LAI estimation across California vineyards using sUAS multi-seasonal multi-spectral, thermal, and elevation information and machine learning

Authors: GAO, R. - Utah State University; TORRES, A. - Utah State University; ABOUTALEBI, M. - E & J Gallo Winery; White, William - Alex; Anderson, Martha; Kustas, William - Bill; AGAM, N. - Ben Gurion University Of Negev; ALSINA, N. - E & J Gallo Winery; Alfieri, Joseph; HIPPS, L.E. - Utah State University; DOKOOZLIAN, N. - E & J Gallo Winery; NIETO, H. - Collaborator; Gao, Feng; McKee, Lynn; Prueger, John; SANCHEZ, L. - E & J Gallo Winery; McElrone, Andrew; BAMBACH, N. - University Of California, Davis; COOPMANS, C. - Utah State University; GOWING, I. - Utah State University

Submitted to: Irrigation Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/22/2022
Publication Date: 3/14/2022
Citation: Gao, R., Torres, A., Aboutalebi, M., White, W.A., Anderson, M.C., Kustas, W.P., Agam, N., Alsina, N., Alfieri, J.G., Hipps, L., Dokoozlian, N., Nieto, H., Gao, F.N., McKee, L.G., Prueger, J.H., Sanchez, L., McElrone, A.J., Bambach, N., Coopmans, C., Gowing, I. 2022. LAI estimation across California vineyards using sUAS multi-seasonal multi-spectral, thermal, and elevation information and machine learning. Irrigation Science. 40:731-759. https://doi.org/10.1007/s00271-022-00776-0.
DOI: https://doi.org/10.1007/s00271-022-00776-0

Interpretive Summary: In agriculture, leaf area index (LAI) is an important variable that describes crop growth and development and strongly correlates to crop water use or evapotranspiration (ET). Current LAI estimation methods using satellites are limited not only by spatial resolution but also by the empirical nature of the retrieval algorithms, mostly based on vegetation indices. Widely used machine learning (ML) algorithms using high-resolution small unmanned aerial system (sUAS) information provide an opportunity for spatial and temporal LAI estimation addressing the spatial and temporal limitations of satellite-based algorithms. The sUAS very fine resolution imagery derived from multiple growing seasons were used with several ML approaches to extract and generate vine-scale information for LAI estimation in commercial vineyards across California. A hybrid ML approach yielded the best agreement with ground-truth LAI measurements and using the Two-Source Energy Balance (TSEB) model produced accurate estimates of ET. The TSEB model also indicated uncertainty in LAI can cause a significant impact on vine ET indicating more robust LAI algorithms for vineyards are required for satellite-based methods using sUAS information for algorithm development

Technical Abstract: In agriculture, leaf area index (LAI) is an important variable that describes occurring biomass and relates to the distribution of energy fluxes and evapotranspiration components. Current LAI estimation methods at subfield scale are limited not only by the characteristics of the spatial data (pixel size and spectral information) but also by the empiricity of developed models, mostly based on vegetation indices, which do not necessarily scale spatiality (among different varieties or planting characteristics) or temporally (need for different LAI models for different phenological stages). Widely used machine learning (ML) algorithms and high-resolution small unmanned aerial system (sUAS) information provide an opportunity for spatial and temporal LAI estimation addressing the spatial and temporal limitations. In this study, considering both accuracy and efficiency, a point-cloud-based feature-extraction approach (Full Approach) and a raster-based feature extraction approach (Fast Approach) using sUAS information were developed based on multiple growing seasons (2014–2019) to extract and generate vine-scale information for LAI estimation in commercial vineyards across California. Three known ML algorithms, Random Forest (RF), eXtreme Gradient Boosting (XGB), and Relevance Vector Machine (RVM), were considered, along with hybrid ML schemes based on those three algorithms, coupled with different feature-extraction approaches. Results showed that the hybrid ML technique using RF and RVM and the Fast Approach with 9 input variables, called RVM-RFFast model, performs better than others in a visual and statistical assessments of the generated LAI being also computationally efficient. Furthermore, using the generated LAI products in the quantification of energy balance using the two-source energy balance Priestley-Taylor version (TSEB-PT) model and EC tower data, the results indicated excellent estimation of net radiation (Rn) and latent heat flux (LE), good estimation of surface heat flux (G), and poor estimation of sensible heat flux (H)