How to rate a single-ply roof with the USACE ROOFER method

What does a USACE ROOFER roof condition rating measure?

Rate each single-ply roof section separately by recording distress type, severity and quantity, converting those quantities to the densities defined in USACERL's April 1993 ROOFER manual, and applying its deduct-value and correction curves. Subtract each component's maximum corrected deduct value from 100 to obtain the membrane condition index, MCI, and flashing condition index, FCI. Keep the measured repair quantities beside both scores and resolve inspector disagreements against the same defect definitions before issuing the rating. [S01, pp. 17 to 21]

USACERL Technical Report FM-93/11, April 1993, produces separate 0 to 100 indexes for observable membrane and flashing condition. Its distress deductions address component function, maintenance need and waterproof integrity. A deduct value is the penalty read from a distress curve; a corrected deduct value adjusts combined penalties using the component's correction graph. [S01, pp. 7 to 10, 20]

Here FCI means flashing condition index. It is distinct from the facility-management ratio of deferred maintenance cost to replacement value. The report discusses EPDM, PVC and Hypalon; it neither names TPO nor provides a TPO-specific validation dataset. Identify the material before choosing a historical score or a documented modern-system adaptation. [S01, pp. 7 to 8]

FM-93/11 directs insulation and combined roof-condition evaluation to M-90/04; averaging MCI and FCI does not produce that overall ROOFER index. ERDC's current ROOFER description evaluates insulation through moisture surveys and gravimetric core analyses. Enter untested insulation as not assessed. [S01; S02, FM-93/11 p. 8]

ASTM C1153-23 covers nighttime infrared location of wet insulation above the deck and in contact with waterproofing, with invasive verification; moisture location does not establish its entry point. ASTM D8231-24a addresses electronic breach detection alongside other inspection methods. Keep those results with the visual rating and use electronic leak detection for the compatible assembly. Carry the resulting work scope into commercial roof capital planning. [S07; S08]

How do you define the roof section and its denominators?

USACERL FM-93/11 recommends sections of 25,000 to 40,000 square feet for large roofs without obvious physical divisions. First separate different levels, systems, construction dates and materially different conditions beneath the roof. Combine similar small canopies only when age, construction, structural system and conditions below the roof match. [S01, p. 10]

Measure the whole section area and complete flashing length, including curbed projections, penthouses and courtyards. Using only distressed runs in the flashing denominator inflates density. The roof measurement guide explains takeoff geometry; a roofing square equals 100 square feet, while this calculation uses square feet. [S01, ch. 3]

Draw a scaled plan with north, boundaries, dimensions, details and access. Keep section IDs and plan revisions stable across surveys.

Keep inspector and recorder roles, material, attachment, installation history, warranty identifier, applicable manufacturer details, weather and visibility exclusions in the header. Mark each quantity as directly measured or expanded from sampling. Figure 7 supplies the historical header above; its July 27, 1992 survey date stays distinct from this page's November 5, 2026 review date. [S01; S03; S05]

How do you survey the entire roof, including ballasted seams?

USACERL FM-93/11 calls for at least two people, an inspector and a recorder, with membrane coverage in strips 10 to 15 feet wide. The recorder assists measurements and observes safety; that role does not establish the roof's fall-protection arrangement. Plan authorized access and protection using the roof fall-protection guide before walking the section. [S01, pp. 14, 16]

Survey perimeter flashing, other flashings and penetrations, then the membrane. Assign each finding a plan location, defect number, severity and quantity; Anvilfield adds a photograph showing the measurement. Postpone the rating walk when snow or extensive ponding conceals the section, and mark inaccessible areas as uninspected. [S01, pp. 14 to 17]

USACERL's April 1993 ballasted-roof procedure begins with five 5-foot seam checks and, when a defect is found, expands to 2% seam sampling for sheets up to 10 feet wide or 4% for wider sheets. With no defects in the initial checks, the manual treats the remaining field seams as satisfactory. The systematic sample uses 2-foot exposures every 100 seam feet for sheets up to 10 feet wide, or every 50 seam feet for wider sheets. Keep screening observations separate from the systematic sample totals. Record actual sampled length and defect length before applying the matching expansion factor; restore the ballast afterward. [S01, pp. 16 to 17, 57]

Selecting only suspect seams does not create a representative density. For broadly uniform distress, the manual permits representative samples, such as 1,000 square feet of membrane or 100 feet of flashing. Map both the sample and the affected population so the expansion can be checked. [S01, p. 16]

Walk the interior and exterior too, recording occupant leak reports, ceiling marks, exterior staining and drainage observations. Connect them to the plan without assigning an unseen membrane defect from a ceiling stain. For exposed-roof seams, Carlisle's June 26, 2026 pre-inspection guide calls for all-seam inspection and additional TPO/PVC probing; keep its installation guidance and the single-ply seam QA procedure distinct from historical ballast sampling. [S01; S12]

How do you classify membrane and flashing distress severity?

Classify the location first: field-sheet distress affects MCI; defects in edge and curb flashings, flashed penetrations and drains affect FCI. Use the named defect and its measurement rule, then assign severity. USACERL FM-93/11 uses the highest severity of the same distress within a localized area; entering that area again at a lower severity duplicates the finding. Preserve the defect number so another inspector can trace the classification. [S01, pp. 14, 17]

Identify the attachment system before classifying securement; system type and edge distance change its thresholds.

For DS, water entry through the seam makes the defect high even when the visible opening is narrow. Three location rules prevent common miscoding. A compatible repair in membrane base flashing is BF low, while a similar-material field repair is not PA. A hole within 2 feet of a drain or roof-level scupper is included in DR high; use the specific detail definition before selecting HL. [S01, pp. 23, 45, 71]

Partially adhered securement distinguishes isolated failed attachment points, medium, from adjacent failed points, high, and measures effective unsupported area. The historical 10-foot perimeter zone in the securement definition is a distress-classification zone. Current wind-design zones and current flashing-height requirements require their own governing details. [S01, p. 65]

Separate splits from mechanically caused cuts, membrane support from securement, and coating loss from membrane loss. Use the full defect lists at the linked curve pages for conditions omitted from the summary table. Record rainfall timing when classifying ponding, and use the penetration-flashing guide to identify the physical detail that needs repair. [S01, pp. 50, 54, 61, 63, 69, 78]

What goes in a measured-distress worksheet?

Keep observation, method quantity and repair quantity in separate columns. A split over an embedded-metal joint contributes one scoring foot, but repair dimensions come from the actual detail. A distressed drain contributes a count divided by total flashing length, rather than a fraction of the roof's drain count. [S01, pp. 36, 45 to 46]

For defective field seams, USACERL FM-93/11 divides distressed seam length in feet by roof-section area in square feet and multiplies by 100, rather than dividing by total seam length. The manual's density label includes mixed units as well as area percentages. Sum entries by component, code and severity only after preserving each mapped raw finding. [S01, pp. 57 to 58]

Use separate statuses for not inspected, not applicable, observed and clear, and measured zero. These are Anvilfield export conventions. Keep distress_code DV, debris and vegetation, separate from deduct_value so the two uses of DV cannot collide in a spreadsheet.

Copyable CSV header

section_id,survey_date,edition,roof_area_sq_ft,total_flashing_ft,item_id,location,component,distress_code,defect_number,severity,raw_quantity,raw_unit,sample_fraction,expansion_factor,method_quantity,density,curve_page,deduct_value,repair_quantity,repair_unit,repair_detail,photo_id,disposition

Convert metric measurements before historical curve lookup. NIST SP 811, 2008 Appendix B, defines the international foot as exactly 0.3048 metre. Keep full conversion precision in the calculation and round displayed measurements to two decimals. [S19]

Anvilfield estimates the published 15-foot DS entry's density as 0.43 using 3,462 square feet. Entering the equivalent 4.57 metres and 321.63 square metres directly gives 1.42, the wrong numeric input for that historical curve. Method note: the area conversion is exactly 0.3048 squared, or 0.09290304 square metre per square foot. Compare 100 times length divided by area in each unit system; length/area ratios change numerically with units. [S01; S19]

Keep density numeric, without an extra spreadsheet percentage conversion.

How do you calculate MCI and FCI from distress density?

Multiply distress quantity by 100 and divide by its dictionary denominator, then read the matching type-and-severity curve in FM-93/11 Appendix A. Keep components separate.

Sort deductions from largest to smallest, then calculate successive groups: the largest alone, the largest two, and so on. For each group, retain the sum and q, the number of included deductions greater than one point. Read the corrected deduction from the component's correction graph; q counts deductions, not densities above 1%. [S01, pp. 20 to 21]

USACERL's April 1993 ROOFER manual calculates membrane and flashing condition indexes separately by subtracting each component's maximum corrected deduct value from 100. The maximum can occur before all deductions have been included. Subtracting the raw sum or keeping only the last corrected result changes the method. [S01, pp. 20 to 21]

Follow density vertically to its severity curve and read horizontally to the deduction axis. Distress-density axes are logarithmic; the correction plots have their own axes. Record graphical interpolation for q values between labeled curves, and retain the plot reference and reading. The published example supplies its own q = 3, 5 and 7 readings. [S01, Appendix A; Table 1]

Curve locator: printed pages in the full manual

Download FM-93/11 with Appendix A. Flashing: BF membrane base and BC coated-metal base, p. 81; MC metal cap and EM embedded edge, p. 82; FP penetrations and PP pitch pans, p. 83; DR drains/scuppers, p. 84; correction, p. 85. Membrane: SP splits and RG ridges, p. 86; HL holes/cuts/abrasions and DS seams, p. 87; SC coating and MD deterioration, p. 88; SS securement and MS support, p. 89; PA patching and DV debris/vegetation, p. 90; EQ supports and PD ponding, p. 91; correction, p. 92.

Some families lack one or more severity curves: SP is high only; RG low/high; SC, MS and PD low/medium; DV medium/high. Use the defined family without inventing a missing curve. An inspected component with no defined distress scores 100. Record inaccessible areas as uninspected, not defect-free; inputs beyond the plots need documented method/software handling. [S01, chs. 2, 4; Appendix A]

Retain the numeric score when it falls exactly on a named band boundary. Figure 1 does not supply a precision convention for assigning that shared boundary. Its descriptive bands do not establish an automatic replacement score; the capital guide's approximate planning cutoffs serve a separate decision. [S01, Fig. 1]

Where do current manufacturer instructions differ from the 1993 method?

Keep historical scoring and current repair compliance in separate fields. An unchanged historical rating preserves its calculation basis; a modified taxonomy needs a named adaptation. For TPO, use the actual manufacturer's inspection criteria and identify any application of the 1993 taxonomy as an adaptation unless applicability has been established. [S01, pp. 7 to 8]

Modern EPDM tape and PVC edge instructions expose specific conflicts with the historical blanket sealant definition.

Use the named product detail and issued warranty for repairs. Sikaplan and SENTINEL have different edge-sealant instructions; transfer neither instruction across PVC products. [S15; S18]

Ponding likewise has three separate records: historical distress classification, drainage-maintenance action and actual warranty coverage. Use ponding diagnosis for the cause, roof warranty terms for coverage, and seam QA for the product's inspection technique. A historical low-severity entry alone does not order replacement of sound modern construction. [S01; S03; S06]

Sealant sources: FM-93/11 p. 57; Carlisle 2020 splice discussion; July 2025 EPDM guide, section 8C.

Worked example: 3,462 square feet, MCI 62 and FCI 67

In the published 3,462-square-foot ROOFER example, maximum corrected deductions of 38 for membrane and 33 for flashing produce MCI 62 and FCI 67, both rated good despite high-severity seam and drain findings. The underlying survey is Building 97, Section B, July 27, 1992, published in April 1993. [S01, Fig. 7; Table 1, pp. 19 to 21]

Quantities and curve deductions above are the report's entries. Anvilfield estimates the displayed densities from those quantities and the 3,462-square-foot and 341-foot denominators, multiplying each ratio by 100. The source used its own rounded density entries for curve reading: 50 square feet produces 1.44 here and appears as 1.4 on the original form. [S01, Fig. 7]

The actionable record includes 15 feet of high-severity seams and one high-severity drain/scupper. Keep those mapped work items beside the good ratings. Figure 7 does not identify the securement system, so its 50-square-foot high SS entry remains the published classification without an inferred attachment mechanism. [S01, Fig. 7]

How do two inspectors compare and reconcile their ratings?

Give both inspectors the same boundaries, edition, reporting units and blank location grid. Each saves an independent detection, classification and measurement record before viewing the other's findings. Match IDs by location afterward, resolve the raw observations, then recompute both indexes. This independent-comparison workflow is Anvilfield's addition to the manual's definitions.

Compare detection, matched type/severity, quantity differences within common units, and final MCI/FCI point differences separately. Anvilfield's proposed issuance rule requires reconciliation of every through-opening disagreement, denominator mismatch and edition difference. A similar score can conceal a different water-entry work list.

A and B are desk-exercise roles. Each row isolates one error while retaining the other published inputs; the score rows use the published deductions. Anvilfield estimates the 25.83% density increase by dividing 341 by 271, subtracting one and multiplying by 100. Changed densities require fresh curve readings before calculating new scores.

Anvilfield estimates detection overlap at 75%, three matched rows divided by the union of four assigned rows, and matched type/severity agreement at 66.67%, two exact matches divided by three matched rows. The four pans form one aggregated exercise row. For field comparisons, use mapped individual defects and state the matched and total counts.

Remeasure disputed endpoints, establish whether a seam has a through-water path, and restore omitted method entries. If access or evidence cannot settle a dispute, retain its unresolved status and the affected work item; issue the rating as provisional. Save both originals and each disposition. Resolve severity before recalculation rather than averaging low/medium/high labels or final scores. Shared marked ballast exposures compare classification; independently selected exposures also introduce sampling variation, so record which design was used.

How do you turn a condition score into repair quantities?

Mapped defects define work quantities; the index describes component condition. Carry direct repairs separately from investigation areas. The published 50 square feet of securement distress requires attachment diagnosis, while its 40 square feet of ponding requires drainage diagnosis; neither quantity establishes wet-insulation area. [S01, Fig. 7; pp. 65 to 66, 78]

A method foot is not automatically a continuous repair foot. The EM worksheet contains six one-foot entries without establishing six damaged joints; BC and MC likewise need the original defect identities and repair geometry. Confirm flashing heights, termination, substrate and compatible materials through the actual flashing detail before ordering. [S01, Fig. 7; ch. 3]

For Carlisle's described improperly installed EPDM tape-splice repair, its July 2025 Field Guide section 8D.3 specifies a minimum 6-inch overlay centered over the splice edge. Anvilfield estimates 12.5 square feet of net strip coverage for 25 linear feet at that width, before end details, intersections and waste. Method note: multiply 25 feet by 0.5 foot; the actual segment layout determines order quantity. This is a separate Carlisle repair-detail calculation from the unidentified Building 97 membrane. [S17, p. 104]

Issue urgent defects beside the score. The owner sequences work by water entry and consequences; use capital planning for renewal economics, with moisture evidence where needed. A score alone determines neither load capacity nor remaining years.

How much survey time should you allow?

Anvilfield estimates eight person-hours for one rating of a small exposed-roof section: two people on the roof for three hours, plus two reporting hours. Two independent passes with reconciliation require 16 person-hours under the planning inputs below.

Method note: these Anvilfield estimates combine FM-93/11's two-person team with assumed field and reporting durations as of November 5, 2026. Multiply crew size by field hours, then add reporting and reconciliation labor. Each additional hour for the two-person team adds two person-hours. [S01, p. 14]

Set the allowance from access arrangements, detail count, ballast movement and the number of sections. Snow and standing water can defer observation; record rainfall timing for ponding and select comparable survey conditions. Scope moisture surveys, cores, repairs, lifts and laboratory work separately.

Define the survey territory and roof scope before setting route capacity. Group the survey territory by membrane material and building size with PorchRocket Ground Control so each route serves a defined roof-survey scope.

How often should you inspect and repeat the rating?

Carlisle's sample Care and Maintenance Guide F0001 Rev 12/24 calls for inspections at least every six months and after specified weather events. JM RS-7395 revision 3-26 specifies at least twice yearly or as the guarantee requires; Sika recommends spring/fall and severe-storm checks. Apply the roof's actual documents through its inspection and maintenance program. [S03; S04; S05]

Anvilfield proposes annual same-method condition scoring for a comparable owner trend, plus rerating after damage or major repair materially changes the section. This is a program choice separate from the OEM maintenance cadence. Keep geometry, edition, sample design and calculation basis consistent across rating dates.

Label measurement corrections, completed work and deterioration as different events. Correcting a denominator changes arithmetic rather than roof condition. Changed boundaries or interpretations require consistent recalculation of both dates before direct comparison.

Record as-found and post-housekeeping condition. FM-93/11 allows redistributed ballast and debris removed during inspection to affect the remaining scored distress; retain the action date and original observation. Calculate score change from comparable ratings, with repaired and new quantities beside it. Recalculate every density and correction after repair; nonlinear curve corrections mean that removing one defect does not add its original deduction directly to the score. [S01, pp. 65, 73]

Field checklist for a checked ROOFER worksheet

Deliver both indexes, unresolved findings, repair details and original notes together.

• ☐ Confirm authorized roof access, protection and walking conditions; record areas that cannot be inspected.
• ☐ Identify membrane, attachment, installed details and warranty; record FM-93/11, April 1993, as the scoring reference.
• ☐ Draw and verify section boundaries, dimensions, north and fixed defect locations.
• ☐ Measure section area and the complete flashing denominator, including curbed projections.
• ☐ Assign an inspector and recorder; retain their roles with the survey record.
• ☐ Complete the documented coverage and ballast-sampling sequence.
• ☐ Record each mapped defect's type, defect number, highest applicable local severity, quantity and units.
• ☐ Photograph the finding and measurement; preserve count, length and area separately.
• ☐ For sampling, save sample locations, population extent, sample quantity and expansion factor; restore moved ballast.
• ☐ Separate measured quantities, method scoring quantities and repair takeoff quantities.
• ☐ Group by component/type/severity, calculate density and identify the exact Appendix A curve.
• ☐ Retain successive sums, q counts and readings from the correct component correction graph.
• ☐ Use the maximum corrected deduction and record both MCI and FCI with the survey date and condition bands.
• ☐ Reconcile differing defect classifications, denominators and calculations; preserve the original entries.
• ☐ Identify current OEM detail differences and separately scope urgent openings, drainage and attachment work.
• ☐ Attach moisture or breach-location evidence where obtained; leave untested insulation unassessed.
• ☐ Record repairs and remeasurements as separate dated events; schedule the next owner-program rating.

Survey and calculation checks follow FM-93/11 chapter 2 and Appendices A/B. Photographic IDs, independent comparison and revision conventions are Anvilfield additions. [S01; S02; S03; S05; S07; S08]

Review record and update log

By Anvilfield. Reviewed 2026-11-05 against the 14 primary sources listed below. Public ASTM scopes support the moisture and breach-detection descriptions; the full acquired FM-93/11 report supports the rating procedure.

2026-11-05: final article review. Rechecked source editions, severity and sampling rules, published calculations and manufacturer details; clarified inspector comparisons and repeat-survey records, and revised survey staffing estimates.

Cite the worked calculation or inspector comparison by its section anchor to preserve context.

How to cite this page: Anvilfield, How to rate a single-ply roof with the USACE ROOFER method, 2026, https://anvilfield.com/field-guides/roofing/roofer-roof-condition-rating-worksheet/

Related tools

FAQ

People also ask

• What is roof asset management?
• How do you test a TPO seam in the field?
• How often should you inspect a commercial roof?
• What is a roofing square?
• How high should base flashing be on a commercial roof?

Sources checked

• S01. USACERL, ROOFER: Membrane and Flashing Condition Indexes for Single-Ply Membrane Roofs, Inspection and Distress Manual, FM-93/11, April 1993, ADA272573 Opened with curl; checked 2026-11-05. Full government report, including Appendix A and published worked example.
• S02. USACE ERDC, Sustainment Management System, ROOFER description, undated, checked 2026-11-05 Opened with curl; checked 2026-11-05.
• S03. Carlisle, Care and Maintenance Guide, F0001 Rev 12/24, within four-page Total Roofing System sample warranty Opened with curl; checked 2026-11-05.
• S04. Johns Manville, Protect Your JM Roof Guarantee / Prevent Roofing Problems, RS-7395 3-26, replaces 10-20 Opened with curl; checked 2026-11-05.
• S05. Sika Sarnafil, Commercial Roof Maintenance, undated webpage checked 2026-11-05 Opened with curl; checked 2026-11-05.
• S06. Sika Sarnafil, Warranty Information, undated webpage checked 2026-11-05 Opened with curl; checked 2026-11-05.
• S07. ASTM International, C1153-23, Standard Practice for Location of Wet Insulation in Roofing Systems Using Infrared Imaging Opened with curl; checked 2026-11-05. Public publisher scope only.
• S08. ASTM International, D8231-24a, Standard Practice for the Use of a Low-Voltage Electronic Scanning System for Detecting and Locating Breaches in Roofing and Waterproofing Membranes Opened with curl; checked 2026-11-05. Public publisher scope only.
• S12. Carlisle, Pre-Inspection Guide / Beat the Heat and Pass Inspection on the First Try, displayed date June 26, 2026 Opened with curl; checked 2026-11-05.
• S15. Sika, Sikaplan Roof System brochure, publication 12.112.5, revision 06/2025 Opened with curl; checked 2026-11-05.
• S16. Carlisle, SpecTips: Why risk your reputation with EPDM splice technology that has not stood the test of time?, displayed date October 14, 2020 Opened with curl; checked 2026-11-05.
• S17. Carlisle, EPDM Field Guide, July 2025 section 8, especially pp. 99 to 104 Opened with curl; checked 2026-11-05.
• S18. SOPREMA, SENTINEL CUT EDGE SEALANT product page, undated, checked 2026-11-05 Opened with curl; checked 2026-11-05.
• S19. NIST, Guide for the Use of the International System of Units (SI), SP 811, 2008 edition, second printing November 2008, Appendix B Opened with curl; checked 2026-11-05.