The Challenge: Scale, Efficiency, and Data Quality
Road condition surveys are not new. But most established methodologies were designed for individual projects — not for multi-disciplinary assessments spanning hundreds of kilometres, twelve bridge structures, and dozens of cross-drainage elements on a tight timeline and with a direct link to a major investment decision.
The Mwanza–Musoma corridor presented additional complexity: the road traverses areas with significant variation in terrain, drainage conditions, and traffic loading. A single assessment protocol would not be appropriate across the full extent.
Survey Team Architecture
We structured the survey into four parallel workstreams, each operating simultaneously to maximise efficiency:
1. Pavement Condition Assessment Team
Two mobile teams using network-level visual condition assessment (VCA) protocols, recording pavement distress types, extent, and severity at 200m intervals using tablet-based forms with embedded GPS. Each team covered approximately 40–50km per day on accessible sections.
2. Bridge and Structure Inspection Team
A dedicated team of two structural engineers and a civil technician conducting principal inspections of all 12 bridge structures in accordance with the Austroads bridge inspection guidelines (selected for compatibility with AfDB requirements). Each structure took between half a day and two days depending on size and condition complexity.
3. Drainage and Cross-Drainage Team
Systematic assessment of all drainage infrastructure — culverts, catchpits, drains, outfall structures — using a simplified condition rating adapted from the PIARC road maintenance guide. Over 340 drainage structures were recorded in the survey.
4. GIS and Data Management Team
A dedicated data management function based at our Mwanza field office, responsible for daily data quality review, GPS track processing, and integration of all workstream data into a unified road database. This function was critical to maintaining data quality under field pressure.
"Parallel workstreams doubled our field productivity — but only because the data management function could absorb the output. Without that backbone, parallel teams create a data reconciliation problem at the end."
Bridge Inspection Methodology
Bridge inspections deserve special attention because the consequences of missing a critical defect are severe. Our inspection approach followed a three-stage process:
- Desk review: Review of any available as-built drawings, previous inspection records, or maintenance history. In practice, documentation for most structures was partial or absent, making the field inspection the primary data source.
- Principal inspection: Close visual examination of all accessible structural elements (deck, superstructure, substructure, bearings, scour protection), supplemented by non-destructive testing (cover meter, rebound hammer) at locations of visible deterioration.
- Condition rating and load assessment: Each structure rated on a 0–5 condition scale for each element, with a load-carrying capacity assessment for structures of concern to inform any immediate load restrictions.
Three of the 12 structures were assigned a Condition Rating 3 (poor), requiring early intervention. One structure was rated CR4 (serious) and flagged for load restriction pending detailed structural assessment — a finding that directly shaped the investment programme's bridge rehabilitation scope.
Data Quality Control
The most important lesson from large-scale surveys is that data quality cannot be inspected in at the end — it must be built in from the beginning. Our QC protocols included:
- Daily review of all submitted forms against GPS tracks, flagging locations where surveyor position and recorded chainage diverged by more than 50m
- 10% back-check of pavement condition ratings by a senior engineer
- Automated database checks for logical inconsistencies (e.g., a bridge recorded as being in better condition than its approach road)
- Weekly client progress updates with preliminary findings, enabling early discussion of any unexpected results
Output: From Data to Investment Decision
The survey output was structured to serve the AfDB's investment programming directly. We delivered:
- GIS-referenced road condition database covering all 340km
- Bridge inspection reports for all 12 structures with photographic records and condition ratings
- Prioritised maintenance and rehabilitation schedule by road section, categorised by treatment type (routine, periodic, rehabilitation)
- Cost model calibrated to current Tanzanian unit rates for each treatment type
- An executive summary presenting the investment case in a format directly usable for AfDB Board documentation
The survey findings were used by TANROADS and the AfDB to define the scope and phasing of the rehabilitation programme, and to justify the investment to the AfDB Board. Two sections were elevated to immediate emergency treatment on the basis of our findings — work that began before the main rehabilitation programme was processed.
What We'd Do Differently
Every large project teaches something. On the Mwanza–Musoma survey, two things we'd approach differently next time:
First, we underestimated the time required for drainage structure inspection in sections with heavy vegetation. Clearing around culverts to allow proper inspection added 15–20% to the drainage team's timeline. Future surveys in similar terrain should build this into the programme explicitly.
Second, the tablet forms we used for pavement condition assessment were adequate but not optimal for capturing photographic evidence at the density we ultimately needed. We're now developing a custom mobile form with automatic photo tagging at defined intervals for future road surveys.
