Epsilon Archive for Student Projects

Abstract

In environmental research, the movement, accumulation and dissolving of pollutants are often of great interest. An important field of investigation is, for instance, the transport of heavy metals - such as cadmium (Cd), lead (Pb) and mercury (Hg) - through the soil. To understand the behaviour and estimate the quantity of those substances in certain subterranean parts, the chemical concentrations as well as hydrological fluxes need to be known. Unfortunately, at this point modern science approaches its limits. It is complicated to accomplish accurate and reliable measurements of hydrological fluxes at a small scale, i.e. percolating soil water.

This problem can be solved by applying a soil hydrological model. WATBAL is such a soil water balance model that can help to estimate the monthly soil water flux values within the rooting zone. The model was developed by Mike Starr (University Helsinki, Finland), but was not adequately tested and verified in the past, especially not applied to Swedish till soils.

In this study, the WATBAL model was intensely examined and slightly improved by changing a few parameters. It was applied to several Swedish sites to investigate its grade of accuracy and plausibility. The results showed that WATBAL is definitely a reasonable soil water model.
The numerical statistical analysis resulted in a moderate grade of accuracy with R2 values between 0.13 and 0.58 combined with MAPE values from 0.02 to 0.21. However, the graphical analysis of the model results revealed a good model fit, so that WATBAL in general is rated as acceptable. The overall average percentage error was estimated as 8.91% ± 1.08% (95% confidence interval).

Furthermore, the obtained results from the WATBAL simulations were used to investigate the heavy metal transport within the soil. With available chemical data and the calculated small scale water fluxes, the mass balance of cadmium, lead, mercury, zinc and copper could be estimated. Even though chemical values were rather incomplete, results were obtained and showed that between 68.37% (cadmium, Kindla) and 97.28% (copper, Kindla) of the incoming heavy metals remained in the soil. Litterfall, throughfall and corresponding absolute accumulation values were much higher in Kindla than in Gammtratten. Besides, a seasonal dependency of monthly absolute accumulation was detected, with much greater values for April to October than for November to March. Based on long-term considerations, the amount of accumulated heavy metals spread over the whole catchment can add up to several kilograms.