Evaluation of genotype x environment interaction using yield and UAV-based vegetation index data from multi-environment trials in chickpea

Authors: UMANI, KINGSLEY - Washington State University; Daba, Sintayehu; PIASKOWSKI, JULIA - University Of Idaho; McGee, Rebecca; Vandemark, George; SANKARAN, SINDHUJA - Washington State University

Submitted to: Journal of Crop Improvement
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/1/2025
Publication Date: 5/5/2025
Citation: Umani, K., Daba, S.D., Piaskowski, J., McGee, R.J., Vandemark, G.J., Sankaran, S. 2025. Evaluation of genotype x environment interaction using yield and UAV-based vegetation index data from multi-environment trials in chickpea. Journal of Crop Improvement. 39(3):225-250. https://doi.org/10.1080/15427528.2025.2489605.
DOI: https://doi.org/10.1080/15427528.2025.2489605

Interpretive Summary: Chickpeas have been a globally important crop for thousands of years and are commercially produced throughout the USA Pacific Northwest and Northern Plains. Developing improved chickpea varieties through plant breeding typically requires careful evaluation of multiple traits across several years and locations. One objective of this research was to evaluate yield of 15 different chickpea varieties and USDA breeding lines across three locations for three years (2017-2019). Two breeding lines, CA13900139C and CA13900151C, consistently had high yields across years and locations. In 2018 the locations at Pullman, Washington and Genesee, Idaho were were found to be ideal for evaluating yield. In 2019 the locations at Fairfield, Washington and Genesee, Idaho were ideal for evaluating yield. These results will be used to identify chickpea breeding lines that may be released as improved varieties based on their consistently high yield, and environments that are ideal for detecting differences in yield between chickpea entries. Another challenge for plant breeding is that traits must be evaluated over the entire life history of a crop, starting at seedling emergence and continuing through harvest, and at times even after harvest for quality traits. Specially trained employees are needed to evaluate field traits such as early season vigor, days to flower, and days to harvest. Employees may evaluate a trait differently and an employee may inconsistently evaluate a trait over the course of a day. One way to overcome these difficulties associated with trait evaluation is to use "remote sensing" technologies, which typically employ a remote-controlled camera to take pictures of plants in the field. There must be a very strong association between the pictures, or images, and actual plant traits for this approach to be effective. In this study we used an unmanned aerial vehicle (UAV) to take images of the research plots described above at various times during the season. The types of images that were most closely associated with seed yield were identified. We determined that images taken at the beginning of flowering or when the seed pods began to develop were most closely associated with seed yield. These results suggest that remote sensing may be useful for identifying plants with high yield potential before harvest, which would improve efficiency of breeding programs.

Technical Abstract: Genotype-by-environment interaction (GEI) significantly impacts the success of breeding strategies for cool-season annual crops such as chickpea (Cicer arietinum L.). In this study, advanced trials of 15 spring chickpea genotypes were tested at three locations (Fairfield, Genesee, and Pullman) in the Pacific Northwest region of the United States from 2017 – 2019. The objective of this research was to determine the effects of environments, genotypes, and GEI on chickpea seed yield and to parallelly evaluate the effectiveness of remote sensing derived vegetation indices to identify genotypes with high seed yield, stability, adaptability, and consistent performance. The normalized difference vegetation index (NDVI), green NDVI (GNDVI), and soil adjusted vegetation index (SAVI) were obtained at flowering and early pod development growth stages. Stable and consistent genotypes were identified using mean performance analysis, additive main effects and multiplicative interaction (AMMI), genotype-environment interaction (GGE), AMMI stability value (ASV), and genotype stability index (GSI). There were significant genotype and environment effects (p<0.001) for seed yield and vegetation indices. GEI effects were significant for seed yield, GNDVI at flowering, and SAVI at early pod development. Seed yield ranged from 1515 (G06) to 2175 kg/ ha-1 (G10). The GNDVI, NDVI, and SAVI ranged between 0.28 to 0.38 at flowering, and 0.33 to 0.47 at early pod development. The highest-yielding and top-performing genotypes were found to be G10, G12, and G09. AMMI, GGE, ASV, and GSI analysis identified G12 and G10 as stable, high-performing genotypes based on seed yield, GNDVI at flowering and SAVI at early pod development data. With such data, several environments including Fairfield (2019), Genesee (2018), Genesee (2019), and Pullman (2018) were identified as ideal based on the average performance and stability across genotypes. The findings revealed that digital traits can potentially be valuable, in addition to seed yield, to measure the genotypic performance of chickpea cultivars in the Pacific Northwest region of the United States.