Page 401 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

376 Receptor Risk Evaluation
fatalities associated with leaks of less than 50 barrels. Since the spill at this location would have no significant impacts on
estimated leak counts included leaks of less than 50 barrels, the drinking water quality. Therefore, under at least 50 percent of
estimated leak rates were reduced by the ratio of reportable to the flow conditions in the river, there would be no impact. The
total leaks. Approximately 56% of the total leaks are below 50 50 percent number is also conservative with respect to the worst
barrels in size. Thus, the leak rates were multiplied by 0.44 to case crossings at Flat Creek and the Pedernales. The 50 percent
obtain the estimated fatality and injury rates. For example, the estimate is also thought to be very conservative in light of other
fatality rate for Case 3 was calculated in the following manner: areas which are currently designated hypersensitive, but for
which more recent modeling suggests that a sensitive/Tier 2
Fatalities = 0.00459 x 69.7 x 0.44 = 0.14 fatalities over the project life. designation would be more appropriate.
Surface water drinking supplies in Tier 2 areas are less vul-
The fatality and injury rates for Case 4 were calculated in a sim- nerable than those in Tier 3 areas. For surface water contami-
ilar manner. The average leak rate for Case 4 was determined as nation in a Tier 2 area to impact public drinking water supply,
described elsewhere in this appendix. very improbable stream flow, soil, and water use (such as
The segment-specific fatality and injury frequencies shown drought stage water needs) would need to occur simultane-
in Table 4 were calculated in much the same manner as those ously. These conditions exist at a lower frequency than is rep-
given in Table 3. The frequencies for the 2500-ft segments were resented by IO percent probability number assigned for Tier 2
produced by reducing the frequencies for the entire pipeline by areas.
the ratio of 2500 ft to 700 miles. For example, the fatality fre- For ground water, a higher probability (relative to the surface
quency for Case 1 was calculated as follows: water case) is assigned to Tier 2 sensitive and Tier 3 areas, in
order to account for a number of factors. These include the
Fatalityfrequency=(0.00459x35~2500)/(700x5280)= 109~ 10" uncertainty about localized ground water flows at every point
along the pipeline, the potential presence of private drinking
water wells which may be impacted, the distance to karst
Drinking water contamination recharge features, the extent of time for which contaminants
could remain in ground water at significant concentrations, and
Contamination of public drinking water resources may occur the variations in ground water flux due to aquifer level and rain-
either from contamination of sensitive ground water or surface fall conditions. However, a major spill in a hypersensitive area
water supplies. does not guarantee impacts to drinking water quality within the
Tier 2 and Tier 3 areas for potential drinlang water contami- associated aquifer. Factors such as uptake by the soil, runoff,
nation were defined by the sensitive and hypersensitive desig- and volatilization from the surface can reduce much of the vol-
nations in Chapter 7 [of the EA-Ref [86]]. The mileage ofTier ume of the product which reaches the aquifer.
2 and Tier 3 areas for ground water and surface water were Additional modeling assumes a case where MTBE is
therefore derived directly fromTables 7.1 and 7.2 [not included removed from the gasoline, and that benzene is the primary
in this book, but Chapter 7 of this book discusses the tier desig- constituent of concern. This modeling indicates that the poten-
nations]. Note that sensitive and hypersensitive areas for tial for significant impacts to drinking water use when MTBE is
ground water and surface water are not mutually exclusive, and removed is far less than one-halfthe potential for spills contain-
therefore some overestimation of overall probability will result. ing MTBE. In order to be conservative, the impact was set at
The assignment of sensitive and hypersensitive areas is one-half of the potential with MTBE. [MTBE refers to a gaso-
based on hydrological and hydrogeological evaluation of the line additive that was being contemplated. This additive makes
characteristics of surface water streams and aquifers which the gasoline more environmentally persistent and hence,
could be impacted by the pipeline. The designation of sensitive increases the chronic product hazard.]
was intended to indicate those areas where it is deemed possible
for damages to occur to a drinking water supply resulting from Edwards aquifer contamination
a release. The designation of an area as hypersensitive suggests
that there is a higher probability of an impact within these areas. The three miles of pipeline crossing hypersensitive recharge
A release to either a sensitive or hypersensitive area does not formations in the Edwards Aquifer/Balcones Fault Zone were
guarantee an impact. There are various location- and time- concluded to represent worst case ground water impacts.
specific determining factors, such as distance to surface water As explained generally in LMC 33, and specifically in the
or karst feature, flow rate in a receiving stream, saturation of Phase I1 BA [biological assessment], LPP will investigate and
soils, temperature, and wind speed, and nature of the event seal off any recharge features within the pipeline ROW while
causing the release. laying new pipe. This should reduce pathways for product
Based on an overview of these factors, the probability of spilled to impact the aquifer by percolating through surface
contaminating drinking water supplies as a result of a major soils to a subsurface recharge feature or flowing overland to a
release along the pipeline were set conservatively at the rates recharge feature.
shown in this report. Fifty percent potential contamination for It is assumed that soils will readily absorb between 500 and
surface water/drinking water contamination was set after 1,500 bbl of a spill: the lower level (500 bbl) is set as the mini-
reviewing modeling results of the most sensitive crossing with mum spill of consequence. The probability of any spill greater
respect to significant drinking water contamination along the than 500 bbl impacting ground water is set at 75 percent, to
pipelinethe crossing of the Pedernales River upstream from reflect the large number of recharge features in the zone. It is
Lake Travis. Modeling exercises conducted to date show that assumed that any contamination of the aquifer will in turn
during mean flow conditions on the Pedernales, a worst case impact drinking water supplies in Sunset Valley.