The pavement-management library

Nine documents, 1999 to 2025, about 760 pages between them. They cover the whole chain — collecting and certifying condition data, modelling how a pavement deteriorates, choosing treatments, measuring how a network is performing, and deciding what to fund. This is a map of them, so you can find the right one in a minute instead of an afternoon.

Everything here is published by FHWA, TRB or PIARC and freely available. None of it is ours and none of it mentions us: it is the background the method is built on, not an endorsement of anything.

Age matters: D9 is 27 years old and its funding/policy content is historical, but its logic (right treatment, right pavement, right time) and its data-tracking list are still current practice.

Thematic map and reading paths

No single document covers the whole chain; D6 (Roadmap) is the closest thing to an index, and D3 is the deepest on analytics. ● = main subject, ○ = treated in part, blank = not covered.

Reading paths

1. New to pavement management (2 hours): D7 → D9 (pages 1–13) → D4 → D5.
2. Building the data pipeline and QA: D8 → D6 sections 2.0 Theme 1 and Theme 4 → D1 (FDOT and TxDOT talks).
3. Building the analytics engine: D4 → D3 chapters 2, 9 and 10 → D3 appendix on the investment candidate file → D9 “Means of Assessing Benefits”.
4. Treatment library and local context: D2 chapters 2–4 → D9 definitions and data-to-track list.
5. Product strategy and feature backlog: D6 chapter 3 action-item tables (Tools, PMS Analysis, Data Quality, Communication).

Document summaries

Common thread: today's PMS tools report condition well but lag on lifecycle cost, financial measures, treatment history and data integration — the gaps a new system can fill.

D3 — Next-generation measures and TAM methodology (final report, 2024)

• Three families of measures: condition (IRI, cracking, rutting, agency index), lifecycle, and financial. Condition measures are lagging; the study looked for leading ones.
• Lifecycle measures: RSI (search all treatment type/timing combinations for the lowest lifecycle cost that keeps service above thresholds), AUCR (programmed EUAC ÷ optimised EUAC) and CAR (NPV of programmed costs ÷ NPV of optimised plan; >1 means missed savings).
• Financial measures: ASI (budget ÷ need), ASR (renewal spend ÷ depreciation), ACR (depreciated value ÷ replacement value), SLR (unfunded backlog ÷ replacement value).
• Pilot results (Idaho, South Dakota, Texas): ASR and SLR were the most useful. In Idaho, condition forecasts showed the network staying Fair or better for 40 years, while ASR/SLR showed the budget failing to offset depreciation after 15 years. ACR added nothing beyond condition.
• Why pilots struggled: commercial PMSs did not output need, depreciation or true lifecycle cost; models ignored pre-treatment condition; RSI runs were too slow; HPMS and PMS segments could not be joined for lack of a common referencing system.
• Cross-asset method: rank investment candidates by benefit-cost ratio, where benefit is social cost saved (agency + safety + mobility user cost) by acting now rather than later; loop through budget scenarios against targets (%Good, %Poor, %Sufficient safety and mobility). Prototype: TA-MAPO spreadsheet with a standard investment candidate file.
• Six PMS enhancements requested: compute need, depreciation and true LCC inside the PMS; evaluate strategies beyond decision trees; use structural (deflection) data; model pre-treatment condition; staff data analytics.

D4 and D5 — TechBriefs (2023)

Eight-page digests of D3. D4 tabulates each measure with its strengths and implementation problems; D5 explains the social-cost tradeoff and names the skills developers need (lifecycle thinking, user-cost models, risk, and efficient algorithms/multithreading for long analyses).

D6 and D7 — Pavement Management Roadmap 2022

• Built from a survey of 61 agencies (all 52 State DOTs) and workshops with 147 participants; 72 action items (46 short-term, 26 long-term), about $30.2M.
• Data gaps: survey-to-result lag of 3–6 months; equipment or vendor changes breaking historical trends; maintenance and construction history missing from PMS; inconsistent distress interpretation; weak geolocation; storage of large image sets.
• Analysis gaps: models without pre-treatment condition; undocumented models and treatment rules; no risk, resilience, carbon or equity in prioritisation; poor feedback from programmed projects back to the PMS; reconciling agency and federal indices.
• Tool items most relevant to software: automated vendor-data checking tool (11A), long-term image access beyond vendors' 5 years (11B), LCP and next-gen measures in vendor software (11C), performance-prediction tool (4F), AI/ML for data reliability (3B), metadata and LRS integration (3E).

D8 — Certify and verify surface condition data (TRB 2020)

• Data quality runs in three stages: before collection (certification), during (blind control sites), after (acceptance checks).
• Manual ratings vary a lot: PCI coefficient of variation reached 35–45% on poor sections over 2013–2019.
• MTC (San Francisco Bay Area) certification: about 24 control sites; automated systems pass if more than 50% of sections are within 5 PCI points of reference and no more than 12% differ by over 15 (raters: 8 and 18).
• Imaging-system guidance: control sites of at least 0.3 mi split into rating intervals of 0.03 mi, ground reference from closed-lane manual rating, then a paired equivalence test per site with limits set by decision impact.
• Transverse-profile targets (90% bounds): rut depth ±2.5 mm, cross slope ±0.4%, edge location ±50 mm.
• Airfield case: average PCI 89 manual, 90 HD camera, 94 3D laser on a runway in good condition; automated methods diverge as distress increases.

D1 — Data-driven strategies (TRB 2025)

• FDOT: the current network forecast is an empirical chance-of-failure curve. A gradient-boosting proof of concept for interstate cracking reached R² = 0.968 using age, truck %, FWD deflection, embankment modulus and latitude. FDOT is integrating 17 datasets across office silos; the goal is a GIS-based pavement digital twin (structure, service conditions, performance history).
• TxDOT: one web system joining construction projects, mix QC/QA, maintenance and condition data — 8 sources, 121 fields, over 128 million rows, Tableau dashboards to benchmark specification items over time.
• Sweden: survival analysis on open data predicts surface life by layer binder type (about 10.5–13 years on one road); network optimisation plus EPDs shows 29% lower maintenance emissions since 2010.

D9 — Preventive maintenance concept (FHWA/NHI 1999)

• Definitions that still anchor the field: routine (reactive), preventive (planned, on structurally sound pavement), preservation (umbrella), rehabilitation, reconstruction.
• Evidence: $1 of preservation defers $4–5 of rehabilitation (idealised); Michigan rehab cost about 14× preventive work per lane-mile; Bedford, Texas microsurfacing $10,270 vs reconstruction $574,000 per lane-mile.
• Network simulations (NY, Michigan, Wisconsin) show “preservation first, then worst roads” beats “worst first” for the same budget.
• Data to track per treatment: pre-treatment condition, weather at placement, design and materials, exact limits (mainline vs shoulder), cost, then condition, ride and friction over time. Benefit/cost is the preferred evaluation method.

D2 — Low-cost pavement systems (PIARC 2025)

• A low first cost is not a low lifecycle cost; options must be ranked by net present worth including maintenance.
• Selection factors: road function and speed, service level, traffic and heavy-vehicle share, climate, budget and staging, local materials and equipment, lifecycle cost.
• Catalogue with service lives: gravel needs regrading every 3–5 years; slurry seal 3–5 years; microsurfacing 4–8 years; dressed stone about 30 years; RCC 18–29% cheaper than asphalt at first cost for AADT 500–5,000 (Turkish study), asphalt cheaper below AADT 20.
• Three-step decision framework: condition/go–no-go → technology shortlist → economic and financial ranking.

Sources

• D1 — TRB Webinar: Data-Driven Strategies for Efficient Pavement Systems (5 Sep 2025), slides.
• D2 — PIARC TC 4.1, Low-Cost Pavement Systems, 2025R01EN (2025).
• D3 — FHWA, Development of Next-Generation Pavement Performance Measures and Asset Management Methodologies To Support MAP-21 Performance Management Objectives, FHWA-HRT-23-102 (Sep 2024).
• D4 — FHWA, Next-Generation Pavement Performance Measures, FHWA-HRT-23-076 (Sep 2023).
• D5 — FHWA, Next-Generation Transportation Asset Management Methodology, FHWA-HRT-23-075 (Sep 2023).
• D6 — FHWA, Pavement Management Roadmap, FHWA-HIF-22-054 (Sep 2022).
• D7 — FHWA, Pavement Management Roadmap — Executive Summary, FHWA-HIF-22-055 (Sep 2022).
• D8 — TRB Webinar: How to Certify and Verify Pavement Surface Condition Data (10 Jun 2020), slides.
• D9 — FHWA/NHI, Pavement Preservation: The Preventive Maintenance Concept — Executive Overview, FHWA-HI-00-006 (Sep 1999).

Every figure on this page comes from the nine documents above and is attributed to the one it came from. The grouping, the reading paths and the judgements about what is useful to an Australian council are ours; none of them are claims made by FHWA, TRB or PIARC, and none of those bodies has reviewed or endorsed anything of ours.