Skip to main content
ARS Home » Plains Area » Fort Collins, Colorado » Center for Agricultural Resources Research » Water Management and Systems Research » Research » Publications at this Location » Publication #374441

Research Project: Improving the Sustainability of Irrigated Farming Systems in Semi-Arid Regions

Location: Water Management and Systems Research

Title: Evaluating the sensitivity of water stressed maize chlorophyll and structure based on UAV derived vegetation indices

Author
item ZHANG, LIYUAN - Northwest A&f University
item HAN, WENTING - Northwest A&f University
item NIU, YAXIAO - Northwest A&f University
item CHÁVEZ, JOSÉ - Colorado State University
item LI, GUANG - Northwest A&f University
item SHO, GUOMIN - Northwest A&f University
item Zhang, Huihui

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/14/2021
Publication Date: 4/30/2021
Citation: Zhang, L., Han, W., Niu, Y., Chávez, J.L., Li, G., Sho, G., Zhang, H. 2021. Evaluating the sensitivity of water stressed maize chlorophyll and structure based on UAV derived vegetation indices. Computers and Electronics in Agriculture. 185. Article e106174. https://doi.org/10.1016/j.compag.2021.106174.
DOI: https://doi.org/10.1016/j.compag.2021.106174

Interpretive Summary: Stomatal conductance has been widely used as water stress reference. In 2018 and 2019 growing seasons, leaf stomatal conductance and leaf area index were measured periodically from an experimental maize field where crop was treated under three irrigation treatments during different growth stages. A time series of unmanned aerial vehicle (UAV) multi-spectral images were acquired while these ground measurements were taken. A few vegetation indices (VIs), including structural VIs and chlorophyll VIs, were derived from multi-spectral images and their sensitivities to stomatal conductance were investigated. Two machine learning algorithms were used to test the correlation between VIs and water stress. The results show that the structure VIs were more sensitive to crop water stress while chlorophyll VIs only responded to crop water stress at severe level. The sensitivity of these VIs to crop water stress can be significantly influenced by different responses of canopy structure and chlorophyll concentration to water stress, and the different spectral resolution of UAV images. Compared to machine learning regression algorithms, simple linear regression was powerful enough to link UAV-based multi-spectral VIs and stomatal conductance. Overall, our results showed the potential of using structural VIs derived from UAV multi-spectral images to estimate maize water stress.

Technical Abstract: To formulate an optimal irrigation schedule and improve water productivity, it is necessary to timely and accurately map crop water stress and monitor its inter-field variability at a farm scale. In this study, maize water stress was estimated based on unmanned aerial vehicle (UAV) multi-spectral images by using stomatal conductance (Gs) as water stress reference, over the entire 2018 and 2019 maize growing seasons. To assess the sensitivity of chlorophyll and structural vegetation indices (VIs) to water stress and to assess the performance of machine learning regression algorithms for estimating maize water stress conditions, two chlorophyll and four structural VIs and two non-linear machine learning regression algorithms, random forest and artificial neural network, were chosen. The results showed that structural VIs had significant correlations (p < 0.001) with Gs with the highest r value of 0.64 in 2018 and 2019. The transformed chlorophyll absorption in reflectance index, Chlorophyll VI, could only estimate severe maize water stress with r value of -0.47 (p <0.001) for the drier 2019. The water stress sensitivity of chlorophyll and structural VIs may be significantly influenced by different responses of canopy structure and chlorophyll concentration to water stress, and the different spectral resolutions of UAV multi-spectral cameras. Compared to non-linear machine learning regression algorithms, simple linear regression was powerful enough to link UAV-based multi-spectral VIs and Gs. Overall, our results showed the potential of using structural VIs derived from UAV multi-spectral images to estimate maize water stress. The next important research topics could be to find a robust relationship between structural VIs and crop water stress reference or to find crop water stress scale of VIs themselves.