Font Size: a A A

Use of thermal desorption/gas chromatography/mass spectrometry (TD/GC/MS), honey bees, and artificial neural networks (ANN) in assessing ecosystem contamination

Posted on:1999-05-12Degree:Ph.DType:Dissertation
University:University of MontanaCandidate:Alnasser, Ghassan HamkiFull Text:PDF
GTID:1461390014472501Subject:Chemistry
Abstract/Summary:PDF Full Text Request
An analytical method has been developed to assess environmental contaminants using honey bees as sample collectors and thermal desorption/gas chromatography/mass spectrometry analysis (TD/GC/MS) to analyze for volatile and semi-volatile organic compounds. Colonies were monitored by passing hive atmospheres through sorption traps (graphetized carbon and molecular sieves). Qualitative results were compiled from samples collected during preliminary field studies in 1995. Bee biomonitoring systems tracked both acute exposure events and chronic exposures during the summers of 1996 and 1997 at a variety of settings in Maryland and Montana. Five chlorinated environmental pollutants were detected in hive atmospheres as bioavailable—perchloroethylene, tetrachloromethane, trichloroethene, hexa-chloroethane, and 1,4-dichlorobenzene. Also detected were BTEX, benzaldehyde, acetophenone and naphthalene. No acute exposures were observed but underlying patterns of contamination emerged.; Challenging aspects to the technique were collecting samples in the adverse environment of a beehive and systematizing the complex array of compounds found in hive atmospheres. Chemicals historically associated with honey bees (pheromones, hive stores, and non-bee sources) have been summarized. Chemical also seen by the new TD/GC/MS method are noted. Finally, the list of compounds present in hive atmospheres have been expanded by adding contaminants emitted by non-bee and non-hive sources such as vehicles, farms, industries and households.; An Artificial Neural Network (ANN) methodology has been developed through which a complex chemical fingerprint can be derived for volatile and semi-volatile compounds in hive atmospheres. Once obtained, the fingerprint can be processed for training a back propagation neural network. Three ANN's were employed to: (1) distinguish individual hives within an instrumented hive cluster; (2) distinguish the field locations of contaminants; and (3) distinguish the suites of organic compounds found in a normal hive versus contaminants found in the external environment.; Two sets of experiments were undertaken in East Missoula to demonstrate a mechanism through which organic contaminants can be transported from the ecosystem to the hive. Perchloroethylene (a liquid phase contaminant) and naphthalene (a solid phase contaminant) were returned to the hive by foragers visiting a watering station.
Keywords/Search Tags:Honey bees, Hive, TD/GC/MS, Contaminants, Neural
PDF Full Text Request
Related items