# A spatially comprehensive canopy cover dataset derived from NASA’s ice, cloud and land elevation satellite-2 (ICESat-2) for the state of Alabama, USA

## TL;DR

This paper introduces a detailed canopy cover dataset for Alabama using NASA's ICESat-2 satellite data, showing how it can be used to assess forest conditions.

## Contribution

The paper presents a novel statewide canopy cover dataset derived from ICESat-2 data, demonstrating its feasibility for large-scale forest monitoring.

## Key findings

- A canopy cover dataset for Alabama was created using ICESat-2 data with 30 m resolution.
- Canopy cover values ranged from 0 to 72% across forested pixels in 2022.
- The dataset was calibrated with airborne lidar and modeled using a Random Forests approach.

## Abstract

NASA’s Ice, Cloud, and land Elevation Satellite-2 (ICESat-2) has already demonstrated an extraordinary capability to assess forests, including providing measurements of canopy heights, and estimating aboveground biomass (AGB) and canopy cover. Despite these advancements, the application of the mission’s data to deriving continuous estimates of canopy cover, as is fundamental parameter for assessing forest conditions, is not well-understood. Here, we present a statewide (135,760 km²) canopy cover dataset at a 30 m scale across mixed temperate forests of the southern United States (US), highlighting feasibility of applying ICESat-2 data for deriving canopy cover, and providing a basis for further upscaling. This dataset provides continuous estimates of canopy cover across forests in the state of Alabama, in the southern United States, for the year 2022. Non-forests were masked out of the final dataset, and canopy cover values within each 30 m pixel, range from 0 to 72 %. Only freely and openly available remote sensing data were used to generate this dataset. ICESat-2′s along-track, vegetation product data were acquired and filtered to retain nighttime-only samples within classified forests, calibrated using reference canopy cover from airborne lidar data, and then extrapolated to achieve wall-to-wall coverage, using a Random Forests (RF) model. The mapped output represents a large-area ICESat-2-derived canopy cover product, highlighting applicability of ICESat-2 for canopy cover information and synergistic use with free and open space-based and other available ancillary products for this information. This dataset is openly accessible through the Open Science Framework.