Seasonal variation and structural influences on indoor radon exposure in residential buildings of Khuzestan Province

Abstract

Radon, a radioactive gas and the second leading cause of lung cancer, poses significant health risks in enclosed spaces. This study examines radon concentrations in residential buildings across Khuzestan Province, Iran, exploring seasonal variations and building-related influences, to establish a baseline radon profile and inform public health strategies. A cross-sectional study measured indoor radon levels in 343 homes using CR-39 solid-state nuclear track detectors during the cold (December–February) and warm (June–August) seasons of 2024. Stratified random sampling ensured representativeness across Khuzestan, varying in geology and urban density. Key variables included building age, height, floor type, and structural framework. Spatial distribution was mapped via Inverse Distance Weighting (IDW) and Ordinary Kriging (OK). Statistical analyses (t-tests, ANOVA) assessed differences (p < 0.05). Mean radon concentrations were higher in the cold season (119.93 ± 113.15 Bq/m²) than in the warm season (72.99 ± 41.16 Bq/m²), with northern cities like Dezful (279.05 Bq/m²) exceeding WHO (100 Bq/m²) and EPA (148 Bq/m²) thresholds. The estimated annual effective doses (AEDs) ranged from 0.78 to 7.04 mSv/year, with higher values during the cold season. Older concrete buildings (> 15 years) and ground-floor units showed significantly elevated levels (p < 0.001). Floor type influenced warm-season concentrations (p = 0.011), with mosaic floors highest (91.32 Bq/m²). Spatial analysis identified northern mountainous regions as radon hotspots, especially in the cold season. Seasonal and structural factors significantly affect radon levels in Khuzestan, highlighting public health risks in high-radon zones. Mitigation strategies, such as ventilation improvements in older concrete homes, are critical. This dataset fills a critical knowledge gap in Iran’s national radon profile and provides an essential baseline for future geospatial health risk modeling, mitigation planning, and policy development.

Introduction

Radon-222 (Rn-222) is an inert, radioactive gas that is naturally produced through the radioactive decay of uranium-238, a common element in soil, rocks, and aquifers¹. Although radon readily disperses in open-air environments where it poses minimal risk, its accumulation in enclosed indoor spaces, particularly in basements and ground-level areas with inadequate ventilation, has become a significant environmental health concern². This gas infiltrates buildings through structural vulnerabilities such as cracks in foundations, wall joints, and permeable construction materials³. Once inhaled, the short-lived decay products of radon emit alpha radiation, which can damage the epithelial lining of the respiratory tract, induce DNA mutations, and increase the likelihood of cancer development^4,5. A robust body of epidemiological evidence has established a clear link between long-term radon exposure and lung cancer incidence^6–8. According to the World Health Organization (WHO), radon ranks as the second leading cause of lung cancer globally, accounting for approximately 3–14% of all cases⁹. This risk is further amplified among smokers due to the synergistic carcinogenic effects of tobacco smoke and radon progeny¹⁰.

In light of these health implications, the global scientific community has increasingly turned its attention to radon-related risks. Recent epidemiological studies conducted across the United States, Europe, and Asi have aimed to quantify the burden of radon-induced diseases, particularly lung cancer and childhood leukemia^11–13. These investigations consistently point to the role of prolonged indoor radon exposure in both adult respiratory malignancies and potential pediatric hematologic cancers^1,12,14,15.

Within Iran, several regional surveys have reported indoor radon levels that, while often below international intervention thresholds, still warrant public health attention. For example, measurements in Isfahan have shown average concentrations of 32 Bq/m² and annual effective doses near 0.5 mSv¹⁶. In contrast, readings from Shiraz and Khorramabad have indicated higher levels, around 43 Bq/m² and up to 1.1 mSv annually. Notably, 6–10% of the monitored spaces, especially those in poorly ventilated basements, exceeded recommended safety limits, underscoring the need for localized mitigation^17,18.

Various international health and regulatory organizations have issued guidelines to limit radon exposure in indoor environments. The WHO advocates for a reference level of 100 Bq/m²⁹while the European Union permits concentrations up to 300 Bq/m²¹⁹. In North America, the United States Environmental Protection Agency (EPA) has set an action level of 148 Bq/m² (4 pCi/L)²⁰and Health Canada recommends a maximum threshold of 200 Bq/m²²¹. These differences reflect not only variations in regulatory frameworks but also a shared global commitment to minimizing radon-related health risks.

Recent investigations in countries such as India, Brazil, and Italy have emphasized the influence of seasonal variation, geological substrates, and construction characteristics on indoor radon levels^22–24. In Iran, the range of measured radon concentrations is broad, from 37 Bq/m² in Damghan²⁵ to 240 Bq/m² in Ardabil²⁶attributable largely to local differences in geology and architectural practices. However, most of these studies are geographically limited and lack methodological standardization or comprehensive national coverage.

Khuzestan Province was selected due to its unique combination of geological diversity and public health relevance. The northern region, part of the High Zagros zone, contains uranium-rich formations (e.g., Pabdeh, Gurpi, Bakhtiari), intersected by active faults that enhance radon migration. In contrast, the southern alluvial plains offer a natural gradient for evaluating spatial variability^27,28. With over 4.65 million residents, many in poorly ventilated buildings, and a lack of comprehensive radon data, Khuzestan remains critically underrepresented in national surveys. These factors underscore the need for region-specific, systematic assessment to inform risk mitigation.

To bridge this gap, the present study aims to:

1. Quantify seasonal variations in indoor radon concentrations across residential buildings in Khuzestan Province and compare them with international reference levels.

2. Examine the influence of structural parameters, such as building height, age, and construction materials, on indoor radon accumulation.

3. Identify high-risk zones and generate radon concentration maps for cold and warm seasons using Inverse Distance Weighting (IDW) and Ordinary Kriging (OK) spatial interpolation methods.

4. Estimate the annual effective dose (AED) from indoor radon exposure across Khuzestan Province

The outcomes of this research are expected to advance understanding of radon exposure in Khuzestan, inform evidence-based mitigation policies, and contribute to the broader field of environmental health risk management.

Materials and methods

Area of study

This study adopted a cross-sectional, quantitative research design to assess indoor radon concentrations in residential buildings across Khuzestan Province, Iran (Fig. 1). The study area covered the entire Khuzestan Province (29° 57′–33° 00′ N, 47° 40′–50° 33′ E), a region spanning 63,633.6 km² with a population exceeding 4.65 million²⁹. Khuzestan is situated at the convergence of the Zagros Fold-Thrust Belt and the Mesopotamian Foredeep, giving rise to a geologically diverse landscape. In the northern regions, uranium-rich sedimentary formations such as the Pabdeh, Gurpi, and Ilam units dominate, while the southern areas are characterized by Quaternary alluvial plains with clay-rich deposits that can retain and gradually release radon. The presence of major active fault systems, including the Main Zagros Reverse Fault (MZRF) and the Dezful Fault, further increases subsurface permeability and facilitates radon migration. This combination of heterogeneous lithology, structural complexity, and tectonic activity makes Khuzestan an ideal setting for investigating spatial variations in indoor radon concentrations^27,28.

Fig. 1.

Study area.

Sampling

A stratified random sampling method was employed to ensure representativeness across the province’s diverse urban and geological settings. Cities were selected based on population density and geological variability, including northern mountainous areas (e.g., Dezful, Andimeshk) and southern coastal plains (e.g., Abadan, Mahshahr). Within each city, residential units were stratified by structural type (concrete vs. steel), age (< 15 years vs. >15 years), and height (ground floor vs. multi-story). A total of 343 households participated in the study. The required sample size was determined using Eq. 1:

1

where n represents the estimated sample size, calculated based on a 99% confidence level using a Z-score of 1.96. The calculation assumes a margin of error of 0.1 µ (mean value), with SD denoting the standard deviation of the target variable³⁰.

Data collection methods

Data collection occurred in two distinct phases to capture seasonal variability. In the cold season (December 2023–February 2024), detectors were deployed for 90 days, coinciding with reduced ventilation due to closed windows and heating system use. The warm season phase (June–August 2024) followed an identical 90-day protocol, reflecting higher temperatures and natural ventilation patterns. Detectors were placed in the main living area of each home, 1–2 m above floor level and at least 50 cm from walls or ventilation sources, adhering to EPA guidelines. Placement avoided kitchens and bathrooms to minimize humidity and airflow biases. Participants were instructed to maintain typical occupancy and ventilation habits to reflect real-world conditions³¹. Each deployment was accompanied by a standardized questionnaire that documented building attributes such as floor type (mosaic, stone, parquet), construction age, height (classified as < 20 m, 20–27 m, > 28 m), structural framework (concrete or steel), and floor level (ground or upper). Geological information- including soil type and proximity to water bodies- was sourced from the Geological Survey of Iran to contextualize spatial findings. After exposure, detectors were collected, sealed in airtight bags, and transported to a laboratory for processing within 48 h to avoid track fading³².

Instruments and equipment

Radon concentrations were determined using CR-39 solid-state nuclear track detectors (SSNTDs), which are a validated passive sampling method for long-term radon evaluation³³. These detectors, constructed from polyallyl diglycol carbonate, effectively capture alpha particle tracks emitted by radon decay products, offering high sensitivity and reliability. After collection, the CR-39 detectors underwent chemical etching in a 6.25 M NaOH solution at 70 °C for 6 h, following standard procedures to expose alpha tracks³⁴. An optical microscope (Olympus BX51) was used at 400× magnification for counting the tracks, with concentrations calculated using a calibration factor of 0.2 tracks/cm² per Bq/m²/day. CR-39 detectors generally exhibit an uncertainty range of around 5%, influenced by factors such as exposure duration and the specifics of the etching process. To mitigate these potential inaccuracies, the detectors were calibrated in advance using a certified radon reference chamber. Background radiation effects were accounted for by deploying control (blank) detectors alongside samples. In addition, certain measurements were repeated to assess consistency, and the etching procedure was conducted under rigorously standardized conditions to maintain uniformity across all samples^35,36. Radon levels were reported in becquerels per cubic meter (Bq/m²), and seasonal averages and standard deviations were computed for each city and building category.

For spatial analysis, ArcGIS Pro software (Esri, USA) was employed, utilizing the Inverse Distance Weighting (IDW) and Ordinary Kriging (OK) interpolation method to visualize radon distribution³⁷. IDW is a deterministic interpolation method that estimates unknown values based on the proximity of known points, assuming that nearby observations have more influence on the prediction. Kriging, a geostatistical method, accounts not only for the spatial autocorrelation of the data but also provides an estimation of uncertainty through semivariogram modeling. Kriging requires the dataset to be normally distributed to ensure reliable predictions; therefore, data normalization was performed before Kriging analysis^2,38,39.

Statistical analyses were performed with SPSS version 27 (IBM, USA), enabling descriptive and inferential computations. Descriptive statistics outlined mean radon concentrations, ranges, and confidence intervals (95% CI). Paired t-tests evaluated seasonal differences, while independent t-tests and one-way ANOVA compared radon levels across various building characteristics (e.g., concrete vs. steel, old vs. new). The threshold for significance was established at p < 0.05 ⁴⁰.

The annual effective dose (AED) resulting from indoor radon exposure was estimated for each city during both cold and warm seasons using the standard formula recommended by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) (Eq. 2) ⁴¹:

2

where C is measured radon concentration (Bq/m³), F is equilibrium factor between radon and its progeny (0.4 for indoor measurement), H is the occupancy factor (0.8 for indoor measurement), T is the hours for a year (8760 h y^− 1), and D is the dose conversion factor (9 × 10^− 6 mSv per Bq h/m³)¹⁸.

Results

Seasonal variations in radon concentrations

As illustrated in Fig. 2, radon concentrations showed clear seasonal variation across cities in Khuzestan province. In most locations, levels were higher during the cold season and dropped in the warm season. However, a few cities experienced the opposite trend. The highest radon levels in the cold season were found in Dezful (279.05 Bq/m²), Andimeshk (258.17 Bq/m²), and Shushtar (257.82 Bq/m²), while Susangerd (31 Bq/m²) and Hamidiyeh (44 Bq/m²) recorded the lowest. Abadan (53.28 Bq/m²) and Khorramshahr (67.33 Bq/m²) had data only for the cold season. In the warm season, Izeh had the highest concentration (170.44 Bq/m²), followed by Ramhormoz (125.5 Bq/m²) and Shadgan (98 Bq/m²). Dezful showed a major decline from its peak in the cold season to 49.9 Bq/m². Increases during the warm season were observed in Hamidiyeh, Hoveyzeh, Izeh, Omidiyeh, Shadgan, and Susangerd. By contrast, cities such as Ahvaz, Andimeshk, Behbahan, Dezful, Mahshahr, Shush, and Shushtar experienced significant reductions during warmer months. These patterns suggest that both climate and local building conditions may influence seasonal radon behavior.

Fig. 2.

Mean radon concentrations by city in cold and warm seasons.

Comparison with international standards across cities

As shown in Fig. 3, indoor radon levels across Khuzestan cities were compared with guidelines from WHO (100 Bq/m²), EPA/Canada (148 Bq/m²), and the EU (300 Bq/m²). In the cold season, several cities—including Dezful (279.05), Andimeshk (258.17), Shushtar (257.82), and Shush (204.4 Bq/m²)—exceeded all major standards. Cities such as Ahvaz and Behbahan were above WHO and EPA thresholds, while Abadan, Hamidiyeh, Hoveyzeh, and Susangerd remained within safe limits. In the warm season, most cities showed a decline in radon levels. However, Izeh (170.44 Bq/m²) and Ramhormoz (125.5 Bq/m²) remained above WHO limits. Notably, Shadgan had a sharp increase (145%) yet stayed under all international thresholds. Hamidiyeh, Hoveyzeh, Omidiyeh, and Susangerd also recorded modest increases but remained within safe limits. Taken together, cold-season values posed a higher risk in several cities, while warm-season concentrations generally aligned better with international standards. This highlights the need for seasonal risk assessments and targeted mitigation in high-risk areas.

Fig. 3.

Comparison of Radon concentrations across Khuzestan cities (Cold and Warm Seasons) with International Standards.

Effective dose assessment and health risk implications

The spatial and seasonal variation in the AED due to indoor radon exposure is illustrated in Fig. 2. Based on the measured radon concentrations, AED values ranged from 0.78 to 7.04 mSv/year across the studied cities and seasons. The highest AED was recorded in Dezful during the cold season (7.04 mSv/year), exceeding the International Commission on Radiological Protection (ICRP)’s reference level (3–10 mSv/year), thereby indicating a potential radiological health hazard^42,43. Conversely, Susangerd exhibited the lowest AED in the same season (0.78 mSv/year), remaining well within internationally acceptable limits. A significant seasonal variation in the AED was observed across the studied cities, with values exceeding 3 mSv/year in at least five locations during the cold season. For instance, AED in Andimeshk sharply declined from 6.51 mSv/year in the cold season to 1.57 mSv/year in the warm season. Similarly, Shushtar and Shush exhibited reductions from 6.50 to 1.45 mSv/year and 5.16 to 1.58 mSv/year, respectively. These pronounced seasonal differences, with cold-season doses reaching up to five times higher than their warm-season counterparts, underscore the significant influence of ventilation and occupancy behaviors.

Fig. 4.

Annual effective dose of radon (mSv/y) across cities in cold and warm seasons.

Radon concentrations by floor type

Radon levels varied by floor type, with significant differences observed in the warm season (Fig. 5A). In the cold season, mean concentrations across floor types were: stone (80.83 ± 52.51 Bq/m²), ceramic (72.65 ± 40.58 Bq/m²), and mosaic (69.69 ± 38.95 Bq/m²). No statistically significant variation was observed among floor types (ANOVA, F = 0.262, p = 0.770). In the warm season, mosaic floors exhibited the highest mean concentration (140.78 ± 125.65 Bq/m²), significantly exceeding stone (62.93 ± 45.11 Bq/m²) (ANOVA, F = 5.040, p = 0.002; post-hoc Tukey, p < 0.05). Ceramic (99.15 ± 97.32 Bq/m²) and parquet (111.45 ± 65.25 Bq/m²) showed intermediate values, with no significant pairwise differences (p > 0.05).

Fig. 5.

Comparative analysis of indoor radon concentrations during cold and warm seasons, stratified by (A) floor type, (B) building height, (C) floor level, and (D) building age.

Radon concentrations by building height

Figure 5B displays radon levels across building height categories defined as follows: low-rise buildings (< 12 m, corresponding to approximately 1–3 floors), mid-rise buildings (12–19 m, approximately 4–6 floors), high-rise buildings (20–27 m, approximately 7–9 floors), and very high-rise buildings (≥ 28 m, corresponding to 10 floors or more). These categories were established based on typical floor heights in Khuzestan, where each floor is approximately 3 m in height.

In the cold season, mean concentrations were 126.82 ± 122.12 Bq/m² (< 12 m), 115.38 ± 98.30 Bq/m² (12–19 m), 82.26 ± 61.69 Bq/m² (20–27 m), and 128.75 ± 107.71 Bq/m² (≥ 28 m). The differences among height groups were not statistically significant (ANOVA, F = 1.757, p = 0.155). In the warm season, radon concentrations were 65.30 ± 42.19 Bq/m² (< 12 m), 76.37 ± 46.55 Bq/m² (12–19 m), 76.24 ± 40.86 Bq/m² (20–27 m), and 72.57 ± 35.38 Bq/m² (≥ 28 m), again with no significant differences observed (ANOVA, F = 0.627, p = 0.598).

Radon concentrations by floor level

Figure 5C illustrates the comparison of radon levels between ground-floor and upper-floor units. During the cold season, ground-floor apartments recorded an average radon concentration of 133.18 ± 125.09 Bq/m² (95% CI 116.75–149.62), which was significantly higher than that of upper-floor units at 94.65 ± 80.52 Bq/m² (95% CI 79.97–109.33) (t = 3.032, df = 341, p = 0.003). In the warm season, ground-floor units had a mean radon level of 64.33 ± 45.08 Bq/m² (95% CI 29.68–98.99), while upper-floor units averaged 73.46 ± 41.03 Bq/m² (95% CI 67.18–79.75), with no statistically significant difference (t = − 0.647, df = 173, p = 0.518).

Radon concentrations by building age

Figure 5D shows radon levels by building age. In the cold season, older buildings (> 15 years) had a significantly higher mean radon concentration of 142.29 ± 123.86 Bq/m² (95% CI 121.05–163.54) compared to newer buildings (< 15 years), which averaged 105.76 ± 103.64 Bq/m² (95% CI 91.66–119.86) (t = − 3.085, df = 341, p = 0.002). In the warm season, buildings older than 15 years also showed significantly higher radon levels (76.29 ± 42.81 Bq/m², 95% CI 69.34–83.25) than those less than 15 years old (54.93 ± 23.95 Bq/m², 95% CI 45.45–64.40) (t = − 3.871, df = 173, p < 0.001).

Radon concentrations by structural framework

Figure 6 illustrates radon concentrations by structural framework. In the cold season, concrete buildings recorded a significantly higher mean concentration of 141.68 ± 121.73 Bq/m² (n = 188, 95% CI 124.16–159.19) compared to steel-framed buildings, which had an average of 93.55 ± 95.71 Bq/m² (n = 155, 95% CI 78.36–108.74) (t = 4.006, df = 341, p < 0.001). During the warm season, no significant difference was observed between concrete structures (71.97 ± 39.80 Bq/m², n = 127, 95% CI 64.98–78.96) and steel structures (75.71 ± 44.90 Bq/m², n = 48, 95% CI 62.67–88.75) (t = − 0.535, df = 173, p = 0.593).

Fig. 6.

Radon concentrations by structural framework (cold and warm seasons).

Spatial distribution

In this study, the spatial distribution of indoor radon concentration was modeled using two interpolation techniques: Inverse Distance Weighting (IDW) (Fig. 7A) and Ordinary Kriging (OK) (Fig. 7B) across two distinct seasons, cold and warm periods. The comparative assessment of the two methods reveals significant spatial and seasonal variability in radon distribution patterns, with distinct concentration gradients across the province. To perform Ordinary Kriging, different semivariogram models were tested. Based on cross-validation and residual analysis, the Stable model was selected for the cold season, and the Exponential model for the warm season. For the cold season, the model showed a nugget of 5092.25, a partial sill of 6676.14, and a range of 2.20 km, indicating moderate spatial autocorrelation. In the warm season, the exponential model had a nugget of 859.20, a partial sill of 832.22, and a range of 2.93 km, suggesting stronger spatial continuity. Both models used 12 lags and isotropic neighborhoods. These configurations provided the most reliable predictions and minimized interpolation errors. The predictive performance of IDW and OK was assessed using standard validation metrics, including Root Mean Square Error (RMSE), Mean Standardized Error, Root Mean Square Standardized Error (RMSS), and Average Standard Error (ASE) (Table 1).

Fig. 7.

Spatial distribution of radon concentration using (A) IDW and (B) OK (Cold and Warm Seasons).

Table 1.

Error metrics for OK and IDW (Cold and warm Seasons).

Method	Season	RMSE	Mean Standardized	RMSS	ASE
OK	Warm	32.5602	0.0119	1.0480	30.9758
OK	Cold	72.9442	0.0187	0.9804	74.3599
IDW	Warm	37.0547	-0.0306	1.0005	37.0374
IDW	Cold	81.1994	0.0735	1.0027	80.9810

The results indicate that OK outperformed IDW in both seasons, as evidenced by lower RMSE and ASE values. The model’s standardized error metrics (Mean Standardized and RMSS) were within acceptable ranges, confirming the reliability of the OK predictions.

During the cold season, both interpolation methods identified the northern regions of Khuzestan, particularly Dezful and Andimeshk, as radon concentration hotspots. The IDW method estimated the highest concentrations within the range of 263.829–266.675 Bq/mv in these areas, whereas regions such as Abadan, Khorramshahr, and Shadegan in the south recorded the lowest values, between 24.696 and 65.5 Bq/m². The spatial pattern derived from Kriging corroborated these findings, indicating elevated levels in Shush, Shushtar, and Masjed Soleyman, albeit with a more spatially smoothed distribution. Kriging produced more homogeneous contours and better spatial continuity, capturing gradual transitions in radon levels across neighboring districts. In contrast, IDW revealed sharper boundaries and more abrupt spatial variations, likely influenced by the clustering and proximity of measurement points. While IDW can highlight local anomalies, Kriging, by incorporating spatial autocorrelation, provides more reliable estimates in areas with sparse data. In the warm season, the radon distribution pattern shifted significantly. The highest concentrations were detected in eastern and southeastern counties, including Izeh, Baghmalek, Behbahan, Ramhormoz, Bandar-e-Mahshahr, and Hendijan. IDW estimated the peak values to be within 105.419–128.896 Bq/m², while the lowest concentrations, ranging from 55.887 to 58.749 Bq/m², were recorded in the northern districts such as Andimeshk and Dezful. The Kriging model showed a similar spatial trend but presented a smoother gradient, indicating a continuous increase in radon concentrations from northwest to southeast.

Discussion

This study comprehensively assesses indoor radon concentrations in residential buildings across Khuzestan Province, Iran, revealing significant seasonal variations and the influence of building characteristics. The observed higher mean radon concentration in the cold season (119.93 ± 113.15 Bq/m²) compared to the warm season (72.99 ± 41.16 Bq/m²) aligns with established patterns of seasonal radon dynamics. This difference (p = 0.002) reflects reduced natural ventilation during colder months, as residents close windows and rely on heating systems, trapping radon emitted from soil and building materials²². In contrast, warmer months facilitate greater airflow, diluting indoor concentrations. City-specific extremes, such as Dezful’s 279.05 Bq/m² in the cold season and Izeh’s 170.44 Bq/m² in the warm season, suggest localized geological and climatic influences. Dezful’s peak likely stems from its northern location amid uranium-rich shales, coupled with minimal ventilation, while Izeh’s warm-season high may relate to soil moisture dynamics, where drying conditions enhance radon exhalation⁴⁴. These findings confirm the hypothesis that seasonal factors significantly modulate indoor radon levels in Khuzestan, necessitating season-specific monitoring strategies. In our study, annual effective dose (AED) estimates ranging from 0.78 to 7.04 mSv/year across different cities in Khuzestan Province. The highest AEDs, observed during winter in Dezful and Andimeshk, exceeded the ICRP’s lower reference level of 3 mSv/year, indicating a potential radiological health concern. These findings are consistent with those reported by Elío and Crowley, who found average indoor radon concentrations in Ireland ranging from 20 to 338 Bq/m², corresponding to AEDs between 0.8 and 13.3 mSv/year³⁸. Their study further quantified the expected burden of radon-induced lung cancer, estimating up to 239 cases per million individuals in high-exposure areas based on a linear no-threshold (LNT) risk model^10,42. Similarly, a study conducted by Akuo-ko et al. in Ghana’s Greater Accra Region found average indoor radon levels of 50.8 Bq/m² (1.3 mSv/year), with values rising to 92.0 ± 5.2 Bq/m² (2.3 mSv/year) in densely populated areas characterized by poor ventilation and limited air circulation². These spatial and seasonal patterns, attributed to behavioral factors such as frequent use of air conditioning and closed windows for security or comfort, closely resemble the winter conditions documented in Khuzestan. The marked differences between cold and warm season exposures in our study, with AEDs in some cities declining by more than fivefold in warmer months, further emphasize the critical role of ventilation practices in modulating indoor radon risk. Building characteristics emerged as critical determinants of radon concentration, with varying impacts across seasons. The significant warm-season difference in floor type (p = 0.002), where mosaic floors (140.78 Bq/m²) exceeded stone (62.93 Bq/m²), suggests porosity and permeability differences. Mosaic’s coarse, unsealed surfaces likely facilitate greater radon diffusion from underlying soil, whereas stone’s density impedes it⁴⁵. The lack of significance in the cold season (p = 0.770) may indicate that ventilation dominates over material effects during winter, masking floor-type variations. Older buildings (> 15 years) consistently showed higher radon levels (cold: 142.29 Bq/m²; warm: 76.29 Bq/m²) than newer ones (p < 0.001), supporting the hypothesis that construction age influences radon accumulation. This trend likely reflects degraded sealing, outdated materials (e.g., uranium-containing aggregates), and poorer ventilation in older structures²⁵. Concrete buildings exhibited significantly higher cold-season concentrations (141.68 Bq/m²) than steel-framed ones (93.55 Bq/m², p < 0.001), possibly due to concrete’s higher radium content and permeability compared to steel’s inert nature⁴⁶. The absence of this difference in the warm season (p = 0.593) suggests ventilation equalizes structural effects in summer. Ground-floor units had higher radon levels (133.18 Bq/m²) than upper floors (94.65 Bq/m²) in the cold season (p = 0.003), consistent with proximity to the soil as the primary radon source⁴⁷. The non-significant warm-season difference (p = 0.518) may reflect increased vertical mixing in multi-story buildings during summer. Building height showed no clear trend (p > 0.05), contradicting expectations of decreasing concentrations with elevation, possibly due to uniform soil radon sources or airflow patterns in Khuzestan’s urban settings⁴⁸. These results resonate with and diverge from prior research, highlighting regional and methodological nuances. The cold-season mean (119.93 Bq/m²) exceeds global averages (40 Bq/m²)⁴⁹ and aligns with elevated levels in Iranian cities like Ardabil (240 Bq/m²)²⁶ and Hamadan (108 Bq/m²)³². Dezful’s peak (279.05 Bq/m²) approaches extreme values reported in Ardabil (2386 Bq/m²)²⁶ underscoring Khuzestan’s potential as a radon hotspot. However, the warm-season mean (72.99 Bq/m²) is closer to Shiraz (57.6 Bq/m²)⁴⁵ and Damghan (< 37 Bq/m²)²⁵ suggesting seasonal moderation akin to less geologically active regions. The floor-type effect mirrors findings by Asgari et al.⁴⁶ where granite interiors (59 Bq/m²) exceeded carbonate ones (11 Bq/m²), though Khuzestan’s mosaic-stone disparity is more pronounced in summer. The age-related trend corroborates Shorgashti et al.²⁵ who linked higher radon in older Damghan homes to construction practices and contrasts with Jafarizadeh et al.²⁹ where moisture-proofing reduced radon significantly. Concrete’s dominance over steel aligns with global studies⁵⁰ but the seasonal shift differs from Fahiminia et al.⁴⁷ where structural effects persisted year-round in Qom. In our analysis, Kriging generated more continuous and spatially consistent radon distribution patterns, successfully depicting gradual changes between neighboring areas. This observation mirrors findings from Liza et al. (2025) in Lima, where Kriging proved more effective than IDW by integrating spatial relationships, minimizing abrupt peaks, and producing more accurate representations of high-risk zones. Conversely, IDW was able to detect localized anomalies but tended to produce abrupt changes around extreme values due to its reliance on distance alone, a limitation also highlighted by Munyati and Sinthumule (2021), who deemed IDW better suited for irregular, sparse distributions such as savannah vegetation. Additionally, Akuo-ko (2020) noted that IDW performs optimally with non-normally distributed data, which may account for its sensitivity to outliers in our dataset. Taken together, these findings support our conclusion that Kriging yields more dependable spatial predictions in areas with sparse sampling, whereas IDW may distort results by overemphasizing isolated high or low values. Spatially, northern Khuzestan’s high concentrations (e.g., Dezful, Andimeshk) parallel radon-prone areas like Punjab, India²³ driven by uranium-rich geology, while southern lows (e.g., Abadan) resemble coastal Brazil⁵⁰ where groundwater suppresses emanation. The ground-floor elevation aligns with Fahiminia et al.⁴⁷and Tiari et al.⁴⁸ though the lack of height effect diverges from expected gradients, possibly reflecting Khuzestan’s flat terrain and uniform soil radon potential.The absence of a height-related trend was unexpected, given prior evidence of decreasing radon with elevation⁴⁷. This may stem from consistent soil radon flux across Khuzestan’s urbanized sampled areas, where building height does not sufficiently isolate upper floors from ground sources. Izeh’s warm-season peak (170.44 Bq/m²), exceeding its cold-season value (150.17 Bq/m²), was also unanticipated. Izeh lies within the High Zagros zone, where uranium-rich geological units, particularly the marl, shale, and limestone layers of the Pabdeh and Gurpi formations, serve as primary sources of natural radon emission. The area is also intersected by active tectonic faults, which promote rock fracturing and increase the permeability of subsurface layers⁵¹. During the warm season, elevated temperatures, reduced soil moisture, and enhanced gas diffusivity facilitate radon migration toward the surface⁵². This elevated radon concentration may also be influenced by seasonal soil drying in the region’s mountainous terrain, further amplifying radon exhalation. In some cases, localized reductions in natural ventilation, undetected in broader measurements, may contribute to observed anomalies²⁸. These findings underscore the complex and site-specific nature of radon behavior in geologically active regions. In the cold season, the highest radon concentrations were found in northern cities like Andimeshk and Dezful, ranging from 226.94 to 512.70 Bq/m², due to geological factors and soil permeability. Shush and Shushtar also had high levels (175.87 to 244.93 Bq/m²), likely influenced by geological composition and fault presence. Ahvaz showed moderate concentrations (84.99 to 149.12 Bq/m²), while Ramhormoz had lower levels (55.34 to 84.99 Bq/m²), related to soil and hydrogeological features. The southern cities (Mahshahr, Abadan, Khorramshahr) had the lowest levels (< 55.34 Bq/m²), likely due to high groundwater levels and coastal sediments. In the warm season, radon concentrations were lower across the province. Northern cities like Andimeshk and Dezful had moderate to high levels (64.14 to 91.05 Bq/m²), while central areas like Ahvaz showed lower levels (55.99 to 61.99 Bq/m²). Southern cities maintained the lowest concentrations (22.58 to 55.93 Bq/m²), with Behbahan in the south showing the highest in the region (97.21 to 144.81 Bq/m²). Radon concentrations decreased in the warm season, likely due to increased temperatures and natural soil ventilation. Radon zoning maps in Khuzestan province indicate that the emission of this gas is influenced by several factors, such as elevation, topography, geology, soil type, and groundwater level. Higher emissions were observed in elevated and drier areas, while concentrations decreased in flat areas closer to water sources.

Conclusion

This study investigated indoor radon concentrations in residential buildings across Khuzestan Province, Iran, revealing significant seasonal and structural influences. Key findings indicate higher radon levels in the cold season (119.93 ± 113.15 Bq/m²) than the warm season (72.99 ± 41.16 Bq/m²), with northern cities like Dezful (279.05 Bq/m²) and Izeh (170.44 Bq/m²) exceeding WHO (100 Bq/m²) and EPA (148 Bq/m²) thresholds. Tte AED estimates varied between 0.78 and 7.04 mSv/year, with elevated levels observed in the cold season, sometimes surpassing global safety thresholds. Older buildings (> 15 years), concrete structures, and ground-floor units consistently exhibited elevated concentrations, while floor type (e.g., mosaic vs. stone) influenced warm-season levels. In our study, although IDW proved effective in capturing localized radon concentration patterns, Kriging yielded more reliable estimates in areas with limited sampling due to its ability to model spatial autocorrelation. Spatial analysis confirmed northern mountainous regions as radon hotspots, contrasting with lower southern coastal values. The findings underscore significant public health risks, particularly in high-radon zones, where mitigation, such as enhanced ventilation and sealing in older concrete homes, is urgently needed. Theoretically, they enrich the understanding of radon dynamics in arid, geologically diverse regions, contributing valuable data to Iran’s national radon mapping efforts. Limitations include the urban focus, uneven sample distribution, and lack of short-term fluctuation data, which may temper generalizability. Future research should extend sampling to rural areas, incorporate meteorological variables, and employ long-term monitoring to refine these insights. This study highlights the necessity of systematic radon assessment and targeted interventions in Khuzestan, offering a framework for addressing indoor radon exposure risks regionally and globally.

Author contributions

N.T. developed the methodology, prepared the spatial distribution maps, and drafted the initial manuscript. A.M., K.N., K.Y., A.S., M.Y., M.A.M., and S.J. contributed to the overall study design, technical validation, and provided scientific input during the research planning phase. M.A.K., M.H., and S.G. performed statistical modeling, data processing, and contributed to data visualization and interpretation of the findings. Z.G.R., Z.Y., M.Ya., N.K., S.K.M., S.Gh., F.J., M.P.F., Z.H., N.M., H.S., N.Sh., M.H.B., and A.G. participated in field implementation, including sampling, data collection, completion of questionnaires, and logistical coordination. N.J. supervised project administration, refined the methodology, critically revised the manuscript for intellectual content, and led the final review and editorial process. All authors reviewed and approved the final manuscript.

Funding

This research received no external funding.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The study was approved by Ahvaz Jundishapur University of Medical Sciences (ethical code of the Ethics Committee are 122 and IR.AJUMS.REC.1398.310, respectively; Grant No. ETRC-9423).

Consent to participate

All participants provided informed consent during the study.

Accordance statement

All experiments were performed in accordance with relevant guidelines and regulations.