Vine water status mapping with multispectral UAV imagery and machine learning

Authors: TANG, ZHEHAN - University Of California, Davis; JIN, YUFANG - University Of California, Davis; ALSINA, MARIA - E & J Gallo Winery; McElrone, Andrew; BAMBACH, NICOLAS - University Of California, Davis; Kustas, William - Bill

Submitted to: Irrigation Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/25/2022
Publication Date: 4/18/2022
Citation: Tang, Z., Jin, Y., Alsina, M.M., McElrone, A.J., Bambach, N.E., Kustas, W.P. 2022. Vine water status mapping with multispectral UAV imagery and machine learning. Irrigation Science. https://doi.org/10.1007/s00271-022-00788-w.
DOI: https://doi.org/10.1007/s00271-022-00788-w

Interpretive Summary:

Technical Abstract: Optimizing water management has become one of the biggest challenges for grapevine growers in California, especially during drought conditions. Monitoring grapevine water status and stress level across the whole vineyard is an essential step for precision irrigation management of vineyards to conserve water. We developed a unified machine learning model to map leaf water potential ('_leaf) over two commercial vineyards, by combining high-resolution multispectral remote sensing imagery and weather data. We conducted six unmanned aerial vehicle (UAV) flights with a five-band multispectral camera from 2018 to 2020, concurrently with ground measurements of sampled vines. Using vegetation indices from the orthomosaiced UAV imagery and weather data as predictors, the random forest (RF) full model can capture 77% of '_leaf variance, with a RMSE of 0.123 MPa, and MAE of 0.100 MPa, based on the validation datasets. The weather variables including air temperature and vapor pressure deficit, and the red edge index such as the normalized difference red edge index (NDRE) were found as the most important variables in estimating '_leaf across space and time. The reduced models excluding weather and red edge indices can only explain 52% to 48% of variance. Our results showed that the estimated '_leaf maps captured well the patterns of both within- and cross-field spatial variability and the temporal change of vine water status, which were consistent with irrigation management and patterns observed from the ground sampling. Our study demonstrated the potential of combining multispectral UAV images and weather data to supplement or scale up the traditional point sampling of '_leaf, and thus provide data-driven information to guide precision irrigation management in vineyards.