← Science Methodology Home
Home → Science → Limitations
SCIENCE · LIMITATIONS

Limitations

Scientific transparency is essential. Here is what CanopySat cannot do — and what it can do.

Known limitations

Persistent cloud cover
In tropical zones with permanent cloud cover (e.g. some areas of the western Amazon), the Sentinel-2 composite may lack recent data. CanopySat uses a 12-month composite to minimize this issue, but heavily cloudy zones may have less recent data.
Spatial resolution
The minimum resolution is 10m (Sentinel-2). Forest disturbances smaller than 10m × 10m may not be detected. For very small zones, results may be less precise.
Analysis zone vs entire forest
CanopySat analyzes a specific GPS zone — not the entire forest. A 50km radius score in the Amazon represents 7,854 km² — a fraction of the total Amazon (5.5M km²). Use well-targeted zones for relevant results.
GEDI LiDAR — partial coverage
NASA GEDI LiDAR covers latitudes between 51.6°S and 51.6°N — northern boreal forests (Canada >51.6°N, Siberia >51.6°N) have no GEDI data. In these zones, canopy height is estimated by other methods.
Update delay
CanopySat uses Sentinel-2 data from the last 12 months. Very recent events (less than 2 weeks) may not yet be integrated into the analysis.

Limitations — Species Climate Suitability (BEA)

Simplified model — not full MaxEnt
CanopySat uses a Bioclimatic Envelope Assessment (BEA) approach based on GBIF species distribution. This method is consistent with MaxEnt for general trends but is not equivalent to a full MaxEnt model integrating all environmental variables.
Temperature estimated from latitude
Species thermal tolerance is estimated from the latitude of their GBIF occurrences via the formula: T = 27 - |lat| × 0.52°C. This approximation does not account for altitude, microclimates or local adaptations.
Variables not accounted for
The current BEA model does not account for: altitude, soil type, interspecific competition, species dispersal, or extreme events (droughts, fires). These factors can significantly influence actual suitability.
Single climate model
2050 projections use a 5-model CMIP6 ensemble (ACCESS-CM2, MIROC6, MPI-ESM1-2-HR, GFDL-ESM4, BCC-CSM2-MR) under the SSP2-4.5 scenario. The ensemble mean reduces uncertainty compared to a single model.
Consistency with MaxEnt confirmed
Comparisons with published MaxEnt studies (e.g. Ceiba pentandra, Frontiers in Forests 2026) show that CanopySat's BEA method produces consistent general trends. It is suitable for rapid global-scale assessment.

What CanopySat does well

Recommended use

Methodology → Validation Report →