Springs Index
- » Spring Water Quality

Spring Water Quality

Green Springs

Accurate and reliable water quality data are required to characterize groundwater chemistry for use in resource management. Major ion chemistry is used as a general indicator of groundwater quality and provides a means to identify regional differences and changes in water quality with time. A spring's water quality is determined by several factors. These include the chemical composition of the water entering the aquifer, the composition and solubility of the rocks with which the water comes into contact along flow paths, the length of time the water is in contact with the rocks as it moves from recharge to discharge areas, and the mixing of fresh groundwater with residual formation water or seawater.

Land use activities in a spring's recharge basin and the upconing of poorer quality water from deeper zones due to groundwater withdrawals may also impact water quality. The most pristine springs in the state occur in the Ocala National Forest, with spring recharge basins that encompass limited land uses. Other springs have recharge basins that encompass urban land uses and well withdrawals that have resulted in spring flow reductions and declining water quality.

The chemical composition and physical properties of a spring sample represent the net effect of all the previous chemical processes that have dissolved, altered, or precipitated the chemical constituents. The major constituents considered are those commonly present in concentrations exceeding 1.0 mg/L and that constitute a major part of the total dissolved solids content. The major cations include calcium, magnesium, sodium, and potassium. The major anions include chloride, and sulfate, carbonate alkalinity. Fluoride, nitrate + nitrite, and phosphate are important minor anions. Chemically related parameters include total dissolved solids and total organic carbon. Field measurement data include temperature, specific conductance, pH, and spring discharge.

The water quality results are summarized in a table for each spring. Descriptive statistics for each water quality variable include the number of samples, period of record, and the minimum, mean, median, and maximum values. The median is the 50th percentile and is a better estimate of the central tendency of the data set than the mean. Concentrations are reported for the total and dissolved fraction, where available. The dissolved fraction represents the concentrations after passing the water through 0.45-micron filters. Only quality assured data were used; data below detection limit or data from samples held beyond holding time before analysis were not used. Total dissolved solids, chloride, sulfate, total nitrogen (nitrate + nitrite), and phosphate are graphed where sufficient measurement data are available.

Statistical Trend Analysis

A temporal trend is the general increase or decrease in observed values over time. Trend analysis is used to determine the significance of a trend in water quality for that spring and to estimate the magnitude of that trend. The Median test, Seasonal Kendall test, and Sen's Slope estimator (Gilbert 1987; Helsel and Hirsch 1992; IDT 1998) were used to evaluate trends in chloride, sulfate, and total nitrogen (nitrate + nitrite) where sufficient measurement data are available. Only data since 1995 were used in this evaluation.

The median test was used as the primary statistical test to determine if the data revealed a stable, decreasing, or increasing trend. The Median test evaluates the difference in the medians between sample populations. The data were grouped into two 5-year periods: 1995-1999 and 2000-2005 (which included samples through May 2005). The test splits each sample population into two groups: those above the median of all observations in the populations tested and those below. The groups are analyzed using the Fisher exact test for comparing two populations. The samples populations were then tested to determine if there is a statistically significant difference at the 80% confidence interval. The test is robust and sample populations may have missing values and unequal numbers of samples, as is the case for most of the spring data sets.

The Sen's Slope estimator is a nonparametric, linear slope estimator that works most effectively on monotonic data. Unlike linear regression, it is not greatly affected by gross data errors, outliers, or missing data. For the chloride and sulfate analyses, a slope of 3 milligrams per liter (mg/L) per year was used as a threshold between a stable trend and an increasing trend; for nitrate + nitrite, a slope of 0.02 mg/L per year was used as a threshold. These thresholds are based on laboratory analysis replication limits. If the Median test resulted in an increasing trend and the Sen's Slope was less than the threshold limit, the increasing trend is considered as insignificant, but these springs would be watched closely in future monitoring.

The Seasonal Kendall test is a nonparametric test utilizing the Mann-Kendall test where the data set is adjusted for seasonality (in this case, the four quarters of the year). The Seasonal Kendall test is not greatly affected by missing data, but does require a minimum number of data points for each season. This test is a linear estimator and works most effectively on monotonic data. Since not all of the spring data sets exhibit strict monotonic trends, the Seasonal Kendall test was used for verification of the median test and as another approach to identify springs that call for ongoing evaluation.

Trend Results

For the chloride data, 25 springs were evaluated for trend. Of these 25 springs, the Median test resulted in 15 springs with no change between the two 5-year periods of 1995-1999 and 2000-2005. A decreasing trend was observed in 4 springs. An increasing trend was found in 6 springs, but all were insignificant trends since the slopes were below the 3.0 mg/L per year slope threshold limit. These 6 springs are Blue (Lake County), Bugg, Miami, Palm (Seminole County), Sanlando, and Wekiwa. These springs will be watched closely in future monitoring.

Chloride Trend Analysis Table

For the sulfate data, 25 springs were evaluated for trend. Of these 25 springs, the Median test resulted in 14 springs with no change between the two periods. A decreasing trend was observed in 5 springs. An increasing trend was found in 6 springs, but all were insignificant trends since the slopes were below the 3.0 mg/L per year slope threshold limit. These 6 springs are Apopka, Blue (Lake County), Bugg, Miami, Palm (Seminole County), and Starbuck. These springs will be watched closely in future monitoring.

Sulfate Trend Analysis Table

For the nitrate + nitrite data, 24 springs were evaluated for trend. Of these 24 springs, the Median test resulted in 18 springs with no change between the two periods. A decreasing trend was observed in 5 springs. Only Gemini Springs showed an increasing trend, with a slope above the 0.02 mg/L per year slope threshold limit, requiring close evaluation in the future.

Nitrate + Nitrite Trend Analysis Table

Definitions of Water Quality Variables