Intercomparison of Multiple High-Resolution LAI Remote Sensing Products Over Neon Forest Sites

Leaf area index (LAI) is an important biophysical variable widely used in ecosystem models. Recent advances in remote sensing technology have led to the development of different high-resolution LAI products but their consistency and accuracy remains unknown. This study compared high-resolution LAI remote sensing products derived from three different types of instruments, including NEON Airborne Observatory Platform Imaging Spectrometer (NIS), Sentinel-2 (S2), and Global Ecosystem Dynamics Investigation (GEDI) lidar products. These products were also compared with in-situ data derived from Digital Hemispheric Photos (DHPs) over 19 NEON forest sites in the US. Results at plot level (i.e. ~20m) suggested there were only moderate agreements between different remote sensing products (r
²
ranging from 0.23 to 0.35) and multiple factors could affect comparison results such as geolocation error and slope. Their agreements were much better at site level (r
²
ranging from 0.59 to 0.76). In comparison to DHPs estimate at site level, GEDI achieved a better agreement (r
²
=0.53) than that of NIS and S2 (r
²
= 0.33 and 0.30 respectively). These results suggest, despite large uncertainties at plot level, there is a great potential of improving LAI estimates at regional landscape levels by fusing different remote sensing products.