Evangelos Tziritis

Forecasting of groundwater level fluctuations using ensemble hybrid multi-wavelet neural network-based models

Accurate prediction of groundwater level (GWL) fluctuations can play an important role in water resources management. The aims of the research are to evaluate the performance of different hybrid wavelet-group method of data handling (WA-GMDH) and wavelet-extreme learning machine (WA-ELM) models and to combine different wavelet based models for forecasting the GWL for one, two and three months step-ahead in the Maragheh-Bonab plain, NW Iran, as a case study. The research used totally 367 monthly GWLs (m) datasets (Sep 1985-Mar 2016) which were split into two subsets; the first 312 datasets (85% of total) were used for model development (training) and the remaining 55 ones (15% of total) for model evaluation (testing). The stepwise selection was used to select appropriate lag times as the inputs of the proposed models. The performance criteria such as coefficient of determination (R2), root mean square error (RMSE) and Nash-Sutcliffe efficiency coefficient (NSC) were used for assessing the efficiency of the models. The results indicated that the ELM models outperformed GMDH models. To construct the hybrid wavelet based models, the inputs and outputs were decomposed into sub-time series employing different maximal overlap discrete wavelet transform (MODWT) functions, namely Daubechies, Symlet, Haar and Dmeyer of different orders at level two. Subsequently, these sub-time series were served in the GMDH and ELM models as an input dataset to forecast the multi-step-ahead GWL. The wavelet based models improved the performances of GMDH and ELM models for multi-step-ahead GWL forecasting. To combine the advantages of different wavelets, a least squares boosting (LSBoost) algorithm was applied. The use of the boosting multi-WA-neural network models provided the best performances for GWL forecasts in comparison with single WA-neural network-based models.

Identification of hydrogeochemical processes and pollution sources of groundwater resources in the Marand plain, northwest of Iran

The main aims of the present study are to identify the major factors affecting hydrogeochemistry of groundwater resources in the Marand plain, NW Iran and to evaluate the potential sources of major and trace elements using multivariate statistical analysis such as hierarchical clustering analysis (HCA) and factor analysis (FA). To achieve these goals, groundwater samples were collected in three sampling periods in September 2013, May 2014 and September 2014 and analyzed with regard to ions (e.g., Ca2+, Mg2+, Na+ and K+, HCO3−, SO42−, Cl−, F− and NO3−) and trace metals (e.g., Cr, Pb, Cd, Mn, Fe, Al and As). The piper diagrams show that the majority of samples belong to Na–Cl water type and are followed by Ca–HCO3 and mixed Ca–Na–HCO3. Cross-plots show that weathering and dissolution of different rocks and minerals, ion exchange, reverse ion exchange and anthropogenic activities, especially agricultural activities, influence the hydrogeochemistry of the study area. The results of the FA demonstrate that 6 factors with 81.7% of total variance are effective in the overall hydrogeochemistry, which are attributed to geogenic and anthropogenic impacts. The HCA categorizes the samples into two clusters. Samples of cluster C1, which appear to have higher values of some trace metals like Pb and As, are spatially located at the eastern and central parts of the plain, while samples of cluster C2, which express the salinization of the groundwater, are situated mainly westward with few local exceptions.

Geochemical status and interactions between soil and groundwater systems in the area of Akrefnio, Central Greece. Risk assessment, under the scope of mankind and natural environment

Runoff forecasting in mountainous regions with processed based models is often difficult and inaccurate due to the complexity of the rainfall-runoff relationships and difficulties involved in obtaining the required data. Machine learning models offer an alternative for runoff forecasting in these regions. This paper explores and compares two machine learning methods, support vector regression (SVR) and wavelet networks (WN) for daily runoff forecasting in the mountainous Sianji watershed located in the Himalayan region of India. The models were based on runoff, antecedent precipitation index, rainfall, and day of the year data collected over the three year period from July 1, 2001 and June 30, 2004. It was found that both the methods provided accurate results, with the best WN model slightly outperforming the best SVR model in accuracy. Both the WN and SVR methods should be tested in other mountainous watershed with limited data to further assess their suitability in forecasting.

Assessing the hydrogeochemistry and water quality of the Aji-Chay River, northwest of Iran

Aji-Chay River is one of the most important surface reservoirs of northwest of Iran, because it passes through Tabriz city and discharges to Urmia Lake, one of the largest permanent salty lakes in the world. The main objectives of the present study are to evaluate its overall water quality and to explore its hydrogeochemical characteristics, including the potential contamination from heavy metals and metalloids such as Co, Pb, Zn, Cd, Cu, Cr, Al and As. For this purpose, 12 water samples were collected from the main river body and its tributaries within Tabriz plain. The Piper diagram classified water samples mainly into Na–Cl and secondary into Ca–HCO3 and mixed Ca–Mg–Cl types, denoting a profound salinization effect. The cross-plots showed that natural geochemical processes including dissolution of minerals (e.g., carbonates, evaporites and silicates), as well as ion exchange, are the predominant factors that contribute to fluvial hydrogeochemistry, while anthropogenic activities (industrial and agricultural) impose supplementary effects. Cluster analysis classified samples into two distinct clusters; samples of cluster B appear to have elevated electrical conductivity (EC) values and trace metals concentrations such as Co, Pb and Cd, while SiO2 and Zn are low in comparison with the samples of the cluster A. The main processes controlling Aji-Chay River hydrogeochemistry and water quality were identified to be salinization and rock weathering. Both are related with geogenic sources which enrich river system with elevated values of Na+, Cl−, Ca2+, Mg2+, K+, SO42− and EC as a direct effect of evaporites leaching and elevated values of Pb and Cd as an impact from the weathering process of volcanic formations. According to the US salinity diagram, all of the water samples are unsuitable for irrigation as having moderate to bad quality.

Hydrogeochemical features of groundwater resources in Tabriz plain, northwest of Iran

Abstract The present study seeks to evaluate the hydrogeochemistry of Tabriz plain in NW Iran, through major ion chemistry and their spatial variations. In order to accomplish these, groundwater sampling from 30 shallow and deep wells in the plain were carried out in July 2012. The water samples were analyzed for various physicochemical parameters such as pH, EC, Na+, Ca2+, K+, Mg2+, Cl−, CO32−, HCO3−, SO42− and NO3−. Chadha’s diagram demonstrates that most of the groundwaters belonged to the Na–Cl and mixed Ca–Mg–Cl hydrochemical facies. The concentrations of some major ions in groundwater are above the permissible limit for drinking and domestic purposes except for a few locations. The results of saturation index computation show that dissolution of gypsum, anhydrite, halite and silicate minerals occurs frequently across the study area, whereas the groundwater is supersaturated with regard to calcite and dolomite. Cross-plots show that weathering and dissolution of different rocks and minerals, ion exchange, reverse ion exchange and anthropogenic activities, especially agricultural activities, are effective in hydrogeochemistry of the study area.

Hydrogeochemistry and water quality of the Kordkandi-Duzduzan plain, NW Iran: application of multivariate statistical analysis and PoS index

Kordkandi-Duzduzan plain is one of the fertile plains of East Azarbaijan Province, NW of Iran. Groundwater is an important resource for drinking and agricultural purposes due to the lack of surface water resources in the region. The main objectives of the present study are to identify the hydrogeochemical processes and the potential sources of major, minor, and trace metals and metalloids such as Cr, Mn, Cd, Fe, Al, and As by using joint hydrogeochemical techniques and multivariate statistical analysis and to evaluate groundwater quality deterioration with the use of PoS environmental index. To achieve these objectives, 23 groundwater samples were collected in September 2015. Piper diagram shows that the mixed Ca–Mg–Cl is the dominant groundwater type, and some of the samples have Ca–HCO3, Ca–Cl, and Na–Cl types. Multivariate statistical analyses indicate that weathering and dissolution of different rocks and minerals, e.g., silicates, gypsum, and halite, ion exchange, and agricultural activities influence the hydrogeochemistry of the study area. The cluster analysis divides the samples into two distinct clusters which are completely different in EC (and its dependent variables such as Na+, K+, Ca2+, Mg2+, SO42−, and Cl−), Cd, and Cr variables according to the ANOVA statistical test. Based on the median values, the concentrations of pH, NO3−, SiO2, and As in cluster 1 are elevated compared with those of cluster 2, while their maximum values occur in cluster 2. According to the PoS index, the dominant parameter that controls quality deterioration is As, with 60% of contribution. Samples of lowest PoS values are located in the southern and northern parts (recharge area) while samples of the highest values are located in the discharge area and the eastern part.

Groundwater and soil geochemistry of the Eastern Kopaida region (Beotia, central Greece)

The Kopaida plain is a cultivated region of Eastern Greece, with specific characteristics related to the paleogeographic evolution and the changes in land use of the area. This study examines the geochemical conditions of the groundwater and soil, and the correlations between them. 70 samples (50 samples of groundwater and 20 samples of soil) were collected in order to asses the geochemical status and the major natural and manmade affecting processes in the region. Extended chemical analyses were carried out including the assessment of 28 parameters for groundwater and 13 for soil samples. The results revealed that groundwater geochemistry is influenced primary by natural processes such as the geological background, and secondary by manmade impact mainly deriving from the extended use of Nitrogen-fertilizers and the over-exploitation of boreholes. Soil geochemistry is influenced exclusively by natural processes, such as weathering of the prevailing geological formations. Chemical analyses and the statistical processing of data revealed that the major factor for the geochemical status of soils is the weathering of the karstic substrate, as well as the existing lateritic horizons and a weak sulfide mineralization.

The geochemistry of Boron-rich groundwater of the Karlovassi Basin, Samos Island, Greece

The upper Miocene of Karlovassi Basin, Samos Island, Greece, contain continental evaporites such as colemanite, ulexite, celestite, gypsum and thenardite. These evaporites are related with volcanic tuffs, diagenetically altered in a saline-alkaline lake environment. The aim of the present paper is to: a) define the impact of the already known and possible buried borates and other evaporites to the geochemistry of the hydrogeological system of Karlovassi Basin, and; b) to assess the correlation between surface and underground evaporite deposits considering the spatial changes in the concentrations of the examined physicochemical parameters. Fieldwork, laboratory measurements and literature data revealed elevated boron values (2136–33012 ⧎/L) in the central part of Karlovassi Basin. In the same area, high amounts of strontium, sodium, lithium and sulfates also occur. It is proposed that these ions originate from the leaching of evaporites and authigenic minerals such as the Sr-rich clinoptilolite and the boron-bearing potassium feldspar. Boron values are abnormally high for freshwater aquifers, and are indicative of the presence of buried evaporites in the basin with unknown significance.

Hydrogeochemical characteristics and genetic implications of Edipsos thermal springs, north Euboea, Greece

Edipsos area, situated in northern Euboea, has been well known since ancient times for the existence of thermal springs. In order to assess the hydrogeochemical conditions, thermal and cold water samples were collected and analyzed by ICP method for major and trace elements. The results revealed the direct impact of seawater, a process which is strongly related to the major tectonic structures of the area. Seawater impact was confirmed by the Cl/Br and Na/Cl ionic ratios, as well as from statistical processing and graphical interpretation of the analytical results, which classified the sampled waters into three groups (two for cold waters and one for the thermal ones). Trace element ranges for thermal waters are: As (44–84 ppb), Pb (23–154 ppb), Ag (1–2 ppb), Mn (31–680 ppb), Cu (61–97 ppb), Cs (66–244 ppb), Se (0–76 ppb), Li (732–3269 ppb), Fe (0–1126 ppb), Sr (14000–34100 ppb), B (4300–9600 ppb).Compared with the chemical composition of other thermal springs from the Hellenic Volcanic Arc, Edipsos thermal waters are enriched in Ca2+, Na+, Cl−, SO42−, Li, B and K+, reflecting the influence from seawater. Cold waters are free of heavy metals compared with other natural waters and are characterized by good quality based on the major element chemistry. Finally, several geothermometers were applied in order to assess the reservoir temperatures, but none of them appear to be applicable, mainly due to the impact of seawater on the initial hydrogeochemistry of the geothermal fluids.

Heavy Metal(loid)s in the Groundwater of Shabestar Area (NW Iran): Source Identification and Health Risk Assessment

The aims of this study are to investigate the potential origin of selected heavy metal(loid)s in the Shabestar plain, NW Iran, by means of multivariate statistical techniques (cluster analysis and factor analysis), as well as to determine the dominant factors that affect groundwater quality and to assess the health risk induced by metal(loid)s using the hazard quotients (HQ). Totally, 29 groundwater samples were collected from wells in August 2016, and the values of 23 parameters, namely pH, electrical conductivity, concentration of major elements (Ca2+, Mg2+, Na+, K+, HCO3−, SO42−, Cl−), minor elements (NO3−, F−, B, and Br−) and heavy metal(loid)s (Fe, Al, Cr, Mn, As, Zn, Pb, Cu, and Ni) were measured. The results indicate that some samples were found with As, Pb, and Zn concentrations exceeding WHO standards for drinking water. Results of correlation coefficients between the measured variables reflect the occurrence of weathering and dissolution of rocks, especially silicates and evaporites, with ion exchange and geochemical characteristics similar to the release of some heavy metal(loid)s. According to hierarchical cluster analysis, samples of cluster 1 are affected by alkalinity and accompanied by elements compatible with alkaline ambience (CO32− and Ni). Samples of subcluster 2-1 demonstrate the effect of salinity, attributed to evaporates, irrigation return flow, and influx of Urmia Lake’s brine, while, samples of sub-cluster 2-2 are influenced by agricultural activities. Factor analysis results illustrate the effects of five factors on the quality of groundwater. The factor analysis accounts for the 71.9% of total variance of groundwater quality for geogenic impacts, while 10% of the groundwater quality variance is controlled by agricultural activities which produce excessive amounts of NO3− along with Zn which is contained in fertilizers and agrochemicals. The results of the human health risk assessment show that As is the most dominant metalloid in inducing maximum noncarcinogenic risk among all the heavy metal(loid)s. Based on HI, 45 and 14% of the samples for children and adults, respectively, are found to be in high risk category.