PDF files can be viewed with the Acrobat® Reader®
Once the objectives and scope of the study have been defined, including a general outline of relevant infrastructure and climate variables, practitioners can begin to assess the vulnerability of their assets. The form, level of effort, and detail of the vulnerability assessment will vary based on the identified objectives, goals, and other factors. The ultimate goal of the assessment is to determine how climate change may impact the transportation assets included in the study. The findings of the assessment can then be integrated into transportation decision-making processes and used to consider and prioritize measures to address the vulnerabilities.
The assessment step in the framework will use the preliminary information and data on the relevant climate variables and assets collected during the scoping of the study. However, data collection begun as a part of project scoping will continue during the vulnerability assessment. Data collection is an iterative, on-going process and is an integral part of the assessment.
Climate change and extreme weather vulnerability in the transportation context is a function of a transportation system's sensitivity to climate effects, exposure to climate effects, and adaptive capacity. Sensitivity refers to how the asset or system fares when exposed to an impact. Related to sensitivity are asset climate thresholds-the specific climate and environmental characteristics such as temperatures, water flow, or precipitation pattern that may warrant changes to the transportation asset design or materials. Section 3.1 discusses developing information on asset sensitivity to climate.
Exposure refers to whether the asset or system is located in an area experiencing direct impacts of climate change, such as temperature and precipitation changes, or indirect impacts, such as sea-level rise. To determine exposure, information on asset location, ideally including elevation information, is combined with information on the extent of the climate impacts of concern. Section 3.2 discusses asset data collection, and Section 3.3 discusses climate inputs.
Adaptive capacity refers to the systems' ability to adjust to cope with existing climate variability or future climate impacts. Are there alternate routes? How easily might impacted assets be modified to adjust to changing climate conditions? Section 3.5 discusses adaptive capacity.
Information for vulnerability assessments can come in many forms ranging from quantitative data driven GIS or spreadsheet computations to qualitative stakeholder engagement analyses based on local knowledge of vulnerabilities. In practice, a vulnerability assessment may incorporate elements of both qualitative and quantitative analyses, and there is no one-size-fits-all approach. Spatial and temporal scale, objectives, and access to resources all may dictate different techniques.
A necessary step in a vulnerability assessment is identifying the ways in which the transportation assets you are studying are sensitive to changes in the climate or to extreme weather events-that is, what are the kinds of impacts they can experience, and at what thresholds are these impacts felt? There are a number of different approaches that have been used to identify these sensitivities.
One source for sensitivity and threshold information are the standards or guidelines developed by State DOTs or other industry organizations, such as standards for designing, constructing, and maintaining infrastructure. Design standards and guidelines can be used to isolate specific climate stressors relevant to a particular asset. The relationships provided within design standards can also be used to provide quantitative indicators of an asset's sensitivity to a particular climate stressor. Transportation infrastructure is designed according to standards and guidelines that are based on detailed sets of empirical tests, asset-specific quantitative models, and other engineering analyses. Engineers use these documents to ensure that a design meets the functional specifications of a project within accepted limits and factors of safety. Design standards and guidance consider relevant climate variables-both typical climate characteristics in which the design will operate, as well as infrequent events such as violent storms, floods, and hurricanes, whose occurrence is estimated through design return periods for each hazard.
Related to design standards are design elements or relationships. For example a narrow, steep stream may be more sensitive to increased flow than a relatively flat stream with wide floodplains. The narrow stream may react more quickly and severely with rapidly rising water surface elevations and increased velocities, whereas a flat wide floodplain may be able to distribute and store the increased flow, effectively dampening the impacts. Relationships such as structure opening area to watershed size, shape and slope may be useful as proxies for broad screens of sensitivity to increasing precipitation. Incised streams, eroding stream banks, scour holes, presence of debris and signs of head cutting or presence of vertical grade control structures can also provide warnings signs of sensitivity to increased flow and instability. The ratio of culvert span or diameter to ponding depth at a roadway sag point may also provide clues as to how much extra capacity that culvert might have available to it. All these "proxy" relationships are useful for the initial sensitivity screening of assets in that they are usually easier to detect or measure without having to perform more complicated hydraulic computations.
Another approach is to draw upon an agency's experiences with system performance in the past, especially during extreme weather conditions such as high temperature, heavy precipitation, flooding, coastal storms, or prolonged lack of precipitation. Examining transportation system performance during historic weather events can aid in understanding the sensitivity of the transportation system to weather extremes and put the projected changes into context, teasing out the types of weather events and thresholds that caused impacts to transportation facilities and operations. This information can be used to gauge impacts associated with future climate conditions.
An assessment of past weather-related disruption and damage might consider:
Some of this information can be gleaned from existing studies; for example, if case studies exist on failures of particular bridges or roads tied to weather extremes, these can be useful to develop an understanding of the variables that caused failure. Where sufficient data are available (and in a readily accessible format), maintenance and engineering logs can be consulted to find out more specifically what types of weather events caused particular failures. This can range from the particular storm surge level that closed a bridge temporarily (for safety concerns) or that caused damage to the bridge; the levels of precipitation and stream flow at particular gages tied to culvert failure. Or, it could involve examining maintenance records, and looking for links between particular changes in the conditions of pavement and heat events.
Finally, district engineers and maintenance personnel are very often quite knowledgeable on the weaknesses or vulnerabilities of the current system. Eliciting their expert opinion on sensitivities of the current system can provide another way to determine how climate can impact the assets in the study. How often does a particular road flood, and what are the weather and environmental conditions that cause it? What temperature levels lead to enhanced pavement rutting and more rapid pavement replacement? Are there certain roads that have been flooded due to coastal storms, and if so, what kind of damage occurred as a result of particular storms? Do particular areas flood regularly at high tide with mild storms?
It is important to recognize, however, that typical historical climate conditions are unlikely to be representative of all future climate conditions. Although analysis of the past can yield useful "analogs" for certain types of weather events and the resulting impacts, the climate is changing and some future climate impacts may go beyond the range of impacts that have occurred in the recent past. Furthermore, it is unlikely that the trends of past decades will persist unchanged into the future; especially on longer timescales (greater than 30 to 40 years) simply extending past trend lines into the future may underestimate future changes. For example, for all parts of the United States, the rates of warming for the 21st century are expected to be greater than the rate of warming between 1900 and 2000. Similarly, sea-level rise rates have increased in recent decades, and they are expected to increase still more in coming decades.
Often the data needed for a vulnerability analysis will be found in multiple data sources that were designed for unrelated purposes and are not easily merged together. Not all data sets are of comparable quality, or in an easily accessible format (e.g., digital, geo-spatially referenced). For instance, facility elevation data may only exist in "as-built" diagrams stored on paper. Different pieces of information on the condition of the system facilities, for instance, might be found in a bridge database, a pavement asset management system, and paper copies of culvert inspection reports. Some data needed may not be collected regularly at all. The challenges of this data integration may be a factor in determining what data to collect for the analysis.
In inventorying State-owned transportation assets, the WSDOT quickly found that the information it wanted to use was found in multiple data sources and that the information varied widely in its level of detail and the extent of descriptive information included. The project team had to allocate additional time to convert the varied data into a format that could be used with other data in the WSDOT's GIS Workbench. Some data, such as information on repeat maintenance events that the team had hoped to use to identify current vulnerabilities, was only available in the form of individual paper maintenance reports and was not practical to use for this effort. Similarly, sea level rise (SLR) analysis proved to be very complex and difficult to determine. Without accurate elevation values for the roadway and infrastructure, it was difficult to determine whether an asset would be affected by the chosen sea level rise scenarios. Given time and resource constraints, GIS analysis was limited to proximity of an asset to the sea-level rise layers used in the database.
The New Jersey pilot gathered data from numerous sources and collected it into a geo-database for querying and analysis. Table 2 details the data included in the geo-database. The pilot team experienced some challenges with the data integration. For instance, their method for determining critical assets was largely dependent on traffic analysis zone (TAZ) information. As the two New Jersey study areas crossed the boundaries of three MPOs, with three separate travel demand models, they needed to create a unified TAZ structure for the analysis. This procedure was time consuming. Agencies with robust asset management systems may have central asset inventories to use as a strong starting point. For example, SEPTA, the transit provider for Philadelphia, had budgeted resources for asset inventorying as part of an FTA-funded climate adaptation pilot. However, because of a recent FTA-funded asset management grant, SEPTA had a robust asset inventory in place and was able to significantly reduce resources devoted to this task.
New Jersey Pilot Data Collection Matrix (for major assets).
Source NJTPA (2012)
|Roadways||NJDOT Congestion Management System (CMS)Network||NJDOT CMS data used for study area coverage and volume data. Primarily higher Functional Class highways, a few county roads and collectors included.|
|NJTPA Model Network||Does not cover the entire study area. Network is simplified highway stick network. Network density is higher than CMS Network and includes lower level functional classes compared to CMS.|
|Bridges||NJDOT Bridge Management Data Tables||2009 BMS data used to create linear bridge location shapefiles. NJDOT's bridge locations are point features and do not include length attribute data.|
|BTS NTAD||National Bridge Inventory from BTS is an additional source of bridge locations.|
|Tunnels||NJDOT Data Tables||Tunnels feature class created from NJDOT's data tables by linear referencing from the NJDOT highway centerline feature class.|
|Passenger Rail||NJ Transit||Feature classes used directly as received from NJ Transit.|
|Amtrak, BTS Data||BTS 2011 has Amtrak's network as a separate shapefile for download.|
|Freight Rail||BTS Data||Active Freight Rail data from Oak Ridge National Laboratory Network.|
|DVRPC||Data made available from DVRPC, regional in extent.|
|Traffic Analysis Zones||NJTPA, SJTPO, and DVRPC||A Unified TAZ structure was created by merging NJTPA, SJTPO, and DVRPC's TAZ shapefiles. NJTPA's 2010 socio-economic data was used from the three data sources.|
|Airports||NJDOT Data Tables, BTS||Runway lengths, airport polygons.|
|Wetlands||NJDEP||Wetlands files provided by NJDEP.|
|Evacuation Routes||NJGIN||NJDOT's Highway Safety Improvement Program Evacuation Route Shapefile.|
|Ports||BTS Data||Port and goods movement data provided by DVPRC.|
|Bus Routes||NJ Transit||Centerline bus routes shapefile from NJ Transit.|
|Signals||NJ Transit shapefiles.|
In addition to identifying infrastructure of interest, other key pieces of information that serve as inputs to the vulnerability assessment include understanding the sensitivity (and adaptive capacity) of the infrastructure to climate/weather events and, importantly, establishing the projected future climate scenarios this infrastructure will be (or is projected to be) exposed to. Development of climate information supports both the historic look at sensitivity and the assessment of future exposure:
The future climate is generally expressed in the form of scenario-based projections, rather than single predictions. These reflect the various uncertainties involved in climate modeling-the amount of greenhouse gases that will be emitted (which depends on the rate and nature of economic growth, technological change, and mitigation policies), climate variability, and model uncertainty. The emission scenarios are generally based on a standard set of scenarios developed for Intergovernmental Panel on Climate Change (IPCC) assessments. Each scenario represents a future path of global societal development, with assumptions regarding population growth, economic growth, and technological change, resulting in different rates of growth in greenhouse gas emissions. In order to reduce the uncertainty and bias resulting from using just one climate model, these scenarios have been used to create outputs from more than twenty major global climate models. As a result, climate projections span a range of values, for which probabilities cannot be easily assigned. This means that in order to better assess future risks, the vulnerability assessment will need to be conducted with a range of climate inputs, rather than a single climate scenario.
Projections of changes in temperature and precipitation can be complex. For general questions or for issues covering large areas, broad geographic region modeling of changes in temperature and precipitation may be sufficient. There are several reports that provide this kind of information and use it to assess impacts on particular regions of the United States and sectors within those regions. Regional climate effects information can also help inform broad planning type questions, such as those where broad trends in temperature and seasonal precipitation are of use. For other types of questions-especially when the fate of specific transportation assets or areas of similar scale are the focus of the analysis-- it is generally preferable to use more detailed information that has been processed to reflect local features/topography and conditions. Such information and procedures for developing it are becoming more available as progress is made on downscaling and regional modeling.
Approaches used in the New Jersey Pilot, the WSDOT Pilot, and Gulf Coast 2:
5-day rainfall during a given year (mm). The project team generated these inputs specific to the study area through analysis of climate models.
Using the statistical methods described above, the estimated future floodplains in the study areas were an average 8 percent, 40 percent, and 59 percent wider in 2050 than the current 1-in-100-year flood plain under the low, medium, and high emissions scenarios, respectively, and an average 17 percent, 80 percent and 178 percent wider than current floodplains in 2100. 81 miles of roadways, 1120 transit bus route miles, and 26 NJ Transit track miles lie in the projected 1-in-100 year floodplain for 2100 under the medium emissions scenario.
Sea-level rise and storm surge are potentially among the most destructive impacts climate change can bring to coastal transportation infrastructure. In order to characterize exposure, most studies of sea-level rise and storm surge risk rely on spatial analysis of projected inundation. The GIS can be used to map inundated areas by analyzing areas of land that fall below increased water levels under different scenarios of sea-level rise (also called the "bathtub model"), while more sophisticated approaches account for erosion, subsidence, and natural and man-made barriers that may protect certain areas from inundation, or even model the flows of water over the landscape.
As described by NOAA (2009), inundation mapping involves the following key steps:
Projections of global, or "eustatic," sea-level rise scenarios can be found in the results from the IPCC or other globally focused studies. Projections of local sea-level rise can be made more sophisticated by accounting for a range of additional factors that influence water levels at regional and local scales, such as:
Although a simple "bathtub" approach may indicate the relative risk among areas, it may not serve as a "prediction" for how the future landscape will appear, and thus may not capture all of the vulnerabilities faced in a particular locale. Typically, such limitations are not critical for identifying areas at risk at broad spatial scales (e.g., regionally or nationally) or communicating these risks to the general public. However, they may be important to keep in mind when such maps are used for local land use planning. The utility and accuracy of a sea-level rise and storm surge assessments depend in part on the resolution of the underlying elevation data. One standard source of elevation data, the USGS National Elevation Dataset, supplies elevation data with a horizontal resolution of 30m and 10m and vertical resolution of around +/-2.4m. However, global projected sea-level rise of up to 2m by the end of the 21st century falls within this +/-2.4m resolution. As such, maps based on the NED will generally not provide accurate predictions of exposure of specific assets. In order to obtain more useful elevation information, local assessments will likely need to rely on digital elevation models derived from high resolution LiDAR (Light Detection and Ranging) data. These data are not available in all locations, and require additional processing to use, including adjustments to the vertical datum to ensure consistency across datasets. 
These five categories are ordered above from those that provide the most potential protection from inundation to the transportation infrastructure behind them to those that provide the least potential for inhibiting inland inundation. The project team grouped the individual shoreline assets into larger systems of protection that protected a certain area. The team examined the sea level rise scenarios under three conditions: high tide, 100-year extreme water level from storms, and 100-year extreme water level coupled with wind waves. The team then analyzed the depth of water overtopping the asset and what percent of the length of the shoreline asset system is overtopped.
Regional Climate Change Effects: Useful Information for Transportation Agencies, FHWA 2010. This document provides basic information on projected future climate change effects over the near term, mid-century and end-of-century by U.S. region.
Global Climate Change Impacts in the United States, US Global Change Research Program 2009. This report summarizes the impacts of climate change on the United States, looking at different regions and economic sectors.
The Use of Climate Information in Vulnerability Assessments, FHWA 2011. This memorandum focuses on the use of climate information when performing a vulnerability assessment. The memorandum includes discussion of using historical climate information and includes information on potential data sources.
Climate Projections FAQ, U.S. Forest Service Rocky Mountain Research Station. This guidebook focuses on understanding and interpreting downscaled climate projections.
Websites for historical weather and climate data available from the NOAA National Climatic Data Center include:
Data sources that have projections of future climate change from many different models for various emissions scenarios include:
Many university climate centers have also developed their own climate projection datasets.
There are multiple ways to combine the climate and asset information to identify potential vulnerabilities, ranging from a desk review GIS or spreadsheet analysis to a stakeholder elicitation based on local knowledge of current vulnerabilities, or a combination of both.
With the desk review approach, data on assets and projected climate are combined via a geospatial map or other analytical tool to identify potential vulnerabilities. Projected climate change impacts are represented in a GIS format along with information on the relevant assets (such as elevation, geographic location, and existing flood protection) to determine potentially vulnerable areas.
With the stakeholder input approach, potential vulnerabilities are identified by stakeholders with intimate knowledge of study area facilities. Through workshops and/or interviews, local transportation practitioners draw from their knowledge and experience to consider how changes in climate may impact transportation facilities within their purview.
Figure 3: Roadways Inundated by 1 meter of sea level rise, coastal study area.
Source: NJTPA (2011), Cambridge Systematics.
Table 3: Roadway miles inundated by 1 meter of sea level rise, coastal study area.
Source NJTPA (2011), Cambridge Systematics
|Roadway Type||Criticality Tier||Increase from 2050|
|NJ Transit Bus Routes||657.88||623.18|
Table 4: The WSDOT Workshop Impact Rating Scale.
Source: WSDOT (2011), Adapted from Oregon Transportation Research and Education Consortium - Risk Assessment Presentation
Results in total loss or ruin of asset. Asset may be available for limited use after at least 60 days and would require major repair or rebuild over an extended period of time."Complete and/or catastrophic failure" typically involves:
May sever some utilities. May damage drainage conveyance or storage systems.
Temporary Operational Failure
Results in minor damage and/or disruption to asset. Asset would be available with either full or limited use within 60 days. "Temporary operational failure" typically involves:
Results in little or negligible impact to asset. Asset would be available with full use within 10 days and has immediate limited use still available.
"Reduced capacity" typically involves:
Adaptive capacity refers to the ability of a system to adjust to climate change (including climate variability and extremes) to moderate potential damages, to take advantage of opportunities, or to cope with the consequences.
One way in which a transportation system can have greater adaptive capacity is if it has redundant routes or modes. For instance, if a particular roadway segment is impassable due to flooding, the availability of parallel routes or alternative modes can continue to enable travel between destinations. The New Jersey pilot considered this aspect of adaptive capacity during its criticality screen. The pilot ranked assets as more highly critical if there were no alternative routes available.
Another relevant criterion is how easily and quickly service can be restored to a segment or asset following a climate-related disruption. The Gulf Coast 2 Study, for instance, is using two measurements to evaluate how quickly service could be restored to a given segment: replacement costs and time duration of disruptions. Replacement costs can provide a rough proxy for the ease in which assets could be repaired or replaced; resources are more easily mobilized for lower cost repairs, and replacement costs indicate overall complexity, size, and expense of the asset itself. Length of time for the disruption to clear (for instance, for debris to be cleared from a rail line following a storm) is an indicator of how well the system can deal with the climate impact. The availability of staffing and resources for preventive measures can also increase adaptive capacity. For instance, SEPTA, the public transportation provider for Philadelphia, provides the example of deploying maintenance crews prior to an expected storm to trim trees along rail lines in order to minimize potential for wind-blown debris from trees to block the tracks.
Key considerations for evaluating adaptive capacity include:
Ultimately, the vulnerability assessment will aid transportation decision-makers in prioritizing actions and determining how to improve the adaptive capacity of the system.
In addition to determining whether an asset may be impacted by climate change, it can also be important to understand just how vulnerable the assets are. Information on risk and the timeframe of the risk are useful for determining whether climate change should be considered in project development, prioritizing actions, and weighing adaptation options. However, assigning risk associated with climate change vulnerability is imprecise. It is not possible to apply a certain likelihood to a climate scenario. And further, climate models do not all agree. This section provides a brief overview of assessing climate risk to transportation assets, discusses some of the challenges to assigning a specific risk, and offers examples of some ways transportation agencies have incorporated likelihood and risk into their vulnerability assessments.
A risk assessment integrates the severity or consequence of an impact with the probability or likelihood that an asset will experience a particular impact. To determine consequence, transportation agencies may wish to consider the level of use of an asset, the degree of redundancy in the system, or the value of an asset (in terms of cost of replacement, economic loss, environmental impacts, cultural value, or loss of life). The consequence of projected impacts may have been a component of the vulnerability assessment discussed in Section 3.4.
Determining the probability of occurrence, or likelihood, of future climate impacts, can be problematic. As mentioned above, it is not possible to apply a particular likelihood to a climate scenario, and climate models do not all agree. With the absence of information on the likelihood of specific climate impacts, climate impact studies will often use separate risk matrices for a range of possibility with a "low emissions" scenario and a "high emissions" scenario, or a range of emissions scenarios, as discussed in Section 3.3. These studies will also use several climate models, often averaging results of those that most accurately model past regional climate, and characterizing impacts that occur under all the models "highly likely." The resultant impact projections will be split into high- to low-likelihood groupings, with the caveat that the climate may be more or less sensitive than we think. Depending on objectives and study scope, the study team may want to choose projections from one scenario or a range of scenarios. To the extent that information is available, impacts will be split into high- to low-likelihood groupings.
In general, there is much more certainty regarding the direction of change than there is regarding the magnitude of change, or the length of time it will take to reach a change. For this reason, an agency may want to consider impacts that are fairly certain to happen at some point this century, even if there is some uncertainty regarding whether the impacts will occur in 2050 or 2080. With information on consequence and likelihood, agencies can categorize assets into groups:
The integrated risk is often represented by a two-dimensional matrix that classifies risks into three categories (low, moderate, high) based on the combined effects of their likelihood and consequence. An example matrix risk rating matrix used by the San Francisco Pilot is provided in Figure 4.
Figure 4: Risk Rating Matrix.
Source: Adapting to Rising Tides: Transportation Vulnerability and Risk Management Assessment Pilot Project, November 2011
Sea-Level Rise Risk Assessment for Honolulu International Airport, Oahu MPO (2011).
Risk Level in Year 2050
TheBus 811 Middle Street
Low Vulnerability, Low Structural Impact
HDOT Highways Division
Oahu District Baseyard
727 Kakoi Street
Low-Moderate Vulnerability, Low Structural Impact
Honolulu International Airport and Access
Low Vulnerability, Low Structural Impact
Risk Level in Year 2100
811 Middle Street
Low-Moderate Vulnerability, Low-Moderate Structural Impact
HDOT Highways Division
Oahu District Baseyard
727 Kakoi Street
High Vulnerability, High Structural Impact
Honolulu International Airport and Access
Low-Moderate Vulnerability, Low Structural Impact
Literature Review: Climate Change Vulnerability Assessment, Risk Assessment, and Adaptation Approaches, FHWA, July 2009. This document details how vulnerability, risk, and adaptation assessments have been or could be used to integrate climate change impacts into transportation decisions and ultimately increase the adaptive capacity of the highway system.