Asset inventory and parcel mapping: the field work everything rests on

Every entitlement in a RAP traces back to a measurement taken once, in a field, by someone working through a queue of parcels before the light goes. If that measurement is wrong, everything computed from it is wrong, and by the time anyone notices the asset has usually been cleared.

Three records, not one

The field exercise produces three distinct things and they should be structured as three linked records rather than one form.

• The parcel — a boundary, an area, a location, and the tenure under which it is held.
• The assets on it — structures with dimensions and construction type, crops and trees by species and maturity, wells, fences, graves.
• The claims against it — every person with an interest, by claim type, which on customary land will be more than one.

Collapsing these into a single flat form is the most common design error in enumeration, and it is the reason so many registers cannot represent a tenant, a sharecropper or a second cultivator.[1]

Boundaries

Handheld GNSS is accurate enough for area computation on agricultural parcels and is not accurate enough to settle a boundary dispute between neighbours. Treat the survey as a record of what was agreed on the day, not as an authoritative cadastral determination:

1. Walk the boundary with the claimant and the adjacent holders present.
2. Record the trace, not just corner points, where the boundary is irregular.
3. Capture who was present, by name, and have them acknowledge the trace on the spot.
4. Photograph boundary markers — planted trees, ditches, stones — with their positions.
5. Flag any boundary the neighbours did not agree, immediately, as a dispute rather than recording a number and moving on.

That last step is the one that saves the programme time. An unresolved boundary recorded as resolved becomes a payment blocked at the tail of the schedule, months later, when the enumerator is gone.

The parcel record is not a survey product. It is evidence of an agreement about a boundary, made in front of witnesses, on a date.

Structures

Structures need enough detail to support a replacement-cost build-up rather than a rate per square metre: external dimensions and number of rooms, wall, roof and floor materials, foundation type, storeys, condition, and the use to which the structure is put.[2] Ancillary works — latrines, kitchens, stores, fences, water tanks, paving — carry real cost and are the items most often omitted, because they are not what the enumerator is looking at.

Photographs matter more here than anywhere else: at least one of each elevation and one of the interior, geolocated and dated. A completion audit that questions a valuation is asking what was actually there, and photographs answer it in a way a dimension table cannot.

Crops and trees

Record species, count and maturity class, and record the crop holder separately from the parcel holder. For dense stands, count by defined sample plots with the method recorded, rather than by an enumerator's estimate — an estimate is unreproducible and will be challenged.[3]

Offline is the normal case

Enumeration happens where there is no network. A system that requires connectivity at the point of capture is unusable for the work, which is why paper persists.

The consequence for tooling is specific: the device must hold the complete local record and reconcile later, and reconciliation must handle conflicts explicitly rather than by overwriting. Two teams enumerating adjacent areas will eventually enumerate the same parcel; last-write-wins silently discards one team's field observation, and nobody discovers it until the counts disagree.

Verification before the data is used

Three checks catch most of what goes wrong, and all three are cheap if run while the team is still mobilised:

• Geometric. Overlapping parcels, parcels outside the corridor, implausible areas, structures whose coordinates fall outside their parcel.
• Logical. Records with no claimant, claimants with no contact details, structures with no dimensions, maturity classes absent on perennials.
• Field re-check. An independent re-measurement of a random sample, by a different enumerator, with the variance reported. This is the only check that detects systematic bias in a team, and it is the one most often skipped.

Give the household its copy

A signed inventory sheet left with the household at enumeration is the cheapest dispute-prevention measure available. It gives the household a contemporaneous record of what was recorded, which converts a later disagreement from a contest between memories into a comparison of two documents.[4]