HR: 17:15h
AN: H32I-06 [PDF]
TI: Analysis Of Multispectral Imagery And Modeling Contaminant Transport
AU: Irvine, J M
EM: john.m.irvine@saic.com
AF: SAIC, 20 Burlington Mall Road
Suite 130, Burlington, MA 01803 United States
AU: Becker, N M
EM: nmb@lanl.gov
AF: Los Alamos National Laboratory, Mailstop D436, Los Alamos, NM 87545 United States
AU: * Brumby, S
EM: brumby@lanl.gov
AF: Los Alamos National Laboratory, Mailstop D436, Los Alamos, NM 87545 United States
AU: David, N A
EM: ndavid@lanl.gov
AF: Los Alamos National Laboratory, Mailstop D436, Los Alamos, NM 87545 United States
AB:
A significant concern in the monitoring of hazardous waste is the potential for contaminants to migrate into locations where
their presence poses greater environmental risks. The transport modeling performed in this study demonstrates the joint use
of remotely sensed multispectral imagery and mathematical modeling to assess the surface migration of contaminants. KINEROS,
an event-driven model of surface runoff and sediment transport, was used to assess uranium transport for various rain events.
While our specific application was uranium transport, the methods apply to surface transport of any substance of concern.
The model inputs include parameters related to the size and slope of watershed components, vegetation, and soil conditions.
One distinct set of model inputs was derived from remotely sensed imagery data and another from site-specific knowledge. To
derive the parameters of the KINEROS model from remotely sensed data, classification analysis was performed on IKONOS
four-band multispectral imagery of the watershed. A system known as GENIE, developed by Los Alamos National Laboratory,
employs genetics algorithms to evolve classifiers based on small, user-selected training samples. The classification
analysis derived by employing GENIE provided insight into the correct KINEROS parameters for various sub-elements of the
watershed. The model results offer valuable information about portions of the watershed that contributed the most to
contaminant transport. These methods are applicable to numerous sites where possible transport of waste materials or other
hazardous substances poses an environmental risk. Consequently, the approach presented here is relevant to homeland security
and emergency response scenarios, as well as long-term environmental monitoring applications. Because the approach rests on
the analysis of remote sensing data, the techniques can be used to monitor a range of sites and can reduce costs of data
collection for model calibration.
DE: 1815 Erosion and sedimentation
DE: 1860 Runoff and streamflow
DE: 1871 Surface water quality
DE: 1894 Instruments and techniques
SC: Hydrology [H]
MN: 2003 Fall Meeting