Routing survey crews when the change tiles don't touch each other

A resurvey cycle used to mean something simple: this quadrangle this year, that one next year, drive the whole thing county by county and update what you find. Routing was easy because the work area was one shape on the map.
Change detection breaks that shape. Once you're working off a list of tiles where the basemap no longer matches imagery, the flagged tiles are scattered across three counties, half a river basin, and the edge of a national forest boundary. None of them touch. A crew that used to drive a tidy loop through a quadrangle is now looking at a dot pattern that makes no geographic sense until you plan a route around it.
Why the old routing logic breaks down
Quadrangle-based routing assumes contiguous coverage. You start at one corner, work a grid, end at the other corner. Every mile driven does double duty: you're moving and surveying at once.
A change tile worklist has no such shape. One flagged tile might be a new subdivision road network forty miles from the next flagged tile, which is a reshaped floodplain after a flood season. If you dispatch crews tile-by-tile in the order they appear on a CSV export, you'll burn two days of driving for four hours of fieldwork. The tiles are real and the updates are necessary. The routing problem is the part nobody solved when they switched from fixed cycles to change-based ones.
Build the route from the worklist, not the map sheet
The fix is to stop thinking in map sheets and start thinking in clusters of flagged tiles, regardless of which quadrangle or county they fall in. A crew lead pulls the current tile index, sorts by region, and groups tiles that sit within a reasonable drive radius of each other even if they span jurisdiction lines that used to define separate projects.
This usually means:
- Grouping flagged tiles by proximity first, administrative boundary second. A tile in the southeast corner of one county and a tile just across the line in the next county are one trip, not two projects.
- Ordering stops by drive time between tiles, not by tile ID or alphabetical sheet name. A change score tells you what changed and how much. It doesn't tell you the fastest way to reach it.
- Holding isolated single-tile flags for a batch run once two or three more appear nearby, unless the change is urgent enough to justify a dedicated trip.
- Checking the change score before committing a crew day. A tile with a marginal score next to three tiles with strong signal might wait a cycle rather than anchor its own trip.
None of this is exotic logistics. It's the same thinking a propane delivery route or a meter-reading circuit uses: visit the stops that need visiting, in an order that doesn't waste the tank.
What this costs you if you skip it
Most agencies that move from calendar-based resurvey to change-based resurvey get the detection part right and the dispatch part wrong. They have a good list of what changed and no plan for visiting it efficiently, so field days get eaten by transit instead of survey work, and the backlog of flagged tiles grows even though fewer tiles need attention than under the old cycle.
The worklist itself matters here too. A tile index that's noisy, that flags shadow and seasonal vegetation change alongside real ground-truth divergence, forces crews to drive to tiles that didn't need a visit at all. Routing only works as well as the list it's built from.
National Change Alerts builds that list from continuous high-resolution coverage compared against your current basemap, delivered as a tile index with change scores, so a crew lead can plan routes around tiles that actually moved instead of guessing which flags are worth the drive.
If your team is still routing crews off a fixed cycle instead of a change tile worklist, that's the place to start.