The adage ‘you can’t manage what you haven’t measured’ applies to almost everything in public works, including pavement condition. In our never-ending attempt to cost-effectively provide a safe and comfortable ride, we’ve embarked on an ongoing asset-management program designed to most effectively target resurfacing dollars.

About 85% all of our 5,781 lane miles are two-lane asphalt, but we’re not sure when they were all installed (more about that later). County population is expected to double by 2030 as orange grove owners sell land to housing and commercial developers.

Florida’s high groundwater table makes protecting the subbase from infiltration absolutely critical. In 2007, we retained local consulting firm Southeastern Surveying and Mapping Corp. to collect georeferenced digital images of all our roads and apply condition data to the corresponding segments using MicroPAVER 7.0. Developed by the U.S. Army Corps of Engineers and administered by the American Public Works Association, MicroPAVER is a subscription-based software program for collecting and analyzing asphalt and concrete condition data that public agencies can use to target limited resources to the most pressing needs.

Before then, our annual preventive maintenance schedule was based on field employee assessments, resident complaints, and pothole repair data. While useful, this approach was often undermined when roads in worse condition were brought to our attention and caused us to make ad hoc changes in priority.

We needed something more “scientific” to justify funding requests, so we joined the hundreds of other agencies worldwide that use MicroPAVER. Based on American Society of Testing and Materials (ASTM) D6433-11: Standard Practice for Roads and Parking Lots Pavement Condition Index Surveys, the software calculates a Pavement Condition Index (PCI) score from zero (very heavily distressed or failed) to 100 (no distress or freshly resurfaced) for individual road segments and a network overall.

The score is based on the extent and severity of 19 surface distresses: alligator cracking, bleeding, block cracking, bumps and sags, corrugation, depression, edge cracking, joint reflection cracking, lane/shoulder drop off, longitudinal and transverse cracking, patching and utility cut patching, polished aggregate, potholes, railroad crossing, rutting, shoving, slippage cracking, swell, weathering and raveling.

Step 1: data for eight maintenance districts
Our consultant’s job was to inventory these distresses to create and populate a database with baseline condition data for our entire network.

The firm’s inspection vehicles collected geo-referenced digital images via roof-mounted cameras that captured two views: forward-looking general streetscape and downward-looking pavement detail. GPS enables condition data to be extracted from the images and assigned to the corresponding pavement segment. This is how we confirmed segment lengths.

Drivers also recorded their observations with voice notes to ensure they’d be good enough for the firm’s GIS technicians to analyze.

Back at the office, the firm loaded the images into a software application so a ‘virtual drive’ could be performed. The GIS technicians viewed the sequential images from both cameras while listening to the voice notes to determine distress level for each road segment. This data was then loaded into MicroPAVER and used to calculate a PCI score for each segment.

The ASTM standard defines sample condition categories, but an agency can assign its own based on whatever criteria they like to determine maintenance and repair strategies.

Our baseline data collection was finished in 2014. We’re now reassessing the network, a process we expect to take five years. This time, however, we’re tweaking the process to make results even more helpful to our asset management efforts.

Step 2: compensate for unintended bias

The first time around, we didn’t realize how pavement surface type, consultant expertise, and image volume affect PCI score.

For example, we used to do a lot of microsurfacing. This non-structural treatment masks initial distress symptoms, so images produce unrealistically high scores. On the other hand, resurfacings that have oxidized produce unrealistically low scores. Images of structural treatments like Type S-III Marshall and Superpave mixes more accurately reflect actual deficiencies.

Similarly, one or two images of a road sample with a small but highly distressed section don’t provide enough data to accurately reflect the segment’s overall integrity.

Therefore, it’s very important that whoever interprets the raw data is knowledgeable about asphalt paving and maintenance.

We decided to visit other local and state agencies to get their input on these issues and learn about their assessment methodology for their resurfacing program. Southeastern Surveying and Mapping spent weeks talking to various county employees to understand their concerns. We spent weeks bringing the firm’s technicians into the field to show them how things like aggregate quality, material, previous resurfacing method, and inspection time affect the appearance of asphalt pavement.

Step 3: process improvements
The beauty of a PMS like MicroPAVER is that it quantifies the long-term impact of funding scenarios and paving methodologies in a way that’s easy for laymen to understand.

Our goal is to develop a five-year plan based on current and projected PCI scores for the entire network. We’ll also be able to estimate rehabilitation and repair costs for various resurfacing methods.

Increase inspection rate. Because road segments vary in size, a simple average doesn’t accurately represent overall condition. ASTM D6433-11 recommends a minimum sample of 20% of a segment regardless of its length. To get a better picture of the overall health of roads with special traffic loading, maintenance, or design issues, we’re increasing the percentage to about 25%, or more, as needed.

We’re also going to inspect additional lanes on roads with more than two lanes. Images will be collected along the rightmost and leftmost through lanes on roads with two or more lanes in each direction. This will also help better identify shoulder drop-offs and curb distresses in lanes closest to medians.

Update daily. We used subdivision platted date as initial road-construction date, but our database doesn’t include work done since then. Our consultant worked with us to develop a pavement deterioration model for estimating future PCI scores, but we need a more complete set of scores and age to confidently calculate a deterioration curve for each road segment.

We’re making a concerted effort to enter the dates of maintenance and/or resurfacing work into the database. We’re also entering the dates that industrial parks, retail strips, and subdivisions are paved.

Enhance quality control. Images taken at high noon, when sunlight is brightest, pick up more damage than those taken in less-intense light. To account for such variables, we’re adding a statistical error range, which we call “noise factor,” to each section’s score to calculate “adjusted PCI.”

It’s important to note that a road segment’s PCI score represents condition at the time data was collected. Since segments vary in size, simple averages don’t accurately represent overall road conditions. Hence, the summary database our consultant provided was weighted by road segment area to produce a composite PCI score.

To validate our new inspection methodology’s accuracy a third quality-control measure will be to randomly reinspect and reassess a percentage of sampled polygons (usually 2,500 square feet). Our paving coordinator, senior foreman, foremen, and paving inspectors are out in the field every day, so they’ll conduct random assessments.

Omit under-construction projects, recently resurfaced roads, and roads scheduled for resurfacing. There’s no sense spending money to evaluate new or resurfaced pavement because, by definition, its PCI score is 100. For that reason, we won’t assess roads resurfaced within the previous three years.

Link our geodatabase to county GIS. The county recently added a roadbase layer to its GIS. By providing updates and revisions for newly built roads and annexed roads, we’ll help create a dynamic, proactive pavement management program that uses taxpayer dollars most effectively.

We believe these changes will improve our ability to extend the life of existing roads and those that will be built over the coming decades.