# SUPPLEMENTARY DATA - METADATA

## FILE 1: SegChar_all.csv  (Table 1 - UAV segment characteristics)

OVERVIEW

* Object-based predictors derived from UAV multispectral imagery (RedEdge-MX).
* One record = one image segment (polygon) used as an analysis object.
* Segment-level spectral, textural, and shape characteristics.
* Each segment linked to a single habitat type (BIO_TYP).

STUDY AND CLASSIFICATION CONTEXT

* Study area: Zdar'ske vrchy Protected Landscape Area, Czech Republic.
* Habitat labelling based on the Habitat Layer of the Czech Republic and Czech Habitat Catalogue (Chytry et al. 2010).
* BIO_TYP consistent with habitat classes / aggregations used in the manuscript (CzHab, aggregated schemes, Natura 2000 crosswalk).
* Variables used as object-based predictors in Random Forest (RF) models (UAV-based classification).

SENSOR AND ACQUISITION

* Platform: eBee fixed-wing UAV.
* Sensor: RedEdge-MX multispectral camera (MicaSense).
* Spatial resolution of orthomosaics: approx. 0.20-0.25 m.
* Seasons: spring and summer UAV campaigns in the same vegetation season.
* Spectral bands (RedEdge-MX, approximate ranges):

  * BLUE: center 475 nm, bandwidth 20 nm (approx. 465-485 nm).
  * GREEN: center 560 nm, bandwidth 20 nm (approx. 550-570 nm).
  * RED: center 668 nm, bandwidth 10 nm (approx. 663-673 nm).
  * RED EDGE (RE): center 717 nm, bandwidth 10 nm (approx. 712-722 nm).
  * NEAR-INFRARED (NIR): center 840 nm, bandwidth 40 nm (approx. 820-860 nm).

UNITS AND GENERAL NOTES

* MEAN_* : unitless surface reflectance (radiometrically calibrated values).
* STD_*  : standard deviation of reflectance within a segment (same units as MEAN_*); proxy for spectral texture.
* PIXELCOUNT_* : number of image pixels per segment (proxy for segment area).
* COMPACTNESS_*, RECTANGULARITY_* : dimensionless shape indices (relative measures).

COLUMN DESCRIPTIONS

* OBJECTID

  * Unique integer identifier of the segment.

* Shape

  * Geometry type from GIS export (polygon).
  * Informational only; not used as a predictor.

* BIO_TYP

  * Habitat type code.
  * Categorical; consistent with habitat classification used in the study.

* loc_jaro

  * Location / segment identifier in the spring UAV dataset.

* loc_leto

  * Location / segment identifier in the summer UAV dataset.

* loc_id

  * Common location ID linking spring and summer segments of the same habitat patch.

* PIXELCOUNT_spring

  * Number of pixels in the spring segment.

* MEAN_BLUE_spring

  * Mean BLUE-band reflectance (spring) within the segment.

* MEAN_GREEN_spring

  * Mean GREEN-band reflectance (spring) within the segment.

* MEAN_RED_spring

  * Mean RED-band reflectance (spring) within the segment.

* MEAN_RE_spring

  * Mean RED EDGE-band reflectance (spring) within the segment.

* MEAN_NIR_spring

  * Mean NIR-band reflectance (spring) within the segment.

* STD_BLUE_spring

  * Standard deviation of BLUE-band reflectance (spring) within the segment.

* STD_GREEN_spring

  * Standard deviation of GREEN-band reflectance (spring) within the segment.

* STD_RED_spring

  * Standard deviation of RED-band reflectance (spring) within the segment.

* STD_RE_spring

  * Standard deviation of RED EDGE-band reflectance (spring) within the segment.

* STD_NIR_spring

  * Standard deviation of NIR-band reflectance (spring) within the segment.

* COMPACTNESS_spring

  * Shape compactness index of the spring segment.

* RECTANGULARITY_spring

  * Shape rectangularity index of the spring segment.

* PIXELCOUNT_summer

  * Number of pixels in the summer segment.

* MEAN_BLUE_summer

  * Mean BLUE-band reflectance (summer) within the segment.

* MEAN_GREEN_summer

  * Mean GREEN-band reflectance (summer) within the segment.

* MEAN_RED_summer

  * Mean RED-band reflectance (summer) within the segment.

* MEAN_RE_summer

  * Mean RED EDGE-band reflectance (summer) within the segment.

* MEAN_NIR_summer

  * Mean NIR-band reflectance (summer) within the segment.

* STD_BLUE_summer

  * Standard deviation of BLUE-band reflectance (summer) within the segment.

* STD_GREEN_summer

  * Standard deviation of GREEN-band reflectance (summer) within the segment.

* STD_RED_summer

  * Standard deviation of RED-band reflectance (summer) within the segment.

* STD_RE_summer

  * Standard deviation of RED EDGE-band reflectance (summer) within the segment.

* STD_NIR_summer

  * Standard deviation of NIR-band reflectance (summer) within the segment.

* COMPACTNESS_summer

  * Shape compactness index of the summer segment.

* RECTANGULARITY_summer

  * Shape rectangularity index of the summer segment.







## FILE 2: PPChar_all_Planets.csv  (Table 2 - PlanetScope per-pixel predictors)

OVERVIEW

* Pixel-based predictors derived from PlanetScope multispectral imagery.
* One record = one reference pixel labelled by BIO_TYP.
* Per-pixel spectral reflectance in eight PlanetScope bands for three seasons (spring, summer, autumn).

STUDY AND CLASSIFICATION CONTEXT

* Same study area and habitat labelling as for UAV data.
* BIO_TYP consistent with the habitat classes / aggregations used in the manuscript.
* Variables used as pixel-based predictors in RF models (PlanetScope benchmark classification).

SENSOR AND ACQUISITION

* Sensor: PlanetScope (SuperDove) 8-band multispectral constellation.
* Spatial resolution: approx. 3 m.
* Seasons: spring, summer, and autumn acquisitions in the same year.
* Product type: bottom-of-atmosphere reflectance Ortho Scenes.
* Spectral bands and approximate ranges (nm):

  * BLUEc  (coastal blue): approx. 431-452 nm.
  * BLUE   (blue): approx. 465-515 nm.
  * GREENi (green I): approx. 513-549 nm.
  * GREEN  (green II): approx. 547-583 nm.
  * YELLOW: approx. 600-620 nm.
  * RED    : approx. 650-680 nm.
  * RE     (red-edge): approx. 697-713 nm.
  * NIR    (near-infrared): approx. 845-885 nm.

UNITS AND GENERAL NOTES

* All band columns: unitless surface reflectance (BOA reflectance, as provided / scaled by PlanetScope).
* Values represent single-pixel reflectance at the time of the corresponding seasonal acquisition.

COLUMN DESCRIPTIONS

* BIO_TYP

  * Habitat type code.
  * Categorical; consistent with habitat classification used in the study.
  * Class label for supervised RF classification.

* loc_id

  * Common location ID linking spring and summer segments of the same habitat patch.

* BLUEc_spring

  * Coastal blue band reflectance (spring acquisition).

* BLUE_spring

  * Blue band reflectance (spring acquisition).

* GREENi_spring

  * Green I band reflectance (spring acquisition).

* GREEN_spring

  * Green II band reflectance (spring acquisition).

* YELLOW_spring

  * Yellow band reflectance (spring acquisition).

* RED_spring

  * Red band reflectance (spring acquisition).

* RE_spring

  * Red-edge band reflectance (spring acquisition).

* NIR_spring

  * Near-infrared band reflectance (spring acquisition).

* BLUEc_summer

  * Coastal blue band reflectance (summer acquisition).

* BLUE_summer

  * Blue band reflectance (summer acquisition).

* GREENi_summer

  * Green I band reflectance (summer acquisition).

* GREEN_summer

  * Green II band reflectance (summer acquisition).

* YELLOW_summer

  * Yellow band reflectance (summer acquisition).

* RED_summer

  * Red band reflectance (summer acquisition).

* RE_summer

  * Red-edge band reflectance (summer acquisition).

* NIR_summer

  * Near-infrared band reflectance (summer acquisition).

* BLUEc_autumn

  * Coastal blue band reflectance (autumn acquisition).

* BLUE_autumn

  * Blue band reflectance (autumn acquisition).

* GREENi_autumn

  * Green I band reflectance (autumn acquisition).

* GREEN_autumn

  * Green II band reflectance (autumn acquisition).

* YELLOW_autumn

  * Yellow band reflectance (autumn acquisition).

* RED_autumn

  * Red band reflectance (autumn acquisition).

* RE_autumn

  * Red-edge band reflectance (autumn acquisition).

* NIR_autumn

  * Near-infrared band reflectance (autumn acquisition).
