Document Type
Article
Publication Title
Remote Sensing of Environment
Publication Date
9-2017
Volume
199
First Page
333
Last Page
349
Abstract
Large-scale crop monitoring and yield estimation are important for both scientific research and practical applications. Satellite remote sensing provides an effective means for regional and global cropland monitoring, particularly in data-sparse regions that lack reliable ground observations and reporting. The conventional approach of using visible and near-infrared based vegetation index (VI) observations has prevailed for decades since the onset of the global satellite era. However, other satellite data encompass diverse spectral ranges that may contain complementary information on crop growth and yield, but have been largely understudied and underused. Here we conducted one of the first attempts at synergizing multiple satellite data spanning a diverse spectral range, including visible, near-infrared, thermal and microwave, into one framework to estimate crop yield for the U.S. Corn Belt, one of the world's most important food baskets. Specifically, we included MODIS Enhanced VI (EVI), estimated Gross Primary Production based on GOME-2 solar-induced fluorescence (SIF-GPP), thermal-based ALEXI Evapotranspiration (ET), QuikSCAT Ku-band radar backscatter, and AMSR-E X-band passive microwave Vegetation Optical Depth (VOD) in this study, benchmarked on USDA county-level crop yield statistics. We used Partial Least Square Regression (PLSR), an effective statistical model for dimension reduction, to distinguish commonly shared and unique individual information from the various satellite data and other ancillary climate information for crop yield estimation. In the PLSR model that includes all of the satellite data and climate variables from 2007 to 2009, we assessed the first two major PLSR components and found that the first component (an integrated proxy of crop aboveground biomass) explained 82% variability of modelled crop yield, and the second component (dominated by environmental stresses) explained 15% variability of modelled crop yield. We found that most of the satellite derived metrics (e.g. SIF-GPP, radar backscatter, EVI, VOD, ALEXI-ET) share common information related to aboveground crop biomass (i.e. the first component). For this shared information, the SIF-GPP and backscatter data contain almost the same amount of information as EVI at the county scale. When removing the above shared component from all of the satellite data, we found that EVI and SIF-GPP do not provide much extra information; instead, Ku-band backscatter, thermal-based ALEXI-ET, and X-band VOD provide unique information on environmental stresses that improves overall crop yield predictive skill. In particular, Ku-band backscatter and associated differences between morning and afternoon overpasses contribute unique information on crop growth and environmental stress. Overall, using satellite data from various spectral bands significantly improves regional crop yield predictions. The additional use of ancillary climate data (e.g. precipitation and temperature) further improves model skill, in part because the crop reproductive stage related to harvest index is highly sensitive to environmental stresses but they are not fully captured by the satellite data used in our study. We conclude that using satellite data across various spectral ranges can improve monitoring of large-scale crop growth and yield beyond what can be achieved from individual sensors. These results also inform the synergistic use and development of current and next generation satellite missions, including NASA ECOSTRESS, SMAP, and OCO-2, for agricultural applications.
Keywords
Optical; Fluorescence; Thermal; Microwave; Radar; Crop yield; Partial least square regression6
DOI
https://doi.org/10.1016/j.rse.2017.06.043
Rights
© 2017 Elsevier Inc.
Recommended Citation
Guan, K., J. Wu, J.S. Kimball, M.C. Anderson, S. Frolking, B. Li, C.R. Hain, and D. Lobell, 2017. The shared and unique values of optical, fluorescence, thermal and microwave satellite data for estimating large-scale crop yields. Remote Sensing of Environment, 199, 333-349, https://doi.org/10.1016/j.rse.2017.06.043
Appendix A. Supplementary Data
Comments
K.G. and D.B.L. have been funded by the U.S. National Science Foundation (NSF-SES-1048946) and a Terman Fellowship from Stanford University. K.G. is also supported by the University of Illinois at Urbana Champaign and the NASA New Investigator Award (NNX16AI56G). S.F. was supported by NSF grant NSF-EF1065074. J.K. was supported by NASA (NNX14AI50G).