Catch basin and drainage structure inspections: site work vs structure work, and how to keep both audit-defensible

Most stormwater coordinators know the difference between a site inspection and a structure inspection in their head. The same distinction does not always show up in their records. That gap is one of the more common reasons a clean field operation looks messy on paper during a state agency review.

This post is about the practical difference between site work and structure work, the structures most municipal MS4 programs are responsible for, the cadences they tend to run on, and what the documentation looks like when it holds up under audit.

Site inspections and structure inspections are different work

A site inspection is an inspection of a place. A construction site, a municipal yard, a commercial parcel with a stormwater treatment system. The unit of work is the parcel or facility. The inspector documents conditions across the whole site: erosion controls, BMP performance, housekeeping, spill response readiness.

A structure inspection is an inspection of an asset. A catch basin, a manhole, an inlet, an oil-water separator, an outfall structure. The unit of work is the individual asset. The inspector documents the condition of that one structure: sediment depth, grate condition, structural integrity, debris, evidence of illicit discharge.

The two overlap. A construction site inspection often includes looking at the catch basins protecting that site. A municipal facility inspection often includes the catch basins on that yard. But the underlying records are not the same record. The site has its own history. The structure has its own history. Treating them as a single record is where audit defensibility gets lost.

Two specific failure patterns come from confusing the two:

The inspector logs catch basin observations against the site, not the structure. Six months later, the structure record shows no inspection in that period. The work happened, but the structure-level audit trail is empty.
The inspector logs site-level findings against a single structure. A year later, the site record reads like one piece of equipment had nine maintenance issues. The site-level pattern is invisible.

The fix is not more discipline. It is having both record types available and clear about which one a given observation belongs to.

The structures most municipal programs are responsible for

The structure inventory varies by jurisdiction, but most MS4 programs are responsible for some combination of:

Catch basins with sumps that need periodic sediment removal
Curb inlets without sumps that need debris removal
Storm drain manholes for inspection, cleaning, and source tracing
Combined sewer overflow structures in older systems
Storm drain outfalls with their own inspection cadence under MCM 3
Oil-water separators at fueling stations and maintenance yards
Trash racks and debris booms at major intakes
Vault-style stormwater treatment systems (cartridge filters, hydrodynamic separators, underground detention)
Trench drains at fueling and washing facilities

Each of these has a different inspection logic. A catch basin sump checks for sediment depth against design capacity. An oil-water separator checks for oil layer thickness and separator function. An outfall structure checks for discharge indicators and physical integrity. A trash rack checks for debris loading.

Programs that try to use one inspection form for all structures end up with forms that are too generic to be useful in the field and too vague to defend in an audit.

Cadences that actually show up in permits

Cadence varies by permit and by structure type. Common patterns for Phase II MS4 programs:

Catch basins: annual inspection, with cleaning when sediment reaches a threshold (often 60 percent of sump capacity, depending on permit). Some programs split this into a screening visit (status check) and a cleaning visit (when triggered).
Storm drain manholes: opportunistic inspection during cleaning, IDDE source tracing, or system maintenance, rather than a fixed annual cadence.
Outfalls: quarterly dry-weather screening on a rotation, with all outfalls touched at some defined interval (often every five years for the full network). Detailed in the outfall and IDDE workflow.
Oil-water separators: quarterly, with maintenance triggered by oil layer depth or separator condition.
Vault treatment systems: annual inspection plus manufacturer-specified maintenance, often with a separate maintenance contractor record.
Trash racks and debris booms: post-storm inspection during the wet season, plus routine cleaning.

The cadence belongs in the system, not in someone’s head. A small program with two hundred catch basins cannot rely on a coordinator remembering which ones were inspected in March. Cadence-aware tracking is one of the places software earns its keep regardless of which tool. The general framing is in the stormwater inspection cadence guide.

What a defensible structure inspection looks like

The minimum useful record for any structure inspection:

Asset identifier for the specific structure, not a free-text description
Date and inspector, with GPS confirmation that the inspector was at the structure
Inspection type (annual, post-storm, complaint-driven, follow-up to deficiency, post-cleaning verification)
Structured findings against a checklist appropriate for that structure type
Measurements where the structure type calls for them (sediment depth, oil layer thickness, debris volume)
Photos of the relevant feature (sump for catch basins, separator outlet for OWS, grate condition for inlets)
Any deficiency identified, with severity and recommended response
Follow-up trigger if the finding requires re-inspection or maintenance work
Inspector sign-off with timestamp

For structures with maintenance triggers tied to measurement thresholds, the inspection should record the measurement, not just narrative. “Sediment depth approximately three quarters of sump capacity” is more defensible than “sediment present.” The measurement is what triggers the next maintenance action.

For more on what good photo evidence looks like specifically, see the photo and GPS evidence guide.

Stenciling, public outreach, and the structure inventory

Some structure work is part of MCM 1 public education, not just MCM 6 housekeeping. Storm drain stenciling (“Drains to stream, no dumping”) is the most common example. Stenciling events touch structures and produce records that belong on the structure as well as on the outreach event.

A clean way to handle this:

The outreach event has its own record under MCM 1 (date, audience, materials distributed, volunteer hours).
The structures touched during the event get a stenciling record that links back to the outreach event.
The structure inventory now shows which assets carry the current stencil pattern. When stencils fade and need refreshing, the structures due for re-stenciling are findable.

Programs that document stenciling only as an event lose the asset-level history. The next coordinator, two years later, has to walk the network to find which structures are stenciled and which are not.

Common failure patterns

Most structure inspection problems trace to a small number of patterns:

Inspections logged at the wrong granularity. The inspector logs “inspected ten catch basins on Pine Street.” Auditors cannot tell which ten, or whether the same ten were inspected last year, or whether structure 04CB17 has been inspected at all in the reporting period.

No structure inventory in the first place. The program inspects catch basins on the streets the inspector remembers. Structures the inspector has never been to are invisible to the inspection program. Recently annexed neighborhoods often have entire blocks of structures that nobody has ever inspected.

Maintenance and inspection treated as the same thing. A cleaning crew empties sumps. The cleaning gets logged but the inspection does not. The inspection logs that should show “sediment was 70% of capacity” get replaced by a maintenance log that shows “cleaned.” Both records are useful. They are not the same record.

Photos that do not identify the structure. The photo shows a catch basin. There are eight thousand catch basins in the city. Without GPS metadata or a visible asset tag in the frame, the photo is illustrative but not evidentiary.

Deficiencies that close without verification. A catch basin is flagged for sediment removal. The cleaning happens. Nobody re-inspects. The closure record says “complete” but the post-cleaning condition is undocumented.

Where software fits

Software does not stop a program from confusing site inspections with structure inspections. What software can do is make both record types easy to access in the field, automatically link photos and GPS to the right record, and surface what is overdue at the structure level.

NPDESTracker keeps site records and structure records as distinct first-class entities, both pinned on the GIS map and both with their own inspection history. A catch basin inspection on a structure pulls up the structure history; a site inspection on a parcel pulls up the site history. The two cross-reference where they overlap (a structure on a parcel knows what site it belongs to, a site shows the structures inside its boundary), but neither is a substitute for the other.

The full structure inventory is part of the GIS workspace, the inspection workflow that captures both record types is on the inspections page, and the audit trail that ties everything to the annual report is on the reporting page.

If your program is currently logging structure observations as part of site inspections (or the other way around), the easiest way to see how the two-record-type model fits is the interactive demo. Sample data, browse only, no signup.