Why canopy matters more than height
New restoration teams often measure success by seedling height. It is easy to record, easy to graph, and easy to report. But height is a lagging indicator. By the time a seedling stops growing, the stress that caused it, waterlogging, salinity spike, herbivory, has already been present for weeks. Canopy condition is the earliest visible signal. Leaves yellow, thin out, or curl before growth slows.
For Rhizophora racemosa, Avicennia germinans, and Laguncularia racemosa stands, the canopy is where photosynthesis, transpiration, and salt exclusion all show their outcomes. A photo of a canopy is a snapshot of every physiological process a mangrove is running. That is why a monitoring protocol built around canopy photography beats a protocol built around a tape measure.
Set up a photo point once, use it forever
The single most important decision is fixing your photo points. A photo point is a precise location, direction, and framing that you return to every visit. Without fixed points, week-to-week photos are not comparable and the whole exercise collapses into impressions.
To set up a photo point:
- Pick a location with a clear line of sight to a representative section of canopy, roughly 5 to 10 metres away.
- Drive a stake or place a permanent marker (a painted rock, a numbered plate on a nearby tree).
- Record the GPS coordinates using your phone's compass or a GNSS app.
- Note the compass bearing you are facing. On the first visit, take a reference photo and save the bearing in the caption or filename.
- Plan for four to eight photo points per hectare, arranged so they cover the range of hydrological conditions on your site, high, mid, and low intertidal.
Shooting protocol
Consistency beats resolution. A blurry photo shot from the same spot every week is more useful than a sharp photo shot from a slightly different angle each time.
Time of day
Shoot within the same two-hour window each visit. Mid-morning (roughly 9:00 to 11:00 local time) gives even light on most West African coasts without harsh shadows. Avoid noon (washed-out colours) and late afternoon (long shadows distort canopy density).
Framing
Hold the phone horizontally at chest height. Frame so the canopy fills the upper two thirds of the image and the ground or water fills the lower third. This gives the analyst (or the AI model) a reference for scale and light.
Focus and exposure
Tap the canopy in the middle of the frame to lock focus and exposure. Do not use HDR, filters, or portrait mode. These modify colour and contrast in ways that are invisible to you but ruin comparability across dates.
Metadata
Keep GPS on. Every modern phone embeds coordinates and timestamps in the photo file (EXIF). CanopyStride AI reads these automatically. If you are archiving manually, add the photo point number, date, and any weather notes to the filename.
Reading the canopy
Look for four categories of signal, in this order:
- Colour. A healthy mangrove canopy is deep, saturated green. Pale yellow-green suggests nitrogen limitation or salinity stress. Bronze or red-tinted leaves in Avicennia are a salt-excretion sign, common but worth tracking if it intensifies.
- Density. Thinning canopy, where you can see through to sky or water through the leaf layer, indicates leaf drop, herbivory, or dieback. Compare against your reference photo from the same point.
- Uniformity. Patchy canopy (healthy leaves next to dead branches on the same tree) usually points to physical damage, storm scarring, boat wash, or debris accumulation.
- New growth. Fresh leaves are lighter green with a slight sheen. A canopy with visible new growth at the branch tips is actively expanding. A canopy with no visible new growth for four consecutive weeks in the growing season warrants investigation.
How often to shoot
Weekly for the first 90 days after planting or a major disturbance. Fortnightly for the following six months. Monthly after that, for as long as the site is under active management. Add an extra visit within 48 hours of any storm, spring tide, or upstream flood event.
Discipline matters more than frequency. A team that reliably shoots monthly for two years will spot slow-onset problems (salinity intrusion, gradual erosion) that a team shooting weekly for the first month and then never again will miss entirely.
Turning photos into signals
Photos are inputs, not outputs. To convert a folder of canopy shots into a monitoring record, you need three things: a fixed vocabulary for what you are describing, a way to aggregate observations across photo points, and a trigger for action when a threshold is crossed.
CanopyStride AI provides the first two automatically. Every uploaded photo is scored on waterlogging, salinity stress proxy, tidal scouring, and canopy health, with a short written note per indicator. Scores aggregate by plot and site so a project manager can scan a dashboard rather than a folder. The third piece, action triggers, is decided by your team: "if canopy health drops two levels in four weeks, dispatch an ecologist" is a reasonable starting rule.
Common pitfalls
- Rotating photographers without a handover. Two people framing the same scene differently produces noise that looks like a change in the canopy. Walk through the protocol together at least once when handing over.
- Skipping the reference photo. Without a first-visit photo, later images have nothing to compare against.
- Shooting only healthy patches. The tempting patches to photograph are the beautiful ones. The valuable patches to photograph are the stressed ones.
- Deleting failed sites. A photo record of a failed restoration is more valuable than no record at all, it tells the next attempt what not to do.
Scaling beyond a single site
Once one site is running well, the protocol scales laterally without much change. Two teams working two sites can compare notes because the vocabulary is the same. A donor reviewing five sites across the initiative can look at a single dashboard because the photo points, cadence, and scoring are consistent. The value of a monitoring protocol compounds with every additional site that adopts it.
That compounding is why we recommend picking a small number of protocols and using them everywhere, rather than reinventing the monitoring approach for each new plot.