Results indicate that more representation in known-source libraries & better classification accuracy would be needed before field application. Thorough reliability assessment of classification is crucial before & during application of MST protocols.
Original working URL: https://pubs.acs.org/doi/abs/10.1021/es0354519