Potential Use of Hyperspectral Reflectance as a High-Throughput Nondestructive Phenotyping Tool for Assessing Salt Tolerance in Advanced Spring Wheat Lines under Field Conditions

Proximal hyperspectral sensing of abiotic stresses in plants

Recent attempts, advances and challenges, as well as future perspectives regarding the application of proximal hyperspectral sensing (where sensors are placed within 10 m above plants, either on land-based platforms or in controlled environments) to assess plant abiotic stresses have been critically reviewed. Abiotic stresses, caused by either physical or chemical reasons such as nutrient deficiency, drought, salinity, heavy metals, herbicides, extreme temperatures, and so on, may be more damaging than biotic stresses (affected by infectious agents such as bacteria, fungi, insects, etc.) on crop yields. The proximal hyperspectral sensing provides images at a sub-millimeter spatial resolution for doing an in-depth study of plant physiology and thus offers a global view of the plant's status and allows for monitoring spatio-temporal variations from large geographical areas reliably and economically. The literature update has been based on 362 research papers in this field, published from 2010, most of which are from four years ago and, in our knowledge, it is the first paper that provides a comprehensive review of the applications of the technique for the detection of various types of abiotic stresses in plants.

Phenotypic Variability of Wheat and Environmental Share in Soil Salinity Stress [3S] Conditions

Through choosing bread wheat genotypes that can be cultivated in less productive areas, one can increase the economic worth of those lands, and increase the area under cultivation for this strategic crop. As a result, more food sources will be available for the growing global population. The phenotypic variation of ear mass and grain mass per ear, as well as the genotype × environment interaction, were studied in 11 wheat (Triticum aestivum L.) cultivars and 1 triticale (Triticosecale W.) cultivar grown under soil salinity stress (3S) during three vegetation seasons. The results of the experiment set on the control variant (solonetz) were compared to the results obtained from soil reclaimed by phosphogypsum in the amount of 25 t × ha−1 and 50 t × ha−1. Using the AMMI analysis of variance, there was found to be a statistically significant influence of additive and non-additive sources of variation on the phenotypic variation of the analyzed traits. Although the local landrace Banatka and the old variety Bankut 1205 did not have high enough genetic capacity to exhibit high values of ear mass, they were well-adapted to 3S. The highest average values of grain mass per ear and the lowest average values of the coefficient of variation were obtained in all test variants under microclimatic condition B. On soil reclaimed by 25 t × ha−1 and 50 t × ha−1 of phosphogypsum, in microclimate C, the genotypes showed the highest stability. The most stable genotypes were Rapsodija and Renesansa. Under 3S, genotype Simonida produced one of the most stable reactions for grain mass per ear.

Combining Hyperspectral Reflectance and Multivariate Regression Models to Estimate Plant Biomass of Advanced Spring Wheat Lines in Diverse Phenological Stages under Salinity Conditions

An area of growing interest in wheat-breeding programs for abiotic stresses is the accurate and expeditious phenotyping of large genotype collections using nondestructive hyperspectral sensing tools. The main goal of this study was to use data from canopy spectral signatures (CSS) in the full-spectrum range (400–2500 nm) to estimate and predict the plant biomass dry weight at booting (BDW-BT) and anthesis (BDW-AN) growth stages, and biological yield (BY) of 64 spring wheat germplasms exposed to 150 mM NaCl using 13 spectral reflectance indices (SRIs, consisting of seven vegetation-related SRIs and six water-related SRIs) and partial least squares regression (PLSR). SRI and PLSR performance in estimating plant traits was evaluated during two years at BT, AN, and early milk grain (EMG) growth stages. Results showed significant genotypic differences between the three traits and SRIs, with highly significant two-way and three-way interactions between genotypes, years, and growth stages for all SRIs. Genotypic differences in CSS and the relationships between the three traits and a single wavelength over the full-spectrum range depended on the growth stage. Water-related SRIs were more strongly correlated with the three traits compared with vegetation-related SRIs at the BT stage; the opposite was found at the EMG stage. Both types of SRIs exhibited comparable associations with the three traits at the AN stage. Principal component analysis indicated that it is possible to assess plant biomass variations at an early stage (BT) through published and modified SRIs. SRIs coupled with PLSR models at the BT stage exhibited good prediction capacity of BDW-BT (57%), BDW-AN (82%), and BY (55%). Overall, results demonstrated that the integration of SRIs and multivariate models may present a feasible tool for plant breeders to increase the efficiency of the evaluation process and to improve the genetics for salt tolerance in wheat-breeding programs.

Combining Hyperspectral Reflectance Indices and Multivariate Analysis to Estimate Different Units of Chlorophyll Content of Spring Wheat under Salinity Conditions

Although plant chlorophyll (Chl) is one of the important elements in monitoring plant stress and reflects the photosynthetic capacity of plants, their measurement in the lab is generally time- and cost-inefficient and based on a small part of the leaf. This study examines the ability of canopy spectral reflectance data for the accurate estimation of the Chl content of two wheat genotypes grown under three salinity levels. The Chl content was quantified as content per area (Chl _area, μg cm^-2), concentration per plant (Chl _plant, mg plant^-1), and SPAD value (Chl _SPAD). The performance of spectral reflectance indices (SRIs) with different algorithm forms, partial least square regression (PLSR), and stepwise multiple linear regression (SMLR) in estimating the three units of Chl content was compared. Results show that most indices within each SRI form performed better with Chl _area and Chl _plant and performed poorly with Chl _SPAD. The PLSR models, based on the four forms of SRIs individually or combined, still performed poorly in estimating Chl _SPAD, while they exhibited a strong relationship with Chl _plant followed by Chl _area in both the calibration (Cal.) and validation (Val.) datasets. The SMLR models extracted three to four indices from each SRI form as the most effective indices and explained 73-79%, 80-84%, and 39-43% of the total variability in Chl _area, Chl _plant, and Chl _SPAD, respectively. The performance of the various predictive models of SMLR for predicting Chl content depended on salinity level, genotype, season, and the units of Chl content. In summary, this study indicates that the Chl content measured in the lab and expressed on content (μg cm^-2) or concentration (mg plant^-1) can be accurately estimated at canopy level using spectral reflectance data.