Evaluation of the temporal scaling variability in forecasting ground-level ozone concentrations obtained from multiple linear regressions

Ozone is a highly unpredictable pollutant which severely affects living conditions in urban and surrounding areas in the Mediterranean basin. This secondary pollutant periodically reaches extremely high concentrations, damaging human health. Multiple linear regression has been widely used in previous works due to the fact that it is a simple and versatile method for forecasting ozone concentrations. However, these models usually prove their validity using fulfillment of statistical constraints, ignoring other intrinsic characteristics existing in the time series, such as the temporal scaling behavior and the data distribution over different time scales. In previous works, it has been demonstrated that observed ozone time series are of a multifractal nature, meaning that the data distribution can be described by using the multifractal spectrum. This work focuses on the capacity of a forecasting model to reproduce the scaling features existing in an observed time series when several chemical and meteorological explanatory variables are introduced following the stepwise procedure. A comparison between the observed spectrum and the simulated ones for each step is used to check which explanatory variables better reproduce the multifractal nature in real ozone time series. It has been confirmed that a model with few explanatory variables allows reproducing the multifractal nature in the simulated time series with an acceptable accuracy without compromising the values of the coefficient of determination and root-mean-squared error, which were used as performance indicators.     

Spatial Prediction of Stream Temperatures Using Top-Kriging with an External Drift

Top-kriging is a method for estimating stream flow and stream flow-related variables on a river network. Top-kriging treats these variables as emerging from a two-dimensional spatially continuous process in the landscape. The top-kriging weights are estimated by a family of variogram models (regularisations) for different catchment areas (kriging support), which accounts for the different scales and the nested nature of the catchments. This assures that kriging weights are distributed to both hydrologically connected and unconnected sites of the stream network according to the data situation: top-kriging gives most weight to close-by sites at the same river system, but when the next hydrologically connected site is far away, more weight is given to a close-by site at an adjacent river system. The distribution of weights is in contrast to ordinary kriging and stream distance-based kriging which does not account for both spatial proximity and network connectivity. We extend the top-kriging method by incorporating an external drift function to account for the deterministic patterns of the spatial variable. We test the method for a comprehensive Austrian stream temperature dataset. The drift is modelled by exponential regression with catchment altitude. Top-kriging is then applied to the regression residuals. The variogram used in top-kriging is fitted by a semiautomatic optimisation procedure. A leave-one-out cross-validation analysis shows that the model performs well for the study domain. The residual mean squared error (cross-validation) decreases by 20 % when using top-kriging in addition to the regression model. For regions where the observed stream temperatures deviate from the expected value of the drift model, top-kriging corrects these regional biases. By exploiting the topological information of the stream network, top-kriging is able to improve the local adjustment of the drift model for the main streams and the tributaries.     

Automated riverine landscape characterization: GIS-based tools for watershed-scale research, assessment, and management

River systems consist of hydrogeomorphic patches (HPs) that emerge at multiple spatiotemporal scales. Functional process zones (FPZs) are HPs that exist at the river valley scale and are important strata for framing whole-watershed research questions and management plans. Hierarchical classification procedures aid in HP identification by grouping sections of river based on their hydrogeomorphic character; however, collecting data required for such procedures with field-based methods is often impractical. We developed a set of GIS-based tools that facilitate rapid, low cost riverine landscape characterization and FPZ classification. Our tools, termed RESonate, consist of a custom toolbox designed for ESRI ArcGIS®. RESonate automatically extracts 13 hydrogeomorphic variables from readily available geospatial datasets and datasets derived from modeling procedures. An advanced 2D flood model, FLDPLN, designed for MATLAB® is used to determine valley morphology by systematically flooding river networks. When used in conjunction with other modeling procedures, RESonate and FLDPLN can assess the character of large river networks quickly and at very low costs. Here we describe tool and model functions in addition to their benefits, limitations, and applications.     

Integrating data-driven ecological models in an expert-based decision support system for water management in the Du river basin (Vietnam)

In this study, classification trees were combined with the Water Framework Directive (WFD)-Explorer, a modular toolbox that supports integrated water management in a river basin to evaluate the impact of different restoration measures on river ecology. First, the WFD-Explorer toolbox analysed the effect of different restoration options on the abiotic river characteristics based on the water and substance balance embedded in the simulation environment. Based on these abiotic characteristics, the biological index Biological Monitoring Working Party for Vietnam was then predicted by classification trees that were trained on biological and abiotic data collected in the Du river basin in northern Vietnam. The ecological status of streams in the basin ranged from nearly pristine headwaters to severely impacted river stretches. Elimination of point sources from ore extraction and decentralised domestic wastewater treatment proved to be the most effective measures to improve the ecological condition of the Du river basin. The combination of the WFD-Explorer results with data-driven models enabled model application in a situation where expert knowledge was lacking. Consequently, this approach appeared promising for decision support in the context of river restoration and conservation management.     

Application of remote sensing data and GIS for landslide risk assessment as an environmental threat to Izmir city (west Turkey)

In this study, landslide risk assessment for Izmir city (west Turkey) was carried out, and the environmental effects of landslides on further urban development were evaluated using geographical information systems and remote sensing techniques. For this purpose, two different data groups, namely conditioning and triggering data, were produced. With the help of conditioning data such as lithology, slope gradient, slope aspect, distance from roads, distance from faults and distance from drainage lines, a landslide susceptibility model was constructed by using logistic regression modelling approach. The accuracy assessment of the susceptibility map was carried out by the area under curvature (AUC) approach, and a 0.810 AUC value was obtained. This value shows that the map obtained is successful. Due to the fact that the study area is located in an active seismic region, earthquake data were considered as primary triggering factor contributing to landslide occurrence. In addition to this, precipitation data were also taken into account as a secondary triggering factor. Considering the susceptibility data and triggering factors, a landslide hazard index was obtained. Furthermore, using the Aster data, a land-cover map was produced with an overall kappa value of 0.94. From this map, settlement areas were extracted, and these extracted data were assessed as elements at risk in the study area. Next, a vulnerability index was created by using these data. Finally, the hazard index and the vulnerability index were combined, and a landslide risk map for Izmir city was obtained. Based on this final risk map, it was observed that especially south and north parts of the Izmir Bay, where urbanization is dense, are threatened to future landsliding. This result can be used for preliminary land use planning by local governmental authorities.     

Defining chemical status of a temporary Mediterranean River

Although the majority of rivers and streams in the Mediterranean area are temporary, no particular attention is being paid for such systems in the Water Framework Directive (WFD). A typical temporal Mediterranean river, draining an intensively cultivated basin, was assessed for its chemical status. Elevated concentrations of nitrates and salts in river water as well as nutrients and heavy metals in river sediments have been attributed to agricultural land uses and practices and point sources of organic pollution. A scheme for the classification of the river's chemical status (within the ecological quality classification procedure) was applied by combining pollution parameters in groups according to related pressures. In light of the temporal hydrological regime and anthropogenic impacts, sediment chemical quality elements were considered, in addition to hydrochemical ones. Despite the extensive agricultural activities in the basin, the majority of the sites examined showed a good quality and only three of them were classified as moderate. For the classification of the chemical quality of temporary water bodies, there is a need to develop ecologically relevant salinity and sediment quality standards.     

Probability-based nitrate contamination map of groundwater in Kinmen

Groundwater supplies over 50 % of drinking water in Kinmen. Approximately 16.8 % of groundwater samples in Kinmen exceed the drinking water quality standard (DWQS) of NO₃ ^?-N (10 mg/L). The residents drinking high nitrate-polluted groundwater pose a potential risk to health. To formulate effective water quality management plan and assure a safe drinking water in Kinmen, the detailed spatial distribution of nitrate–N in groundwater is a prerequisite. The aim of this study is to develop an efficient scheme for evaluating spatial distribution of nitrate–N in residential well water using logistic regression (LR) model. A probability-based nitrate–N contamination map in Kinmen is constructed. The LR model predicted the binary occurrence probability of groundwater nitrate–N concentrations exceeding DWQS by simple measurement variables as independent variables, including sampling season, soil type, water table depth, pH, EC, DO, and Eh. The analyzed results reveal that three statistically significant explanatory variables, soil type, pH, and EC, are selected for the forward stepwise LR analysis. The total ratio of correct classification reaches 92.7 %. The highest probability of nitrate–N contamination map presents in the central zone, indicating that groundwater in the central zone should not be used for drinking purposes. Furthermore, a handy EC–pH-probability curve of nitrate–N exceeding the threshold of DWQS was developed. This curve can be used for preliminary screening of nitrate–N contamination in Kinmen groundwater. This study recommended that the local agency should implement the best management practice strategies to control nonpoint nitrogen sources and carry out a systematic monitoring of groundwater quality in residential wells of the high nitrate–N contamination zones.     

The relative influence of nutrients and habitat on stream metabolism in agricultural streams

Stream metabolism was measured in 33 streams across a gradient of nutrient concentrations in four agricultural areas of the USA to determine the relative influence of nutrient concentrations and habitat on primary production (GPP) and respiration (CR-24). In conjunction with the stream metabolism estimates, water quality and algal biomass samples were collected, as was an assessment of habitat in the sampling reach. When data for all study areas were combined, there were no statistically significant relations between gross primary production or community respiration and any of the independent variables. However, significant regression models were developed for three study areas for GPP (r ² = 0.79–0.91) and CR-24 (r ² = 0.76–0.77). Various forms of nutrients (total phosphorus and area-weighted total nitrogen loading) were significant for predicting GPP in two study areas, with habitat variables important in seven significant models. Important physical variables included light availability, precipitation, basin area, and in-stream habitat cover. Both benthic and seston chlorophyll were not found to be important explanatory variables in any of the models; however, benthic ash-free dry weight was important in two models for GPP.     

Landslides susceptibility change over time according to terrain conditions in a mountain area of the tropic region

Susceptibility to landslides in mountain areas results from the interaction of various factors related to relief formation and soil development. The assessment of landslide susceptibility has generally taken into account individual events, or it has been aimed at establishing relationships between landslide-inventory maps and maps of environmental factors, without considering that such relationships can change in space and time. In this work, temporal and space changes in landslides were analysed in six different combinations of date and geomorphological conditions, including two different geological units, in a mountainous area in the north-centre of Venezuela, in northern South America. Landslide inventories from different years were compared with a number of environmental factors by means of logistic regression analysis. The resulting equations predicted landslide susceptibility from a range of geomorphometric parameters and a vegetation index, with diverse accuracy, in the study area. The variation of the obtained models and their prediction accuracy between geological units and dates suggests that the complexity of the landslide processes and their explanatory factors changed over space and time in the studied area. This calls into question the use of a single model to evaluate landslide susceptibility over large regions.     

Vegetation recovery patterns assessment at landslides caused by catastrophic earthquake: A case study in central Taiwan

The catastrophic earthquake, 7.3 on the Richter scale, occurred on September 21, 1999 in Central Taiwan. Much of standing vegetation on slopes was eliminated and massive, scattered landslides were induced at the Jou-Jou Mountain area of the Wu-Chi basin in Nantou County. We evaluated three methods for assessing landslide hazard and vegetation recovery conditions. (1) Self-organizing map (SOM) neural network coupled with fuzzy technique was used to quickly extract the landslide. (2) The NDVI-based vegetation recovery index derived from multi-temporal SPOT satellite images was used to evaluate vegetation recovery rate in the denudation sites. (3) The spatial distribution index (SDI) based on land-cover topographic location was employed to analyze vegetation recovery patterns, including the invading, surviving and mixed patterns at the Jou-Jou Mountain area. On September 27, 1999, there were 849.20 ha of landslide area extracted using the self-organizing map and fuzzy technique combined model. After six years of natural vegetation succession, the landslide has gradually restored, and vegetation recovery rate reached up to 86%. On-site observation shows that many native pioneer plants have invaded onto the denudation sites even if disturbed by several typhoons. Two native surviving plants, Arundo formosana Hack and Pinus taiwanensis Hayata, play a vital role in natural vegetation succession in this area, especially for the sites on ridgeline and steep slopes.     

Contaminated area instability along Ångermanälven River,northern Sweden

Industrially utilized river basins are frequently exposed to contaminants originating from polluting activities. However, the physical instability and probability of mass movement mobilization of contaminated soil into rivers have only received little attention. In this study, we present a GIS-based method to produce a regional overview of where and how contaminated areas are potentially exposed to slope instability. A landslide susceptibility-index was used to study the degree and distribution of overlap between contaminated sites and unstable ground. A contaminated area instability hazard classification was produced integrating slope instability and contamination risk classification. Our results indicate that mass movement can be tied mainly to a slope gradient ≥16°, a proximity to the river that is <500 m, a distance of <500 m from roads, concave surface curvature, and sand- and silt soils. Forty-six (22%) of all considered contaminated sites are located within areas with a non-negligible slope instability, of which a majority, 30 sites (14%) are situated on ground with a low or moderate instability. Three sites with a class 2 contamination risk (the 2nd highest class) are located on ground with a very high slope instability.     

Assessing possible visitor-use impacts on water quality in Yosemite National Park, California

There is concern that visitor-use associated activities, such as bathing, dish washing, wastewater production, and stock animal use near lakes and streams, could cause degradation of water quality in Yosemite National Park. A study was conducted during 2004-2007 to assess patterns in nutrient and Escherichia coli (E. coli) concentrations in the Merced and Tuolumne Rivers and characterize natural background concentrations of nutrients in the park. Results indicated that nutrient and E. coli concentrations were low, even compared to other undeveloped sites in the United States. A multiple linear regression approach was used to model natural background concentrations of nutrients, with basin characteristics as explanatory variables. Modeled nitrogen concentrations increased with elevation, and modeled phosphorus concentrations increased with basin size. Observed concentrations (±uncertainty) were compared to modeled concentrations (±uncertainty) to identify sites that might be impacted by point sources of nutrients, as indicated by large model residuals. Statistically significant differences in observed and modeled concentrations were observed at only a few locations, indicating that most sites were representative of natural background conditions. The empirical modeling approach used in this study can be used to estimate natural background conditions at any point along a study reach in areas minimally impacted by development, and may be useful for setting water-quality standards in many national parks.     

A statistical approach for the assessment and redesign of the Nile Delta drainage system water-quality-monitoring locations

There are several deficiencies in the statistical approaches proposed in the literature for the assessment and redesign of surface water-quality-monitoring locations. These deficiencies vary from one approach to another, but generally include: (i) ignoring the attributes of the basin being monitored; (ii) handling multivariate water quality data sequentially rather than simultaneously; (iii) focusing mainly on locations to be discontinued; and (iv) ignoring the reconstitution of information at discontinued locations. In this paper, a methodology that overcomes these deficiencies is proposed. In the proposed methodology, the basin being monitored is divided into sub-basins, and a hybrid-cluster analysis is employed to identify groups of sub-basins with similar attributes. A stratified optimum sampling strategy is then employed to identify the optimum number of monitoring locations at each of the sub-basin groups. An aggregate information index is employed to identify the optimal combination of locations to be discontinued. The proposed approach is applied for the assessment and redesign of the Nile Delta drainage water quality monitoring locations in Egypt. Results indicate that the proposed methodology allows the identification of (i) the optimal combination of locations to be discontinued, (ii) the locations to be continuously measured and (iii) the sub-basins where monitoring locations should be added. To reconstitute information about the water quality variables at discontinued locations, regression, artificial neural network (ANN) and maintenance of variance extension (MOVE) techniques are employed. The MOVE record extension technique is shown to result in a better performance than regression or ANN for the estimation of information about water quality variables at discontinued locations.     

Towards generalised reference condition models for environmental assessment: a case study on rivers in Atlantic Canada

Evaluation of the ecological status of river sites in Canada is supported by building models using the reference condition approach. However, geography, data scarcity and inter-operability constraints have frustrated attempts to monitor national-scale status and trends. This issue is particularly true in Atlantic Canada, where no ecological assessment system is currently available. Here, we present a reference condition model based on the River Invertebrate Prediction and Classification System approach with regional-scale applicability. To achieve this, we used biological monitoring data collected from wadeable streams across Atlantic Canada together with freely available, nationally consistent geographic information system (GIS) environmental data layers. For the first time, we demonstrated that it is possible to use data generated from different studies, even when collected using different sampling methods, to generate a robust predictive model. This model was successfully generated and tested using GIS-based rather than local habitat variables and showed improved performance when compared to a null model. In addition, ecological quality ratio data derived from the model responded to observed stressors in a test dataset. Implications for future large-scale implementation of river biomonitoring using a standardised approach with global application are presented.     

Hydrometeorological variables predict fecal indicator bacteria densities in freshwater: data-driven methods for variable selection

Statistical models of microbial water quality inform risk management for water recreation. Current research focuses on resource-intensive, location-specific data collection and water quality modeling, but this approach may be cost-prohibitive for risk managers responsible for numerous recreation sites. As an alternative, we tested the ability of two data-driven models, tree regression and random forests with conditional inference trees, to select readily available hydrometeorological variables for use in linear mixed effects (LME) models predicting bacterial density. The study included the Chicago Area Waterway System (CAWS) and Lake Michigan beaches and harbors in Chicago, Illinois, at which Escherichia coli and enterococci were measured seasonally in 2007–2009. Tree regression node variables reduced data dimensionality by >50 %. Variable importance ranks from random forests were used in a forward-step selection based on R ² and root mean squared prediction error (RMSPE). We found two to three variables explained bacteria densities well relative to random forests with all variables. LME models with tree- or forest-selected variables performed reasonably well (0.335?<?R ²?<?0.658). LME models for Lake Michigan had good prediction accuracy with respect to the single sample maximum standard (72–77 %), but limited sensitivity (23–62 %). Results suggest that our alternative approach is feasible and performs similarly to more resource-intensive approaches.     

Dynamic regression modeling of daily nitrate-nitrogen concentrations in a large agricultural watershed

Nitrate-nitrogen concentrations in rivers represent challenges for water supplies that use surface water sources. Nitrate concentrations are often modeled using time-series approaches, but previous efforts have typically relied on monthly time steps. In this study, we developed a dynamic regression model of daily nitrate concentrations in the Raccoon River, Iowa, that incorporated contemporaneous and lags of precipitation and discharge occurring at several locations around the basin. Results suggested that 95 % of the variation in daily nitrate concentrations measured at the outlet of a large agricultural watershed can be explained by time-series patterns of precipitation and discharge occurring in the basin. Discharge was found to be a more important regression variable than precipitation in our model but both regression parameters were strongly correlated with nitrate concentrations. The time-series model was consistent with known patterns of nitrate behavior in the watershed, successfully identifying contemporaneous dilution mechanisms from higher relief and urban areas of the basin while incorporating the delayed contribution of nitrate from tile-drained regions in a lagged response. The first difference of the model errors were modeled as an AR(16) process and suggest that daily nitrate concentration changes remain temporally correlated for more than 2 weeks although temporal correlation was stronger in the first few days before tapering off. Consequently, daily nitrate concentrations are non-stationary, i.e. of strong memory. Using time-series models to reliably forecast daily nitrate concentrations in a river based on patterns of precipitation and discharge occurring in its basin may be of great interest to water suppliers.     

PLS and PCR Methods in the Assessment of Coastal Water Quality

Partial least squares regression analysis (PLS) and principal component regression analysis (PCR) were examined asmethodological procedures for assessing the quality of coastalwaters in a tourist area. Four variables related to the trophicstate of waters, namely nitrate, nitrite, ammonia and phosphate were analyzed. The models resulting from PLS and PCR were verysimilar. Both defined three groups of water masses characterizedby different nutrient loadings. These groups were in accordancewith those obtained by numerical classification. The PLS methodwas selected as the optimal model, on the basis of its lowerprediction errors (lower Press and Rmsd values). For managementpurposes, this statistical model allows mesotrophic conditions,reflecting some nutrient enrichment over background conditions,to be characterized and the successful diagnosis of additionalsamples within this context.     

Monitoring, assessment and modelling using water quality data in the Saale River Basin, Germany

The European Water Framework Directive (WFD) is the overall driver for this environmental study and currently requires the identification of patterns and sources of pollution (monitoring) to support objective ecological sound decision making and specific measures to enhance river water quality (modelling). The purpose of this paper is to demonstrate in a case study the interrelationship between (1) hydrologic and water quality monitoring data for river basin characterization and (2) modelling applications to assess resources management alternatives. The study deals with monitoring assessment and modelling of river water quality data of the main stem Saale River and its principal tributaries. For a period of 6 years the data, which was sampled by Environmental Agencies of the German states of Thuringia, Saxony and Saxony-Anhalt, was investigated to assess sources and indicators of pollution. In addition to the assessment a modelling exercise of the routing of different pollutants was carried out in the lower part of the test basin. The modelling is a tool to facilitate the evaluation of alternative measures to reduce contaminant loadings and improve ecological status of a water body as required by WFD. The transport of suspended solids, salts and heavy metals was modelled along a selected Saale reach under strong anthropogenic influence in terms of contaminants and river morphology between the city of Halle and the confluence with the Elbe River. The simulations were carried out with the model WASP5 which is a dynamic flood-routing and water quality model package developed by the US Environmental Protection Agency.     

Linking Hydrogeological and Ecological Tools for an Integrated River Catchment Assessment

Ecological (biological and hydrochemical assessment) and hydrogeological (vulnerability and pollution risk mapping) tools have been combined to assess the ecological quality and hydrogeological vulnerability of an agricultural river basin. In addition, the applicability of the recently developed vulnerability assessment approach (COP method) in the particular environmental conditions was tested by comparing its results with hydroecological assessment tools (i.e., pollution metrics). Five sampling sites were selected and sampled for benthic macroinvertebrates and physicochemical variables during summer and spring. Overall, sites ranged from moderate to poor ecological quality. The results illustrated that 26% of the study area was of moderate pollution risk, while 65% was classified as of low and very low risk zones. However, the higher elevation zones where calcareous rock formations are encountered presented moderate to high pollution risk that was accredited by the ecological quality assessment. Pollution metrics facilitated from hydrochemical analysis indicated a significant association with groundwater vulnerability, thus validating vulnerability and risk estimations. This study indicated that the particular groundwater pollution risk mapping methodology and the water quality assessment indices can be well combined to provide an integrated evaluation tool at a catchment scale.