Quantifying remediation effectiveness under variable external forcing using contaminant rating curves.

Remediation efforts are typically assessed through before-and-after comparisons of contaminant concentrations or loads. These comparisons can be misleading when external drivers, such as weather conditions, differ between the pre- and postremediation monitoring periods. Here, we show that remediation effectiveness may be better assessed by comparing pre- and postremediation contaminant rating curves, which permit "all else equal" comparisons of pre- and postremediation contaminant concentrations and loads under at any specified external forcing. We illustrate this approach with a remediation case study at an abandoned mercury mine in Northern California. Measured mercury loads in the stream draining the mine site were a factor of 1000 smaller after the remediation than before, superficially suggesting that the cleanup was 99.9% effective, but rainstorms were weaker and less frequent during the postremediation monitoring period. Our analysis shows that this difference in weather conditions alone reduced mercury loads at our site by a factor of 73-85, with a further factor of 12.6-14.5 being attributable to the remediation itself, implying that the cleanup was 92-93% (rather than 99.9%) effective. Our results illustrate the need to account for external confounding drivers when assessing remediation efforts, particularly in systems with highly episodic forcing.