Regrowth vs. replanting: flagging the difference in your extent report

A gain polygon is not self-explanatory. When this year's canopy boundary shows more mangrove than last year's, that area could be two very different things: seedlings that came up on their own after a dieback, or mangrove your team put in the ground on purpose. Lump them together in the annual report and a verifier is going to ask why.

Why the registry cares about the split

Most methodologies treat planted area and naturally regenerated area as separate claims, with separate evidence requirements. A planting polygon needs a planting record behind it: species, stock, planting date, density, who did the work. A regrowth polygon needs none of that, because nobody planted it, but it still needs to be justified as recovery rather than, say, mangrove that was already there and just missed in a prior year's digitizing pass.

If your gain polygons are one undifferentiated layer, you can't produce either kind of evidence cleanly. The verifier ends up asking you to go back and split the file, which means re-running the comparison you should have run the first time, under a deadline.

Telling the two apart in the canopy boundary

Start with the planting record, not the imagery. If your restoration team keeps a shapefile or even a spreadsheet with planting block boundaries and dates, that's your ground truth for one half of the split. Any gain polygon that falls inside a planting block, planted before this year's observation date, gets flagged as replanting.

Everything left over is a candidate for regrowth, but check it before you label it that way. A few things to look for:

• Shape and pattern. Planted stock tends to show up in rows or blocks that match how the crew worked the site. Natural regrowth is patchier, following the old creek edges, dieback scars, or wherever propagules actually settled.
• Timing against a known disturbance. If the area was mapped as loss two or three years ago after a storm or a die-off, and it's now showing canopy again with no planting record attached, that's a strong regrowth signal.
• Edge continuity. Regrowth usually grows out from an existing stand it's attached to the standing canopy boundary rather than appearing as an isolated patch in open mudflat, which is more typical of a planted trial plot.

None of these checks require new imagery acquisitions. They're a desk exercise against the canopy boundary you already have for this year and last, run alongside whatever planting records your restoration side maintains.

Keeping the split in the deliverable, not just your head

The distinction only helps if it survives into the file you hand the registry. Two ways to do it:

1. Separate vector layers for planted-gain and regrowth-gain, each dated to the observation year.
2. A single gain layer with an attribute field ("source: planted" / "source: regrowth" / "source: unconfirmed") and a reference to the planting record ID where one applies.

Either works. What doesn't work is a single "gain 2026" polygon set with no attribute telling anyone which parts came from which process. Pick one approach before your first annual cycle and stay with it, because the verifier will be comparing this year's layer structure against last year's, and reshuffling the schema partway through a crediting period is its own headache.

One more thing worth carrying forward: regrowth areas can revert to loss faster than mature planted stands, especially in the first couple of years. If you're tracking them as a separate layer, you'll see that pattern early instead of discovering it when the next loss polygon shows up in the same spot you called "gain" two reports ago.

A clean year-over-year canopy boundary with gain and loss already dated and delineated is what makes this split fast instead of a weeks-long digitizing exercise, which is the layer Mangrove Monitor produces once a year on your reporting cycle. If your next annual report is coming up, it's worth seeing what that layer looks like against your own project boundary.