Rural households dependent on rain-fed agriculture face growing climate risks, yet the role of industrial wage employment as an adaptation pathway remains unclear. This study examines whether and how industrial employment functions as a rural household climate adaptation pathway in Central Ethiopia by assessing its adaptation modality (intentional versus indirect), temporal dimension (ongoing versus prospective), and the underlying drivers shaping household participation. Combining survey data from 625 households with climate data, an Endogenous Switching Regression model was applied to adjust for selection bias, and Structural Equation Modeling was used to identify pathways. Households with industrial employment were 64% less sensitive to climate shocks than farm-only households, yielding an average treatment effect of 0.49 standard deviations (SD) on a composite food security index. The benefit follows a sequential mediation pathway led by income stabilization (47%), which supports dietary diversity (28%) and agricultural input investment (25%). Path decomposition further shows that income stabilization indirectly reinforces both mechanisms, accounting for an additional 0.08 SD of the total treatment effect through the inter-mediator pathway. While this labor reallocation functions as an indirect adaptation response to agricultural stress, structural factors, including younger and more educated household heads, better road access, and social network connections, primarily enable transition into factory work. The effect is largest for female-headed households (0.58 SD), land-scarce households (0.55 SD), and those near industrial parks. Importantly, climate exposure does not drive industrial job uptake; entry is instead governed by education, social networks, and land constraints. Hence, the adaptive benefit is an indirect outcome of structural employment access rather than a deliberate climate management decision. Policies integrating industrial job creation with rural development can reinforce household resilience alongside traditional on-farm adaptation.
Empirical claims about subnational green-transition determinants can appear robust to model specification while remaining fragile to measurement. Using a balanced panel of Vietnam's 63 provinces for 2010 - 2024, this study re-audits the relationship between inequality and a composite Green Growth Index (GGI) after concerns about instrumental-variable identification, inequality measurement, index construction, and spatial inference. The archived analytical inequality series produces a stable negative but marginal two-way fixed-effects association (β = -0.202, SE = 0.119, p = 0.094), with negative coefficients across all control subsets and leave-one-province/year checks. However, an independent remeasurement using 441 directly published National Statistics Office provincial income-Gini observations for 2018 - 2024 does not reproduce that relationship (β = 0.044, SE = 0.063, p = 0.490); annual first differences are also null (β = 0.020, p = 0.713). The archived and official inequality series correlate only moderately (r = 0.573). Raw province-year histories for the three GGI components are unavailable; sharp partial-identification bounds show that component-specific coefficient signs cannot be inferred from the composite alone. Official inequality remains spatially clustered, but local and neighboring GGI associations are imprecise across alternative weight matrices. The evidence therefore does not support a robust directional or causal inequality effect. Its main implication is methodological: specification robustness cannot substitute for measurement validity and reproducibility.
Reliable climate projections at the watershed scale are particularly difficult in complex terrain because model biases can obscure local climate signals. This study develops bias-corrected, performance- and independence-weighted CMIP6 projections for the Omo-Kuraz Watershed in southwestern Ethiopia using daily observations from 11 ENACTS stations and five CMIP6 models. Model performance and inter-model dependence were evaluated for 1985–2014, followed by empirical quantile mapping and ensemble weighting. Changes in temperature, precipitation, climate extremes, return periods, projection uncertainty, and climate-signal emergence were assessed under SSP2-4.5 and SSP5-8.5. By the 2090s, basin-wide mean maximum temperature is projected to rise by 2.21°C under SSP2-4.5 and 3.85°C under SSP5-8.5, while minimum temperature increases by 2.30°C and 4.02°C, respectively. Elevation-dependent warming reaches 1.20 and 2.04°C per 1000 m (R² = 0.965). Annual precipitation increases by 3.5% and 5.8%, but climate extremes change much more strongly under SSP5-8.5. The annual count of days exceeding the historical 90th-percentile Tmax threshold increases from 18 to 98 days yr⁻¹ (+444%), while R95p, R99p, Rx1day, and R25mm increase by 45%, 67%, 38%, and 125%, respectively. A historical 100-year Rx1day event could recur approximately every 25 years by the 2090s. Model uncertainty dominates before about 2065, after which scenario uncertainty becomes more important. The warming signal emerges around 2045 under SSP5-8.5 and 2062 under SSP2-4.5. These findings point to increasing hydroclimatic risk and provide a basis for climate-resilient water-resource and infrastructure planning.
Although many studies have found that downscaling soil moisture is effective, research on how the accuracy of the downscaled soil moisture varies as a function of spatial resolution is limited. To address this gap, this study downscales soil moisture to different resolutions that are necessary for agriculture, hydrology, and climatology applications, and assesses how the accuracy varies as a function of spatial resolution. This study applied Random Forest (RF) to downscale NASA's Soil Moisture Active Passive (SMAP) mission and model-based soil moisture from the North American Land Data Assimilation System (NLDAS) to 2-km, 1-km, 700-m, and 400-m for soil moisture volumetric water content (VWC m3/m3) and percentiles (%). Both SMAP and NLDAS downscaled products exhibit high accuracy across regions and resolutions, with mean absolute error (MAE) of VWC (m3/m3) consistently below 0.1 in most areas. The MAE for SMAP VWC (m3/m3) downscaling at 2-km, 1-km, 700-m, and 400-m are 0.0804 m3/m3, 0.0798 m3/m3, 0.0794 m3/m3, and 0.0787 m3/m3, respectively. A 400-m resolution SMAP product achieved a national MAE of 0.0787 m3/m3, which is similar to NLDAS (MAE = 0.0805 m3/m3). Based on correlations between the downscaled soil moisture and in situ observations, SMAP outperformed NLDAS in several regions, particularly the WestNorthCentral and SouthEast. Using the multi-resolution framework developed in this study, soil moisture products can be developed that accurately capture the spatial heterogeneity pertinent to land-surface processes. These products provide improved hydrologic modeling, drought monitoring, and water resource decision-making by providing soil moisture estimates at the appropriate field, watershed, and regional scales. Therefore, this study improves our understanding of how accuracy varies as a function of the spatial resolution of downscaled soil moisture datasets and supports practical application of these data in hydrologic and agricultural analyses.
Emerging economies must improve resource efficiency while managing energy demand, emissions pressure, industrial upgrading, and uneven subnational governance capacity. This study examines provincial digital transformation as environmental implementation capacity - the ability to mobilize information, coordination, adoption, and innovation for environmental management - rather than treating digitalization as inherently green. Using a balanced baseline panel of 63 Vietnamese provinces over 2010-2023 (882 province-year observations), we construct a five-dimensional Digital Transformation Indicator (DTI) and estimate two-way fixed-effects models, Spatial Durbin Models, an innovation-capacity pathway model, and Hansen threshold regressions. Provincial final energy use is harmonized from sectoral activity data and national energy balances; Green Total Factor Productivity and exclusion of high-imputation provinces provide sensitivity checks. In the preferred controlled fixed-effects model, DTI is positively associated with energy efficiency (β = 0.334, p < 0.01). Spatial decomposition attributes 37.5% of the total association to cross-provincial spillovers. The bootstrapped innovation-capacity pathway estimate is 0.144 (95% CI [0.089, 0.212]), consistent with a material but noncausal innovation-capacity pathway. A threshold at DTI = 0.487, near the 60th percentile, separates a lower-slope regime (0.187) from a higher-slope regime (0.412). Results remain qualitatively stable across alternative DTI constructions, spatial matrices, outcomes, a reduced-sample two-period DTI lag, and endogeneity-oriented robustness checks. The findings support a conditional, noncausal interpretation of digital capability as an environmental implementation resource.
Across the world's oceans, lakes, and rivers, climate change is reshaping fisheries and aquaculture systems on which billions of people depend for food and millions rely for their livelihoods. This narrative review summarizes peer-reviewed and selected literature published between 2008 and 2026 on climate impacts, adaptation options, and critical knowledge gaps in marine and freshwater fisheries, with a particular focus on tropical and African inland fisheries. This review shows that warming, acidification, deoxygenation, and altered hydrology are disrupting fish physiology; shifting the distribution of marine species poleward by over 70 km per decade, at rates that vary greatly by taxon, depth, and area; and destroying critical habitats. Although this rapid poleward shift is well documented in open marine environments, it contrasts sharply with landlocked or fragmented freshwater systems, where horizontal migration is physically constrained, a contrast central to the concept of climate entrapment developed in this review. In scenarios characterized by high emissions, tropical fisheries may see a reduction of up to 40% in their maximum catch potential, a statistic relevant to specific tropical Exclusive Economic Zones based on bioclimate-envelope forecasts, which should not be interpreted as a global or universal result. This figure is a model-derived projection under a bioclimate-envelope framework rather than an observed trend, and the realized outcome will depend on future emission pathways. These biophysical changes compound already serious pressures, overfishing, habitat loss, and pollution, and the burden falls heaviest on those least responsible: small-scale fishers, women, and indigenous communities. We introduce the concept of climate entrapment to describe the distinct vulnerability of freshwater fisheries: trapped by fragmented habitats and hydrological barriers, fish stocks cannot migrate to more secure waters as conditions deteriorate. While there are promising adaptation strategies, major gaps remain in understanding what works, for whom, and under what conditions. We conclude with priority recommendations for research, policy, and practice.
Heavy metal contamination by copper (Cu) and Zinc (Zn) in environmental water and benthic sediments was characterised within an urban industrial waterbody. The potential for metal bioaccumulation within the soft and hard tissue (shell) of the bivalve Corbicula fluminea was examined using laboratory-scale mesocosm experiments. Each mesocosm was dosed with specific concentrations of Cu (1.3 mg/l, 2.6 mg/l, 3.9 mg/l), Zn (5 mg/l, 10 mg/l, 15 mg/l) and combined Cu and Zn (1.3 mg/l & 5 mg/l; 2.6 mg/l & 10 mg/l; 3.9 mg/l & 15 mg/l) under ambient conditions (17°C). Bioaccumulation in soft tissue increased proportionally with heavy metal concentrations in water, although observed levels remained within World Health Organisation (WHO) guidelines and did not vary significantly over time. Cu and Zn bioaccumulation followed a first-order kinetic association, with indicative agreement between observed and estimated values. Cu exhibited a tendency towards higher accumulation than Zn under the conditions investigated. In hard tissue (shell), metal concentrations were consistently higher than in soft tissue for both control and experimental treatments, demonstrating proportional bioaccumulation relative to metal concentrations in the aquatic phase. The high levels in the control reflect the fact that the calms had been accumulating metals within their shells throughout their lives prior to the experiment in the natural environment. Cu and Zn bioaccumulation in the shell also followed a first-order kinetic association. This study highlights the need to identify appropriate bioindicators for monitoring and managing heavy metal contamination. The results also indicate that Corbicula fluminea may serve as a potential bio-indicator of heavy metal contamination and pollutant transfer pathways in freshwater ecosystems.
The rhizosphere of medicinal plants represents a dynamic microbial hotspot harbouring taxonomically and functionally diverse fungal communities that drive nutrient cycling, organic matter decomposition, soil biogeochemical processes, and plant health through complex mutualistic, saprotrophic, and pathogenic interactions. This study comprehensively investigated the rhizospheric fungal diversity associated with Bergenia ciliata across three different sampling sites in the Kashmir Himalaya, using an integrated approach combining culture-dependent isolation and Illumina MiSeq metabarcoding. Both approaches revealed clear spatial structuring of fungal communities across the sampling sites, indicating site-associated variation in fungal community composition across the sites. Although environmental variables were not directly quantified in the present study, the observed community turnover and differences in putative functional guild composition suggest that site-associated conditions may contribute to the observed variation in fungal community structure. Metabarcoding analysis revealed substantially greater taxonomic diversity, including rare and unculturable taxa, and identified a Basidiomycota-dominated fungal community, with Wallemia as the most abundant genus across sites. In contrast, culture-based isolation captured only a subset of this diversity, primarily fast-growing filamentous fungi such as Aspergillus, Fusarium, Penicillium, and Trichoderma, underscoring the complementary nature and inherent biases of the two approaches. Integrated alpha and beta diversity analyses from both metabarcoding and culture-based approaches revealed clear site-specific variation in rhizospheric fungal communities associated with B. ciliata, with distinct community composition at Gulmarg and intermediate compositional characteristics relative to Aharbal and Sinthan Pass. FUNGuild-based functional annotation indicated that putative saprotrophs represented the dominant predicted trophic guild across all sampling sites, with substantial contributions from putative pathotroph–saprotroph taxa. Although Illumina metabarcoding revealed a broadly similar distribution of predicted trophic guilds among the three sampling sites, culture-dependent analyses showed greater site-associated variation and a higher representation of putative symbiotrophic and multifunctional guilds. These complementary approaches together provided a broader assessment of the potential functional composition of rhizospheric fungal communities associated with B. ciliata. Overall, integrating culture-dependent and metabarcoding approaches improved the resolution of rhizospheric fungal diversity and putative functional guild composition, demonstrating clear site-associated variation in fungal communities and providing a foundation for future investigations on fungal community ecology and plant–microbe interactions in Himalayan medicinal plants.
Myanmar migrants residing in Thailand may face diverse barriers—such as financial difficulties, housing condition limitations, and time constraints—when performing indoor air quality (IAQ) management practices. These barriers may lessen the motivation to perform such practices, even though Myanmar migrants have risk perceptions, sufficient knowledge, and skills related to IAQ management. This study aims to examine how IAQ literacy, perceived barriers, and protection motivation theory (PMT) variables—such as perceived threat (perceived susceptibility and severity) and perceived capacity (response efficacy and self-efficacy)—are associated with IAQ management practices. It also aims to examine how perceived barriers moderate the association of IAQ literacy and PMT constructs with IAQ management practices. A cross-sectional study was conducted using questionnaire surveys among Myanmar migrant communities in Samut Sakhon Province, Thailand, from 15 to 30 June 2025. The participants included 421 Myanmar migrants residing in Samut Sakhon Province for at least one year at the time of data collection. The analysis of the measurement model was conducted using confirmatory factor analysis to test the reliability of the measurements. Questionnaire items with low factor loadings were removed to enhance convergent validity. Then, a hierarchical linear regression model was analyzed to test the association of IAQ literacy and PMT constructs with IAQ management practices. In addition, the moderating effect of perceived barriers was evaluated. The results revealed that PMT constructs—such as self-efficacy and perceived severity—were significantly associated with IAQ management practices only when they interacted with perceived barriers. The association of self-efficacy and perceived severity with IAQ management practices was strong when perceived barriers were high. In addition, perceived barriers negatively moderated the positive relationship between IAQ literacy and management practices, indicating that this positive relationship weakened when perceived barriers were high. The results have implications for the development of communication strategies to enhance motivation to perform IAQ management practices when faced with diverse barriers.
The decarbonization of agro-industrial systems is critical for reducing greenhouse gas (GHG) emissions and improving energy efficiency. This study investigated the transition of a large-scale rice mill in Thailand from a conventional to a low-carbon system through the integration of solar photovoltaic (PV) energy, electrification of internal transportation, and process optimization. Using five years of operational data (2020–2024), energy consumption and CO₂ emissions were evaluated across four scenarios representing progressive stages of low-carbon implementation. The analysis applied Thailand Greenhouse Gas Management Organization (TGO) emission factors and used specific energy consumption (SEC) and emission intensity as key performance indicators. Results showed that renewable energy integration significantly improved performance, reducing SEC from 0.439 to 0.365 kWh/kg in the early transition stages. However, full electrification and increased production demand led to a slight rebound in SEC (up to 0.383 kWh/kg), highlighting the importance of integrated energy management. CO₂ emission intensity consistently declined across scenarios, with the most advanced configuration achieving a 26.99% reduction compared to the baseline and reducing total emissions by over 2.4 million kgCO₂e. The findings demonstrate that effective decarbonization in rice milling requires a system-level approach integrating renewable energy, electrification, and operational optimization to achieve sustained environmental benefits.
Climate change poses a major threat to agricultural production in Sub-Saharan Africa, with Ethiopia being particularly vulnerable due to its strong dependence on rain-fed agriculture. Understanding farmers’ perceptions of climate change, including gender differences in these perceptions, is critical for designing effective and inclusive adaptation strategies. Therefore this study examined historical climate trends and gender-based perceptions of climate change among smallholder farmers in the East Shewa Zone, Ethiopia. A cross-sectional research design was employed, with data collected from 446 households through structured surveys, focus group discussions (FGDs), and key informant interviews. Historical temperature and rainfall records (1992–2024) were obtained from the National Meteorological Agency. Climate trends were analyzed using the Mann–Kendall trend test and Sen’s slope estimator, while household perceptions were assessed using a Likert scale and Severity Index (SI). Qualitative data were analyzed using narrative analysis. The results revealed a significant warming trend across the study area, with mean annual temperature increasing by 0.057°C year⁻¹ (p < 0.001). The highest warming rates were observed in Boset and Fantale, where maximum temperatures increased by up to 0.141°C year⁻¹, indicating increasing exposure to heat stress. Rainfall analysis showed an overall declining trend across monthly, seasonal, and annual timescales. Annual rainfall declined significantly by 4.78 mm year⁻¹ based on Sen’s slope estimator (p = 0.034). Kiremt rainfall also showed a significant declining trend (−3.80 mm year⁻¹; p = 0.040), while Bega (−1.45 mm year⁻¹) and Belg (−0.90 mm year⁻¹) exhibited declining but statistically non-significant trends, reflecting substantial interannual variability. Significant rainfall reductions were also observed during May and June, coinciding with the onset of the main cropping season. Farmers’ perceptions closely corresponded with observed climatic trends, with 90.1% of respondents perceiving increasing temperatures and 77.6% perceiving declining rainfall. Significant gender differences were observed (p < 0.05), with male-headed households reporting greater awareness of increasing temperature and declining rainfall. Therefore, strengthening gender-responsive extension services, climate information systems, water harvesting, supplementary irrigation, drought-tolerant crop varieties, and soil and water conservation practices is essential for enhancing the resilience of smallholder farming systems in the East Shewa Zone.
Irrigated agricultural landscapes are increasingly recognized as critical interfaces where nutrient enrichment and trace metal (PTEs) accumulation may threaten soil health, food safety, and downstream ecosystems. Yet, quantitative assessments of sediment geochemistry in intensifying irrigation systems of sub-Saharan Africa remain scarce. Here, we investigate the spatial distribution and environmental risk of PTEs in irrigation sediments across Usangu Basin, one of Southern Tanzania’s most important rice-producing area. A total of 48 composite sediment samples were collected from 7 irrigation schemes i.e., Mubuyuni, Igalako, Chang'ombe, Uturo, Kapunga, Mahongole, and Ukwavile in Mbarali district. Schemes were classified into two groups (Group I dominated by agricultural areas versus Group II dominated by residential and agricultural areas) and analyzed for physicochemical properties and PTEs (Cr, Pb, Cd, and Co) using aqua regia digestion followed by ICP-OES and ICP-MS analysis. Statistical comparisons between scheme groups employed both parametric (independent samples t-test) and non-parametric (Mann-Whitney U) methods based on normality testing, supplemented by effect size calculations (Cohen's d and rank-biserial correlation r) to assess practical magnitude of observed differences. Sediment pH ranged from 4.94 to 6.96 (mean 5.66 ± 0.57), reflecting moderately acidic conditions typical of tropical paddy soils. Mean Cr, Pb, Cd, and Co concentrations were 0.006, 0.756, 0.101, and 1.224 mg/kg, respectively. Cd exhibited most pronounced spatial variation, with concentrations 45% higher in irrigation schemes influenced by mixed residential-agricultural land use (0.132 mg/kg) compared with predominantly agricultural schemes (0.091 mg/kg). Despite these variations, measured concentrations remained well below international sediment quality guideline limits, indicating limited current contamination. Pollution indices further confirmed low environmental risk, with contamination factors (CF < 1), geo-accumulation indices indicating uncontaminated sediments (Igeo < 0), and an overall pollution load index of 0.0028. Correlation analysis suggests that nutrient enrichment and PTEs occurrence may partly reflect combined influences of fertilizer inputs, irrigation sediment transport, and local land-use interactions. Although current contamination levels remain low, observed Cd enrichment highlights potential PTEs gradual accumulation in intensively irrigated rice systems. These results provide the first basin-scale baseline assessment of irrigation sediment geochemistry in Usangu basin, offering critical evidence for environmental monitoring, sustainable fertilizer management, and long-term protection of soil and water resources in Tanzanian irrigation landscapes.
Agriculture, forestry and other land use (AFOLU) is simultaneously a source of greenhouse-gas emissions and a means of atmospheric carbon removal, yet these two functions are often assessed separately and projected without consistent accounting identities. We constructed a 31-province annual framework for mainland China, covering 2005-2023 historically and 2024-2050 under SSP1-2.6, SSP2-4.5 and SSP3-7.0. Crop emissions, livestock emissions and forest removals were estimated separately and organized through five prespecified system states. Six STIRPAT, random-forest (RF) and residual/stacked candidates were compared using a prespecified 2019-2023 temporal test block, rolling-origin tests and leave-one-region-out validation. Gross AFOLU emissions increased from 978.1 Mt CO₂e in 2005 to 1,067.6 Mt CO₂e in 2023, while estimated forest removals increased from 352.2 to 675.4 Mt CO₂e and reduced net AFOLU emissions from 625.8 to 392.2 Mt CO₂e. Livestock accounted for the largest share of gross emissions, with enteric fermentation contributing approximately 522.1 Mt CO₂e in 2023. No single model dominated all outcomes: shrinkage residual STIRPAT-RF performed best for gross emissions, whereas standalone RF performed best for net emissions and land carbon-emission intensity. Under joint accounting, residual, model-weight and scenario uncertainty, all three pathways had negative median net balances in 2050, but their 90% intervals crossed zero; the probability of a negative national balance ranged from 0.857 to 0.933. These results identify a high-probability but conditional transition toward net-negative AFOLU, contingent on sustained forest removals, effective livestock mitigation and the continued validity of model relationships beyond the historical domain.
Coal remains a widely used energy source across multiple industrial sectors, but its combustion contributes significantly to atmospheric emissions of nitrogen oxides (NOₓ), polycyclic aromatic hydrocarbons (PAHs), heavy metals, and other pollutants. Among available post-combustion control technologies, Selective Catalytic Reduction (SCR) stands out as the most effective method for NOₓ reduction due to its high efficiency, operational flexibility, and compatibility with existing flue gas treatment systems. This review presents a comprehensive assessment of research advances in SCR for coal-fired plants, with emphasis on catalyst innovations, technological improvements, predictive modeling techniques, and integration with multi-pollutant control strategies. The review explores recent developments in catalyst formulation such as elemental doping, surface engineering, and multifunctional materials that enhance low-temperature activity, resistance to deactivation, and co-removal of pollutants like VOCs and Hg⁰. In addition, advances in computational modeling, including CFD, neural networks, and hybrid AI algorithms, are evaluated for their ability to optimize NOₓ reduction and ammonia injection strategies under fluctuating load conditions. Emerging alternatives to SCR such as SNCR, plasma-assisted systems, and wet scrubbing are also critically examined. Persistent challenges, including catalyst deactivation, ammonia slip, and ineffective gas-phase PAH removal, highlight the need for more robust, intelligent, and environmentally sustainable emission control solutions. This review identifies key knowledge gaps and outlines future research directions to support the development of next-generation SCR systems within a circular and cleaner coal combustion framework.
This study critically examines the application of the Theory of Planned Behavior (TPB) in explaining pro-environmental intentions and behaviors across forest-related contexts. While TPB has been widely adopted as a dominant framework in environmental research, its consistency and explanatory capacity in the forestry domain remain unclear. To address this gap, this study synthesizes empirical evidence from a broad range of TPB-based studies conducted in different forest settings, including urban forests, natural forests, and community-managed landscapes. The findings reveal substantial variability in the influence of TPB’s core constructs. Attitude, subjective norms, and perceived behavioral control do not exhibit stable or uniform effects on behavioral intention, indicating that the model operates in a context-dependent manner. In several cases, perceived behavioral control or subjective norms outperform attitude, while in others, their effects are negligible. Moreover, extended TPB models generally demonstrate improved explanatory power; however, these extensions are fragmented and lack theoretical coherence. Overall, the results highlight the need for more context-sensitive and theoretically integrated approaches to better understand human behavior in forest and environmental management contexts.
This study assessed heavy metal contamination in surface waters of Lake Hawassa, the Boicha River and effluents from textile, brewery and soft-drink factories. Samples from 10 sites were analyzed for As, Cd, Hg, Pb, Cr, Ni, Cu and Zn using ICP-OES and compared with Ethiopian Standards Agency (ESA) and World Health Organization (WHO) guideline value. Human-health risks were assessed using the Hazard Quotient (HQ), Hazard Index (HI), and Incremental Lifetime Cancer Risk (ILCR). Results revealed higher contamination in riverine and industrial sites than lake sites. Most lake sites remained below guideline values, except Tikurwuha, where As and Pb exceeded WHO limits. Cr, Cd, Pb, and As exceeded WHO limits at several riverine and industrial sites. Cd and Hg exceeded FAO irrigation standards, indicating potential risks of soil contamination, crop uptake, and food-chain transfer. HI values exceeded the threshold of 1 for children (1.78–10.66), whereas those for adults ranged from 0.76 to 4.55, indicating non-carcinogenic health risks. Carcinogenic risk estimates for As and Cr ranged from 0.76 × 10-4 to 3.99 × 10-4 in children and from 3.82 × 10-4 to 19.95 × 10-4 in adults, with values exceeding the USEPA acceptable limit of 1 × 10-4 at most sites, indicating elevated lifetime cancer risk. Severe contamination occurred at riverine and industrial sites, with elevated metal levels at Tikurwuha inflow zone of Lake Hawassa. Spatial patterns suggest anthropogenic contributions, but tracer-based and isotopic studies are needed to confirm pathways. Single-season sampling and limited exposure pathways indicate current rather than year-round conditions, supporting pollution control, remediation, and monitoring.
Background Reliable assessment of long-term fine particulate matter (PM2.5) and World Health Organization (WHO) guideline non-attainment remains challenging in sparse, evolving monitoring networks. Urban emissions and seasonal biomass burning may obscure persistent spatial patterns. We developed a reproducible framework for uncertainty-qualified annual screening surfaces using routine regulatory data. Methods Hourly PM2.5 data (2017–2023) from Northeastern (NE) Thailand were standardized into quality-controlled (QC) station-level time series. Short gaps (≤24 h) were interpolated; longer gaps were preserved; extreme hourly values were winsorized at the 5th and 95th percentiles (P05-P95). Annual means were computed under two completeness criteria (≥270 valid days/year, strict; ≥120 valid days/year, mapping support). Annual concentration surfaces (5-km grid) were generated using inverse distance weighting. Reliability was evaluated via leave-one-out cross-validation (LOOCV) for p = 1–3; p = 2 residuals were assessed using exploratory Global Moran’s I. Non-attainment was classified using the WHO 2021 interim target 3 (WHO IT3; 15 µg/m³). Results Substantial spatial heterogeneity was observed, with station medians ranging from 13.70 to 28.09 µg/m³ and a maximum daily mean of 65.25 µg/m³. Network-level annual means declined from 29.97 µg/m³ in 2018 to 21.42 µg/m³ in 2022, before increasing to 25.67 µg/m³ in 2023. Harmonized raster classification showed non-attainment across 100.00% of the classified domain in 2019–2020, 92.68% in 2021, 89.52% in 2022, and 98.88% in 2023. Under p = 2, pooled LOOCV root mean square error (RMSE) and mean absolute error (MAE) were 6.85 and 5.70 µg/m³, respectively. Negative R² values indicated limited predictive performance, while p = 1 performed best and no significant residual clustering was detected during 2020–2023. Conclusions Persistent, near-domain-wide WHO IT3 PM2.5 non-attainment was observed amid sparse, evolving monitoring. By integrating QC transparency, completeness regimes, and spatial uncertainty diagnostics, this framework provides reproducible screening-level spatial products for regional PM2.5 assessment and monitoring network evaluation in data-limited tropical regions.
