How to specify a road condition survey using TMH 13 and TMH 22

A road condition survey specification is the contract text that tells a survey contractor what to deliver so the data load into a road authority's register without rework. To specify a survey, write the road classes, device classes, survey frequency, reporting interval, kilometre datum, validation evidence and file format into the contract before the survey starts. The specification draws on the South African road manuals TMH 13 Part A (Committee Draft Final, May 2016) and TMH 22 Part D (Draft, March 2013) (COTO, 2016d) (COTO, 2013).

Key facts

• The South African road manual TMH 13 Part A describes in section A.4 how survey objectives become contract specifications, and states that its guidelines are not a complete specification.
• TMH 22 clause D.4.3 ties the permitted roughness device class to the road class of a paved road, and clause D.5.3 sets the years between surveys for each road class.
• In the South African road manual TMH 13 Part A, clause A.4.3.3 asks for a pre-validation report holding calibration certificates for all survey equipment, including reference devices, before validation starts.
• TMH 13 Part A clause A.4.4 says validation sections are selected and profiled before the survey, ideally by the agency or by a contractor other than the survey contractor.
• TMH 18, the South African road asset data exchange formats, has Draft Version 4 (October 2014) and Committee Draft 5.0 (February 2016), so a specification names the version it expects.
• COTO Standard Specifications (Draft Standard, October 2020), clause A20.1.5.5, set a six-month validation age for roughness, rutting and texture on road and bridge works unless the contract says otherwise.

What does a road condition survey specification have to settle?

A specification has to settle road class, device class, survey frequency, reporting interval, kilometre datum, validation evidence and file format. A missing line is hard to repair later, because the contractor chooses any datum, device or file format left open.

TMH (Technical Methods for Highways) is a series of South African road manuals now issued under COTO, the Committee of Transport Officials. TMH 22 is the road asset management manual and TMH 13 covers automated road condition assessments (COTO, 2013) (COTO, 2016d). Part A of TMH 13 says its guidelines are not a complete specification; they help a network manager check that no aspect is missing (COTO, 2016d).

A road class is a grade of road by function. In TMH 13, a device class is a grade of measuring instrument. A reporting interval is the length of road over which readings are averaged and stored as one record. A kilometre datum is the zero point and direction from which kilometres are counted. A validation section is a stretch of road measured by a reference device and then by the survey vehicle.

Which manual decides which line of the specification?

Four manuals supply the lines: road class, survey requirements, contract wording and file format.

• TRH 26 (Technical Recommendations for Highways), the South African Road Classification and Access Management Manual (Version 1.0, August 2012), classifies each road by function (COTO, 2012).
• TMH 22 Part D sets survey frequency and roughness device class by road class, and lists interval, lane and wheelpath per data type (COTO, 2013).
• TMH 13 Part A turns those requirements into contract text, including the pre-validation report and the validation sections (COTO, 2016d).
• TMH 13 Parts B to G (Committee Draft Final, May 2016) give equipment, validation and control guidance for positioning, roughness, rutting, skid resistance and texture, deflection and imaging. The explainer on the seven parts maps them.
• TMH 18 (Draft Version 4, October 2014; Committee Draft 5.0, February 2016) sets out the road asset data exchange formats in which the data arrive (COTO, 2014) (COTO, 2016i).

Narrated film · 1:30

Specifying a road condition survey

Ten specification lines are written one by one into an outline, then a map shows TMH 22 Part D requirements flowing into TMH 13 Part A and on to Parts B to G.

Read the film transcript

What must a South African road condition survey specification say, so the data load into the register without rework? It settles road class, device class, frequency, interval, kilometre datum, validation evidence and file format. Ten lines follow: the decision, each road's class, the frequency, the device class, then interval, lane and wheelpath. Then the kilometre datum, the pre-validation report, the validation sections and the validation age. The last two lines name the delivery format and TMH 18 version, then the reports and acceptance checks. TRH 26 classifies the road, TMH 22 Part D lists the requirements, and TMH 13 Part A turns them into contract text. Parts B to G of TMH 13 give equipment and validation guidance, and TMH 18 sets out the file formats. A kilometre datum is the zero point and direction from which kilometres are counted. In this synthetic example, the run starts three hundred metres early but is labelled kilometre twelve, the start of Road S1. Every value shifts, yet the count still reads sixty-four of sixty-four, so check position as well as count. TMH 13 Part A sets no acceptance limits or penalties, so the contract states the figures and every edition it adopts. A missing line is hard to repair later, because the contractor chooses whatever the contract leaves open. Next, open TMH 13 Part C on validating a roughness device, and TMH 18 on the returned files.

How do you specify a road condition survey, step by step?

Specify the survey in ten steps; each cites a named clause and reprints no manual table.

1. Name the decision the data will support. State whether the survey feeds network-level planning or project-level design, which roads it covers and whether sampling is intended (TMH 13 A.4.1). The answer sets the precision the device must offer (COTO, 2016d).
2. Give every link its TRH 26 class. Classify each road link, a uniform stretch of road, by function, using classes such as R1 to R5 for rural and U1 to U5 for urban roads (TRH 26 sections 4.1 and 4.3), and print the class beside the link in the survey schedule. TMH 22 B.3.1 and Part D use the same classes (COTO, 2012) (COTO, 2013).
3. Fix the survey frequency from the class. Take the years between surveys by road class and data type from TMH 22 D.5.3 (Tables D-4 and D-5). Add the season: TMH 13 A.4.1.4 advises surveying a section at about the same time each year to limit seasonal effects (COTO, 2013) (COTO, 2016d).
4. Choose the device class for each measurement. TMH 22 D.4.3 ties the permitted roughness device class to the road class of a paved road. Name one device type for the whole contract, ideally a multi-year one, because results from different types may not be interchangeable (TMH 13 A.4.2, A.4.3) (COTO, 2013) (COTO, 2016d).
5. Fix the reporting interval, lane, wheelpath and direction. TMH 22 D.5.4 and Table D-8 list, for each data type, the survey direction, lane, wheelpath, test interval and stored interval; copy the entries for the surveyed data types. D.5.4 sends visual segment lengths to TMH 9; name its edition, because TMH 22 predates the May 2016 Committee Draft Final. TMH 13 A.4.3.5 adds start and end locations (COTO, 2013) (COTO, 2016d).
6. Fix the kilometre datum and direction. State whether files use the register's kilometres or the surveyed distance, whether kilometres restart at zero on each road section, where the distance measuring instrument (DMI) is reset, which landmarks are checked and that line direction matches kilometre direction. TMH 22 B.1.3 and B.1.4 record that register kilometres can differ from surveyed ones (COTO, 2013), and TMH 13 B.3.3.3 covers DMI validation (COTO, 2016c).
7. Require the pre-validation report. Ask for it before validation starts, with current calibration certificates for all survey equipment including reference devices. Validation begins only once the network manager is satisfied that calibration is current and the device is capable of the required accuracy (TMH 13 A.4.3.3) (COTO, 2016d).
8. Name the validation sections and the validation age. State who selects and profiles the sections before the survey, ideally not the survey contractor (A.4.4), the acceptance criteria and the validation age (A.4.3.3). For road and bridge works, Chapter 20 of the COTO Standard Specifications (Draft Standard, October 2020), clause A20.1.5.5, sets six months for roughness, rutting and texture validation unless the contract says otherwise; adopt it or name another period. Add control testing (A.4.3.4) (COTO, 2016d) (COTO, 2020a).
9. Name the delivery format and the TMH 18 version. Write "TMH 18 Draft Version 4, October 2014" or "TMH 18 Committee Draft 5.0, February 2016", then the file types, required fields and the unit of every field. Require the lane code, which TMH 18 leaves empty only for data that are not lane specific. The combined instrument data (.cid) file holds roughness (International Roughness Index, IRI), rut depth and texture (mean profile depth, MPD) measured at the same time and place (section 5.6 of both versions) (COTO, 2014) (COTO, 2016i).
10. Require the plan and the reports, and state the acceptance checks. Ask for a contract quality plan (A.4.3.6), pre-validation, validation, summary and progress reports with submission deadlines (A.4.3.8), and safety arrangements (A.4.3.2). TMH 22 D.5.7 and D.5.8 describe data quality control and data quality acceptance as separate activities (COTO, 2016d) (COTO, 2013).

What happens when the kilometre datum is wrong?

A wrong kilometre datum can shift every value along the road while the record count still passes. Synthetic example, every value invented: Road S1 is defined in the network definition as 6.400 km long with increasing kilometres. Assume a 100 m reporting interval; the real interval comes from TMH 22 D.5.4. The expected count is 6,400 m / 100 m = 64 records per direction.

Quantity	Synthetic value
Register road	km 12.000 to km 18.400
Expected records, one direction	64
Run starts at the previous junction	km 11.700, labelled km 12.000
Label error	300 m, or 3 intervals
Records inside the road window	64

Because the contractor labelled the run start km 12.000, every value sits 300 m further along the road than where it was measured. Section 5 of both TMH 18 Draft Version 4 and Committee Draft 5.0 says values outside a road's start and end kilometres are discarded. The last 300 m of real data are therefore dropped, and the first 300 m of the file describe the road before S1 (COTO, 2014) (COTO, 2016i).

The count still reads 64 of 64, so a count check passes. A position check, using a bridge at a known register kilometre or record coordinates against the road geometry, exposes it; step 6 guards against the shift. How a run becomes a record is followed in from a profiler run to a register value.

What do TMH 13, TMH 22 and TMH 18 not decide?

TMH 13 Part A does not set acceptance limits, validation statistics or a penalty scheme; A.4.3 lists what a specification should cover (COTO, 2016d). The measurement parts give guideline figures, such as Part C Table C.9 for roughness validation, and A.4.3.3 has the specification let the network manager amend certain requirements under special conditions, so the contract states the figures it adopts (COTO, 2016a) (COTO, 2016d).

The TMH series asks for the latest edition but does not decide which edition a contract adopts. Committee drafts carry no legal standing (Document Versions, TMH 13 Part A), so the contract must name every edition, including the TMH 18 version (COTO, 2016d). Nor do the manuals decide which TMH 18 version the receiving system reads; ask that system's owner. Contract wording is for the client's legal adviser.

What are the common misreadings when specifying a survey?

Guard against five misreadings: class, device-class coverage, the name TMH 20, deflection units, and repeatability versus accuracy.

Three meanings of "class". A TRH 26 road class describes function, not traffic volume, speed or condition (sections 4.1 and 4.2). A device class grades an instrument. The older pavement management classes A to D, which TRH 26 section 3.14 traces to documents including TRH 22 (Draft, 1994), should no longer be used for that purpose (COTO, 2012).

Where device classes exist. TMH 13 classifies roughness devices in Table C.1, texture profilers in Table E.1 and imaging systems in Tables G.1 and G.4. Part F instead gives a minimum equipment specification for deflectometers, and F.2.5 has the road owner select the equipment type first. A supplier's description of a device is not validation evidence (COTO, 2016a) (COTO, 2016f) (COTO, 2016h) (COTO, 2016g).

"TMH 20". TMH 22 D.4.3 and TMH 13 B.4.2 point to a "TMH 20" for reporting formats, which the TMH 22 reference list titles "Standard for Road Network Data", the running head of TMH 18 Draft Version 4. TMH 13 C.5.3 points roughness reporting to TMH 18. Write "TMH 18" and its version (COTO, 2013) (COTO, 2016c) (COTO, 2016a) (COTO, 2014).

Deflection units. TMH 22 D.4.5 reports deflections in micrometres, while the TMH 18 field definitions in both versions are labelled in nanometres. State the unit for every deflection field (COTO, 2013) (COTO, 2014) (COTO, 2016i).

Repeatability is not accuracy. TMH 13 B.3.1 names both as the aspects validated. A specification that asks only for repeat runs leaves bias untested, as the insight on repeatability, bias and uncertainty explains (COTO, 2016c).

What evidence should be kept from specification to delivery?

Keep six records:

• The specification, naming each edition adopted and the TMH 18 version.
• The survey schedule: link, TRH 26 class, start and end kilometre, direction and lane codes.
• A table of device class, frequency, interval, lane and wheelpath per data type.
• The datum note: kilometre basis, reset point and landmark list with register kilometres.
• The pre-validation report, calibration certificates and validation sections with reference survey dates.
• Validation, control-test and summary reports, each dated.

The insight on survey acceptance evidence covers the receiving end.

The procedure, step by step

1. Name the decision the data will support

   State whether the survey feeds network-level planning or project-level design, which roads it covers and whether sampling is intended (TMH 13 A.4.1). The answer sets the precision the device must offer.

2. Give every link its TRH 26 class

   Classify each road link, a uniform stretch of road, by function with TRH 26 classes such as R1 to R5 and U1 to U5 (sections 4.1, 4.3). Print the class in the survey schedule; TMH 22 B.3.1 uses them.

3. Fix the survey frequency from the class

   Take the years between surveys by road class and data type from TMH 22 D.5.3 (Tables D-4 and D-5). Add the season: TMH 13 A.4.1.4 advises surveying a section at about the same time of year to limit seasonal effects.

4. Choose the device class for each measurement

   Choose the roughness device class that TMH 22 D.4.3 permits for each paved road class. Name one device type for the whole contract, ideally a multi-year one, because results from different types may not be interchangeable (TMH 13 A.4.2, A.4.3).

5. Fix the reporting interval, lane, wheelpath and direction

   Copy the direction, lane, wheelpath, test and stored intervals for each data type (TMH 22 D.5.4, Table D-8). Name the TMH 9 edition for visual segment lengths; TMH 22 predates May 2016. Add start and end locations (TMH 13 A.4.3.5).

6. Fix the kilometre datum and direction

   State the kilometre basis (register or surveyed distance), whether kilometres restart at zero on each section, where the distance measuring instrument (DMI) is reset, the landmarks checked, and matching line and kilometre direction (TMH 22 B.1.3, B.1.4; TMH 13 B.3.3.3).

7. Require the pre-validation report

   Ask for the report before validation starts, with current calibration certificates for all survey equipment, including reference devices. Validation begins only when the network manager is satisfied calibration is current and the device can reach required accuracy (TMH 13 A.4.3.3).

8. Name the validation sections and the validation age

   Name who selects and profiles the sections, ideally not the survey contractor (A.4.4), acceptance criteria and validation age (A.4.3.3). COTO Standard Specifications A20.1.5.5 sets a six-month default for roughness, rutting and texture on road and bridge works. Add control testing (A.4.3.4).

9. Name the delivery format and the TMH 18 version

   Write TMH 18 Draft Version 4 (October 2014) or Committee Draft 5.0 (February 2016), file types, required fields and every field's unit. Require the lane code, which TMH 18 leaves empty only for data that are not lane specific.

10. Require the plan and the reports, and state the acceptance checks

    Ask for a contract quality plan (A.4.3.6), pre-validation, validation, summary and progress reports with submission deadlines (A.4.3.8), and safety arrangements (A.4.3.2). TMH 22 D.5.7 and D.5.8 describe data quality control and data quality acceptance as separate activities.

Frequently asked

What goes into a TMH 13 survey specification?

A TMH 13 survey specification, the contract text for an automated road condition survey in South Africa, covers equipment, safety, calibration and validation, measurement control, survey procedures, a contract quality plan, the data reporting format and reports. TMH 13 Part A, section A.4.3, lists these aspects but states that its guidelines are not a complete specification, so the contract states the figures it adopts.

Which device class does a road class need under TMH 22?

TMH 22, the South African road asset management manual (Draft, March 2013), clause D.4.3, ties the permitted roughness device class to the road class of a paved road, so the road class must be fixed first. The road class itself comes from TRH 26, the South African road classification manual, which classifies roads by function. Read D.4.3 for the pairing; this guide does not reprint it.

What is a pre-validation report in a TMH 13 survey?

A pre-validation report is a document the survey contractor supplies before validation starts. TMH 13 Part A, the South African road manual on automated condition assessments, clause A.4.3.3, says it should hold calibration certificates for all survey equipment, including reference devices, and that validation should begin only once the network manager is satisfied that calibration is current and the device can reach the accuracy the specification needs.

How long does a road survey validation stay valid?

The South African road manual TMH 13 Part A, clause A.4.3.3, asks the specification to state how long a validation remains valid but gives no figure. For tests on pavements in road and bridge works, the Committee of Transport Officials (COTO) Standard Specifications (Draft Standard, October 2020), Chapter 20 clause A20.1.5.5, set six months for roughness, rutting and texture validation unless the contract documentation specifies otherwise. Name the period in the survey contract.

Which TMH 18 version should a survey specification name?

TMH 18, the South African road asset data exchange formats, exists as Draft Version 4 (October 2014) and Committee Draft 5.0 (February 2016). Name the version the receiving system reads. The two differ in file extensions, geometry format and fields, for example the authority identifier in the combined instrument data (.cid) file. TMH 18 itself does not say which version a receiver uses.

Why must the kilometre datum be written into the specification?

A kilometre datum is the zero point and direction from which kilometres are counted, and register kilometres can differ from surveyed distance, as clauses B.1.3 and B.1.4 of the South African manual TMH 22 note. Without a stated datum, reset point and landmark checks, every record can arrive shifted while the count still matches. TMH 18 discards values outside a road's start and end kilometres, so a shifted file can silently lose real data.

Where this connects

Next, open TMH 13 Part C on validating a roughness device against reference sections, and TMH 18 on the files the survey returns. The lane and wheelpath conventions fixed in step 5 are compared in lane and wheelpath conventions across the manuals. Last reviewed 21 October 2026 against the editions named above.

References

1. COTO (2016d). TMH 13: Automated Road Condition Assessments. Part A: General. Committee Draft Final, May 2016. Original source.Sections, tables and clauses are cited inline in each guide
2. COTO (2013). Draft TMH 22: Road Asset Management Manual. Draft, March 2013. Original source.Sections, tables and clauses are cited inline in each guide
3. COTO (2012). TRH 26: South African Road Classification and Access Management Manual (RCAM). Version 1.0, August 2012. Original source.Sections, tables and clauses are cited inline in each guide
4. COTO (2014). TMH 18: Road Asset Data Electronic Exchange Formats. Draft Version 4, October 2014. Original source.Sections, tables and clauses are cited inline in each guide
5. COTO (2016i). TMH 18: Road Asset Data Electronic Exchange Formats. Committee Draft 5.0, February 2016. Original source.Sections, tables and clauses are cited inline in each guide
6. COTO (2020a). Standard Specifications for Road and Bridge Works for South African Road Authorities. Draft Standard, October 2020. Original source.Sections, tables and clauses are cited inline in each guide
7. COTO (2016c). TMH13: Automated Road Condition Assessments. Part B: Positioning. Committee Draft Final, May 2016. Original source.Positioning and distance measurement guidance
8. COTO (2016a). TMH13: Automated Road Condition Assessments. Part C: Roughness. Committee Draft Final, May 2016. Original source.Sections C.2–C.5; Table C.5 is for calibration-section selection
9. COTO (2016f). TMH 13: Automated Road Condition Assessments. Part E: Skid Resistance and Texture. Committee Draft Final, May 2016. Original source.Sections, tables and clauses are cited inline in each guide
10. COTO (2016g). TMH 13: Automated Road Condition Assessments. Part F: Surface Deflection. Committee Draft Final, May 2016. Original source.Sections, tables and clauses are cited inline in each guide
11. COTO (2016h). TMH 13: Automated Road Condition Assessments. Part G: Imaging. Committee Draft Final, May 2016. Original source.Sections, tables and clauses are cited inline in each guide

Use the manual edition and project requirements that apply to your work. Worked examples use synthetic values and are not site-specific design advice.