The spectral‐based photochemical reflectance index (PRI) and leaf surface temperature (Tleaf) derived from thermal imaging are two indicative metrics of plant functioning. The relationship of PRI with radiation‐use efficiency (RUE) and Tleaf with leaf transpiration could be leveraged to monitor crop photosynthesis and water use from space. Yet, it is unclear how such relationships will change under future high carbon dioxide concentrations ([CO2]) and drought. Here we established an [CO2] enrichment experiment in which three wheat genotypes were grown at ambient (400 ppm) and elevated (550 ppm) [CO2] and exposed to well‐watered and drought conditions in two glasshouse rooms in two replicates. Leaf transpiration (Tr) and latent heat flux (LE) were derived to assess evaporative cooling, and RUE was calculated from assimilation and radiation measurements on several dates along the season. Simultaneous hyperspectral and thermal images were taken at 1.5 m from the plants to derive PRI and the temperature difference between the leaf and its surrounding air (Tleaf−air). We found significant PRI and RUE and Tleaf−air and Tr correlations, with no significant differences among the genotypes. A PRI–RUE decoupling was observed under drought at ambient [CO2] but not at elevated [CO2], likely due to changes in photorespiration. For a LE range of 350 W m–2, the ΔTleaf−air range was 10°C at ambient [CO2] and only 4°C at elevated [CO2]. Thicker leaves in plants grown at elevated [CO2] suggest higher leaf water content and consequently more efficient thermoregulation at high [CO2] conditions. In general, Tleaf was maintained closer to the ambient temperature at elevated [CO2], even under drought. PRI, RUE, ΔTleaf−air, and Tr decreased linearly with canopy depth, displaying a single PRI‐RUE and ΔTleaf−airTr model through the canopy layers. Our study shows the utility of these sensing metrics in detecting wheat responses to future environmental changes.
Vertical green living walls (VGWs)—growing plants on vertical walls inside or outside buildings—have been suggested as a nature-based solution to improve air quality and comfort in modern cities. However, as with other greenery systems (e.g., agriculture), managing VGW systems requires adequate temporal and spatial monitoring of the plants as well as the surrounding environment. Remote sensing cameras and small, low-cost sensors have become increasingly valuable for conventional vegetation monitoring; nevertheless, they have rarely been used in VGWs. In this descriptive paper, we present a first-of-its-kind remote sensing high-throughput monitoring system in a VGW workplace. The system includes low- and high-cost sensors, thermal and hyperspectral remote sensing cameras, and in situ gas-exchange measurements. In addition, air temperature, relative humidity, and carbon dioxide concentrations are constantly monitored in the operating workplace room (scientific computer lab) where the VGW is established, while data are continuously streamed online to an analytical and visualization web application. Artificial Intelligence is used to automatically monitor changes across the living wall. Preliminary results of our unique monitoring system are presented under actual working room conditions while discussing future directions and potential applications of such a high-throughput remote sensing VGW system.
The combination of a future rise in atmospheric carbon dioxide concentration ([CO2]) and drought will significantly impact wheat production and quality. Genotype phenology is likely to play an essential role in such an effect. Yet, its response to elevated [CO2] and drought has not been studied before. Here we conducted a temperature‐controlled glasshouse [CO2] enrichment experiment in which two wheat cultivars with differing maturity timings and life cycle lengths were grown under ambient (aCO2 approximately 400 μmol mol–1) and elevated (eCO2 approximately 550 μmol mol–1) [CO2]. The two cultivars, bred under dry and warm Mediterranean conditions, were well‐watered or exposed to drought at 40% pot holding capacity. We aimed to explore water × [CO2] × genotype interaction in terms of phenology, physiology, and agronomic trait response. Our results show that eCO2 had a significant effect on plants grown under drought. eCO2 boosted the booting stage of the late‐maturing genotype (cv. Ruta), thereby prolonging its booting‐to‐anthesis period by approximately 3 days (p < 0.05) while unaffecting the phenological timing of the early‐maturing genotype (cv. Zahir). The prolonged period resulted in a much higher carbon assimilation rate, particularly during pre‐anthesis (+87% for Ruta vs. +22% for Zahir under eCO2). Surprisingly, there was no eCO2 effect on transpiration rate and grain protein content in both cultivars and under both water conditions. The higher photosynthesis (and transpiration efficiency) of Ruta was not translated into higher aboveground biomass or grain yield, whereas both cultivars showed a similar increase of approximately 20% in these two traits at eCO2 under drought. Overall, Zahir, the cultivar that responded the least to eCO2, had a more efficient source‐to‐sink balance with a lower sink limitation than Ruta. The complex water × [CO2] × genotype interaction found in this study implies that future projections should account for multifactor interactive effects in modeling wheat response to future climate.
To meet the ever-growing global population necessities, integrating climate-change-relevant plant traits into breeding programs is required. Developing new tools for fast and accurate estimation of chlorophyll parameters, chlorophyll a (Chl-a) content, chlorophyll b (Chl-b) content, and their ratio (Chl-a/b), can promote breeding programs of wheat with enhanced climate adaptability. Spectral reflectance of leaves is affected by changes in pigment concentration and can be used to estimate chlorophyll parameters. The current study identified and validated the top known spectral indices and developed new vegetation indices (VIs) for Chl-a and Chl-b content estimation and used them to non-destructively estimate Chl-a/b values and compare them to hyperspectral estimations. Three wild emmer introgression lines, with contrasting drought stress responsiveness dynamics, were selected. Well-watered and water-limited irrigation regimes were applied. The wheat leaves were spectrally measured with a handheld spectrometer to acquire their reflectance in the 330 to 790 nm range. Regression models based on calculated VIs as well as all hyperspectral curves were calibrated and validated against chlorophyll extracted values. The developed normalized difference spectral indices (NDSIs) resulted in high accuracy of Chl-a (NDSI415,614) and Chl-b (NDSI406,525) estimation, allowing for indirect non-destructive estimation of Chl-a/b with root mean square error (RMSE) values that could fit 6 to 10 times in the range of the measured values. They also performed similarly to the hyperspectral models. Altogether, we present here a new tool for a non-destructive estimation of Chl-a/b, which can serve as a basis for future breeding efforts of climate-resilient wheat as well as other crops.
19 June 2025 in Agricultural and Forest Meteorology
Leaf water potential (leaf) is a key indicator of plant water status, but its measurement is labor-intensive and limited in spatial coverage. While remote sensing has emerged as a useful tool for estimating vegetation water status, leaf remains unexplored, particularly in mixed forests. Here, we use spectral indices derived from unmanned aerial vehicle-based hyperspectral imaging and machine learning algorithms to assess leaf in a mixed, multi-species Mediterranean forest comprised of five key woody species: Pinus halepensis, Quercus calliprinos, Cupressus sempervirens, Ceratonia siliqua, and Pistacia lentiscus. Hyperspectral images (400–1000 nm) were acquired monthly over one year, concurrent with leaf measurements in each species. Twelve spectral indices and thousands of normalized difference spectral index (NDSI) combinations were evaluated. Three machine learning algorithms—random forest (RF), extreme gradient boosting (XGBoost), and support vector machine (SVM)—were used to model leaf. We compared the machine learning model results with linear models based on spectral indices and the NDSI. SVM, using species information as a feature, performed the best with a relatively good leaf assessment (R2 = 0.53; RMSE = 0.67 MPa; rRMSE = 28%), especially considering the small seasonal variance in leaf ( = 0.8 MPa). Predictions were best for Cupressus sempervirens (R2 = 0.80) and Pistacia lentiscus (R2 = 0.49), which had the largest leaf variances ( > 1 MPa). Aggregating data at the plot scale in a ‘general’ model markedly improved the leaf model (R2 = 0.79, RMSE = 0.31 MPa; rRMSE = 13%), providing a promising tool for monitoring mixed forest leaf. The fact that a non-species-specific, ‘general’ model could predict leaf implies that such a model can also be used with coarser resolution satellite data. Our study demonstrates the potential of combining hyperspectral imagery with machine learning for non-invasive leaf estimation in mixed forests while highlighting challenges in capturing interspecies variability.