Main content area

An assessment of long-term soil acidification trends in Alberta, Canada

Cho, Sunny, Dinwoodie, Gordon, Fu, Yiyang, Abboud, Salim, Turchenek, Larry
Ecological indicators 2019 v.98 pp. 712-722
acidification, aluminum, atmospheric deposition, base saturation, cations, chemical bases, environmental indicators, grasslands, linear models, monitoring, oil sands, oils, pH, regression analysis, sandy soils, soil acidification, Alberta
Long-term soil acidification monitoring was initiated in 1981 to track potential effects of acidic atmospheric deposition on soils in Alberta, Canada. Six of eight long-term sampling sites were located in forest ecosystems on sandy soils in northern and central Alberta, and two were situated in grasslands on sandy and fine-loamy soils in southern Alberta. Sampling was undertaken every four or five years, with 12 replicate samples taken from seven depth increments from 0–60 cm at each site. The statistical analysis of pH, exchangeable base saturation (BS) percentage, and solution base cation (BC) to aluminum (Al) ratio was conducted in the topmost soil layers (0–2, 2–5 and 5–10 cm). Analysis of variance (ANOVA) and linear regression indicated decreasing trends in pH and BS percentage in some of the sites. Comparison of the initial and most recent monitoring data indicated decreases of up to 0.5 pH units in these layers at five of the forested sites. At two forested sites located in the vicinity of oil sands and heavy oil processing plants, decreasing BS was observed early in the monitoring period, but the trend showed more variability since 2000, even though pH has continued to decline since 1981. BC to Al ratio showed statistically significant variations within some sites over time, but except for a downward trend at one central Alberta site, there were no overall patterns evident at the other sites over time. Both grassland sites had much higher BC to Al ratio and exchangeable BS percentages than the forested sites, and acidification trends were not detected at either of these sites. ANOVA and linear modelling approaches showed significant differences even in instances where pH differed by 0.1–0.2 units, or where other parameters showed very small differences. It is therefore necessary to distinguish between statistically important and biologically important differences in this monitoring program, which is intended to continue for several more years.