Hydrochemical evaluation of the Voltaian system--the Afram Plains area, Ghana

Inverse geochemical modeling from PHREEQC, and multivariate statistical methods were jointly used to define the genetic origin of chemical parameters of groundwater from the Voltaian aquifers in the Afram Plains area. The study finds, from hierarchical cluster analysis that there are two main hydrochemical facies namely the calcium-sodium-chloride-bicarbonate waters and the magnesium-potassium-sulfate-nitrate waters in the northern and southern sections, respectively, of the Afram Plains area. This facies differentiation is confirmed by the distribution of the SO(4)(2-)/Cl(-) ratio, which associates groundwater from the northern and southern sections to areas influenced by contact with evaporites and seawater, respectively. Principal component analysis (PCA) with varimax rotation using the Kaiser criterion identifies four principal sources of variation in the hydrochemistry. Mineral saturation indices calculated from both major ions and trace elements, indicate saturation-supersaturation with respect to calcite, aragonite, k-mica, chlorite, rhodochrosite, kaolinite, sepiolite, and talc, and undersaturation with respect to albite, anorthite, and gypsum in the area. Inverse geochemical modeling along groundwater flowpaths indicates the dissolution of albite, anorthite and gypsum and the precipitation of kaolinite, k-mica, talc, and quartz. Both the PCA and inverse geochemical modeling identify the incongruent weathering of feldspars as the principal factors controlling the hydrochemistry in the Afram Plains area. General phase transfer equations have been developed to characterize the geochemical evolution of groundwater in the area. A very good relationship has been established between calcite and aragonite saturation indices in the Afram Plains area, with R(2)=1.00.     

Cluster analysis and quality assessment of logged water at an irrigation project, eastern Saudi Arabia

A multivariate statistical technique, cluster analysis, was used to assess the logged surface water quality at an irrigation project at Al-Fadhley, Eastern Province, Saudi Arabia. The principal idea behind using the technique was to utilize all available hydrochemical variables in the quality assessment including trace elements and other ions which are not considered in conventional techniques for water quality assessments like Stiff and Piper diagrams. Furthermore, the area belongs to an irrigation project where water contamination associated with the use of fertilizers, insecticides and pesticides is expected. This quality assessment study was carried out on a total of 34 surface/logged water samples. To gain a greater insight in terms of the seasonal variation of water quality, 17 samples were collected from both summer and winter seasons. The collected samples were analyzed for a total of 23 water quality parameters including pH, TDS, conductivity, alkalinity, sulfate, chloride, bicarbonate, nitrate, phosphate, bromide, fluoride, calcium, magnesium, sodium, potassium, arsenic, boron, copper, cobalt, iron, lithium, manganese, molybdenum, nickel, selenium, mercury and zinc. Cluster analysis in both Q and R modes was used. Q-mode analysis resulted in three distinct water types for both the summer and winter seasons. Q-mode analysis also showed the spatial as well as temporal variation in water quality. R-mode cluster analysis led to the conclusion that there are two major sources of contamination for the surface/shallow groundwater in the area: fertilizers, micronutrients, pesticides, and insecticides used in agricultural activities, and non-point natural sources.     

Assessment of surface water quality of selected estuaries of Malaysia: multivariate statistical techniques

Multivariate statistical techniques such as cluster analysis (CA), factor analysis (FA) were used for the evaluation of spatial variations and the interpretation of a large complex water quality data set of two selected estuaries of Malaysia. The two locations of interest with 10 sites in each location were Kuala Juru (Juru estuary) and Bukit Tambun (Jejawi estuary). Cluster analysis showed that some sites in both locations have similar sources of pollution from point or non-point sources whereas FA yielded four factors which are responsible for water quality variations explaining more than 80% of the total variance of the data set and allowed to group the selected water quality. Correlation analysis of the data showed that some parameters have strong association with other parameters and they share a common origin source. This study illustrates the usefulness of multivariate statistical analysis for evaluation and interpretation of complex data sets to get better information about the pollution sources/factors and understanding the behavior of the parameters in water quality for effective river water quality management.     

STRATIFICATION OF VARIABILITY IN RUNOFF AND SEDIMENT YIELD BASED ON VEGETATION CHARACTERISTICS1

ABSTRACT: Runoff and sediment yield were collected from 100 plots during simulated rainfalls (100 mm/hr for 15 minutes) at antecedent soil moisture conditions. A clustering technique was used to stratify the variability of a single data set within a sagebrush‐grass community into four groups based on vegetation life form and amount of cover. The four cluster groups were grass, grass/shrub, shrub, and forb/grass and were found to be significantly different in plant height, surface roughness, soil bulk density, and soil organic matter. Stepwise multiple regression analyses were performed on the single data set and each cluster group. Results for individual groups resulted in more robust predictive equations for runoff (r²= 0.65–0.73) and sediment yield (r²= 0.37–0.91) than for equations developed from the single data set (r²= 0.56 for runoff and r²= 0.27 for sediment yield). The standard errors of the cluster group regression equations were also improved in three of the four group equations for both runoff and sediment yield compared to the single data set. Runoff was found to be significantly less (p >0.01) in the forb/grass group compared with other vegetation cluster groups, but this was influenced by four plots that produced little or no runoff. Sediment yield was not found to be significantly different among any cluster groups. Discriminant analysis was then used to identify important variables and develop a model to classify plots into one of the four cluster groups. The discriminant model could be incorporated into rangeland hydrology and erosion models. The percentage cover of grasses, shrubs, litter, and bare ground effectively stratified about 12 percent of the variation observed in runoff and 26 percent of the variability for sediment yield as determined by r².     

Identifying relationships between baseflow geochemistry and land use with synoptic sampling and R-mode factor analysis

The relationship between land use and stream chemistry is often explored through synoptic sampling of rivers at baseflow conditions. However, baseflow chemistry is likely to vary temporally and spatially with land use. The purpose of our study is to examine the usefulness of the synoptic sampling approach for identifying the relationship between complex land use configurations and stream water quality. This study compares biogeochemical data from three synoptic sampling events representing the temporal variability of baseflow chemistry and land use using R-mode factor analysis. Separate R-mode factor analyses of the data from individual sampling events yielded only two consistent factors. Agricultural activity was associated with elevated levels of Ca2+, Mg2+, alkalinity, and frequently K+, SO4(2-), and NO3-. Urban areas were associated with higher concentrations of Na+, K+, and Cl-. Other retained factors were not consistent among sampling events, and some factors were difficult to interpret in the context of biogeochemical sources and processes. When all data were combined, further associations were revealed such as an inverse relationship between the proportion of wetlands and stream nitrate concentrations. We also found that barren lands were associated with elevated sulfate levels. This research suggests that an individual sampling event is unlikely to characterize adequately the complex processes controlling interactions between land use and stream chemistry. Combining data collected over two years during three synoptic sampling events appears to enhance our ability to understand processes linking stream chemistry and land use.     

Elemental composition of total dry particulate matter deposited in the selected industrial areas within Lagos state

The total particulate matter (PM) deposited within 17 selected industrial areas in Lagos state during the dry season (December 2015 to January 2016) was studied. Deposition gauges measuring 0.2 meters (m) in diameter by 0.15 m in depth were placed at the sampling locations for a period of one month to collect the total deposited PM. The PM was then characterized using energy‐dispersive X‐ray florescence (EDXRF). The sources of the heavy metals were evaluated using enrichment factor (EF) analysis. Factor analysis (FA) was then used to determine the correlations between the identified heavy metals. Twenty‐three elements—sodium (Na), silicon (Si), phosphorous (P), sulfur (S), chlorine (Cl), potassium (K), calcium (Ca), titanium (Ti), vanadium (V), chromium (Cr), manganese (Mn), iron (Fe), nickel (Ni), copper (Cu), zinc (Zn), germanium (Ge), arsenic (As), strontium (Sr), zirconium (Zr), lead (Pb), tungsten (W), polonium (Po), and actinium (Ac)—were characterized in the PM collected at the sampling sites. The iron elemental ratio ranged from 0.0003 to 3.8848. The EF ranged from 0.0015 to 1697.47, including at the control location. The FA, using principal component analysis techniques, revealed seven factor loadings with 90.03% cumulative, which suggests that the sources are anthropogenic, such as from industrial activities, vehicular emissions, and the combustion of fuel.     

TRENDS IN LEAD CONCENTRATIONS IN MAJOR U.S. RIVERS AND THEIR RELATION TO HISTORICAL CHANGES IN GASOLINE-LEAD CONSUMPTION1

ABSTRACT: Declines in concentrations of dissolved lead occurred at nearly two-thirds of 306 locations on major U.S. rivers from 1974 to 1985. Declines in dissolved lead concentrations are statistically significant (p < 0.10) at approximately one-third of the sampling locations. Statistically significant increases in dissolved lead concentrations occurred at only 6 percent of the sites, but are clustered in the Texas-Gulf and Lower Mississippi regions. Possible explanations for the observed trends in lead concentrations are tested through comparisons with (1) records of lead discharges from major sources including leaded-gasoline consumption and municipal- and industrial-point source discharges, (2) trends in various water-quality constituents such as pH and total alkalinity, and (3) basin characteristics such as drainage area. Statistically significant declines in lead concentrations in streams and gasoline lead (i.e., the largest source of lead at these sites) are highly coincident for the 1979 to 1980 period at most sampling locations. The greatest amount of decline in gasoline lead occurred at sites showing statistically significant downtrends in stream concentrations of lead from 1974 to 1985. No more than 5 percent of the trends in stream lead are influenced by municipal- and industrial-point sources of lead. Factors that affect the transport of dissolved lead, including lead solubility, suspended sediment, and basin characteristics such as drainage basin size, are not significantly related to trends in dissolved lead. Trends in streamflow explain no more than 7 percent of the downtrends in concentrations of lead and may partly explain the frequent increases in lead concentrations in the Texas-Gulf and Lower Mississippi regions.     

Determination of arsenic levels in the water resources of Aksaray Province, Turkey

Arsenic levels were determined in 62 stations utilized as drinking and potable water resources by local community for Turkey's Aksaray Province (4589 km(2); 980 m above sea level). The samplings were implemented every two months for 1 year. The arsenic values were found to be ranging between 10 and 50 μg/L in 22 points and were found to be >50 μg/L in 5 stations, according to the mean value of the 6 samples. WHO and the Turkish Standards have permitted an arsenic concentration of 10 μg/L in drinking waters. The multivariate statistical technique, cluster analysis (CA), followed by principal component analysis (PCA) were applied to the data on 17 water quality parameters in 47 stations that are used for drinking and other domestic resources. Two significant sampling locations were detected based on the similarity of their water quality. The chemical correlations were observed in the two sub-sampling locations by Principal Component Analysis.     

Sources and Transport of Phosphorus to Rivers in California and Adjacent States,U.S., as Determined by SPARROW Modeling

The SPARROW (SPAtially Referenced Regression on Watershed attributes) model was used to simulate annual phosphorus loads and concentrations in unmonitored stream reaches in California, U.S., and portions of Nevada and Oregon. The model was calibrated using de‐trended streamflow and phosphorus concentration data at 80 locations. The model explained 91% of the variability in loads and 51% of the variability in yields for a base year of 2002. Point sources, geological background, and cultivated land were significant sources. Variables used to explain delivery of phosphorus from land to water were precipitation and soil clay content. Aquatic loss of phosphorus was significant in streams of all sizes, with the greatest decay predicted in small‐ and intermediate‐sized streams. Geological sources, including volcanic rocks and shales, were the principal control on concentrations and loads in many regions. Some localized formations such as the Monterey shale of southern California are important sources of phosphorus and may contribute to elevated stream concentrations. Many of the larger point source facilities were located in downstream areas, near the ocean, and do not affect inland streams except for a few locations. Large areas of cultivated land result in phosphorus load increases, but do not necessarily increase the loads above those of geological background in some cases because of local hydrology, which limits the potential of phosphorus transport from land to streams.     

COMPARING METHODS OF HYDROLOGIC REGIONALIZATION1

ABSTRACT: Data splitting is used to compare methods of determining “homogeneous” hydrologic regions. The methods compared use cluster analysis based on similarity of hydrologic characteristics or similarity of characteristics of a stream's drainage basin. Data for 221 stations in Arizona are used to show that the methods, which are a modification of DeCoursey's scheme for defining regions, improve the fit of estimation data to the model, but that is is necessary to have an independent measure of predictive accuracy, such as that provided by data splitting, to demonstrate improved predictive accuracy. The methods used the complete linkage algorithm for cluster analysis and computed weighted average estimates of hydrologic characteristics at ungaged sites.     

Macroinvertebrate Community Structure Along the Longitudinal Gradient of an Agriculturally Impacted Stream

/ Lapwai Creek, an agriculturally impacted stream in northern Idaho, was sampled seasonally over a two-year period to determine if macroinvertebrate community composition changed along the longitudinal gradient and if changes followed predictions of the river continuum concept. Possible relationships between changes in food resource availability and community structure were also examined. Benthic invertebrates were collected at eight locations along the longitudinal gradient of Lapwai Creek using a Hess sampler. Random skewer analysis suggested there was no longitudinal gradient for either number of individuals or functional feeding group composition. Cluster analysis revealed that all locations, excluding a site receiving outflow from a small, eutrophic reservoir, had a similar community structure, further suggesting that invertebrate community composition remained consistent along the longitudinal gradient of the stream. The community was dominated at all sites, excluding the site below the reservoir, by functionalgrazers. Shredders were rare throughout Lapwai Creek, even in areas where healthy riparian vegetation still remained. Studies of other streams within the drainage basin show that many species found in the upper reaches of these streams, where agricultural impacts are low, were absent throughout the length of Lapwai Creek. Data collected concurrently with macroinvertebrates indicated that the input, storage, and transport of particulate organic matter was low throughout the stream, whereas periphyton abundance was high. The absence of longitudinal changes, despite flowing through three distinct geomorphological regions, and the grouping of all sites except one by cluster analysis for both dominant taxa and functional feeding groups suggest that agricultural alteration has influenced community structure of Lapwai Creek, resulting in a relatively homogeneous assemblage of macroinvertebrates capable of tolerating agricultural nonpoint source pollution. Additional support for this hypothesis is the high abundance of one food source, periphyton, and the small quantities of terrestrially derived organic matter. The abundance of the former and the rarity of the latter can be attributed to alteration of the drainage basin resulting from agricultural activities through inputs of fertilizers that generated high nutrient concentrations and the removal of riparian vegetation to clear more land for agriculture and provide increase access to the stream.KEY WORDS: Agriculture; Longitudinal patterns; Macroinvertebrates; Nonpoint source; River continuum     

Assessment of water quality and identification of pollution sources of plateau lakes in Yunnan (China)

The plateau lakes of Yunnan are important both ecologically and economically in China. Nevertheless, the human impact on water quality in these lakes has become increasingly highlighted. The water quality of 10 plateau lakes was monitored regularly over the period of 2000 through 2004 for 24 parameters. Multivariate statistical techniques, including cluster analysis (CA), factor analysis (FA), and principal component analysis (PCA), were employed to better interpret information about the water quality and its pollution sources. No obvious data reduction from CA/FA was found because three principal components (PCs) needed 14 variables to explain 85.01% of the total variance. However, three latent factors accounted for pollution mainly from the following sources: agricultural activities, residential activities and anthropogenic-toxic pollution from industrial effluents, or other special activities. Box-whiskers plots were employed to visually interpret the spatiotemporal variations of water quality variables, which were highly correlated with three PCs. Three types of water quality (i.e., low-, medium-, and high-polluted lakes) were determined through CA based on the similarity of water quality variables. Our results may provide helpful information for the authorities to effectively manage the water quality and make sound policies.     

MULTIVARIATE CLASSIFICATION OF SMALL ORDER WATERSHEDS IN THE QUABBIN RESERVOIR BASIN,MASSACHUSEVFS1

ABSTRACT: A multivariate approach was used to analyze hydrologic, geologic, geographic, and water-chemistry data from small order watersheds in the Quabbin Reservoir Basin in central Massachusetts. Eighty three small order watersheds were delineated and landscape attributes defining hydrologic, geologic, and geographic features of the watersheds were compiled from geographic information system data layers. Principal components analysis was used to evaluate 11 chemical constituents collected bi-weekly for 1 year at 15 surface-water stations in order to subdivide the basin into subbasins comprised of watersheds with similar water quality characteristics. Three principal components accounted for about 90 percent of the variance in water chemistry data. The principal components were defined as a biogeochemical variable related to wet. land density, an acid-neutralization variable, and a road-salt variable related to density of primary roads. Three subbasins were identified. Analysis of variance and multiple comparisons of means were used to identify significant differences in stream water chemistry and landscape attributes among subbasins. All stream water constituents were significantly different among subbasins. Multiple regression techniques were used to relate stream water chemistry to landscape attributes. Important differences in landscape attributes were related to wetlands, slope, and soil type.     

Conceptual classification model for Sustainable Flood Retention Basins

The aim of this paper is to recommend a rapid conceptual classification model for Sustainable Flood Retention Basins (SFRB) used to control runoff in a temperate climate. An SFRB is an aesthetically pleasing retention basin predominantly used for flood protection adhering to sustainable drainage and best management practices. The classification model was developed on the basis of a database of 141 SFRB using the River Rhine catchment in Baden (part of Baden-Württemberg, Germany) as a case study. It is based on an agglomerative cluster analysis and is intended to be used by engineers and scientists to adequately classify the following different types of SFRB: Hydraulic Flood Retention Basin, Traditional Flood Retention Basin, Sustainable Flood Retention Wetland, Aesthetic Flood Retention Wetland, Integrated Flood Retention Wetland and Natural Flood Retention Wetland. The selection of classification variables was supported by a principal component analysis. The identification of SFRB in the data set was based on a Ward cluster analysis of 34 weighted classification variables. Scoring tables were defined to enable the assignment of the six SFRB definitions to retention basins in the data set. The efficiency of these tables was based on a scoring system which gave the conceptual model for the example case study sites an overall efficiency of approximately 60% (as opposed to 17% by chance). This conceptual classification model should be utilized to improve communication by providing definitions for SFRB types. The classification definitions are likely to be applicable for other regions with both temperate oceanic and temperate continental climates.     

Targeted Application of Seasonal Load Duration Curves Using Multivariate Analysis in Two Watersheds Flowing Into Lake Houston1

Teague, Aarin, Philip B. Bedient, and Birnur Guven, 2011. Targeted Application of Seasonal Load Duration Curves Using Multivariate Analysis in Two Watersheds Flowing Into Lake Houston. Journal of the American Water Resources Association (JAWRA) 47(3):620‐634. DOI: 10.1111/j.1752‐1688.2011.00529.x Abstract: Water quality is a problem in Lake Houston, the primary source of drinking water for the City of Houston, Texas, due to pollutant loads coming from the influent watersheds, including Spring Creek and Cypress Creek. Statistical analysis of the historic water quality data was developed to understand the source characterization and seasonality of the watershed. Multivariate analysis including principal component, cluster, and discriminant analysis provided a custom seasonal assessment of the watersheds so that loading curves may be targeted for season specific pollutant source characterization. The load duration curves have been analyzed using data collected by the U.S. Geologic Survey with corresponding City of Houston water quality data at the sites to characterize the behavior of the pollutant sources and watersheds. Custom seasons were determined for Spring Creek and Cypress Creek watersheds and pollutant source characterization compared between the seasons and watersheds.     

Toward an improved ecosystem based management approach: incorporating catchment characteristics into better management and planning of the Great Barrier Reef marine ecosystem

Discharge from the Great Barrier Reef Catchment (GBRC) is considered the second most serious threat to the Great Barrier Reef, Australia. Utilising principal component analysis (PCA) and cluster analysis (CA), this research aimed to assess the variability and co-variation of 28 water basins (WBs) within the GBRC, in order to improve the institutional arrangements and regulation of water quality and increase collaboration horizontally between management organisations, and vertically between government tiers. Water basin variability was measured by nine variables: size (ha), population, agricultural land use (ha), number of major water storages, major rivers and major towns, total nitrogen exported (T/yr), total phosphorus exported (T/yr) and herbicide use (ha). The Fitzroy WB, with PC scores of 7.0081, 2.2897 and ?1.6504, was identified as the most dissimilar and therefore needing to be managed differently. Many WBs within the same regions were very dissimilar to each other, indicating that current management practices, based largely on geographic location, are unlikely to be the most efficient and effective. Instead, managing groups of WBs with similar geo-political properties (determined by the CA) could be more effective and efficient. Coordination and collaboration are key to successful ecosystem based management, therefore managing similar WBs together through inter-NRM (natural resource management) agreements, irrespective of their geographical location, facilitates management bodies building strong, cooperative working relationships.     

Management of Sedimentation in Tropical Watersheds

/ The sedimentation of reservoirs is a serious problem throughout the tropics, yet most attempts to control sedimentation in large river basins have not been very successful. Reliable information on erosion rates and sources of sediments has been lacking. In regions where geologically unstable terrain combines with high rainfall, natural erosion rates might be so high that the effects of human activity are limited. Estimates of natural erosion in these situations often have been poor because of the episodic nature of most erosion during large storms and because mass-wasting may supply much of the sediment. The predominance of mass-wasting in some watersheds can result in an unexpectedly high ratio of bedload to suspended load, shifting sedimentation to "live" rather than "dead" storage within reservoirs. Furthermore, the inappropriate use of the Universal Soil Loss Equation to assess the effectiveness of erosion control measures has led to inaccurate estimates of the sediment reduction benefits that could accrue to watershed treatment efforts. Although reducing erosion from cultivated areas is desirable for other reasons, efforts aimed at reducing reservoir sedimentation by controlling agricultural sources of erosion may have limited benefits if the principal sources are of natural origin or are associated with construction of the dams and reservoirs and with rural roads and trails. Finally, the most appropriate locations for watershed rehabilitation depend on the magnitude of temporary storage of colluvium and alluvium within the river basin: Where storage volume is large and residence time of sediment very long, reducing agricultural erosion may have limited impacts on sedimentation within the expected life of a reservoir. Systematic development and analysis of sediment budgets for representative watersheds is needed to address these limitations and thereby improve both the planning of river basin development schemes and the allocation of resources towards reducing sedimentation. When sedimentation of reservoirs is the key issue, sediment budgets must focus especially on channel transport rates and sediment delivery from hillsides. Sediment budgets are especially critical for tropical areas where project funds and technical help are limited. Once sediment budgets are available, watershed managers will be able to direct erosion control programs towards locations where they will be most effective. KEY WORDS: Tropical watersheds; Sedimentation; Reservoirs; Erosion control     

GROUPWISE MODELING STUDY OF BACTERIALLY IMPAIRED WATERSHEDS IN TEXAS: CLUSTERING ANALYSIS1

ABSTRACT: Under the Clean Water Act (CWA) program, the Texas Commission on Environmental Quality (TCEQ) listed 110 stream segments in the year 2000 with pathogenic bacteria impairment. A study was conducted to evaluate the probable sources of pollution and characterize the watersheds associated with these impaired water bodies. The primary aim of the study was to group the water bodies into clusters having similar watershed characteristics and to examine the possibility of studying them as a group by choosing models for total maximum daily load (TMDL) development based on their characteristics. This approach will help to identify possible sources and determine appropriate models and hence reduce the number of required TMDL studies. This in turn will help in reducing the effort required to restore the health of the impaired water bodies in Texas. The main characteristics considered for the classification of water bodies were land use distribution within the watershed, density of stream network, average distance of land of a particular use to the closest stream, household population, density of on‐site sewage facilities (OSSFs), bacterial loading from different types of farm animals and wildlife, and average climatic conditions. The climatic data and observed instream fecal coliform bacteria concentrations were analyzed to evaluate seasonal variability of instream water quality. The grouping of water bodies was carried out using the multivariate statistical techniques of factor analysis/principal component analysis, cluster analysis, and discriminant analysis. The multivariate statistical analysis resulted in six clusters of water bodies. The main factors that differentiated the clusters were found to be bacterial contribution from farm animals and wildlife, density of OSSFs, density of households connected to public sewers, and land use distribution.     

ROLE OF CARBONATE ROCKS IN MODIFYING FLOOD FLOW BEHAVIOR1

ABSTRACT. The interrelationships between the runoff characteristics of watersheds (expressed as the mean annual flood), standard basin parameters (area, drainage properties, and relief), and the parameters which describe the solutional modification of the basins (carbonate rock fractions, sinkhole development, and measures of internal drainage) were used to group 62 carbonate watersheds. Simple binary correlations were obtained by direct plotting of the data. This was followed by multivariate analyses: factor and cluster analyses. Following the cluster analysis, which separated the basins into three groups, the variance within each group was examined again by binary correlations and by factor analysis. Prediction equations for those basins underlain by dolomite rock [QBAR = 12.4 TOT^1.01] and for those basins underlain by carbonate rock with very little surface expression [QBAR = 43.5 TOT^0.87] were proposed. Basins underlain by karstic limestone had a large amount of variance within the data set; therefore no prediction equation could be obtained. (QBAR = mean annual flood, cfs; TOT = total length of all blue lines shown on topographic maps, miles.)     

Multivariate Analysis of the Ecoregion Delineation for Aquatic Systems

The ecoregion concept is a popular method of understanding the spatial distribution of the environment', however, it has yet to be adequately demonstrated that the environment is distributed in accordance with these bounded units. In this paper, we generated a testable hypothesis based on the current usage of ecoregions: the ecoregion classification will allow for discrimination between lakes of different water quality. The ecoregion classification should also be more effective better than a comparably scaled classification based on political boundaries, land-use class, or random grouping. To test this hypothesis we used the Environmental Monitoring and Assessment Program (EMAP) lake water chemistry data from the northeast United States. The water chemistry data were reduced to four components using principal component analysis. For comparison to an optimal grouping of these data we used K-means cluster analysis to define the extent at which these lakes could be segregated into distinct classes. Jackknifed discriminant analysis was used to determine the classification rate of ecoregions, the three alternative spatial classification methods, and the clustering algorithm. The classification based on ecoregions was successful for 35% of the lakes included in this study, in comparison to the clustered groups accuracy of 98%. These results suggest that the large scale spatial distribution of ecosystem types is more complicated than that suggested by the present ecoregion boundaries. Further tests of ecoregion delineations are needed and alternative large-scale management strategies should be investigated.