A filled terrain void is an estimated value, not a new observation. When planning a UG73 terrain deliverable, request the original invalid-data mask, a record of the filling operation and an explicit rule for where estimated values may be used. A continuous-looking surface can be easier to display while still containing areas that need a different acceptance decision.

A complete picture is not a complete evidence record

A reviewer opens a hillshade and sees an uninterrupted surface. Earlier in the workflow, the same area contained a gap. Has the project acquired better evidence, or has a processing operation supplied an estimate? The final image alone may not answer. This is an original review scenario, not a description of a UG73 deployment or an assertion that any particular deliverable was processed incorrectly.

The procurement issue is not whether interpolation is always undesirable. A customer may deliberately request an estimated surface for a defined purpose. The issue is whether the customer can identify that decision afterward. If the only surviving artifact is the final surface, the distinction between the original information and the later estimate becomes harder to review. Make that distinction part of the scope before accepting the package.

What a fill operation establishes

The GDAL raster fill-nodata documentation describes interpolation into invalid pixels using surrounding valid pixels. It exposes choices including a mask, search distance and interpolation method; smoothing is also configurable. These are processing controls, not evidence that the filled terrain matches the physical surface. The documentation applies to that GDAL command, not to an assumed software package supplied with UG73.

Even the values present before filling need their own provenance and quality assessment. In a derived terrain product, valid does not automatically mean directly measured or independently verified. Our buyer recommendation is therefore to label the additional fill operation precisely, rather than presenting the original raster as unquestionable ground truth. The useful question is what changed, why it changed and what evidence supports the intended use.

Decide how each gap should be handled

A gap can lead to several different scope decisions. The buyer may leave it explicitly unavailable, accept a bounded estimate, request additional evidence, or exclude an area from the intended analysis. These choices have different consequences for the final report. Avoid allowing a default software setting to make that decision silently on behalf of the customer.

Start by asking what the recipient plans to do at the gap location. A contextual illustration and a consequential terrain calculation do not necessarily justify the same acceptance approach. The responsible specialist should define the applicable quality requirements. This article does not prescribe a universal maximum gap size, search distance, terrain tolerance or engineering use; those values cannot be inferred from an aircraft model name.

Ask for a paired sample

A useful acceptance sample contains the pre-fill surface, the post-fill surface and a way to locate the cells affected by the operation. Identify the sample's purpose and limitations in plain language. The recipient should not need to reconstruct the entire processing history simply to determine whether a point was affected. Preserve the files and their relationship as a small, reviewable package.

Choose a sample containing a real edge condition relevant to the project, or an explicitly synthetic gap for a workflow demonstration. Do not describe a synthetic test as surveyed evidence. Ask the processor to explain why the example is representative of the proposed treatment and what it cannot demonstrate. A test that proves file handling is not automatically a validation of the terrain estimate.

Agree who keeps this paired sample after the project closes. A file stored only in an individual analyst's working folder may be difficult for the recipient to retrieve later. Name the delivery location and responsible custodian in the handover note, and confirm that the final package includes the approved revisions rather than an earlier trial.

Record the actual operation and relevant settings rather than only a note saying that the surface was repaired. The record should identify the input and output revisions and any associated mask. If several processing steps were applied, distinguish the fill step from other changes. This keeps the review focused on a specific transformation instead of attributing every visible difference to a single command.

Review affected locations before the overview

Our editorial review lesson is to begin with the before-and-after mask, then inspect selected affected locations, and only afterward view the polished overview. This order makes the estimate visible as a decision. Starting with a smooth final rendering can make the later provenance discussion feel like a minor technical appendix, even when the gap lies in an important part of the customer's analysis.

Invite the intended recipient to identify which affected areas matter most. A processor may prioritize visual continuity while a customer cares about a specific location or comparison. Capture that difference before final approval. The reviewer should be able to state whether an affected area is accepted, restricted to a particular use, or still unresolved, with a named person responsible for the decision.

Keep the mask with downstream products

A later recipient may receive only a derived map or report. Ask how the interpolation limitation will remain visible at that stage. Depending on the agreed deliverable, the answer might be an accompanying layer, a location-specific note or an explicit exclusion in the report. The format is a scope decision; the important requirement is that the limitation does not disappear when the surface leaves the original processing workspace.

Also identify which downstream outputs need review after a filled surface changes. Do not assume that replacing one raster file updates every previously exported figure or table. Keep a short dependency list for the agreed deliverables. This is a recommendation for managing the handover, not a claim that a particular software system automatically tracks or regenerates those outputs.

Do not confuse filling with clipping

A file can look irregular because it follows the intended project boundary, not because the project contains an unresolved gap. Conversely, a rectangular export can include locations outside the agreed analysis area. Before discussing filling, establish whether the location belongs in the deliverable at all. The UVH1 guide to polygon and bounding-box clipping addresses that separate inclusion decision.

Where terrain results feed a wider monitoring report, preserve the population and location definitions as well. Changing which sites contribute to a headline can change its interpretation independently of terrain processing. See the UG62 discussion of changing site mix for a separate reporting problem. Neither an attractive surface nor a favorable aggregate trend removes the need to understand the underlying evidence.

What to ask in a UG73 configuration discussion

The UG73 product listing establishes the product being considered. It does not establish an included terrain sensor, processing license, interpolation service or validated accuracy for a particular site. Ask the proposed supplier to identify the aircraft configuration, payload integration and processing responsibilities separately. When a specification or service inclusion is not confirmed, contact us for configuration details.

Explain the required outputs and acceptance restrictions before asking whether the package is suitable. Include the intended receiving organization, the type of review and any areas where estimates would be unacceptable. US and Irish buyers can use these scoping questions without assuming equivalent operating permissions or technical obligations. Local operations and professional acceptance requirements remain separate matters for the responsible project team.

Questions about continuous terrain surfaces

Is every filled cell a problem?

No. The buyer may accept an estimate for a clearly defined purpose. The acceptance should identify the affected area and the basis for that decision. Do not describe the estimate as an additional observation or use visual smoothness as the sole reason for approval.

Does retaining the mask validate accuracy?

No. It makes the processing scope inspectable. Accuracy or suitability requires the relevant evidence and review for the intended use. Provenance is useful because it lets the reviewer ask the right questions, not because it supplies the answers automatically.

Should the original gap be deleted after approval?

Retain the original artifact and its relationship to the approved output under the project's agreed records policy. A later reviewer may need to distinguish source limitations from processing choices. Approval should create a clear decision record, not erase the evidence that led to it.

Ask for provenance with the terrain proposal

Compare aircraft candidates in the VTOL and fixed-wing drone collection alongside a separate description of the intended terrain product. Prepare one sample gap or exclusion, the receiving team's intended use and the decision owner. State whether estimates are permitted and what evidence is required before accepting them.

Submit that brief through the UNITED UAV project inquiry channel when discussing UG73. Request a paired original-mask and filled-surface example, with processing ownership and unresolved assumptions listed. A useful proposal should help you identify where information was added by computation and who accepts that addition, rather than simply promising a map without visible holes.