Prediction of wheat grain yield by measuring root electrical capacitance at anthesis

Larger root system size (RSS) is critical for increased early vigour and water use, it contributes to enhanced grain yield (GY) in crops (Fageria, 2013), thus emphasizing the importance of applying field root phenotyping techniques in breeding programmes (Postic et al., 2019). Nevertheless, as conventional root investigation methods are generally laborious and destructive, and the isolation of the intact root system from field soil is practically impossible, the investigation of roots is often neglected compared to those of shoots. The measurement of root electrical capacitance (CR) is a promising, rapid in situ technique capable of screening numerous plants at different growth stages. Moreover, the sampled plants can be harvested at maturity to determine GY and can also be used for reproduction (Středa et al., 2020). The CR method was successfully applied in the field to evaluate the effect of dwarfing genes on the RSS of barley (Chloupek et al., 2006), in order to select barley and wheat genotypes for higher RSS and drought tolerance (Chloupek et al., 2010; Svačina et al., 2014; Heřmanská et al., 2015), to assess the root diversity and water use of wheat varieties (Středa et al., 2012; Nakhforoosh et al., 2014), and to estimate canola RSS in relation to lodging resistance (Wu and Ma, 2016). Some of these studies demonstrated significant relationships between the CR-based root size and individual GY, particularly in dry environments. © 2021 Institute of Agrophysics, Polish Academy of Sciences I. CSERESNYÉS et al. 160 The measurement technique is based on the correlation between RSS variables and the CR detected between a ground electrode (inserted into the soil) and a plant electrode (fixed on the stem) using a low-frequency alternating current (AC) signal (Chloupek, 1972). Conceptual models consider the roots to be imperfect cylindrical capacitors, in which the amount of electric charge stored by the polarizable membrane dielectrics depends on the root-soil interfacial area (Dalton, 1995). Even though some of the underlying biophysical principles are still unclear and there are uncertainties about the relative contribution of proximal and distal (fine) roots to the magnitude of the CR detected (Dietrich et al., 2012; Ellis et al., 2013; Cseresnyés et al., 2020; Peruzzo et al., 2020), several pot and field trials have convincingly demonstrated the efficiency of the capacitance method (Středa et al., 2020). One advantage of the technique is that, as the CR value is affected not only by the size but also by the histological properties of the roots (e.g. suberization), the method characterizes both root physiological status and its functionality (Ellis et al., 2013; Cseresnyés et al., 2018). Even though the measured capacitance is very sensitive to soil water content (SWC), this effect can be taken into account by converting the measured CR to saturation (apparent) capacitance, CR*, which was detected in experiments on water-saturated soil (Cseresnyés et al., 2018). This adaptation allows us to compare the field data collected at different dates (under different SWC), which was previously considered to be a serious limitation for the capacitance technique (Chloupek et al., 2010; Středa et al., 2012). In this manner, field monitoring revealed that CR, as a proxy of root activity, peaked during flowering in maize and soybean (Cseresnyés et al., 2018). Minirhizotron and soil core studies verified that wheat root biomass and root length reached a maximum around anthesis, in parallel with the peaks of leaf area, transpiration and water use, and were also significantly correlated with stand GY (Wang et al., 2014; Yang et al., 2018; Postic et al., 2019). A methodological field study involving three winter wheat cultivars is presented here. As intercropping systems have gained increasing attention in organic farming worldwide due to more efficient, complementary resource use (Bedoussac and Justes, 2011; Lithourgidis et al., 2011), wheat-pea mixtures were tested to compare them with wheat sole crops. Focusing on wheat, RSS was assessed merely on the basis of CR* measured in situ at anthesis. The specific aims of the study were (i) to study the correlation of the individual CR* values with the total aboveground biomass (TAB) at maturity and also with GY for each wheat cultivar in order to validate the stand-scale results, (ii) to evaluate the effect of pea intercropping with halved wheat density on mean CR* and the corresponding GY using a stand scale over a three-year period, and (iii) to analyse the relationship between mean CR* and GY across the cropping systems and years. In brief, the study examined the relevance of the capacitance method in the field, or more precisely, the efficiency with which wheat grain yield may be predicted by measuring the saturation root capacitance (CR*) at anthesis under different cultivation and climatic conditions. MATERIALS AND METHODS The field study was conducted during three winter wheat growing seasons from 2017 to 2020 (referred to as harvest years 2018, 2019 and 2020) in a certified organic field in Martonvásár, Central Hungary (N 47°18’, E 18°47’, 109 m a.s.l.). The soil was a Haplic Chernozem (36% sand, 41% silt, 23% clay) with a pH value of 7.66, 1.61% CaCO3, 3.22% humus, 1887/361/445 mg kg total N/P/K and 0.309 cm cm water content at field capacity. The climate is continental with a mean (1987-2016) annual temperature of 11.0°C (January: –1.0°C, July: 21.2°C) and annual precipitation of 548 mm, of which 193 mm falls during the main crop growing season (March-June; Fig. 1). There were optimal rainfall conditions in 2018. By contrast, late-winter and Fig. 1. Monthly rainfall (mm, columns) and mean air temperature (°C, lines) at the experimental site (Martonvásár, Hungary) during the winter wheat growing seasons. The long-term (1987–2016) average is displayed as a reference. WHEAT YIELD PREDICTION BY ROOT ELECTRICAL CAPACITANCE 161 spring droughts occurred in the next two seasons with sufficient precipitation only occurring from early May (flowering stage) in 2019 and from late May (milk stage) in 2020. Winter wheat (Triticum aestivum L.) cultivars ‘Mv Nádor’ (“N”) and ‘Mv Kolompos’ (“K”) and the YQCCP composite population (“C”) were sown in October each year in 6 × 1 m plots with 12 cm row spacing as sole crops (“0”) at a density of 300 seeds m, and at half that density (150 seeds m) intercrops (“P”) with winter pea (Pisum sativum L., cv. Aviron; 50 seeds m). The three replications of each treatment were randomly arranged in the same field, with each one being surrounded by a 1 m border strip, but in slightly different places each year. Natural fertilizers and artificial chemicals were not used directly, which latter is even banned in organic agriculture. At the time of anthesis (in early to mid-May, depending on the cultivar and year) 15 wheat plants were randomly selected from the inner rows of each plot. SWC was measured in the 0-12 cm layer 5 cm away from each sample plant (equal to the depth and position of the CR ground electrode) with a calibrated CS620 portable TDR meter (Campbell Sci. Ltd., Loughborough, UK). The relative water saturation (θrel) value was calculated by dividing the measured volumetric SWC values (cm cm) by the predetermined saturation water content of 0.476 cm cm (Cseresnyés et al., 2018). Thereafter, parallel CR was recorded for the selected plants with a U1733C handheld LCR meter (Agilent Co. Ltd., Penang, Malaysia) at 1 kHz, 1 V AC. The ground electrode was a stainless steel rod 15 cm in length and 6 mm in diameter (303S31; RS Pro GmbH., Gmünd, Austria), pushed vertically into the soil 5 cm from the stem to a depth of 12 cm. The plant electrode was clamped to all of the basal parts of the plant 15 mm above the soil (Svačina et al., 2014) after smearing them with conductivity gel. In order to eliminate the SWC effect, all of the CR data were converted into CR*, according to the empirical function: CR* = CR 5.807erel, using the relevant θrel values (for a detailed calculation, see Cseresnyés et al., 2018). After the CR measurements were complete, five randomly selected wheat plants per plot were cut at ground level, and oven-dried at 70°C until a constant weight was achieved in order to determine shoot dry mass (SDM; ±0.001 g). In the last year (2020) the plants chosen for measuring CR were individually tagged. At maturity (in early July), the tagged plants were hand harvested and oven-dried to determine TAB, after which they were hand threshed to obtain plant GY. Thereafter, the plots were harvested mechanically, and the wheat grains were separated from the peas and weighed. The mean plant GY was determined for each plot on the basis of wheat seedling density. The data were analysed with Statistica 13.0 software (StatSoft Inc., Tulsa, OK, USA). The unpaired t-test or one-way ANOVA with Tukey’s posthoc test was performed to compare the means of CR*, SDM and GY (p < 0.05). If the F-test or Bartlett’s test indicated unequal variances, Welch’s t-test or Kruskal-Wallis with Dunn’s posthoc test was used. Linear regression analysis was applied to relate CR* to TAB, SDM and GY. The resultant regressions were compared using a linear analysis of covariance (ANCOVA).