Trends In Water Quality Data Research Articles

Analyses of long-term trends in water quality data of the Plitvice Lakes National Park

Water quality management requires monitoring of the condition of natural waters by both biologicaland chemical indicators to assess the ecological health of a catchment. General water quality characteristics areextracted through a low-frequency dataset, which allows analyses of long-term trends and seasonality. PlitviceLakes National Park freshwater ecosystem includes springs, rivers, streams and lakes with its specific process oftufa formation. The aim of this study was to determine, based on long-term water quality data, potential changes inphysico-chemical parameters required for ongoing tufa deposition processes. The dataset encompasses a 12-yearperiod (2006 – 2017), with measurements performed from April to October per each year on a weekly basis, coveringthirteen sites belonging to karst springs, streams or lotic biotopes. Positive trends were observed for dissolvedoxygen, conductivity, water hardness, alkalinity, chemical oxygen demand (COD) and nitrites. Negative trendswere observed for temperature, pH and most nutrients (nitrates, ammonia and orthophosphates). Temperature valuesshowed different trend slopes among monitored types. Negative correlations between water level and/or dischargewith conductivity and alkalinity probably indicate the effect of dilution. Conversely, nutrients and CODhave a positive correlation with discharge, possibly indicating resuspension from sediment during high flows. Theanalyses of spatial and temporal variations in hydrochemistry through a longer period imply annual alterations,correlations between hydrological components and water chemistry, and significant differences among monitoredtypes, while the tufa deposition process is stable. Long-term monitoring and continuous data analyses can help infuture planning of monitoring programs and the improved understanding of climate changes.

Trends in water quality data in the Hawkesbury–Nepean River System, Australia

The Hawkesbury–Nepean River System (HNRS) is one of the most important inland river systems in Australia, which supplies over 90% of Sydney's potable water. In this paper, 25 water quality parameters from nine sampling stations in the HNRS covering a period of 12 years are used to examine the trends in the water quality data in the HNRS. It has been found that there is an overall increasing trend of turbidity, chlorophyll-a, alkalinity, total iron, total aluminium, total manganese and reactive silicate, indicating an overall water quality deterioration in the HNRS during the last decade. The parameters such as phosphorus, suspended solids and ammonical nitrogen do not show any marked change over the period of study. Although an improvement in water quality can be seen at some stations downstream of the undisturbed parts of the catchment, there is a clear trend of increased chemical and physical water quality deterioration at many locations in the HNRS. Better land use planning is recommended to achieve an overall improvement in the water quality of the HNRS in future.

Application of a Nonparametric Approach for Monitoring and Detecting Trends in Water Quality Data for the St

Abstract To overcome some of the restrictions due to the irregular structure of water quality data, a nonparametric approach has been developed based on Kendall’s T and a seasonal adjustment model, which enables one to test for the significance of trends as well as to monitor the trend variations with time. This methodology has been applied to water quality data for the St. Lawrence river obtained from Environment Canada and the municipality of Varennes. The results from the first source indicate that a positive trend is developing for conductivity and pH downstream from the Island of Montreal, whereas upstream from the Island pH is the only parameter indicating a positive trend. Data from the municipality of Varennes, due to their excellent quality, were used to verify the validity of the previous results as well as the reliability of the applied methodology. Analysis of the Varennes data with a parametric method indicated that a positive trend in pH is indeed evident. This finding will be of consequence for future studies to ascertain the effect on the river of the new Montreal Urban Community Sewage Treatment Plant.

Methods of discharge compensation as an aid to the evaluation of water quality trends

Two new methods are described for compensating for discharge when evaluating trends in water quality data. One method, discharge normalization, adjusts daily discharges using a central value calculated for the period of record and recalculates daily specific conductance from the adjusted discharges and discharge versus specific conductance regressions. Normalized concentrations for many constituents can then be calculated from linear relationships between specific conductance and constituent concentrations. The second method, discharge‐frequency weighting, weights each observed concentration by a fraction of the total area underneath the discharge‐frequency distribution of the period of record. This fraction is determined using the stream discharge at the time of sampling and the discharge‐frequency distribution for the period of record. The weighted concentrations are summed for each year. Both normalized values and weighted values can be plotted against time to produce trends essentially independent from discharge effects. Results from the methods are statistically similar to each other and to results from other trend detection techniques.

Detection of trends in water quality data from records with dependent observations

Classical statistical tests for trend, both parametric and nonparametric, assume independence of observations, a condition rarely encountered in time series obtained by using moderate to high sample frequencies. A method is developed for summarizing the power of the parametric t tests and the nonparametric Spearman's rho test and Mann‐Whitney's test against step and linear trends in a dimensionless ‘trend number’ which is a function of trend magnitude, standard deviation of the time series, and sample size. For the case of dependent observations, use of an equivalent independent sample size rather than the actual sample size is shown to enable use of the same trend number developed for the independent case. An important related result is the existence of an upper limit on power (trend detectability) over a fixed time horizon, regardless of the number of samples taken, for a lag 1 Markov process.