<< Previous | Contents | Next >> |

## Evaluation of Highway Performance Measures for a Multi-Study Corridor - A Pilot Study

### 4.0 Analysis and Evaluation of Alternative Condition Indicators

This section contains the analysis of condition measure data, both bridge and pavement, received from the participating states. The analysis begins with background on the traditional measures used to evaluate these assets before presenting a statistical breakdown of the specific data captured for this project. Condition is one measure of performance. Other performance measures may relate to functional adequacy or overall importance of a given asset.

Following the analysis, CS presents information obtained from the participating states on the processes used to capture bridge and pavement condition information, including methods employed to validate the quality of the data. This section also includes feedback on the state-specific criteria used to categorize asset condition (e.g., as "good," "fair," or "poor").

Finally, CS presents conclusions on the adequacy of the existing measures for bridges and pavement, highlighting some of the results of the data analysis. CS also makes recommendations on alternative indicators and discusses the issues related to adopting different measures.

#### 4.1 Data Analysis

Section 3 discusses the bridge and pavement data that were collected from the states participating in this study. Due to time and budget constraints as well as the fact that Maryland provided data only for Interstate 95, CS generally restricted its analysis to assets present on I-95.

##### Review of Statistical Functions

This section briefly reviews some of the statistical functions used in the data analysis.

**Correlation coefficient** indicates the correlation or linear dependence between two variables. The correlation coefficient is defined as the covariance of the variables divided by the product of their standard deviations. It yields a value between -1 and +1. If the correlation coefficient is equal to 0, it means that there is no linear correlation between the variables. A value of +1 or -1 means that there is a linear equation that describes the relationship between the variables. The correlation coefficient is positive if the variables simultaneously are greater than or less than their respective means. Conversely, the correlation coefficient is negative if the variables lie on opposite sides of their respective means.

For example, the correlation coefficient for health index and sufficiency rating describes the linear dependence between these two variables. If the variables are correlated (i.e., their correlation coefficient is high), a predictive relationship between them can be inferred. On the other hand, if the correlation coefficient is small, no linear relationship exists.

**Probability distribution** describes the range of possible values that a variable can attain and the probability that the value of the variable is within any subset of that range. Probability distributions are computed using the average and standard deviation for a series of data points. The distributions in this section show possible values for measures such as sufficiency rating, health index, and IRI as well as the probability of certain values occurring. It is important to note that the probability distribution function assumes a normal (i.e., "bell shaped") distribution and does not take into account the fact that sufficiency rating and health index cannot exceed 100.

**Cumulative distribution** is another method of graphically representing the range of values that a variable can attain. Cumulative distributions also are computed using the average and the standard deviation. However, cumulative distributions run from 0 to 1 and, for any given value, represent the percentage of the total population having that value or less. This calculation also assumes a normal distribution of values and does not reflect the maximum value of 100 for sufficiency rating and health index.

##### Bridge Analysis

For the bridge data analysis, CS reviewed:

- Delaware data - 62 bridges with element data obtained from Pontis;
- Maryland data - 118 bridges with element data obtained from the Maryland bridge management system and 118 bridges without element data obtained from the Maryland National Bridge Inventory (NBI) file; and
- Virginia data - 385 bridges with element data obtained from Pontis.

CS did review a recent Pontis database for Washington, D.C. However, this database did not contain any structures on I-95. Therefore, bridges for the District of Columbia were not included in this analysis.

The bridge analysis focused on the following key data items, which are available from the NBI or may be calculated using Pontis:

- Deck Rating (NBI Element 58);
- Superstructure Rating (NBI Element 59);
- Substructure Rating (NBI Element 60);
- Sufficiency Rating (calculated);
- Health Index for deck elements only (calculated);
- Health Index for superstructure elements only (calculated);
- Health Index for substructure elements only (calculated); and
- Health Index for all elements.

Historically, the sufficiency rating and structurally deficient/functionally obsolete (SD/FO) values have been used as the measure of bridge condition. The development in recent years of the health index in Pontis provides a different measure of the structural condition of a bridge. It is important to note that health index does not attempt to measure the functional adequacy of a bridge.

###### Discussion of Current Measures

Although states are able to calculate sufficiency rating for themselves, the official calculation is performed by FHWA using NBI data submitted annually by states. The sufficiency rating uses four separate factors to obtain a numeric value that indicates whether a bridge is sufficient to remain in service. The result is a percentage in which 100 percent represents an entirely sufficient bridge and zero percent represents an entirely insufficient or deficient bridge. The sufficiency rating is never less than 0 or more than 100.

A bridge's sufficiency rating affects its eligibility for Federal funding for maintenance, rehabilitation, or replacement activities. For bridges to qualify for Federal replacement funds, they must have a rating of 50 or less. To qualify for Federal rehabilitation funding, a bridge must have a sufficiency rating of 80 or less.

The sufficiency rating factors are:

- S1, the structural adequacy and safety factor;
- S2, the serviceability and functional obsolescence factor;
- S3, the essentiality for public use factor; and
- S4, the special reductions factor.

S1 is a function of the lowest rating code of Item 59 (Superstructure Rating), Item 60 (Substructure Rating) or Item 62 (Culvert Rating), and Item 66 (Inventory Rating).

S2 is a function of Item 58 (Deck Rating), Item 67 (Structural Evaluation), Item 68 (Deck Geometry), Item 69 (Underclearances), Item 71 (Waterway Adequacy), Item 72 (Approach and Road Alignment), Item 29 (ADT), Item 51 (Bridge Roadway Width), Item 28 (Lanes on the Structure), Item 100 (Defense Highway), Item 32 (Approach Roadway Width), Item 43 (Structure Type), and Item 53 (Vertical Clearance Over Deck).

S3 is a function of S1, S2, Item 29 (ADT), Item 100 (Defense Highway), and Item 19 (Detour Length).

To obtain the sufficiency rating, S4 is subtracted from the sum of S1, S2, and S3. S4 is only used when the sum of S1, S2, and S3 is greater or equal to 50. S4 is a function of Item 19 (Detour Length), Item 36 (Traffic Safety Features), and Item 43 (Structure Type).

Similar to sufficiency rating, health index provides a single numeric indicator of the structural health of the bridge. This indicator is expressed as a percentage from zero, which corresponds to the worst possible condition, to 100, which is the best condition. Health index is a function of the fractional distribution of the bridge element quantities across the range of their applicable condition states. Concretely, the health index value of an entire bridge is calculated as a weighted average of the health indexes of its elements, where elements are weighted by their total quantity and relative importance (i.e., failure cost). Consequently, the bridge health index is a function of the failure cost of each element of the bridge, quantity of each element on the bridge, and condition state of each element. Likewise, the health index of each element is a function of its unit failure cost, its quantity on the bridge, and its condition state on the bridge.

Though the sufficiency rating provides a consistent standard for the evaluation of sufficiency to remain in service, it is not comprehensive enough to provide performance-based information on each element of the bridge. For example:

- Sufficiency rating focuses on the overall condition of the bridge, making it irrelevant for maintenance decision-making.
- Sufficiency rating emphasizes the functionality and geometric characteristics of the bridge rather than an element-based view of the bridge. (Factor S2)
- Sufficiency rating assumes that the worst distress between the superstructure, substructure, or culvert will represent the overall structural adequacy and safety of the bridge (Factor S1). This means sufficiency rating cannot account for localized problems or safety issues.
- The superstructure, substructure, and culvert ratings are on a 0-9 scale by severity of deterioration. These ratings do not give a picture of the deterioration process at work or the extent of deterioration.
- Sufficiency rating focuses only on the major parts of the bridge: superstructure, substructure, deck, and culverts. The lack of detail makes it impossible to estimate the cost of rehabilitation or maintenance.

On the other hand, the health index indicates the condition of each element at a given time. The bridge health index aggregates element-level health indexes where weights, which embody the economic consequences of failure, are assigned to each element. The main advantage of the health index is that it links the condition of bridge to the resources allocated. Consequently, decision-makers can evaluate the impact of several resource allocation scenarios on the future condition of a bridge network. Also, maintenance and rehabilitation decision-making is facilitated since the measures are detailed enough to represent localized problems.

###### Analysis of Current Measures

The first part of the analysis of bridge data involved the calculation of basic statistical information by state and across all states. These basic statistics, not weighted by bridge deck area, are presented in Table 4.1.

State | Statistic | Deck Rating | Superstructure Rating | Substructure Rating | Sufficiency Rating | Health Index - Deck | Health Index - Superstructure | Health Index - Substructure | Health Index - Overall |
---|---|---|---|---|---|---|---|---|---|

DE | Count | 45 | 45 | 45 | 61 | 44 | 45 | 61 | 61 |

DE | Min | 5 | 5 | 4 | 53 | 50 | 61 | 67.5 | 67.5 |

DE | Max | 8 | 8 | 8 | 98 | 100 | 100 | 100 | 100 |

DE | Average | 6.56 | 6.40 | 6.33 | 84.47 | 89.14 | 88.63 | 94.87 | 92.44 |

DE | Median | 7 | 6 | 6 | 85 | 99.85 | 92.5 | 97.4 | 93.8 |

DE | Std Dev | 0.66 | 0.65 | 0.77 | 10.46 | 14.65 | 10.70 | 7.46 | 6.44 |

DE | Variance | 0.43 | 0.43 | 0.59 | 109.36 | 214.64 | 114.44 | 55.62 | 41.44 |

MD | Count | 107 | 108 | 108 | 140 | 98 | 108 | 118 | 118 |

MD | Min | 5 | 5 | 5 | 50.2 | 25 | 67 | 46.2 | 49.3 |

MD | Max | 8 | 9 | 9 | 100 | 100 | 100 | 100 | 100 |

MD | Average | 6.64 | 6.36 | 6.19 | 81.20 | 84.44 | 96.63 | 90.18 | 88.96 |

MD | Median | 7 | 6 | 6 | 83.60 | 75 | 99.7 | 94.35 | 93.05 |

MD | Std Dev | 0.62 | 0.81 | 0.78 | 10.10 | 17.03 | 6.56 | 11.39 | 11.51 |

MD | Variance | 0.38 | 0.66 | 0.61 | 102.02 | 290.15 | 43.01 | 129.78 | 132.41 |

VA | Count | 223 | 223 | 223 | 373 | 220 | 222 | 370 | 372 |

VA | Min | 4 | 4 | 4 | 25.3 | 25 | 48.3 | 12.8 | 24.4 |

VA | Max | 9 | 9 | 9 | 100 | 100 | 100 | 100 | 100 |

VA | Average | 6.57 | 6.57 | 6.28 | 81.97 | 87.54 | 92.39 | 89.77 | 89.69 |

VA | Median | 7 | 7 | 6 | 83 | 100 | 97.8 | 95.3 | 94.2 |

VA | Std Dev | 0.94 | 1.08 | 1.07 | 12.52 | 15.13 | 11.38 | 12.75 | 11.56 |

VA | Variance | 0.88 | 1.16 | 1.14 | 156.78 | 228.79 | 129.59 | 162.65 | 133.53 |

ALL | Count | 375 | 376 | 376 | 574 | 362 | 375 | 549 | 551 |

ALL | Min | 4 | 4 | 4 | 25.3 | 25 | 48.3 | 12.8 | 24.4 |

ALL | Max | 9 | 9 | 9 | 100 | 100 | 100 | 100 | 100 |

ALL | Average | 6.59 | 6.49 | 6.26 | 82.05 | 86.89 | 93.16 | 90.42 | 89.84 |

ALL | Median | 7 | 6 | 6 | 83.00 | 100 | 98 | 95.5 | 94 |

ALL | Std Dev | 0.83 | 0.97 | 0.96 | 11.78 | 15.65 | 10.42 | 12.08 | 11.12 |

ALL | Variance | 0.68 | 0.93 | 0.92 | 138.79 | 244.83 | 108.64 | 145.82 | 123.69 |

The following are some observations regarding the basic statistics:

- Counts only include assets where the measure is not missing. Differences in the counts can be attributed primarily to the absence of certain types of elements. For example, culverts generally do not have deck elements and, therefore, would not have either a deck rating or deck health index.
- The low standard deviation and variance for the NBI ratings compared to the equivalent health index measures can be attributed to the small number of values the NBI rating can hold. For these statistics, only ratings from zero to nine were considered.
- The combined statistics for all states do not vary dramatically from the statistics for any individual state. The ratings and the component-based health indexes for Virginia vary more than in the other states. There is no reason to attribute this to anything other than actual variability in the condition of the corresponding structures.

Table 4.2 presents some of the same basic statistics, except that values have been weighted by bridge deck area. Deck area is one of the primary metrics by which bridges generally are categorized. Only the statistics necessary to support calculations of probably and cumulative distributions were calculated.

State | Statistic | Deck Rating | Superstructure Rating | Substructure Rating | Sufficiency Rating | Health Index - Deck | Health Index - Superstructure | Health Index - Substructure | Health Index - Overall |
---|---|---|---|---|---|---|---|---|---|

DE | Count | 45 | 45 | 45 | 61 | 44 | 45 | 61 | 61 |

DE | Average | 6.20 | 6.03 | 6.63 | 78.45 | 78.87 | 94.91 | 95.31 | 88.96 |

DE | Std Dev | 0.62 | 0.68 | 0.75 | 7.81 | 11.93 | 8.36 | 8.14 | 5.05 |

DE | Variance | 0.38 | 0.46 | 0.57 | 61.07 | 142.31 | 69.89 | 66.32 | 25.55 |

MD | Count | 107 | 108 | 108 | 113 | 98 | 103 | 103 | 103 |

MD | Average | 6.61 | 6.40 | 6.14 | 79.66 | 82.36 | 95.70 | 92.23 | 90.55 |

MD | Std Dev | 0.53 | 0.65 | 0.69 | 8.55 | 17.99 | 7.56 | 9.53 | 9.76 |

MD | Variance | 0.28 | 0.42 | 0.48 | 73.17 | 323.57 | 57.14 | 90.80 | 95.19 |

VA | Count | 213 | 213 | 213 | 215 | 211 | 211 | 212 | 212 |

VA | Average | 6.81 | 6.73 | 6.62 | 81.47 | 88.42 | 93.79 | 89.94 | 90.58 |

VA | Std Dev | 1.12 | 1.38 | 1.30 | 12.82 | 14.63 | 10.26 | 12.14 | 9.89 |

VA | Variance | 1.25 | 1.91 | 1.70 | 164.34 | 213.93 | 105.29 | 147.42 | 97.78 |

ALL | Count | 365 | 366 | 366 | 389 | 353 | 359 | 376 | 376 |

ALL | Average | 6.61 | 6.46 | 6.39 | 80.21 | 84.75 | 94.48 | 91.81 | 90.16 |

ALL | Std Dev | 0.80 | 0.98 | 0.96 | 10.37 | 14.76 | 9.25 | 10.68 | 8.85 |

ALL | Variance | 0.64 | 0.96 | 0.93 | 107.49 | 217.85 | 85.57 | 114.03 | 78.35 |

The following are some observations regarding the weighted statistics:

- The count of assets goes down in many cases because bridges without a deck area (e.g., culverts) are removed from the results.
- Some of the other statistics (e.g., average, standard deviation, and variance) are reduced when weighted by deck area. This reflects both the smaller set of assets being considered as well as a general trend observed in other states that health index weighted by deck area is lower than the average health index.
^{(1)} - Despite the differences observed, the statistics generally are very close for bridges that have and have not been weighted by deck area. For the data covered by this analysis, this indicates that the values are well distributed across the spectrum of possible results.

For the second phase of the analysis, CS measured the correlation between different measures. The closer the correlation coefficient is to 1 or -1, the greater the statistical relationship between two sets of values. It is important to note that the correlation coefficient does not make any determination regarding accuracy of the values nor does it infer any cause/effect relationship between the values. Table 4.3 presents the correlation coefficients for bridge measures within individual states and across all states.

State | Statistic | Superstructure Rating | Substructure Rating | Sufficiency Rating | Health Index - Deck | Health Index - Superstructure | Health Index - Substructure | Health Index - Overall |
---|---|---|---|---|---|---|---|---|

DE | Deck Rating | 0.05 | 0.39 | 0.47 | 0.62 | 0.08 | 0.23 | 0.60 |

DE | Superstructure Rating | Not Applicable | 0.09 | 0.13 | 0.05 | 0.43 | -0.03 | -0.03 |

DE | Substructure Rating | Not Applicable | Not Applicable | 0.51 | 0.20 | -0.12 | 0.67 | 0.37 |

DE | Sufficiency Rating | Not Applicable | Not Applicable | Not Applicable | 0.37 | 0.32 | 0.38 | 0.25 |

DE | Health Index - Deck | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.12 | 0.20 | 0.60 |

DE | Health Index - Superstructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | -0.11 | 0.63 |

DE | Health Index - Substructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.67 |

MD | Deck Rating | 0.51 | 0.38 | 0.16 | 0.48 | 0.18 | -0.10 | 0.09 |

MD | Superstructure Rating | Not Applicable | 0.58 | 0.32 | -0.03 | -0.03 | 0.00 | 0.04 |

MD | Substructure Rating | Not Applicable | Not Applicable | 0.52 | 0.17 | -0.14 | 0.02 | 0.45 |

MD | Sufficiency Rating | Not Applicable | Not Applicable | Not Applicable | 0.13 | -0.19 | 0.04 | 0.43 |

MD | Health Index - Deck | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.26 | 0.31 | 0.36 |

MD | Health Index - Superstructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.27 | 0.33 |

MD | Health Index - Substructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.99 |

VA | Deck Rating | 0.74 | 0.70 | 0.40 | 0.59 | 0.54 | 0.49 | 0.35 |

VA | Superstructure Rating | Not Applicable | 0.69 | 0.47 | 0.47 | 0.52 | 0.36 | 0.25 |

VA | Substructure Rating | Not Applicable | Not Applicable | 0.41 | 0.47 | 0.47 | 0.62 | 0.28 |

VA | Sufficiency Rating | Not Applicable | Not Applicable | Not Applicable | 0.67 | 0.68 | 0.64 | 0.29 |

VA | Health Index - Deck | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.38 | 0.32 | 0.74 |

VA | Health Index - Superstructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.39 | 0.77 |

VA | Health Index - Substructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.84 |

ALL | Deck Rating | 0.65 | 0.62 | 0.36 | 0.58 | 0.45 | 0.43 | 0.36 |

ALL | Superstructure Rating | Not Applicable | 0.63 | 0.41 | 0.38 | 0.46 | 0.29 | 0.18 |

ALL | Substructure Rating | Not Applicable | Not Applicable | 0.44 | 0.43 | 0.38 | 0.60 | 0.31 |

ALL | Sufficiency Rating | Not Applicable | Not Applicable | Not Applicable | 0.62 | 0.60 | 0.60 | 0.30 |

ALL | Health Index - Deck | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.28 | 0.30 | 0.59 |

ALL | Health Index - Superstructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.30 | 0.62 |

ALL | Health Index - Substructure | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | Not Applicable | 0.86 |

The following are some observations regarding the correlation coefficients:

- There are no firm rules regarding how close a coefficient must be to 1 or 1 in order to be considered significant. For purposes of this analysis, CS has chosen 0.70 as a reasonable threshold to identify values that are well correlated.
- Theoretically, both sufficiency rating and overall health index represent the structural condition of a bridge. If these measures were equally good for this purpose, we would expect a relatively high correlation between these values. However, as shown in Table 4.3, the correlation coefficients are DE=0.25, MD=0.43, VA=0.29, and ALL=0.30. At least part of this difference can be explained by the fact that the sufficiency rating is a combination of structural adequacy (55 percent), serviceability (30 percent), and essentiality (15 percent).
^{(2)} - The same reasoning applies to Deck Rating versus Health Index - Deck, Superstructure Rating versus Health Index - Superstructure, and Substructure Rating versus Health Index - Substructure. Some of these component values are significantly better correlated than the overall measures. For example, in Delaware the correlations of the deck and substructure measures are 0.62 and 0.67, respectively. These relatively high correlations indicate that the NBI ratings can be reasonable component-based measures of condition once the nonstructural aspects of the sufficiency rating calculation are removed.
- However, some component-based correlations are much worse than the overall measures. For example, in Maryland, the superstructure and substructure measures are -0.03 and 0.02, respectively. This may indicate a systemic problem with the way in which inspectors capture either element-level data or NBI ratings. CS is aware of similar issues in other states, who report that the FHWA NBI Translator built into Pontis, which calculates NBI ratings from element data, is giving significantly different values than the NBI ratings entered by inspectors.
- The highlighted values in Table 4.3 are associated with correlations of the Health Index to subcomponents of the Health Index or one NBI rating with another. They are not associated with correlations of Health Index and NBI values. The highest correlations between Health Index and NBI are found for the substructure ratings in Delaware and Virginia.

In addition to computing the correlation coefficients, CS also prepared probability and cumulative distributions and associated graphs. A large number of graphs were produced and representative samples are shown below. Figure 4.1 shows probability distributions across all states for Health Index - Overall, Health Index - Deck, Sufficiency Rating, and Deck Rating. The graphs for other measures and individual states generally are not significantly different.

The similarity between Health Index - Overall and Sufficiency Rating graphs reflects the modest differences in the mean and variance for these measures. The exception is Health Index - Overall for Delaware. A significantly lower variance resulted in a narrower distribution. This trend is demonstrated by the extremely sharp curve for Deck Rating. This graph, which repeats for the other NBI ratings, is driven by the small number of values that an individual rating can assume and the correspondingly lower variance for this measure.

Figure 4.1 Bridge Probability Distributions - All States

Figure 4.2 shows cumulative distributions across all states for Health Index - Overall, Health Index - Deck, Sufficiency Rating, and Deck Rating. The graphs for other measures and individual states generally are not significantly different. Again, the sharp curve for Deck Rating reflects the small number of values and correspondingly small variance for this measure.

Figure 4.2 Bridge Cumulative Distributions - All States

Figures 4.1 and 4.2 show values not weighted by deck area. Figure 4.3 shows Health Index - Overall and Sufficiency Rating comparing values that are weighted and not weighted.

Figure 4.3 Weighted Bridge Distributions - All States

Generally, the weighted measures are sharper (i.e., the probability distributions are narrower and the peaks higher). This reflects the fact that fewer assets are included in the calculations because bridges without deck areas are not included. Otherwise, the curves are very similar, which reinforces conclusions associated with Table 4.2.

##### Pavement Analysis

For the pavement data analysis, CS reviewed:

- Delaware data - 193 segments on I-95 totaling approximately 37 miles;
- Maryland data - 2,179 segments on I-95 totaling approximately 217 miles; and
- Virginia data - 192 segments on I-95 totaling approximately 371 miles.

Note that these mileage totals generally do not match the official number of I-95 miles by state.^{(3)} The totals reported above were calculated by summing the difference of beginning and ending mile points for each segment. For Maryland and Virginia, the data include separate records for northbound and southbound roadways. The data for Delaware appear to cover one direction only. Also, although CS originally believed that Maryland did not provide data for the section of I-95 that lies with the city limits of Baltimore, a more thorough review of the data determined that these records were present but there was a mismatch between county name and county code. CS corrected these data and matched the records to the correct road section.

The pavement analysis focused on the following key data elements:

- International Roughness Index (IRI), which was provided by all states;
- Overall Pavement Condition (OPC), which was reported by Delaware and calculated for Virginia by CS;
- Critical Condition Index (CCI), which was reported by Virginia; and
- Other distress index data (e.g., cracking, rutting, etc.) as appropriate.

Historically, IRI has been used as the measure of pavement condition while Present Serviceability Rating (PSR) has measured the ability of the pavement to service expected traffic. These values are provided by states as part of their annual Highway Performance Monitoring System (HPMS) report. Changes to the HPMS reporting process are part of FHWA's HPMS Reassessment 2010+ initiative. These changes include improving the consistency IRI measurement and reporting as well as submitting more data elements (e.g., rutting, faulting, cracking, overlay information).

###### Discussion of Current Measures

Pavement roughness is defined in accordance with the American Society for Testing and Materials (ASTM) standard E867 as "the deviation of a surface from a true planar surface with characteristic dimensions that affect vehicle dynamics and ride quality." IRI was chosen by the World Bank in the 1980s to quantify roughness. After a detailed study of various methodologies and road profiling statistics, IRI was chosen as the HPMS standard reference roughness index. The HPMS data reporting unit for IRI is meters/kilometer (inches/mile).

IRI is the amount of roughness in a measured longitudinal profile. Lower values for IRI indicate smoother pavement. IRI is based on average rectified slope (ARS), which is a filtered ratio of a standard vehicle's accumulated suspension motion (e.g., millimeters or inches) divided by the distance traveled by the vehicle during the measurement. IRI is equal to ARS multiplied by 1,000. IRI is computed from a single longitudinal profile measured with a road profiler in both the inside and outside wheel paths of the pavement. The average of these two IRI measurements is reported as the roughness of the pavement section.

However, IRI only captures road smoothness. Some states use other indicators, such as OPC or CCI, to describe the general health of the pavement. Indeed, the pavement may be very smooth and yet have deep rutting in the wheel path or cracking that allows water to enter and cause deterioration. OPC and CCI both are composite values that combine several distress ratings to produce an overall pavement condition measure.

Delaware uses the following process to calculate OPC for asphalt pavement:

- Convert five distress measures into numeric indexes using the tables shown in Figure 4.4;

Figure 4.4 Delaware Pavement Distress Conversation Tables

- Calculate the average (avg) and the standard deviation (stdev) of the five numeric indexes; and
- Calculate OPC using the formula OPC = avg - (1.25 * stdev).

Generally, OPC is a number between 0 and 100. In some cases, OPC is divided by 20 and reported as a number between 0 and 5.

Virginia uses the following process^{(4)} to calculate CCI for asphalt pavement:

- Calculate a load distress index (LDR) to describe distresses related to wheel loads (e.g., alligator cracking, delaminations, patching, potholes and rutting);
- Calculate a nonload distress index (NDR) to describe distresses related to weathering (e.g., bleeding, block cracking, linear cracking and reflection cracking); and
- Define CCI as the lower of the LDR and NDR index values.

Both LDR and NDR start at a base value of 100. Points are deducted based on the severity and frequency of occurrence of each distress. Some distresses are classified as more detrimental to pavement and are weighted more heavily. The deductions are based on the deduct curves in the PAVER pavement management system. The specifics of these calculations are beyond the scope of this document but are available from the Virginia Department of Transportation. Like OPC, CCI is a number between 0 and 100. Note that IRI is not one of the inputs into the CCI calculation.

###### Analysis of Current Measures

As with bridges, the first part of the pavement analysis involved the calculation of basic statistical information by state and across all states. These basic statistics, not weighted by pavement segment length, are presented in Tables 4.4 and 4.5.

State | Statistic | IRI - Left | IRI - Right | IRI - Average | Transverse Cracks - Severity 1 (linear feet) | Transverse Cracks - Severity 2 (linear feet) | Longitudinal Cracks - Severity 1 (linear feet) | Longitudinal Cracks - Severity 2 (linear feet) | Alligator Cracks - Severity 1 (square feet) | Alligator Cracks - Severity 2 (square feet) | Alligator Cracks - Severity 3 (square feet) |
---|---|---|---|---|---|---|---|---|---|---|---|

DE | Count | 91 | 91 | 91 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Min | 49 | 49 | 49 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Max | 368 | 354 | 358 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Average | 144 | 156 | 151 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Median | 122 | 135 | 138 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Std Dev | 76 | 80 | 74 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Variance | 5731 | 6334 | 5458 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Count | N/A | N/A | 2179 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Min | N/A | N/A | 31 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Max | N/A | N/A | 482 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Average | N/A | N/A | 83 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Median | N/A | N/A | 67 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Std Dev | N/A | N/A | 47 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Variance | N/A | N/A | 2182 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

VA | Count | 192 | 192 | 192 | 192 | 192 | 192 | 192 | 192 | 192 | 192 |

VA | Min | 44 | 41 | 45 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |

VA | Max | 160 | 181 | 160 | 12589 | 9767 | 5918 | 9010 | 15866 | 31659 | 10407 |

VA | Average | 86 | 90 | 88 | 746 | 299 | 180 | 222 | 1284 | 1793 | 368 |

VA | Median | 86 | 88 | 88 | 12 | 1 | 0 | 0 | 416 | 414 | 15 |

VA | Std Dev | 21 | 23 | 21 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

VA | Variance | 436 | 523 | 446 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Count | N/A | N/A | 2462 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Min | N/A | N/A | 31 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Max | N/A | N/A | 482 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Average | N/A | N/A | 86 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Median | N/A | N/A | 70 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Std Dev | N/A | N/A | 48 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Variance | N/A | N/A | 2330 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

State | Statistic | Patching - Wheel Path (square feet) | Patching - Non-wheel Path (square feet) | Number of Potholes | Rut Depth | CCI | OPC | IRI Condition Index | Rut Count | Rut Condition Index | Friction Number | Friction Condition Index | Cracking Index | Cracking Condition Index |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|

DE | Count | N/A | N/A | N/A | N/A | N/A | 193 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Min | N/A | N/A | N/A | N/A | N/A | 28 | NN/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Max | N/A | N/A | N/A | N/A | N/A | 100 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Average | N/A | N/A | N/A | N/A | N/A | 71 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Median | N/A | N/A | N/A | N/A | N/A | 71 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Std Dev | N/A | N/A | N/A | N/A | N/A | 15 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

DE | Variance | N/A | N/A | N/A | N/A | N/A | 225 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

MD | Count | N/A | N/A | N/A | 2179 | N/A | N/A | 2179 | 2179 | 2179 | 398 | 398 | 1956 | 1956 |

MD | Min | N/A | N/A | N/A | 0.05 | N/A | N/A | 1 | 0 | 1 | 10 | 1 | 55 | 1 |

MD | Max | N/A | N/A | N/A | 0.37 | N/A | N/A | 5 | 48 | 3 | 63 | 3 | 100 | 4 |

MD | Average | N/A | N/A | N/A | 0.16 | N/A | N/A | 1.95 | 5 | 1.28 | 45 | 2.81 | 95 | 1.18 |

MD | Median | N/A | N/A | N/A | 0.14 | N/A | N/A | 2 | 0 | 1 | 46 | 3 | 98 | 1 |

MD | Std Dev | N/A | N/A | N/A | 0.07 | N/A | N/A | 0.96 | 11 | 0.66 | 6 | 0.47 | 7 | 0.51 |

MD | Variance | N/A | N/A | N/A | 0.00 | N/A | N/A | 0.92 | 124 | 0.44 | 35 | 0.22 | 43 | 0.26 |

VA | Count | 192 | 192 | 192 | 192 | 190 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

VA | Min | 0 | 0 | 0 | 0.1 | 16 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

VA | Max | 9190 | 11858 | 3 | 0.48 | 100 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

VA | Average | 411 | 393 | 0.08 | 0.19 | 73 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

VA | Median | 17 | 16 | 0 | 0.19 | 77 | N/A | N/A | N/A | N/A | N/A | N/Ae | N/A | N/A |

VA | Std Dev | N/A | N/A | 0.41 | 0.06 | 20 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

VA | Variance | N/A | N/A | 0.17 | 0.00 | 384 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Count | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Min | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Max | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Average | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Median | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Std Dev | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

ALL | Variance | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |

1. *2008-2011 Draft Statewide Transportation Improvement Program: Evaluation of the State Bridge Program,* Oregon Department of Transportation, Bridge Engineering Section, April 2007.

2. *Recording and Coding Guide for the Structure Inventory and Appraisal of the Nation's Bridges,* FHWA-PD-96-001, December 1995.

3. http://en.wikipedia.org/wiki/Interstate_95.

4. McGhee, K. H., *Development and Implementation of Pavement Condition Indices for the Virginia Department of Transportation, Phase 1: Flexible Pavement,* Virginia Department of Transportation, Maintenance Division, September 2002.

<< Previous | Contents | Next >> |