VARIATION IN THE QUALITY OF TENSION LYSIMETER SOIL WATER SAMPLES FROM A FINNISH FOREST SOIL

Abstract

This paper examines sources of variation in the chemical quality of soil water collected during a 9-wk period (July to September 1981) with six tension lysimeters installed at a forest harvested site in central Finland. The soil was an iron podsol, and the lysimeters were installed to a depth of 25 cm, i.e., within the B horizon. Three lysimeters were installed at 2− to 3-m spacings in the center of two plots, some 200 m apart. The lysimeters were fitted with 6 cm long x 3 cm o.d. sintered glass cups of pore size 10 to 16 μm. Replicate samples, usually five, were drawn regularly each week with each lysimeter, using a tension pressure of ca. 42 kPa. The mean soil water tension at the 15− and 35-cm depths was 4.1 and 3.1 kPa, respectively. Chemical analysis (conductivity and H+], NH4+], PO43-], and K+] and Ca2+] ion concentrations) of the unfiltered samples was completed within 72 h of collection. The sampling error among replicates was found to decrease in the order: NH4+], H+], PO43-], K+] and Ca2+], conductivity. The mean and coefficient of variation of successive replicates decreased progressively and tended toward uniformity in the order in which the replicates had been taken. Significant differences (P = 0.05) in lysimeter water quality within the same plot were found for all determinations except PO43-]. Significant differences among plots were found for all determinations except NH4+] and PO43-]. The amount of spatial variation at each sampling date varied, but the differences among lysimeters were consistent throughout, and all six lysimeters showed generally similar trends. The overall mean soil water quality was: conductivity, 64.2 μS/cm; H+], 14.4 μM; NH4+], 2.4 μM; PO43-], 0.2 μM; K+], 50.1 μM; and Ca2+], 184.5 μM. The mean lysimeter/rain quality ratios were: conductivity, 1.9; H+], 1.1; NH4+], 0.04; PO43-], 0.2; K+], 6.3; and Ca2+], 29.3. With the exception of conductivity, lysimeter water quality showed no, or only weak, rainfall-related dilution effects. Peak concentrations of H+] and Ca2+] followed rainfall with extreme concentrations and loads of these ions. I concluded that tension lysimeters offer a simple means of obtaining soil water samples for chemical analysis, but their value may be limited in Finnish conditions where there is low rainfall and high potential evapotranspiration during the summer months, resulting in little percolating water. The soil water collected is probably derived from deeper in the profile, and it would tend to show less variation than infiltrating water because of longer soil contact time. Because of the considerable spatial variation in soil water quality, it is important that data from the same lysimeters contribute to the mean value for each sampling date if lysimeters are to be used to monitor trends in quality over time.