Water quality variations in different climates of Iran: toward modeling total dissolved solid using soft computing techniques

In present paper, wavelet analysis of total dissolved solid that monitored at Nazlu Chay (northwest of Iran), Tajan (north of Iran), Zayandeh Rud (central of Iran) and Helleh (south of Iran) basins with various climatic conditions, have been studied. Daubechies wavelet at suitable level (db4) has been calculated for TDS of each selected basins. The performance of artificial neural networks (ANN), two different adaptive-neurofuzzy inference system (ANFIS) including ANFIS with grid partition (ANFIS-GP) and ANFIS with subtractive clustering (ANFIS-SC), gene expression programming (GEP), wavelet-ANN, wavelet-ANFIS and wavelet-GEP in predicting TDS of mentioned basins were assessed over a period of 20 years at twelve different hydrometric stations. EC (μmhos/cm), Na (meq L^?1) and Cl (meq L^?1) parameters were selected (based on Pearson correlation) as input variables to forecast amount of TDS in four studied basins. To develop hybrid wavelet-AI models, the original observed data series was decomposed into sub-time series using Daubechies wavelets at suitable level for each basin. Based on the statistical criteria of correlation coefficient (R), root mean square error (RMSE) and mean absolute error (MAE), the hybrid wavelet-AI models performance were better than single AI models in all basins. A comparison was made between these artificial intelligence approaches which emphasized the superiority of wavelet-GEP over the other intelligent models with amount of RMSE 18.978, 6.774, 9.639 and 318.363 mg/l, in Nazlu Chay, Tajan, Zayandeh Rud and Helleh basins, respectively.     

Evaporation estimation using artificial neural networks and adaptive neuro-fuzzy inference system techniques

Evaporation, as a major component of the hydrologic cycle, plays a key role in water resources development and management in arid and semi-arid climatic regions. Although there are empirical formulas available, their performances are not all satisfactory due to the complicated nature of the evaporation process and the data availability. This paper explores evaporation estimation methods based on artificial neural networks (ANN) and adaptive neuro-fuzzy inference system (ANFIS) techniques. It has been found that ANN and ANFIS techniques have much better performances than the empirical formulas (for the test data set, ANN R² = 0.97, ANFIS R² = 0.92 and Marciano R² = 0.54). Between ANN and ANFIS, ANN model is slightly better albeit the difference is small. Although ANN and ANFIS techniques seem to be powerful, their data input selection process is quite complicated. In this research, the Gamma test (GT) has been used to tackle the problem of the best input data combination and how many data points should be used in the model calibration. More studies are needed to gain wider experience about this data selection tool and how it could be used in assessing the validation data.     

Adaptive neuro-fuzzy inference system coupled with shuffled frog leaping algorithm for predicting river streamflow time series

ABSTRACT

Accurate runoff forecasting plays a key role in catchment water management and water resources system planning. To improve the prediction accuracy, one needs to strive to develop a reliable and accurate forecasting model for streamflow. In this study, the novel combination of the adaptive neuro-fuzzy inference system (ANFIS) model with the shuffled frog-leaping algorithm (SFLA) is proposed. Historical streamflow data of two different rivers were collected to examine the performance of the proposed model. To evaluate the performance of the proposed ANFIS-SFLA model, six different scenarios for the model input–output architecture were investigated. The results show that the proposed ANFIS-SFLA model (R² = 0.88; NS = 0.88; RMSE = 142.30 (m³/s); MAE = 88.94 (m³/s); MAPE = 35.19%) significantly improved the forecasting accuracy and outperformed the classic ANFIS model (R² = 0.83; NS = 0.83; RMSE = 167.81; MAE = 115.83 (m³/s); MAPE = 45.97%). The proposed model could be generalized and applied in different rivers worldwide.     

Regional flood frequency analysis using support vector regression in arid and semi-arid regions of Iran

Regional flood frequency analysis (RFFA) was carried out on data for 55 hydrometric stations in Namak Lake basin, Iran, for the period 1992–2012. Flood discharge of specific return periods was computed based on the log Pearson Type III distribution, selected as the best regional distribution. Independent variables, including physiographic, meteorological, geological and land-use variables, were derived and, using three strategies – gamma test (GT), GT plus classification and expert opinion – the best input combination was selected. To select the best technique for regionalization, support vector regression (SVR), adaptive neuro-fuzzy inference system (ANFIS), artificial neural network (ANN) and nonlinear regression (NLR) techniques were applied to predict peak flood discharge for 2-, 5-, 10-, 25-, 50- and 100-year return periods. The GT + ANFIS and GT + SVR models gave better performance than the ANN and NLR models in the RFFA. The results of the input variable selection showed that the GT technique improved the model performance.     

A minimalistic approach for evapotranspiration estimation using the Prophet model

ABSTRACT

This study aimed to evaluate the potential of the recently introduced Prophet model for estimating reference evapotranspiration (ETo). A comparative study was conducted for benchmarking the model results with support vector regression (SVR) and temperature-based empirical models (Thornthwaite and Hargreaves) in southern Japan. The performance of the Prophet, SVR and temperature-based empirical models was evaluated by Nash–Sutcliffe efficiency (NSE) and coefficient of determination (R²). The results indicate that temperature-based Prophet and SVR models have greater accuracy than the empirical models. The Prophet model with sole input of relative humidity, sunshine hours or windspeed showed acceptable accuracy (NSE > 0.80; R² > 0.80), while SVR models with similar inputs showed greater errors. Accuracy improved with increasing number of input parameters, giving excellent performance (NSE > 0.95; R² > 0.95) with all input parameters. Hence, the Prophet model is a new promising approach for modelling ETo with limited input variables.     

Multi-step water quality forecasting using a boosting ensemble multi-wavelet extreme learning machine model

The use of electrical conductivity (EC) as a water quality indicator is useful for estimating the mineralization and salinity of water. The objectives of this study were to explore, for the first time, extreme learning machine (ELM) and wavelet-extreme learning machine hybrid (WA-ELM) models to forecast multi-step-ahead EC and to employ an integrated method to combine the advantages of WA-ELM models, which utilized the boosting ensemble method. For comparative purposes, an adaptive neuro-fuzzy inference system (ANFIS) model, and a WA-ANFIS model, were also developed. The study area was the Aji-Chay River at the Akhula hydrometric station in Northwestern Iran. A total of 315 monthly EC (µS/cm) datasets (1984–2011) were used, in which the first 284 datasets (90% of total datasets) were considered for training and the remaining 31 (10% of total datasets) were used for model testing. Autocorrelation function (ACF) and partial autocorrelation function (PACF) demonstrated that the 6-month lags were potential input time lags. The results illustrated that the single ELM and ANFIS models were unable to forecast the multi-step-ahead EC in terms of root mean square error (RMSE), coefficient of determination (R²) and Nash–Sutcliffe model efficiency coefficient (NSC). To develop the hybrid WA-ELM and WA-ANFIS models, the original time series of lags as inputs, and time series of 1, 2 and 3 month-step-ahead EC values as outputs, were decomposed into several sub-time series using different maximal overlap discrete wavelet transform (MODWT) functions, namely Daubechies, Symlet, Haar and Coiflet of different orders at level three. These sub-time series were then used in the ELM and ANFIS models as an input dataset to forecast the multi-step-ahead EC. The results indicated that single WA-ELM and WA-ANFIS models performed better than any ELM and ANFIS models. Also, WA-ELM models outperformed WA-ANFIS models. To develop the boosting multi-WA-ELM and multi-WA-ANFIS ensemble models, a least squares boosting (LSBoost) algorithm was used. The results showed that boosting multi-WA-ELM and multi-WA-ANFIS ensemble models outperformed the individual WA-ELM and WA-ANFIS models.     

Electromagnetic induction prediction of soil salinity and groundwater properties in a Tunisian Saharan oasis

Abstract

Electromagnetic induction measurements (EM) were taken in a saline gypsiferous soil of the Saharan-climate Fatnassa oasis (Tunisia) to predict the electrical conductivity of saturated soil extract (ECe) and shallow groundwater properties (depth, Dgw, and electrical conductivity, ECgw) using various models. The soil profile was sampled at 0.2 m depth intervals to 1.2 m for physical and chemical analysis. The best input to predict the log-transformed soil salinity (lnECe) in surface (0–0.2 m) soil was the EMh/EMv ratio. For the 0–0.6 m soil depth interval, the performance of multiple linear regression (MLR) models to predict lnECe was weaker using data collected over various seasons and years (R _a ² = 0.66 and MSE = 0.083 dS m^-1) as compared to those collected during the same period (R _a ² = 0.97, MSE = 0.007 dS m^-1). For similar seasonal conditions, for the Dgw–EMv relationship, R ² was 0.88 and the MSE was 0.02 m for Dgw prediction. For a validation subset, the R ² was 0.85 and the MSE was 0.03 m. Soil salinity was predicted more accurately when groundwater properties were used instead of soil moisture with EM variables as input in the MLR.

Editor D. Koutsoyiannis; Associate editor K. Heal

Citation Bouksila, F., Persson, M., Bahri, A., and Berndtsson, R., 2012. Electromagnetic induction predictions of soil salinity and groundwater properties in a Tunisian Saharan oasis. Hydrological Sciences Journal, 57 (7), 1473–1486.     

Estimation of monthly reference evapotranspiration using novel hybrid machine learning approaches

ABSTRACT

In this research, five hybrid novel machine learning approaches, artificial neural network (ANN)-embedded grey wolf optimizer (ANN-GWO), multi-verse optimizer (ANN-MVO), particle swarm optimizer (ANN-PSO), whale optimization algorithm (ANN-WOA) and ant lion optimizer (ANN-ALO), were applied for modelling monthly reference evapotranspiration (ET_o) at Ranichauri (India) and Dar El Beida (Algeria) stations. The estimates yielded by hybrid machine learning models were compared against three models, Valiantzas-1, 2 and 3 based on root mean square error (RMSE), Nash-Sutcliffe efficiency (NSE), Pearson correlation coefficient (PCC) and Willmott index (WI). The results of comparison show that the ANN-GWO-1 model with five input variables (T_min, T_max, RH, U_s, R_s) provides better estimates at both study stations (RMSE = 0.0592/0.0808, NSE = 0.9972/0.9956, PCC = 0.9986/0.9978, and WI = 0.9993/0.9989). Also, the adopted modelling strategy can build a truthful expert intelligent system for estimating the monthly ET_o at study stations.     

Variations in turbulent kinetic energy at a point source submarine groundwater discharge in a reef lagoon

The influence of sea level variations due to tides and wave setup on turbulent kinetic energy (TKE) was observed at a point source submarine groundwater discharge in a fringing coral reef lagoon. Tidal and wave setup variations modulated speed, TKE, TKE dissipation, and water temperature and salinity at the buoyant jet. The primary driver of jet TKE and speed variations was tides, while wave setup was a minor contributor. An inverse relationship between surface elevation and TKE was explained with an exponential equation based on sea level variations. During low tides, peak jet speeds (up to 0.3 m s^?1) and TKE per unit mass (up to 0.4 m² s^?2) were observed. As high tide approached, the jet produced minimum TKE of ~0.003 m² s^?2 and TKE dissipation ranged from 2 to 8×10^?4 m² s^?3. This demonstrated the sensitivity of the jet discharge to tides despite the small tidal range (<20 cm). Jet temperatures and salinities displayed semidiurnal oscillations with minimum salinity and temperature values during maximum discharge. Jet salinities increased throughout low tides while temperatures decreased. This pattern suggested the jet conduit was connected to a stratified cavity within the aquifer containing cool fresh water over cool salty water. As low tides progressed, jet outflow increased in salinity because of the mixing within the conduit, while lower jet temperatures suggested water coming from further or deeper in the aquifer. The presence of such a cavity has been recently confirmed by divers.     

Seasonal Short‐Term Prediction of Dissolved Oxygen in Rivers via Nature‐Inspired Algorithms

This study challenges the use of three nature‐inspired algorithms as learning frameworks of the adaptive‐neuro‐fuzzy inference system (ANFIS) machine learning model for short‐term modeling of dissolved oxygen (DO) concentrations. Particle swarm optimization (PSO), butterfly optimization algorithm (BOA), and biogeography‐based optimization (BBO) are employed for developing predictive ANFIS models using seasonal 15 min data collected from the Rock Creek River in Washington, DC. Four independent variables are used as model inputs including water temperature (T), river discharge (Q), specific conductance (SC), and pH. The Mallow's Cp and R² parameters are used for choosing the best input parameters for the models. The models are assessed by several statistics such as the coefficient of determination (R²), root‐mean‐square error (RMSE), Nash–Sutcliffe efficiency, mean absolute error, and the percent bias. The results indicate that the performance of all‐nature‐inspired algorithms is close to each other. However, based on the calculated RMSE, they enhance the accuracy of standard ANFIS in the spring, summer, fall, and winter around 13.79%, 15.94%, 6.25%, and 12.74%, respectively. Overall, the ANFIS‐PSO and ANFIS‐BOA provide slightly better results than the other ANFIS models.     

Viriobenthos in aquatic sediments: variability in abundance and production and impact on the C-cycle

In the last 15 years viruses have been acknowledged as important components of the benthic microbial community, but our understanding of their role in the functioning of aquatic systems remains poor. Viruses can affect bacterial assemblages and mineralization activities, but the extent of their influence remains unclear. We synthesised available data on viriobenthos dynamics to understand which factors drive the variability in their abundance and production and to quantify their influence on the benthic carbon cycle. Results highlighted a large variability in viral abundance (from 2 × 10⁸ to 7 × 10⁹ virus ml^?1) and production estimates (from 1 × 10⁷ to 5 × 10⁸ virus ml^?1 h^?1) obtained with different techniques. This variability limits the comparability of data across studies and indicates the need to improve protocols and develop standard methods. The dynamics of viruses infecting prokaryotes appeared linked to prokaryotic metabolism, supporting the hypothesis that benthic viruses originate directly in the sediment as a result of infection events rather than sinking from the water column. Sediment characteristics (porosity, temperature, depth) appeared to effect viral production, mostly indirectly by influencing bacterial productivity and abundance, but possibly also interfering with the rate of virus–host encounter. Conversely, trophic status appeared unrelated to viral parameters. Viral contribution to carbon turnover appeared low and unrelated to temperature, water depth, trophic status and salinity. More detailed studies are needed to understand the pelagic contribution to the viriobenthos and the extent to which dissolved organic carbon released by viruses is effectively used by bacteria.     

Modelling of infiltration using artificial intelligence techniques in semi-arid Iran

ABSTRACT

Infiltration plays a fundamental role in streamflow, groundwater recharge, subsurface flow, and surface and subsurface water quality and quantity. In this study, adaptive neuro-fuzzy inference system (ANFIS), support vector machine (SVM) and random forest (RF) models were used to determine cumulative infiltration and infiltration rate in arid areas in Iran. The input data were sand, clay, silt, density of soil and soil moisture, while the output data were cumulative infiltration and infiltration rate, the latter measured using a double-ring infiltrometer at 16 locations. The results show that SVM with radial basis kernel function better estimated cumulative infiltration (RMSE = 0.2791 cm) compared to the other models. Also, SVM with M4 radial basis kernel function better estimated the infiltration rate (RMSE = 0.0633 cm/h) than the ANFIS and RF models. Thus, SVM was found to be the most suitable model for modelling infiltration in the study area.     

Modification of the deep salinity-maximum in the Southern Ocean by circulation in the Antarctic Circumpolar Current and the Weddell Gyre

The evolution of the deep salinity-maximum associated with the Lower Circumpolar Deep Water (LCDW) is assessed using a set of 37 hydrographic sections collected over a 20-year period in the Southern Ocean as part of the WOCE/CLIVAR programme. A circumpolar decrease in the value of the salinity-maximum is observed eastwards from the North Atlantic Deep Water (NADW) in the Atlantic sector of the Southern Ocean through the Indian and Pacific sectors to Drake Passage. Isopycnal mixing processes are limited by circumpolar fronts, and in the Atlantic sector, this acts to limit the direct poleward propagation of the salinity signal. Limited entrainment occurs into the Weddell Gyre, with LCDW entering primarily through the eddy-dominated eastern limb. A vertical mixing coefficient, κ_V of (2.86 ± 1.06) × 10^?4 m² s^?1 and an isopycnal mixing coefficient, κ_I of (8.97 ± 1.67) × 10² m² s^?1 are calculated for the eastern Indian and Pacific sectors of the Antarctic Circumpolar Current (ACC). A κ_V of (2.39 ± 2.83) × 10^?5 m² s^?1, an order of magnitude smaller, and a κ_I of (2.47 ± 0.63) × 10² m² s^?1, three times smaller, are calculated for the southern and eastern Weddell Gyre reflecting a more turbulent regime in the ACC and a less turbulent regime in the Weddell Gyre. In agreement with other studies, we conclude that the ACC acts as a barrier to direct meridional transport and mixing in the Atlantic sector evidenced by the eastward propagation of the deep salinity-maximum signal, insulating the Weddell Gyre from short-term changes in NADW characteristics.     

Forced oscillation measurements of seismic attenuation in fluid saturated sandstone

Adopting the method of forced oscillation, attenuation was studied in Fontainebleau sandstone (porosity 10%, permeability 10 mD) at seismic frequencies (1–100 Hz). Confining pressures of 5, 10, and 15 MPa were chosen to simulate reservoir conditions. First, the strain effect on attenuation was investigated in the dry sample for 11 different strains across the range 1 × 10^?6–8 × 10^?6, at the confining pressure of 5 MPa. The comparison showed that a strain of at least 5 × 10^?6 is necessary to obtain a good signal to noise ratio. These results also indicate that nonlinear effects are absent for strains up to 8 × 10^?6. For all the confining pressures, attenuation in the dry rock was low, while partial (90%) and full (100%) saturation with water yielded a higher magnitude and frequency dependence of attenuation. The observed high and frequency dependent attenuation was interpreted as being caused by squirt flow.     

Bayesian network-based risk assessment for hazmat transportation on the Middle Route of the South-to-North Water Transfer Project in China

A Bayesian network-based risk assessment (BN-RA) model was developed to assess the risk of hazmat transportation by identifying, modeling, and quantitatively calculating the risk on the Middle Route of the South-to-North Water Transfer Project (MRSNWTP) in China. First, we selected seven parameters from five categories of impact factors (i.e., human, vehicle, tank, weather, and road environment) as quintessential risk factors for accidents. Second, we used the developed BN-RA model to predict the probability of accidents. Third, using bidirectional inference in the BN approach, we analyzed and ranked the importance of the effects of these factors. The developed model was subsequently applied to assess the risks of major bridges crossing canals with different pavement grades and traffic flow levels both at present and in the future for the Beijing-Shijiazhuang Section of the MRSNWTP. The results indicated the following: (1) Although the overall potential risk of hazmat transportation accidents on all bridges in the Beijing-Shijiazhuang Section was fairly low (e.g., 0.08 %), the impacts cannot be ignored because of the potential for huge losses. (2) According to the analysis of many factors that may affect accidents, the driving patterns of drivers exerted the strongest influence on the probability of an accident, followed by vehicle conditions and lighting conditions. (3) If a vehicle were to fail, the highest probability (0.17 %) of an accident would arise if it were traveling on a road with no street lighting and poor road conditions at night. (4) Assuming that a vehicle was in good condition, the highest probability (0.12 %) of an accident arised when the vehicle suddenly encountered poor road conditions with no lights on a foggy night. (5) The predicted probabilities of accidents on Bridge TCWRR (short for the Tang County West Ring Road Bridge) in the short (i.e., the year 2017), medium (i.e., the year 2022) and long terms (i.e., the year 2027) were 3.25 × 10^?4, 5.37 × 10^?4, and 8.89 × 10^?4, respectively. For Bridge DNR (short for the Dian Bridge on the North Road), these values were 8.64 × 10^?6, 1.02 × 10^?5, and 1.21 × 10^?5, respectively. Based on the risk assessment results, to lower the accident probability and avoid the serious consequences resulting from hazmat transportation accidents, we developed an appropriate emergency response program to reduce potential hazards. This research resolved the problems of randomness and uncertainty associated with hazmat transportation in the MRSNWTP and can provide a reference for the effective prevention of hazmat transportation accidents and scientific decision-making in risk management.     

Cancer risk estimation from dietary arsenic,a new approach from longitudinal cohort study

Inorganic arsenic is a carcinogen and consumption in low dose may lead to cancer. We estimated the cancer risk of the participants from arsenic endemic regions of West Bengal, India. The probable cancer risk was estimated following the assessment of daily inorganic arsenic intake through drinking water and diets of 20 participants for three consecutive years who had been using low arsenic water in the Indian context (median arsenic concentration in the study Years-I, II and III were 22, 16, 13 µg/l respectively). Probable cancer risk of the population was 2.80 × 10^?4, 2.94 × 10^?4, 3.12 × 10^?4 in the three respective study years (Year-I, II and III); just higher than the US EPA risk level of concern. The arsenic species content of the paired raw, cooked rice and urine was estimated in the as is taken basis. The major diet component, rice contained 72–86% inorganic arsenic whereas urine contains 70% organic arsenic on an average. The cancer risk assessment has been proposed to be modified by inclusion of urine arsenic release, considering the fact of arsenic release through urine. The risk became 1.28 × 10^?5, 1.13 × 10^?5, 1.01 × 10^?5 in the study Year-I, II and III respectively, considering urinary arsenic release, attributed the consideration of urine arsenic release into probable cancer risk estimation.     

Absolute18O content of standard mean ocean water

The absolute values of the¹⁸O/¹⁶O ratio (R_s) and the relative¹⁸O content (X_s) in SMOW have been determined by comparing SMOW mass spectrometrically with well-defined synthetic mixtures of pure D₂¹⁸O and H₂¹⁶O. The results are:R_S = (2005.20 ± 0.45) × 10^?6, X_S = (2000.45 ± 0.45) × 10^?6     

Multiple linear regression,multi-layer perceptron network and adaptive neuro-fuzzy inference system for forecasting precipitation based on large-scale climate signals

ABSTRACT

Nowadays, mathematical models are widely used to predict climate processes, but little has been done to compare the models. In this study, multiple linear regression (MLR), multi-layer perceptron (MLP) network and adaptive neuro-fuzzy inference system (ANFIS) models were compared for precipitation forecasting. The large-scale climate signals were considered as inputs to the applied models. After selecting the most effective climate indices, the effects of large-scale climate signals on the seasonal standardized precipitation index (SPI) of the Maharlu-Bakhtaran catchment, Iran, simultaneously and with a delay, was analysed using a cross-correlation function. Hence, the SPI time series was forecasted up to four time intervals using MLR, MLP and ANFIS. The results showed that most of the indices were significant with SPI of different lag times. Comparison of the SPI forecast results by MLR, MLP and ANFIS models showed better performance for the MLP network than the other two models (RMSE = 0.86, MAE = 0.74 for the first step ahead of SPI forecasting).

Editor D. Koutsoyiannis; Associate editor F. Pappenberger     

Stratigraphy, chemistry, and eruptive dynamics of the 12.4 ka plinian eruption of Apoyeque volcano, Managua, Nicaragua

Apoyeque volcano, located 9 km northwest of Managua city, erupted explosively at 12.4 ka. The Plinian eruption deposited a widespread pumice fall deposit known as the Upper Apoyeque Tephra (UAq). The UAq is massive, reversely graded, and consists of white juvenile pumice (~78 vol.%), a variety of cognate lithics and accidental altered lithics. The whole-rock pumice composition is rhyodacitic (SiO₂?=?66.9–68.5 wt.%) with a mineral paragenesis of plagioclase, orthopyroxene, clinopyroxene, amphibole, titanomagnetite, and ilmenite in a rhyolitic glass groundmass (SiO₂?=?74.4?±?0.6 wt.%). The deposit’s dispersal axis is to the south, with the deposit covering a minimum area of 877 km² within the 50 cm isopach and has a total volume of 3 km³ (dense rock equivalent, 1.15 km³). The eruption column reached a maximum height of ca.28 km. The eruption ejected a total mass of 3?×?10¹² kg at an average rate of 2?×?10⁸ kg/s, and based on available models, we infer duration of almost 4 h. Petrographic and geochemical characteristics suggest that the eruption was triggered by magma mixing.     

Atmospheric concentrations of polycyclic aromatic hydrocarbons (PAHs) in an urban traffic site in Erzurum,Turkey

This study presents daily and seasonal variations of PAH concentrations in Erzurum atmosphere in summer season of 2008 and in winter seasons of 2008 and 2009. Sampling location at Erzurum urban center was selected to represent the effects of traffic (University junction). 18 PAH compounds were analyzed by GC–MS. Average total PAH concentration (gas + particulate) of 18 PAH compounds were measured during 2008 winter (431 ngm^?3) and summer (103 ngm^?3) seasons at the University junction. Daily and seasonal variations of PAH compounds were investigated and compared with other urban centers in the literature. Multiple linear regression and artificial neural network (ANN) models were constructed to determine the impacts of meteorological parameters on measured individual PAH concentrations. Results of the multiple linear regression and ANN models indicated that wind speed, wind direction and intensity of total solar radiation were the most significant factors for the measured concentrations of PAH compounds.