Full document / COSOC-W-86-002 - the National Sea Grant Library

Estuarine andCoastal Management - Tools ofthe Abstract only
Trade. Proceedings ofthe Tenth National Conference
of The CoastalSociety. October 12-15.1986. New
Orleans, LA. Copyright by TheCoastalSociety
1987.
IMPROVED QUANTITATIVE ASSESSMENTS
OF ENVIRONMENTAL HAZARDS
Donald T. Resio
Offshore & CoastalTechnologies.Inc.
911 Clay Street
Vicksburg. MS 39180
One of the majorproblemsfacing decision-makers todayrelates to the lack of quantitative
guidancethat they receive fromscientistsandengineers. Measurements from varioussourcesare
frequently piecemeal anddo not provideanoverviewof theenvironmental concerns. Sophisticated
models aredeveloped and then appliedto only a small rangeof selectedconditions. Conflicting
hypotheses are oftenput forward based onthese sparce data andinterpretations of environmental
interactions. This often leaves the decision-making process primarilydictated by bureaucratic
considerations.
A statistical methodology for obtaining a better quantitative characterization of
environmental hazards is used to obtain improved quantitative estimates of risk. This statistical
technique is basedon aconvolution of spatial fields andcanbe usedto obtainimprovedestimates
of univariate and multivariate distributions of parameters. In turn, these distributions can be
integrated toobtain risksof threshold exceedance, durations of exceedance conditions, andrange
ofconditions leading up to the exceedance.
In overview of the FEMA-defined flood zones applied in southern Louisiana it was
shown that the older methods used to define these levels appearto be quite over-conservative.
Second, ananalysis of theuseof thistechnique to evaluate thehazards dueto encountering spills
(oilor toxicmaterials) hasshownthatthese riskscanbequantitatively evaluated in a fashion that
can provideuseful guidanceto decision makers.
One of the majoradvantages of this new methodology lies in the increased abilityto
predict environmental extremes. Inapplications tooffshore wave-related hazards intheGulfof
Mexico, it has been shown to provide approximately 10 times more information than a
conventional site-by-site analysis. Another advantage is inherent in theability of thistechnique to
estimatenonlinear multivariate relationships amongseveral environmental parameters. Standard
multivariate statistical methods (such as multipleregression, canonical correlation analysis,
principal componentanalysis, and factor analysis) areallbasedon linearalgebra andaredifficult to
apply in situationswhere interrelations arehighly nonlinear(such as the relationships between a
direction quantity and a speed quantity).
655