Stochastic modelling of phytoremediation

Leaching of heavy metals and other contaminants from soils poses a significant environmental threat as it affects the quality of downstream water bodies. Quantifying these losses is particularly important when employing phytoremediation approaches to reduce soil contamination, as contaminant escaping the system through leaching cannot be taken up by vegetation. Despite its undoubted importance, the role of such hydrologic forcing has seldom been fully considered in models describing the long-term contaminant mass balance during phytoremediation. The partitioning of contaminants between leaching and vegetation uptake is controlled by a number of biophysical processes as well as rainfall variability. Here, we develop a novel stochastic framework that provides analytical expressions to quantify the partitioning of contaminants between leaching and plant uptake and the probability of phytoremediation duration as a function of rainfall statistics and soil and vegetation characteristics. Simple expressions for the mean phytoremediation duration and effectiveness (defined as the fraction of contaminant that is recovered in plant biomass) are derived. The proposed framework can be employed to estimate under which conditions phytoremediation is more efficient, as well as to design phytoremediation projects that maximize contaminant recovery and minimize the duration of the remediation process.