Sensitivity is a component of vulnerability that is difficult to define and assess. Exposure and adaptive capacity, the two other components of vulnerability, can be generalized more easily with indicators across a large region. For example, analyses often use a combination of models to estimate exposure over a wide region, such as the Gulf Coast. Similarly, climate change studies often use indicators of adaptive capacity (such as wealth, technology, and education indicators)1 in order to estimate overall adaptive capacity. However, sensitivity is locally-defined and depends on contexts and relationships that are difficult to generalize.
The sensitivity of an asset is the change in the condition of the asset when it is exposed to a change in a climate variable (e.g., higher temperatures). Sensitive assets will experience a large degree of impact if the climate varies even a small amount. At the opposite extreme, assets that are not particularly sensitive could withstand high levels of climate variation before exhibiting any response.
Distinguishing between exposure and sensitivity
Exposure indicates whether or not, and to what extent, a given asset experiences a climate variation.
Sensitivity refers to the degree of response exhibited by the asset if it is exposed. However, the two concepts are interrelated since if an asset is continuously exposed over time, its sensitivity to the climate impact may increase. For example, pavement that is continually exposed to precipitation will weather over time and become more sensitive.
Quantitatively, sensitivity is the slope of the functional relationship between the condition of an asset and a particular climate variable. For example, Figure 1 hypothetically describes the relationship between coastal bridges (the assets) and storm surge (the climate variable). The slope of this curve represents the sensitivity of bridges to storm surge. Figure 2 provides a mathematical description of this hypothetical relationship.
In the case illustrated in Figure 1 , the sensitivity (or slope of the line) does not have a single value, but instead varies as a function of increased storm surge elevation. In other words, the damage associated with storm surge escalates dramatically once the surge elevation reaches the height of the low-chord bridge elevation.2 This important threshold point (indicated by the red dot in Figure 1) could occur when bridges are not designed to withstand direct wave action on the bridge superstructure.3 Bridges that are more highly prone to damage at this threshold (and thus become highly sensitive) are graphically represented by the steeply sloping tail of the sensitivity curve. Bridges that are better able to withstand surge that exceeds the low chord elevation are graphically represented by the more gradually sloping sensitivity curve. The example function in Figure 1 is only a hypothetical possibility of how damage to a bridge might vary as a function of storm surge elevation.
In this particular example the impacts are deleterious. However, it is also possible that changes in climate variables could benefit certain transportation assets. In other words, sensitivity, or the slope of the damage function, can be positive or negative. For example, roadways may experience fewer frost heaves and potholes if increases in seasonal temperatures reduce the frequency of freeze-thaw cycles.
Figure 1 : Possible Damage Function Representing the Relationship between Coastal Bridges and Storm Surge
The sensitivity is the slope of the line (or derivative of the function) measured at a specific point. The red dot indicates the threshold, or the place where damage to the bridge begins to escalate.
Figure 2 : Equation Estimating the Sensitivity of Bridges to Storm Surge Elevation between Two Levels of Exposure
In some types of environmental impact analyses, it makes sense to articulate the concept of sensitivity through a quantitative damage function, such as in the case of the dose-response functions that are a cornerstone of air quality and groundwater quality regulation. However, in general, the information available at the present time in the transportation sector does not lend itself to representation in the form of continuous damage functions, as shown in Figure 1, that relate climate stressors to transportation impacts.4 There are a few reasons for this:
To overcome the obstacles associated with developing quantitative damage functions, an alternate approach to assessing sensitivity was taken. This approach qualitatively identified the nature of relationships between asset classes and stressors, indicated thresholds (including quantitative thresholds, where possible), and provided indicators that can be used to quickly assess the sensitivity of a particular asset. The Sensitivity Matrix framework presents the relationship between four climate variables and six major transportation modes: bridges, roads and highways, railroads, airports and heliports, oil and gas pipelines, and marine ports, terminals and waterways. In addition to these six transportation modes, electrical power systems were included in this analysis because the reliance of transportation modes on electrical systems can be a determining factor for sensitivity.
The Sensitivity Matrix is organized according to the four climate variable groupings that were identified to be most important in the region: waves and storm surge, wind, temperature, and precipitation. The matrix contains two tables for each variable: a main summary table and a second table containing any additional detail or notes. The following four columns make up each variable table: Important Impact-Asset Relationships, Threshold, Mobile-specific Detail, and Potential Indicators of Sensitivity.
Table 1 contains an illustrative excerpt of the Sensitivity Matrix describing the sensitivity of bridges, roads, and highways to wind. Starting from the left in the Matrix, the Important Impact-Asset Relationship column qualitatively describes stressors between the climate variable and the sub-mode. The Threshold column includes any specific information about the exposure level at which damage to the sub-mode may begin increasing. Historical context relevant to Mobile is placed in the Mobile-Specific Detail column, and the Potential Indicators of Sensitivity column contains a list of indicators that have been associated with increased sensitivity to that climate stressor in the past.
In the illustrative excerpt below, the Matrix reveals that wind can damage bridges directly or indirectly by increasing wave action or damaging operator houses. The ASCE 7-05 design standard recommends using a wind design speed of 130-150 for the majority of coastal Mobile County, indicating that damage to bridges may begin to increase at around that threshold.5 ASCE 7-10 recommends a three-second gust wind speed of 140-175 mph, depending on the structure's risk category (ATC 2012). However, the threshold column also points out that service on bridges is likely to stop at around 56 mph, potentially causing degradation of service due to traffic delays and freight disruptions. The Mobile-Specific Detail column notes that in the past, damage to bridges has occurred from debris-related collisions, rather than the direct impacts of wind on infrastructure. Finally, the indicators column states that the presence of operator houses and the age of the bridge may indicate increased sensitivity to wind in certain contexts.
Table 1 : Section of the Sensitivity Matrix Presenting the Sensitivity of Bridges, Roads, and Highways to Wind
|Asset Categories||Wind - Summary|
|Mode||Sub-Mode||Important Impact-Asset Relationships||Threshold||Mobile-Specific Detail||Potential Indicators of Sensitivity|
|Bridges||Bridge (Superstructure)||Winds stress bridges with additional horizontal loading. Bridges are designed with a certain amount of wind loading accounted for in the structure design.  Strong winds create more powerful waves which can stress the bridge superstructure and substructure. ||AASHTO LRFD bridge design specifications are based on a base design wind velocity of 100 mph, although the base design wind velocity investigated for tall structures to account for local wind speed conditions [79, 80]. ASCE 7-05 recommends using a wind design speed between 130-150 mph for majority of Mobile County . ASCE 7-10 recommends a three-second gust wind speed of 140-175 mph, depending on the structure's risk category . Higher wind speeds are correlated with larger waves and, to a lesser extent, more frequent wave periods. The effect is continuous, so the threshold at which damage occurs depends on the elevation of the bridge deck and other conditions. ||During Hurricane Katrina, strong winds helped break a 13,000 ton semi-submersible drilling platform free from its dry-dock moorings. The drilling platform collided into the Cochrane-Africatown bridge. ||- Presence of operator houses with electrical and mechanical equipment
|Bridge (Substructure)||Strong winds create high flow velocities (including high wave impact energy), which can lead to bridge scour. ||Higher wind speeds are correlated with higher kinetic energy of the water. The effect is continuous, so the threshold at which damage occurs depends on factors such as the substrate type, and the depth and geometry of footings. |
|Operator Houses (movable bridges) and electrical parts||Wind damage to operator houses causes damage to the electrical and mechanical equipment of the bridge, and may exacerbate rain damage. ||Higher wind speeds are correlated with larger waves and, to a lesser extent, more frequent wave periods. The effect is continuous, so the threshold at which damage occurs depends on the elevation of the bridge deck and other conditions. |
|Roads and Highways||Paved road surface||Wind does not directly damage the physical structure of the road, but can severely disrupt road traffic and other service activities.||No documented relationship.||No documented relationship.||- Structure elevation|
|Road substructure (gravel base, substructure)||No documented relationship.|
|Unpaved roads||Wind can stir up dust from unpaved roads, causing eye irritation to residents and other health issues. ||Unknown|
|Stormwater drainage (culverts, side drains, etc)||Wind damages trees, buildings, and other structures. Debris from this destruction can clog the stormwater drainage system, resulting in flooding impacts to the surrounding area. ||Non linear|
|Highway, road and street signs and traffic lights||Winds can blow over highway, street, and road signs.||The Alabama AASHTO wind design speed is 140 mph. If street signs (such as stop signs) are not buried deeply in strong soils they may not be in compliance with design standard and may fail at (much) lower wind speeds. ||During Hurricanes Katrina and Wilma, a large proportion of street signs failed in Miami-Dade county and the vicinity. The majority of these street signs failed at their foundations - mostly by leaning more than 15 degrees sideways or falling over completely. |
|Highway and road traffic and service||High winds cause safety risks and travel delays, a loss of visibility, impaired mobility, loss of communications and power, freight/cargo damage risk, increased risk of collisions/spills of hazardous cargo, and transport schedule delays. ||Winds become dangerous to road maintenance, truck operations, and other road users at around 39 mph and are very dangerous at 74 mph.  AASHTO LRFD wind load provisions assume that no traffic will be present on a bridge when wind speed exceeds 56 mph. |
 Chen and Duan (1999)
 Douglass and Krolak (2008)
 Ghosn et al. (2003)
 Lin and Fabregas (2007)
 Hitchcock et al. (2008)
 Mitchell (2010)
 OFCM (2002)
 Padgett et al. (2009)
Although the Sensitivity Matrix does not quantitatively evaluate the sensitivity of an asset exposed to a climate stressor, this approach has advantages compared to damage functions. The matrix does not require extensive damage information across a range of climate stressors to identify key impact-asset relationships. It can accommodate the full range of available data from available literature, design standards and guidelines, analogues from historical data and case studies, and expert consultations. It is accessible to planners and policy makers, and can be used as a screen to identify important relationships between climate stressors and critical assets. Finally, it is designed to be transferable to other regions-both in terms of the approach itself, and the bulk of the information summarized within the Sensitivity Matrix.
The Sensitivity Matrix is limited to describing direct sensitivities of transportation infrastructure to climate change. There are, however, indirect ways in which climate change can affect infrastructure. For example, marshes may not be able to keep pace with long-term sea-level rise, depending on the marshes' vertical accretion rate and the rate of sea-level rise. Elimination of marshes would tend to amplify the storm surge impacts on transportation because the marshes would no longer help to buffer a storm's energy. Similarly, barrier islands may be affected by multiple factors associated with climate change and their storm buffering capacity may also be impaired. Any impact and/or risk analysis performed using the Sensitivity Matrix should consider such whole-system interactions.
The Sensitivity Matrix is an important step toward a more comprehensive understanding of the relationships between climate stressors and transportation assets. Due to the matrix's reliance on qualitative data, however, it can be difficult to concisely summarize the information for planners and decision makers. To make the information in the Sensitivity Matrix more accessible, a Sensitivity Screen was developed from the relationships and thresholds identified in the Sensitivity Matrix to provide a tool that can be used to quickly assess whether transportation assets may be sensitive to certain climate stressors.
1 Although these adaptive capacity indicators are not directly applicable to transportation assets, indicators such as capital turnover cycles, maintenance periods, monitoring and reporting, redundancy, excess capacity, and the modularity of assets could be used to assess adaptive capacity in transportation.
2 The bridge's low chord elevation refers to the elevation of the portion of the bridge closest to the water (often the bottom of the girders).
3 During Hurricanes Rita and Katrina in 2004 and 2005, many bridges along the Gulf Coast experienced damage because their superstructures were directly exposed to wave forces. This exposure to wave loading was particularly damaging because in most cases, the bridges had not been designed to withstand that type of exposure. In order to improve design standards, the Federal Highway Administration and ten states sponsored the development of new guidance for bridges vulnerable to coastal storms. An AASHTO/FHWA Wave Task Force was established to oversee the development of the guidance. The resulting guidance includes new methods for calculating wave forces on superstructures that had not been presented in previous design provisions. The new guidance specifies that, "whenever practical, the vertical clearance of highway bridges should be sufficient to provide at least 1 foot of clearance over the 100-year design wave crest elevation, which includes the design storm water elevation." The guidance also notes that where it is not possible to provide this level of vertical clearance, strategies such as venting cells that could entrap air, using large holes in concrete diaphragm, constructing continuous superstructures, and using solid or voided slab bridges could all help increase the resilience of the bridge to wave forces acting on the superstructure (AASHTO, 2008).
4 Exceptions to this statement exist. For example, see Powell and Reinhold (2007) for an example of a damage function that relates residential wind damage (claim to insured value ratio) to 10-m open terrain wind speed. See Padgett et al. (2009) for an analysis of bridge damage at different levels of storm surge during Hurricane Katrina.
5 AASHTO Load and Resistance Factor Design (LRFD) Bridge Specifications use a base design wind velocity of 100 mph, although they require that the base wind design velocity be investigated for tall structures to account for variations in local conditions (AASHTO 2012). The LRFD specifications were initially based on ASCE 7-88 wind load provisions, which have since been updated to reflect 3-second-gust wind speeds. NCHRP has indicated that there is a need to update the AASHTO LRFD wind load provisions to provide consistent reliability across different regions and locations (NCHRP 2012).