \frac{\mathrm{L}\mathrm{e}\mathrm{q}1}{10}}}\mathrm{t}\mathrm{n}))$$

(1)

Additionally, a frequency spectrum analysis in octave bands was performed on a subsample of four monitoring points using A- and C-weighting scales, categorized according to the defined street types.

The reference noise limits used in this study were based on Ecuador’s Ministerial Agreement 097-A, Annex 5, which specifies 60 dBA for commercial land use during the day, and the World Health Organization guidelines, which recommend 55 dBA for the daytime period.

The spatial representation of the measured sound pressure levels was performed using Software ArcMap 10.4.1. Thematic noise maps were generated using the Inverse Distance Weighted (IDW) interpolation method. This approach was selected because it provides a consistent and realistic representation of spatial variability when working with a limited number of sampling points. In this study, the total of 20 monitoring locations did not meet the minimum data requirements for reliable semi variogram estimation, making the application of Ordinary Kriging unsuitable. Therefore, IDW was adopted as an appropriate deterministic method for environmental noise representation, assuming that closer measurements exert a stronger influence on the predicted values than those farther away. The interpolation parameters were configured as default with a power of 2, a fixed search radius including the 12 nearest points, and an output cell size of 5 m. The interpolation extent was restricted to the 0.71 km² study area, using as input the mean LAeq, 10min values averaged from six replicates per site for each experimental condition (with and without generator operation). The resulting raster layers were classified into five categories to allow a clearer visualization of spatial variations in noise levels, using a continuous color gradient from green (lower sound levels) to red (higher sound levels). This configuration preserved local variability and ensured that the final thematic map reflected the spatial distribution of measured environmental noise levels.

A violin plot was used to visualize the relationship between LAeq, 10min, and generator presence (with and without the generator). A U Mann–Whitney test was applied to compare mean LAeq, 10min values with and without the generator, following confirmation of no normal distribution using the Shapiro–Wilk test (p value < 0.05). A dispersion plot was used to visualize the relationship between number of generators, vehicular traffic, average building height, street typology, street width, sidewalk width, and commercial activity intensity and LAeq, 10min. The influence of number of generators, vehicular traffic, average building height, street typology, street width, sidewalk width, and commercial activity intensity and on LAeq, 10min was examined using generalized linear models (GLMs). The LAeq, 10min was modeled assuming a gamma error distribution with a log link function, which is appropriate for continuous data. Model selection was guided by Akaike’s Information Criterion (AIC), which ensured the identification of a minimally adequate model. All statistical analyses were conducted using R version 3.2.2. [50].

3. Results and Discussion

The violin plot showed that the highest LAeq, 10min values were found in zones with generator presence, decreasing in zones without generator presence (Figure 3). The broader sections of both violin plots indicate that most recorded measurements exceeded the maximum permissible limits set at 55 to 60 dBA,10min. Similarly, the Mann–Whitney test revealed a statistically significant difference between the medians of LAeq, 10min in zones with generator presence and those without a generator. In accordance with this, previous studies recorded similar values of LAeq, 10min when the generators were active [24].

Figure 3. Violin plot for LAeq, 10min in zones with and without generator presence. Boxes span the first to third quartiles; the horizontal line inside the boxes represents the median and the whiskers show a 1.5 × interquartile range (IQR) in the rest of the data. On each side of the black line, a kernel density estimation illustrates the shape of the data distribution. Wider sections of the violin plot indicate a higher concentration of observations, whereas narrower sections reflect lower density.

The scatter plot showed a positive relationship between the number of diesel generators and LAeq, 10min (Figure 4A). This relationship was confirmed with the GLMs, with generator number and commercial activity intensity positively influencing LAeq, 10min. Additionally, street width had a significant and negative effect on LAeq, 10min (Figure 4B, Table 3). Similarly, Giwa et al. [24] reported that noise levels from the generators were significantly higher, indicating that their operation contributed to noise pollution. The scatter plots representing vehicular traffic and average building height also exhibited a positive relationship pattern; however, it was less pronounced (Figure 4B,C).

Figure 4. A scatter plot showed the relationship between LAeq, 10min and quantitative road infrastructure variables. (A) generators vs. LAeq, 10min (positive significant relationship); (B) vehicular traffic vs. LAeq, 10min; (C) average building height vs. LAeq, 10min; (D) street typology vs. LAeq, 10min; (E) street width vs. LAeq, 10min (negative significant relationship); (F) sidewalk width vs. LAeq, 10min and (G) commercial activity intensity vs. LAeq, 10min (positive significant relationship).

Table 3. Summary of the Generalized Linear Models (GLMs) of the LAeq, 10min modeled with vehicular traffic, generators, average building height, street typology, street width, sidewalk width, and commercial activity intensity as predictors. SE = Standard error.

Factor	Coefficient	SE	t-Value	p-Value
(Intercept)	4.3016304	0.0432308	99.504	<0.0001
Generators	0.0240462	0.0030656	7.844	<0.0001
Vehicular traffic	0.0004021	0.0002229	1.804	0.07404
Average building height	−0.0078215	0.0057081	−1.37	0.17342
Street typology	−0.0612243	0.0533588	−1.147	0.25373
Street width	−0.0075989	0.0027241	−2.79	0.00623
Sidewalk width	0.0154361	0.0107814	1.432	0.15508
Commercial activity intensity	0.1073906	0.0500134	2.147	0.03399

When comparing locations according to urban configuration, narrower streets (width <10 m) systematically showed higher noise levels, frequently exceeding 80 dBA, while wider streets (>10 m) provided noticeable acoustic relief. This pattern confirms that street width is one of the strongest determinants of noise propagation, as confined corridors increase reverberation and reduce energy dissipation. In contrast, the relationship between generator density and LAeq, 10min was more variable. Although sites with more units tended to register slightly higher sound levels, the increase was not proportional to the number of generators, suggesting that urban geometry and surrounding activity play a greater role than the quantity of sources themselves. For instance, locations along Bolívar and Bernardo Valdivieso—characterized by narrow streets (3–5 m) and dense façades —showed LAeq, 10min values of 82–85 dBA, even with moderate generator counts, while broader corridors, such as Juan de Salinas or Colón–Av. Universitaria remained below 78 dBA despite comparable or higher generator presence.

On the other hand, the line in the scatter plot showed a slight downward slope with commercial activity intensity. This observation indicates a weak negative association between the categories of commercial density and the recorded LAeq, 10min values (Figure 4G). The results showed that as commercial density decreases (particularly in low-density areas, classified as having a density of 0.8), the LAeq, 10min values also show a slight decrease. However, there is a notable vertical spread across the different commercial density categories, with numerous points representing varying LAeq, 10min values. This suggests high variability in sound level measurements across all groups or conditions.

Reported studies have not examined the noise impact associated with diesel generators in relation to the effects of street width and street typology. However, it has been proven that the influence of road width, traffic density, and building arrangement has a significant effect on urban noise levels. Broader streets generate more annoyance among residents than narrow streets [36]. Noise can be mitigated more effectively when buildings are distributed throughout urban areas rather than clustered. High traffic volumes and increased road densities exacerbate traffic noise in these environments [38]. The results of the present study align with reported cases. For example, street typology showed a slight negative relationship, indicating that streets with no adjacent buildings recorded the highest noise values. Other factors showed a neutral relationship.

The analysis of octave frequency distribution (Figure 5), using A-weighting, shows that in the absence of generators, noise levels remain below 60 dBA across the entire analyzed frequency spectrum (25 Hz–4 kHz), with frequency peaks around 500 Hz (Figure 5A,C,D) and 1 kHz (Figure 5B). In contrast, during generator use, frequencies between 250 Hz and 4 kHz exceed 60 dBA at all measurement points, with dominant frequencies ranging from 1 kHz (Figure 5A,B) to 2 kHz (Figure 5C,D). C-weighted frequencies, on the other hand, remained below 68.5 dBA across the 25 Hz–4 kHz range in the absence of generators, with dominant peaks observed at 25 Hz (Figure 5C), 63 Hz (Figure 5B,D), and 125 Hz (Figure 5A). However, during generator operation, dominant frequencies shifted to the 32 Hz band (Figure 5D) and 63 Hz band (Figure 5A–C), with noise levels exceeding 75 dBA.

Figure 5. Comparison of A- and C-weighted frequencies across different street types in central area of the city of Loja. (A): P10–U–shaped street; (B): P7–L–shaped street; (C): P11–J–shaped street; and (D): P4–J–shaped street.

The frequency spectrum analysis confirmed the substantial contribution of generator use to elevated noise levels, particularly in the mid-frequency range (A-weighting) and low-frequency range (C-weighting). Although elevated, results remained below the 4–6 kHz frequency band typically associated with noise-induced hearing loss (NIHL). Nevertheless, existing literature suggests that prolonged exposure to high noise levels can extend the range of auditory damage from 500 Hz to 2 kHz [6], leading over time to progressive and irreversible hearing impairment [51]. This phenomenon is primarily attributed to the continuous noise generated by operating generators, which impairs the auditory system’s ability to recover from acoustic stimuli [52]. Consequently, this can increase stress levels, feelings of fatigue, and lead to other adverse health effects such as anxiety, hypertension, cardiac arrhythmias, and tinnitus—which is considered a clinical manifestation of auditory damage [7,8,53,54]. Beyond overall sound pressure levels, the frequency spectrum analysis showed that generator operation primarily amplified mid- and low-frequency bands (250 Hz–2 kHz under A-weighting and 32–125 Hz under C-weighting). These components interact with the reflective facades of narrow streets, prolonging resonance and increasing perceived loudness even when measured LAeq values differ only slightly. Such frequencies are linked to annoyance, sleep disturbance, and cardiovascular stress, indicating that health effects may stem as much from frequency composition and urban confinement as from sound levels.

Low-frequency noise presents a distinct problem due to its ability to propagate through the ground, atmosphere, and built structures, leaving residual low-frequency vibrations in the environment that are often imperceptible to the human ear but capable of inducing structural and environmental degradation [55,56]. This study showed that the highest amplitude peaks were consistently recorded across all sites at 32 Hz and 63 Hz, reaching a maximum of 84 dBA. These levels are comparable to occupational exposures in factory environments where workers are subjected to low-frequency noise (≤500 Hz) at 90 dBA and have reported symptoms including ear and head pain, cognitive decline, and vibroacoustic disease (VAD), a condition characterized by abnormal extracellular matrix growth leading to the deterioration of blood vessels, the heart, lungs, and kidneys [57,58,59].

The thematic sound map during periods when the generators were inactive showed lower noise levels compared to periods of generator operation (79.6 dBA), with values around 71.5 dBA (Figure 6A), concentrated along busy streets and avenues, particularly in the southeastern and western areas of the city. Conversely, when the generators were operational, a prominent noise hotspot appeared in the central region, exceeding 79.6 dBA (Figure 6B). At the same time, the peripheral areas to the north, east, and south exhibited lower levels, ranging below 76.5 dBA. These spatial patterns align with the point-based measurements obtained during the study, which indicated that noise levels during generator operation ranged from 71.2 to 82.4 dBA, with an average of 77.6 dBA. This represents a significant increase in environmental noise exposure related to generator use compared to baseline data from 2023, as reported by del Pozo et al. [46], which recorded values between 64.3 and 72.7 dBA (mean: 66.8 dBA) in the commercial areas of downtown Loja. This represents a significant increase in environmental noise exposure linked to generator use. Similarly, two previous studies conducted in Ecuador reported a rise in environmental problems (e.g., air pollution) in the city of Quito during the nationwide electricity crisis, primarily due to the widespread use of diesel generators [18,19].

Figure 6. Thematic sound map showing spatial variability in central area of the city of Loja. (A) Without generators; and (B) with generators.

Finally, the results provide valuable input for municipal zoning and regulatory frameworks, supporting the establishment of criteria or buffer distances for generator placement in historical or densely built sectors. Moreover, it highlights the need to regulate both the temporary and permanent location of generators during electricity shortages to prevent cumulative impacts on environmental quality. In a broader context, these results serve as an essential reference for future scenarios, which are likely to recur as climate change intensifies drought events in countries like Ecuador, which heavily depend on hydropower generation. Integrating this spatial understanding with urban planning, traffic control, and the promotion of renewable micro-grids would help reduce cumulative noise exposure during future energy crises.

Some limitations should be acknowledged. The simultaneous presence of traffic, pedestrian activity, and other mechanical sources made it difficult to fully isolate the acoustic contribution of the diesel generators. Moreover, the study focused exclusively on daytime periods during the electricity crisis, which limits temporal representativeness and does not capture potential nighttime exposure. Spatially, monitoring was restricted to the central commercial district, characterized by narrow streets, dense building facades, and confined sound propagation; therefore, the findings may not be directly applicable to more open or residential sectors.

Future research should focus on long-term acoustic monitoring to capture behavioral variations in generator use, especially during recurring energy shortages. Expanding the study to other cities in Ecuador and Latin America would enable a comparative analysis across urban morphologies and socioeconomic contexts. Additionally, evaluating the effectiveness of noise mitigation strategies—such as acoustic barriers, zoning adjustments, and public awareness campaigns—could inform targeted interventions.

4. Conclusions

This study reveals that the extended use of diesel generators significantly worsened noise pollution in Loja, Ecuador, during the 2024–2025 hydroelectricity crisis. When generators were operating, a distinct noise hotspot exceeding 79.6 dBA emerged in the central area, exposing hundreds of people to levels above the limits established by the WHO and Ecuadorian environmental regulations. Although generator use clearly contributed to the increase in ambient sound levels, the results indicate that urban morphology, street width, and traffic intensity played an even greater role in shaping overall noise propagation. In several cases, narrow streets and continuous vehicular flow amplified background noise, regardless of the presence of a generator, highlighting the complex interaction of multiple urban sound sources.