『Abstract
　Runoff non-point source pollution from phosphate mining areas poses a potential risk to ecosystems in many parts of the world. Mining sand tailings in Central Florida, which still contain apatite (phosphate rock), have shaped the landscape in reclaimed lands at the upper Peace River basin. The objective of this study is to model the efficiency of vegetative filter strips for controlling surface runoff pollution from phosphate mining sand tailings. The numerical model VFSMOD-W is used to predict overland flow and sediment trapping within the filter and is linked to a simplified phosphorus (P) transport algorithm based on experimental data to predict total P (TP), particulate P (PP) and dissolved P (DP) fractions in the filter outflow. An advanced global inverse optimization technique is used for the model calibration process, and the uncertainty of the measured data is considered in goodness-of-fit indicators. The VFSMOD-W can predict hydrology and sediment transport well (Nash-Sutcliffe coefficient of efficiency ＞0.6) for calibration and validation events with peak outflow rate from VFS greater than 0.0004 m³/s. The good prediction in runoff and sediment resulted also in good predictions of PP and TP transport since apatite is a main component of sediment. A good prediction of DP was found by considering the rainfall impact on DP dissolved from apatite in surface soil. The uncertainty of measured data included in the goodness-of-fit indicators is a more realistic method to evaluate model performance and data sets. VFSMOD-W combined with the simplified P modeling approach successfully predicted runoff, sediment, and P transport in phosphate mining sand tailings, which provides management agencies a design tool for controlling runoff and P transport using vegetative filter strips.

Keywords: Sediment; Phosphorus; Mining sand tailings; Apatite; Vegetative filter strips; Uncertainty』

Introduction
Materials and methods
　Field experiments
　Simplified phosphorus modeling
　Inverse calibration methodology and model validation
　Calibration procedure
　Selected calibrated parameters
　Measurements of selected parameters
　Validation procedure
　Goodness-of-fit indicators
　Consideration of measured data uncertainty in the model evaluation
Results and discussion
　Optimized value of calibrated parameters
　Evaluation of model calibration and validation results
　　Hydrology component
　　Sediment component
　　Phosphorus prediction
Conclusions
Acknowledgments
Appendix
　Nash and Sutcliffe coefficient of efficiency (Ceff)
　Modified form of Ceff (Ceff_m)
　Root mean square error (RMSE)
References