Dam records: separate hazard potential and condition ratings
A dam classified high hazard is not necessarily in poor condition. Hazard potential concerns consequences if a failure or misoperation occurs; condition concerns the assessed state of the structure under the applicable criteria. Research should preserve both axes and their dates instead of turning one alarming word into a failure probability.
The focused deliverable is a US dam evidence card linked to a company’s actual ownership, operation or supply dependency. This example uses Hawaii definitions as a stated jurisdictional illustration, with sources read on October 25, 2026. It does not provide a structural assessment, engineering design or expected-loss estimate.
Locate definitions and data fields
Hawaii DLNR’s Hazard Classification page states that hazard potential describes incremental consequences and does not indicate failure likelihood or current condition. Its high, significant and low categories therefore belong on a consequence axis, not a probability scale.
Hawaii DLNR’s Condition Assessment page separately defines Satisfactory, Fair, Poor and Unsatisfactory using inspection-based criteria. Preserve the actual term and its context rather than translate every favourable-looking category into a blanket safe designation.
USACE’s National Inventory of Dams public-layer fields include NIDID, HAZARD_POTENTIAL, CONDITION_ASSESSMENT, CONDITION_ASSESS_DATE, LAST_INSPECTION_DATE and emergency-plan fields. The field list establishes what can be recorded; it does not supply the rating of the fictional dams below or certify that every public record is complete and current.
Build the two-dam main case
Assume a fictional company depends on Dam A for 20% of its annual electricity supply and Dam B for 5%. The dependency shares are explicitly invented and concern matched delivered-energy volumes, not ownership or generation capacity. The documented hypothetical records say A has high hazard potential and a recent Satisfactory condition assessment; B has low hazard potential and a Poor assessment.
| Dam | Hazard potential | Condition | Illustrative company dependency |
|---|---|---|---|
| A | High | Satisfactory | 20% of matched annual electricity |
| B | Low | Poor | 5% of matched annual electricity |
A’s consequence classification does not override its condition entry, and B’s lower consequence classification does not erase the recognised deficiency implied by its stated condition category. The card should retain separate questions about A’s consequences and B’s condition evidence.
Neither record supplies a numerical failure probability or restoration duration. Multiplying invented scores of one to three by 20% or 5% would not produce a defensible expected loss. The shares measure business dependency under the case assumptions, not the chance that the dam will fail.
Match dates before comparing condition
Record the condition-assessment date, last inspection, database update and research extraction separately. DATA_UPDATED is not automatically the date of an engineering assessment. A new database edit can change an address without establishing that a fresh structural review occurred.
For a boundary case, suppose B’s Poor rating is old and a company presentation says repairs are complete. Preserve the old rating and the dated repair claim, then request the relevant engineering or regulator follow-up. A completion statement alone does not establish that the official condition assessment has changed.
Conversely, absence of a recent public update does not prove no inspection occurred. Use condition not located or current assessment unverified rather than convert missing data into either Satisfactory or Poor. A rating’s age is an evidence limitation, not a self-generated deterioration rate.
Separate emergency planning and actual condition
An emergency-plan-prepared field can indicate a recorded planning status. It is not a substitute for the condition assessment or proof that a plan has been tested, covers every current downstream exposure or will achieve a specific recovery time. Obtain the relevant date and supporting document for the question being asked.
Do not infer an operating shutdown, repair cost or legal breach from one classification. Those conclusions require the actual applicable orders, engineering records, contracts or budgets. This card can identify a dependency needing further evidence without pretending to perform the engineer’s or regulator’s assessment.
Map the company connection
Match the dam identifier, location and legal entity to the issuer’s disclosed role. Owner, operator, water customer and electricity purchaser are different relationships. A company using power associated with a reservoir does not automatically own the dam or bear its entire repair obligation.
The hypothetical 20% and 5% shares can only be combined to 25% if they measure non-overlapping supply for the same period and company boundary. They are not interchangeable with percentage of asset value or revenue. Alternative supply and contract treatment remain unknown unless separately documented.
Use the evidence-card check
| Field | Obtain | Check |
|---|---|---|
| Dam identity | NIDID and agency record | Name variants and associated structures |
| Hazard | Applicable jurisdictional classification | Consequence distinct from probability |
| Condition | Rating and assessment record | Exact category, scope and date |
| Planning | Emergency-plan status and revision | Planning field not treated as structural proof |
| Company role | Ownership, operation or supply records | Actual duty and dependency scope |
| Missing evidence | Engineering or regulator follow-up | No invented failure, cost or downtime |
When state, inventory and company records differ, compare dates, structure identifiers and the purpose of each record. A main dam and an associated structure can be separate entries. Do not average labels or silently choose the most recent corporate statement over a different regulatory record.
Update after an actual new assessment, inspection evidence, documented repair review, ownership change or revised dependency. Preserve old and new versions with their authors. The conclusion should state consequence classification, assessed condition and company relationship separately, with specific evidence gaps instead of a false numerical risk ranking.
Make the evidence card usable in a portfolio review
Add a relationship identifier to each dependency entry. If the same annual electricity is purchased through a supplier and also described as reservoir-linked power, those are potentially two descriptions of one supply path. They cannot both be added to exposure merely because they appear in separate disclosures. The 25% total in the case requires distinct delivered-energy streams and a common denominator.
Create a dated follow-up queue for the two dams. For A, the next research question may be whether the consequence classification and emergency information cover the same structure and downstream area. For B, it may be whether a post-repair assessment exists. Different questions lead to different requested documents; ordering them by the word high alone could overlook B’s missing condition evidence.
A later rating should be entered as a new observation. Suppose B receives an actual new Satisfactory assessment after the identified repair. The condition axis changes from the old Poor observation to the newly documented category, with its date and author. The low hazard classification and the 5% company dependency do not change automatically. Conversely, a revised supply contract can change the dependency while leaving both engineering fields untouched.
This separation makes updates selective and checkable. It also prevents an apparent improvement in a combined score from concealing whether the underlying change was structural evidence, consequence classification, or the company’s purchasing boundary.
FAQ
Does high hazard mean high probability of failure?
No. The cited Hawaii definition explicitly separates those concepts.
Is an unreported condition the same as Satisfactory?
No. Missing or outdated evidence remains unknown.