Not all contamination originates from a single source. It could be a release that mixes with background sources or maybe multiple contaminant sources or plumes mixing together in the environment. Environmental forensics can identify the sources or the timing of the release but it is sometimes necessary to figure out which contaminants are coming from which source. This is referred to as source apportionment. When assigning which party is responsible for a proportion of the contamination, this is called source allocation.
Source apportionment is completed by statistical models called receptor models that can determine the source fingerprints (patterns) and can calculate how much of each source is in each sample in the investigation. There are many different types of models available that can do these calculations. Each model can have advantages and disadvantages. Some example models include: positive matrix factorization (PMF), non-negative matrix factorization (NMF), alternating least squares regression (ALS), Bayesian modelling and Unmix.
Chemistry Matters Inc. has developed a proprietary geospatial multi-model data visualization approach that provides multi-receptor model interpretations (runs multiple models at the same time) coupled with geospatial analysis to aid in the analysis and communication of source apportionment results. Chemistry Matters Inc. has applied these models to many different environmental contaminants such as polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs), polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs), volatile organic compounds (VOCs) and petroleum hydrocarbons (PHCs).
What does this mean for our clients and their source apportionment requirements?
Multiple assessments to determine optimal model conditions coupled with interactive, visual and easy to communicate interface. The data visualization tool can be used in the court room for expert witness testimony, or with the clients and stakeholders directly to provide real time analysis and interpretation of data. This provides unparalleled transparency in what is usually a black box technique. The visual and interactive tool allows for convincing conversations and testimony showing pictures and real time data analysis to help explain complex science and statistics.