Pooja Anand

1
A GIS Analysis of the Exposure of Soil and
Groundwater to Spills of Hazardous Materials
Transported by Rail
Presented at the 85th Annual Meeting of the
Transportation Research Board 01/25/2006

• Pooja Anand Christopher P.L. Barkan
• AAR Affiliated LaboratoryUniversity of Illinois
  at Urbana-Champaign
• Railroad Engineering Program

2
Introduction

• A large amount of hazardous material is
  transported in railroad tank cars (1.32 million
  shipments in 2004).
• Accident-caused-releases from tank cars are
  uncommon events, but,
• if released, the chemical can adversely impact
  the environment.
• The environmental impact of a hazardous material
  spill is a complex function of
• the properties of the spilled material, and
• the local environmental conditions in which it is
  spilled.

3
Background

• A series of costly environmental cleanup
  accidents in late 1990s involving
  non-halogenated hydrocarbons such as benzene,
  phenol etc.
• AAR initiated the current project to estimate
  environmental risk from transporting chemicals
  and consider if they were candidates for enhanced
  packaging.
• This presentation deals with the exposure
  assessment of environmental parameters to spills
  of hazardous materials.

• This is the first study that has categorized
  environmental features
• and quantified their exposure for risk
  assessment purposes,
• similar to population density categories for
  risk to humans.

4
Acknowledgements

• AAR Initiation and funding of the project
• David Schaeffer - Environmental toxicology and
  risk
• Charles Werth - Pollutant fate transport and
  remediation
• Barbara Minsker - Environmental risk management
  and optimal groundwater treatment management
• Jonathan Lorig, Robert Anderson, Mohd. Rapik
  Saat, Ze Ziong Chua, Adrienne Dumas and John
  Zeman

5
Environmental Parameters of Concern

• Rail line exposure to environmental features
  varies,consequently, so does environmental risk.

• Variables under study
• Type of Soil Various soil characteristics, such
  as grain size, reactivity, infiltration rate
  etc., all affect the fate and transport of a
  spilled chemical in soil
• Groundwater Depth Whether the chemical reaches
  groundwater also depends on the depth of the
  water table.

6
Objectives of Study

• Generate a manageable set of environmental
  scenarios from various possible combinations of
  values for soil type and groundwater depth.
• Develop a geographical probability distribution
  of soil type and groundwater depth, in the
  vicinity of rail network, in the 48 contiguous
  states.
• Develop probabilistic estimates of exposure of
  all environmental scenarios, to a potential
  hazardous material spill.

• Results from this study have been used in
  conjunction with
• consequence analysis of hazardous material
  spills on soils and
• groundwater, to assess environmental risk.

7
Comparison of Different Spill Scenarios

• Sandy Soil (loose)

Silty Soil (tight)
8
Soil Exposure Assessment

• Soil Database State Soil Geographic database
  (STATSGO), from the Natural Resources
  Conservation Service (NRCS).
• STATSGO is a GIS database organized on a
  state-by-state basis.
• Surface area of each state is divided into small
  segments of land called map units.
• Surface area of a map unit is further divided
  into sectors called components.
• Each map unit can have 1-21 components.
• These components are associated phases of soil
  series.
• Vertical section of a component is divided into
  layers.
• Each component can have 1-6 layers.

9
Soil Type Categorization

• To create a representative set of all nationwide
  soil types, soil type is categorized into
• Clay,
• Silt,
• Sand
• Categorization Scheme Followed (Hillel, 1982)

• Attributes of importance the attribute
  perml present at the
• layer level, the attribute comppct, which
  represents the area
• of a component as a percent of the map
  units area.

10
Example
Soil Type Silt
11
Soil Rail Overlay
Is rail line location independent of the soil
distribution?

• However, the distribution of soil types beneath
  rail lines is not significantly
• different from the overall distribution of
  soil types in the 48 contiguous states.

12
Groundwater Exposure Assessment

• Groundwater Database Real time groundwater
  monitoring database maintained by USGS.
• Attributes of importance
• Depth of the water table below the surface
• Latitude and longitude information
• Limitations with the groundwater database
• Presence of confined aquifers cannot be
  completely segregated.
• Seasonal variations in the depth No significant
  differences on paired t-test analysis.

13
Geographic Bias in Well Distribution

• Geographic bias in the well distribution
  Reducing geographic bias
• by using states area as the weighing factor
  and random selection of wells.

14
Groundwater Depth Categories Probability of
Exposure
Proportion of five groundwater depth categories,
in the 48 states
Most Common

• The choice of groundwater depths did not
  significantly affect the risk results.
• It was assumed that the lines traverse
  groundwater depth regions randomly and
  independently.
• Hence, distribution of groundwater depths under
  rail lines is assumed to be the same as their
  distribution nationwide.

15
Probability Distribution of the Soil-Groundwater
Matrix
16
Principal Conclusions

• Development of a representative set of values for
  soil type and groundwater depth occurring
  nationwide.
• Consequence analysis can be conducted for these
  chosen values of environmental features.
• Spill simulations result in different degrees of
  contamination for the various scenarios.
• Chemical does not reach groundwater for clay,
• reaches groundwater for silt only if the water
  table is less than 50 ft. deep
• reaches groundwater for sand even if the water
  table is as deep as 200 ft.

17
Principal Conclusions

• Soil type and groundwater depth are not
  independent of each other
• Clays are positively associated with groundwater
  depths gt 125 ft. and negatively
  associated with lt 25 ft.
• Silts are positively associated with groundwater
  depths 75-125 ft.
• Areal percentage of soil in the 48-states is not
  significantly different from the soil
  distribution under rail lines.
• This suggests that future nationwide
  environmental risk studies can simply use
  nationwide statistics
• since highway network is denser than rail, a
  truck transport analysis could use the same
  distribution
• However, route-specific studies will require data
  specific to that route, which can be obtained
  from our overlay map

18
Principal Conclusions

• Nationwide geographic probability distribution of
  soil type and groundwater depth
• Estimates of exposure probability for
  environmental scenarios.
• Exposure estimates used to develop environmental
  risk per car-mile metric.
• The underlying costing model for railroads does
  not account for this risk.
• This is a new metric that will help railroads
  better understand their risk, and prioritize
  safety improvements.
• The methodology developed in this paper can be
  applied to other studies of environmental
  transportation risk.

19
Thank You!
Questions