Recorded disturbance / projected state
Looks the same at 0.1 m.
Carries a different history.
A 10 m forest cell has the same four-number display signature before and after the 2020 Creek Fire. Exact LiDAR values, return counts, sensor protocols, and event context still differ.
01 / Source-locked result
One tile, two surveys, one bounded comparison
The analysis retains high-vegetation LiDAR returns, subtracts one shared terrain model, and summarizes four height quantiles in 10 m cells. The map below contains every cell that passes the declared return threshold in both years.
02 / Recorded chronology
Before, event context, after
The event records and the tutorial's fire-affected-tile interpretation are attributable evidence. Their temporal order does not become a causal edge.
03 / Flagship cell
Equal after rounding is not equal in full
Among two matched cells with the same 0.1 m four-quantile signature in both years, the deterministic rule selects the lowest cell ID.
Rounded display signature / metresSAME
04 / Tile-level projection
All four median height summaries are lower in 2021
This is a descriptive before/after result over the selected tile. Different sensors and the missing exact event-perimeter join prevent a causal estimate.
05 / Same-grid comparison
Inspect the matched surface
Switch between the 2019 and 2021 P90 projections or view their difference. Blank cells did not satisfy the declared eligibility rule in both years.
loadingloading
Loading cell grid.
06 / Before and after
The result, metric by metric
Each card reports the median paired change and the number of matched cells whose projected height was lower in 2021.
07 / Equivalence is always qualified
Six questions produce different answers
The word “same” is useful only after the comparison regime is named. An unresolved full ecosystem identity is not silently converted to “no.”
08 / History window
What you include changes what you can say
A snapshot, a survey pair, and a survey pair with recorded event context are three different evidence windows.
09 / Reachability
An after-state is observed. A recovery path is not.
The 2021 survey is one recorded later state. With two survey dates, no recovery trajectory, future forecast, or counterfactual path is identified.
10 / Historical Load
Undefined is the result
—NOT EVALUATED
null, not zero11 / Reading discipline
What this release does not claim
Vegetation-height projection
Four quantiles summarize retained high-vegetation returns. They are not the complete ecosystem state.
Creek Fire context
Four source records and a published tutorial interpretation establish context, not a direct spatial causal join.
Measurement protocol
The 2019 and 2021 surveys use different sensor systems. Protocol identity is false by construction.
Causal effect
No control tile, exact perimeter intersection, or adjustment design supports attribution to fire.
Recovery trajectory
Two survey dates do not establish a continuous path, a future state, or a recovery curve.