Comparison and analysis of data upscaling schemes for predicting crop leaf area index

Estimating crop leaf area index (LAI) based on remotely sensed observations is a normal way for regional crop monitoring and yield estimation in precision agriculture. However, the spatial heterogeneity in crop canopies and the nonlinearity of model for LAI estimation make differences between estimated-upscaled LAI (baseline data) and upscaled-estimated LAI. Then, aiming to reduce these differences, i.e. the scale effects of crop LAI, we propose a data upscaling scheme to construct new upscaled-estimated LAI. Barley in Inner Mongolia in 2010 is selected as the object for numerical experiments to compare and analyze the differences between baseline data and upscaled-estimated LAI datasets calculated with different data upscaling schemes. Taking Advanced Land Observing Satellite (ALOS) multispectral image as data source, theory analyses and numerical practices confirm the feasibility and validity of this proposed data upscaling scheme in reducing scale effects of crop LAI, and also manifest its stability and robustness in application.