SELECTING TREND TESTS FOR WATER QUALITY SERIES WITH SERIAL CORRELATION AND MISSING VALUES1

ABSTRACT: The use of nonparametric tests for monotonic trend has flourished in recent years to support routine water quality data analyses. The validity of an assumption of independent, identically distributed error terms is an important concern in selecting the appropriate nonparametric test, as is the presence of missing values. Decision rules are needed for choosing between alternative tests and for deciding whether and how to pre-process data before trend testing. Several data pre-processing procedures in conjunction with the Mann-Kendall tau and the Seasonal Kendall test (with and without serial correlation correction) are evaluated using synthetic time series with generated serial correlation and missing data. A composite test (pre-testing for serial correlation followed by one of two trend tests) is evaluated and was found to perform satisfactorily.     

Classification of local- and landscape-scale ecological types in the southern Appalachian Mountains

Five local ecological types based on vegetative communities and two landscape types based on groups of communities, were identified by integrating landform, soil, and vegetation components using multivariate techniques. Elevation and several topographic and soil variables were highly correlated with types of both scales. Landscape ecological types based only on landform and soil variables without vegetation did not correspond with types developed using vegetation. Models developed from these relationships could allow classification and mapping of extensive areas using geographic information systems.     

Floristics, performance and prognosis of historical trees in the urban forest of Guangzhou City (China)

Outstanding historical trees embedded in cities constitute pertinent environmental assets, yet they are widely threatened in third-world cities. Inadequate understanding of this valuable natural-cum-cultural heritage hinders proper conservation. A case study of Guangzhou in south China evaluated floristic composition, age profile and biomass structure of historical trees, assessed their performance in major habitats (institutional, park and roadside), and established a prognosis for future growth and management. The 348 historical trees examined belonged to only 25 species, vis-à -vis 254 trees in the entire urban forest, dominated by five species and native members. Roadside had more trees, followed by institutional and park, with merely the most common four species shared by all habitats. The limited commonality reflected tree-performance differentiation by habitats exerting selection pressure on species. The institutional growth-regime was more conducive to nurturing high-caliber specimens, whereas park is less capable. Individual species achievement by habitats, derived from tree-count ranking and relative-abundance indices, could inform species choice and tree conservation. Few trees exceeded 300 years of age in the millennium-old city, echoing a history of intense tree—city conflicts. Potential life-span, trunk and crown diameters indicated ample opportunities for further expansion of biomass and landscape impacts, which would be straitjacketed by the tightening urban fabric.     

Interlaboratory evaluation of endotoxin analyses in agricultural dusts--comparison of LAL assay and mass spectrometry

Endotoxin exposure is associated with wheeze and asthma morbidity, while early life exposure may reduce risk of allergy and asthma. Unfortunately, it is difficult to compare endotoxin results from different laboratories and environments. We undertook this study to determine if lipopolysaccharide (LPS) extraction efficiency could account for differences among laboratories. We generated and collected aerosols from chicken and swine barns, and corn processing. We randomly allocated side-by-side filter samples to five laboratories for Limulus assay of endotoxin. Lyophilized aliquots of filter extracts were analyzed for 3-hydroxy fatty acids (3-OHFAs) as a marker of LPS using gas chromatography-mass spectrometry. There were significant differences in endotoxin assay and GC-MS (LPS) results between laboratories for all dust types (p < 0.01). Patterns of differences between labs varied by dust type. Relationships between assay and GC/MS results also depended on dust type. The percentages of individual 3-OHFA chain lengths varied across labs (p < 0.0001) suggesting that each lab recovered a different fraction of the LPS available. The presence of large amounts of particle associated LPS and absence of a freezing thawing cycle were associated with lower correlations between LPS and bioactivity, consistent with an absence of Limulus response to cell-bound endotoxin. These data suggest that extraction methods affect endotoxin measurements. The LAL methods may be most suitable when comparing exposures within similar environments; GC-MS offers additional information helpful in optimizing sample treatment and extraction. GC-MS may be of use when comparing across heterogeneous environments and should be considered for inclusion in future studies of human health outcomes.     

Characterization of microbial communities from coastal waters using microarrays

Molecular methods, including DNA probes, were used to identify and enumerate pathogenic Vibrio species in the Chesapeake Bay; our data indicated that Vibrio vulnificus exhibits seasonal fluctuations in number. Our work included a characterization of total microbial communities from the Bay; development of microarrays that identify and quantify the diversity of those communities; and observation of temporal changes in those communities. To identify members of the microbial community, we amplified the 16S rDNA gene from community DNA isolated from a biofilm sample collected from the Chesapeake Bay in February, 2000. The resultant 75 sequences were 95% or more similar to 7 species including two recently described Shewanella species, baltica and frigidimarina, that have not been previously isolated from the Chesapeake. When the genera of bacteria from biofilm after culturing are compared to those detected by subcloning amplified 16S fragments from community DNA, the cultured sample exhibited a strong bias. In oysters collected in February, the most common bacteria were previously unknown. Based on our 16S findings, we are developing microarrays to detect these and other microbial species in these estuarine communities. The microarrays will detect each species using four distinct loci, with the multiple loci serving as an internal control. The accuracy of the microarray will be measured using sentinel species such as Aeromonas species, Escherichia coli, and Vibrio vulnificus. Using microarrays, it should be possible to determine the annual fluctuations of bacterial species (culturable and non-culturable, pathogenic and non-pathogenic). The data may be applied to understanding patterns of environmental change; assessing the health of the Bay; and evaluating the risk of human illness associated with exposure to and ingestion of water and shellfish.     

24 h diffusive sampling of toxic VOCs in air onto Carbopack X solid adsorbent followed by thermal desorption/GC/MS analysis-laboratory studies

Diffusive sampling of a mixture of 42 volatile organic compounds (VOCs) in humidified, purified air onto the solid adsorbent Carbopack X was evaluated under controlled laboratory conditions. The evaluation included variations in sample air temperature, relative humidity and ozone concentration. Linearity of samples with loading was examined both for a constant concentration with time varied up to 24 h and for different concentrations over 24 h. Reverse diffusion and its increase with accumulation of sample were determined for all compounds. Tubes were examined for blank levels, change of blanks with storage time, and variability of blanks. Method detection limits were determined based on seven replicate samples. Based on this evaluation, 27 VOCs were selected for quantitative monitoring in the concentration range from approximately 0.1 to 4 ppbv. Comparison results of active and diffusive samples taken over 24 h and under the same simulated ambient conditions at a constant 2 ppbv were interpreted to estimate the effective diffusive sampling rates (ml min(-1)) and their uncertainties and to calculate the corresponding diffusive uptake rates (ng ppmv(-1) min(-1)).     

Prenatal ultrasound diagnosis of fetal scoliosis with termination of the pregnancy: Case report

Gross scoliosis of the fetal thoracic spine was diagnosed at 18 weeks gestation. The pregnancy was terminated and the fetus found to have webbing of the neck and an imperforate anus in addition to vertebral defects.     

Cost-effective selection of sampling frequencies for regulatory water quality monitoring

A dynamic programming code was formulated for the purpose of assigning sampling frequencies throughout a regulatory water quality monitoring network in order to optimize the statistical performance of the network while operating within a fixed budgetary constraint. The statistical objective is to achieve the greatest possible station to station uniformity in confidence interval widths about annual geometric means of the measured water quality variables and to keep the average confidence width reasonably small. The objective function is the sum (over several selected variables and all stations) of the normalized positive deviations of the predicted confidence interval widths from preselected design confidence interval widths. The code was designed to account for the effects of deterministic seasonal variation and serial correlation of the water quality observations by incorporating the results of the time series analysis of historical quality data. The economic constraint ensures that the annual operating cost of the system, including direct costs of travel and laboratory analysis, will not exceed the allowable budget. As an example situation, the dynamic programming code was used to assign sampling frequencies to the nine stations in Illinois from which historical quality data had been obtained and analyzed. Using five design quality constituents and representative travel and laboratory costs, an “optimal” design was produced. The optimal design achieved a 10% improvement in uniformity (standard deviation) of confidence interval widths when compared to a more traditional design based on the same budget and using identical sampling frequencies at every station.