B037-0005
Characterization of Soil-Plant Spatial Relationships and their Impact on Crop Yield Using Remote Sensing and Geophysics from Crop to Farm Scale

Wednesday, 9 December 2020
Poster
Karina Nugent1, Haruko M Wainwright2, Baptiste Dafflon2, Craig Ulrich1, Florian Soom3, Jay McEntire4, James Bentley Brown2,4 and Nicola Falco2, (1)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences, Berkeley, CA, United States, (2)Lawrence Berkeley National Laboratory, Berkeley, CA, United States, (3)Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA, United States, (4)ARVA Intelligence, Salt Lake City, UT, United States
Abstract:
Understanding soil-plant interactions and their spatial-temporal variability in agro-ecosystems is critical to improve crop development and yield as well as optimize sustainable farming practices. Satellite imagery has proven to be useful for capturing plant dynamics at a large scale, while near-surface geophysical surveys can be used to derive soil properties (such as soil texture, soil moisture, porosity). The aim of this study is to investigate the effects of soil spatial heterogeneity on plant growth by integrating remote sensing and geophysical data, in order to develop a tractable methodology that is able to identify spatial regions characterized by specific soil-plant interactions at farm scale. This study was conducted over multiple soybean fields, on a single farm in Humphrey, Arkansas.


We computed vegetation indices (VIs) such as normalized difference vegetation index (NDVI) and green chromatic coordinate (GCC) from a time-series of satellite imagery (Planet Lab and Sentinel-2) to characterize the temporal and spatial variability of plant growth. We implemented a cluster analysis based on the VI time-series to identify regions with specific plant behavior. We then investigated the co-variability of soil electrical conductivity (EC), VIs, and crop yield using simple statistical methods such as principal component analysis (PCA). From our analysis, we have determined that soil spatial heterogeneity has a distinct impact both on plant development and crop yield. In particular, we observed that the identified clusters computed using the VIs were also attributed to areas with specific soil EC. Going forward, the developed methodology will be used to characterize soil-plant interactions at large scale and will incorporate other crops such as rice and corn, as we expect that such interactions are plant-specific. Furthermore, we will evaluate the breadth of this methodology, using only plant development information as an input.

<< Previous Abstract | Next Abstract >>