← Papers

Paper record

Cubesat Constellations Provide Enhanced Crop Phenology And Digital Agricultural Insights Using Daily Leaf Area Index Retrievals.

Johansen K, Ziliani MG, Houborg R, Franz TE, McCabe MF.

Research Square Platform LLC · 13 Dec 2021 · 10.21203/rs.3.rs-1134595/v1

Abstract

Abstract Satellite remote sensing has great potential to deliver on the promise of a data-driven agricultural revolution, with emerging space-based platforms providing spatiotemporal insights into precision-level attributes such as crop water use, vegetation health and condition and crop response to management practices. Using a harmonized collection of high-resolution Planet CubeSat, Sentinel-2, Landsat-8 and additional coarser resolution imagery from MODIS and VIIRS, we exploit a multi-satellite data fusion and machine learning approach to deliver a radiometrically calibrated and gap-filled time-series of daily leaf area index (LAI) at an unprecedented spatial resolution of 3 m. The insights available from such high-resolution CubeSat-based LAI data are demonstrated through tracking the growth cycle of a maize crop and identifying observable within-field spatial and temporal variations across key phenological stages. Daily LAI retrievals peaked at the tasseling stage, demonstrating their value for fertilizer and irrigation scheduling. An evaluation of satellite-based retrievals against field-measured LAI data collected from both rain-fed and irrigated fields shows high correlation and captures the spatiotemporal development of intra- and inter-field variations. Novel agricultural insights related to individual vegetative and reproductive growth stages were obtained, showcasing the capacity for new high-resolution CubeSat platforms to deliver actionable intelligence for precision agricultural and related applications.

Code and data availability

The paper's core assets (Planet Fusion imagery and daily 3 m LAI retrievals over the Nebraska maize fields) are not publicly deposited; the authors state they can be provided by KAUST upon request pending scientific review. No public code, trained models, or datasets with a public URL are described. The only supplement

No evidence-backed public reproduction asset is currently recorded.