Frequently Asked Questions (FAQ)¶

What is the purpose of cross-sensor calibration?¶

To make reflectance from NEON, drones, and satellites comparable by removing geometric artifacts and aligning spectral responses.

Why does NEON use scaled reflectance?¶

NEON stores reflectance as scaled integers for performance. The export stage automatically converts these into floating-point reflectance values using NEON’s scale factors.

Do I need both topographic and BRDF correction?¶

Yes, for most ecological analyses.
Topographic correction removes slope/aspect artifacts; BRDF correction normalizes view/sun geometry.

Why do some bands look noisy?¶

Low-SNR wavelength regions (especially in the SWIR) are more sensitive to BRDF fitting errors.

What sensors can I harmonize to?¶

Currently:

Landsat OLI / OLI-2
MicaSense RedEdge (via regression)

More sensors can be added as SRF tables become available.

Why does the pipeline take so long?¶

NEON tiles are very large (tens of GB).
BRDF and ENVI export steps are computationally expensive.

Parallelization helps, but memory constraints matter.

Why do some pixels have reflectance > 1?¶

BRDF or DEM correction may amplify noise at certain angles or terrain slopes.
Such pixels are reported in QA metrics and usually masked.

Can I run analyses without touching ENVI files?¶

Yes—use the Parquet outputs (*_merged_pixel_extraction.parquet) for large-scale analysis.