The Project

RoadTek Mackay (TMR) required a GPR survey of Marion Creek Bridge, Ilbilbie, to evaluate the relative asphalt thickness across the bridge deck. Rather than using traditional destructive coring, a ground penetrating radar survey was carried out across the full length of the bridge using a Proceq GP8000 GPR system.

Due to the short overall bridge length comprising four spans, two longitudinal scan lines were assessed within each span to provide improved spatial representation of asphalt thickness variability. The interpreted thicknesses represent averaged values derived from these longitudinal assessments.

The goal sounded simple — measure asphalt thickness — but the conditions on site told a more complex story.

What the Scan Revealed

Early in the survey, it became clear that the asphalt thickness was far from uniform. In most areas, thickness ranged between 30 mm and 45 mm, with some sections dropping as low as 23 mm. In other isolated locations, thickness approached 49 mm. For a bridge of this size, that level of variation is significant.

Unlike Alligator Creek Bridge, where asphalt thickness was generally consistent, Marion Creek Bridge exhibited significant intra-span variability. Thickness fluctuations were noticeable even within individual spans, and became particularly evident approaching the relieving slab. This variation does not present as a systematic progressive thinning trend but instead reflects localised construction variability and surface condition.

How We Worked

What complicated things further was the condition of the surface itself. The asphalt was heavily weathered, with visible cracking and moisture ingress following recent rainfall. Water accumulation within cracks created signal interference, making it more difficult to clearly distinguish the boundary between the asphalt layer and the underlying concrete deck.

To manage this, the data was not interpreted from a single view. Both B-scan and A-scan analysis were used to cross-check results. Where reflections were unclear in one dataset, waveform behaviour in the other was used to confirm layer boundaries. While not perfect, this approach provided a reliable interpretation of the structure.

Velocity calibration was performed using hyperbola matching on the reinforcement layer to maintain internal depth consistency. A representative dielectric constant of 6.5 was adopted to reflect the aged and heat-exposed asphalt conditions typical of Mackay's climate.

In several areas, the asphalt–concrete interface was not consistently distinguishable in B-scan data and required confirmation through detailed A-scan waveform analysis. Moisture pockets and surface cracking occasionally produced early signal responses between the surface reflection and the true interface, further complicating interpretation.

Understanding the Limitations

It is important to note that GPR is best used as a diagnostic tool rather than an exact measurement method. Results typically carry an accuracy range of around ±10 percent. This is more than sufficient for planning and risk assessment, but where precise thickness values are required, physical verification such as core sampling should be undertaken.

The accuracy of GPR-derived depth estimates can be influenced by factors such as moisture content, material heterogeneity, surface condition, and signal attenuation. In some locations, horizontal horizons associated with asphalt and concrete interfaces were clearly defined, whereas in other areas they were more difficult to resolve.

While reinforcement depth correlation supports the general validity of the adopted velocity model, the results at Marion Creek Bridge should be considered a representative snapshot of relative asphalt thickness rather than a uniformly consistent overlay profile.

The Outcome

Another key observation was the lack of any consistent pattern in thickness variation. This was not a gradual change along the bridge — thickness fluctuated unpredictably, even within the same span. This typically points to past patching, inconsistent resurfacing, or variable compaction during previous works.

The final dataset provided a realistic picture of the bridge surface condition. The combination of thin areas and inconsistent thickness raises concerns around long-term durability, particularly in sections already approaching minimum cover.

• Full asphalt thickness profile mapped across all four bridge spans non-destructively
• Localised thin zones (23–25 mm) identified for priority maintenance attention
• Significant intra-span variability documented — not visible from surface inspection alone
• Both B-scan and A-scan analysis used to confirm results in difficult signal conditions
• Results suitable for maintenance planning and resurfacing assessment
• No bridge closures or coring required to obtain the dataset