Peanut is a highly nutritious and high-valued crop worldwide; but achieving competitive yield and quality is often challenged by biotic and abiotic stresses as well as relatively high production requirements compared with other crops. To successfully address these challenges, it is important to understand peanut physiology and its essential requirements for growth and development.

Water deficit stress is responsible for more crop loss than any other abiotic or biotic stress. In peanut, recent work is focusing on understanding physiological responses associated with plant dehydration (Borém et al., 2012). Examination of these physiological responses is often done under the context of identifying specific plant traits governing possible tolerance to “drought” conditions. In general, water deficit tolerance refers to crop traits providing adaptations of dehydration avoidance or tolerance strategies to water limitation (Blum, 1988; Pallardy, 2007), while drought refers to the severity and timing of water deficit events occurring during crop development in a given environment. Avoidance is different from tolerance; avoidance can be accomplished by deeper rooting to sustain water uptake, while tolerance is more related to survival via osmotic adjustment. An intermediate strategy can be that of reducing transpiration (whether caused by stomatal response to VPD or naturally low hydraulic conductance) to maintain relatively high turgor and less stress, while forgoing some carbon gain from photosynthesis whereby some yield may be lost.

In the United States, peanut production is primarily in the southeast (GA, FL, AL, MS), southwest (TX, OK, NM), and the Virginia-Carolina (VC) region (VA, NC, SC). The Ogallala aquifer is the primary source of irrigation for production of peanuts, cotton, corn, sorghum, and wheat in the southwestern U.S. This aquifer has been used heavily since the 1950s and, although irrigation efficiency has improved, the current rate of decline in the water table is 0.8 – 1.1 feet per year and it is estimated that the aquifer could be depleted within 30 to 40 years (Paxton, personal communication). Decline in some locales is already reducing the ability to irrigate when plant demand is the highest, during the midsummer months from flowering to mid-pod fill when air temperatures routinely exceed 40 C. For the southeastern U.S. peanut producing region, there are also issues of water scarcity compounded with high growing season temperatures. This region is not inherently thought of as having water resource issues, but because rainfall is typically erratic during the growing season and groundwater supplies are increasingly being diverted to rapidly growing major urban areas, the issues related to water quantity are reaching critical proportions. Similarly, in the VC region, precipitation amount and distribution are erratic, and, unlike the other regions, irrigation is unavailable for over 90% of the peanut growing land. In addition to the obvious implications for yield, a significant concern is that water deficit stress promotes production of the highly carcinogenic aflatoxin mycotoxins (Butts et al., 2023; Williams et al., 2004), which can be minimized by irrigation.

Recent examples abound of abiotic stress affecting U.S. peanut production. Approximately one-third of Texas peanut acreage was not harvested in 2000 and again in 2011 due to drought and heat stress. In Georgia, drought and aflatoxin contamination of the crop resulted in an inability to market large quantities of peanuts in 2002 and 2019. In Virginia, severe drought in 2010 caused a 60% decline of the state average yield. The severe drought in 2006 and 2007 caused a 40% decline in southwest peanut production. On a global scale, the effects of these abiotic stresses on peanut crops are increasingly more severe and will continue to affect directly and indirectly through sparking of new diseases and pest insect cycles as the result of climate change (Prasad et al., 2000; Ramsey, Tack & Balota, 2020; Paredes et al., 2024; Porcuna-Ferrer et al., 2024). Particularly in Africa and Asia, where irrigation water quantity and quality are limited, the effects of water stress are compounded by the effects of high temperature. In these regions, drought tolerant indigenous crops, including peanut, seem to lose their resilience in the face of climate change (Porcuna-Ferrer et al., 2024).

Droughts and high temperatures reduce not just yield but also peanut seed size and quality, with negative effects on crop revenue (Balota, 2020; Ramsey, Tack & Balota, 2020). For example, Balota (2020) showed that a relatively moderate drought reduced sound mature kernel (SMK) content by 7.5% while severe drought depressed SMK by 25%. The SMK is an important grade factor with significant contribution to the revenue, i.e. the higher the SMK the higher the revenue. Similarly, Ramsey et al. (2020) predicted an approximately 11% revenue decline for Virginia-type peanut producers with 1 C temperature increase due to both yield and seed size reduction. These authors examined the varietal improvement as an adaptation strategy to heat and dryland conditions using data from a unique multi-state project, the Peanut Variety and Quality Evaluation (PVQE), collected on Virginia-type peanut accessions for over 35 years in the Virginia-Carolina peanut region of the USA. They concluded that breeding could offset moderate climate warming, but is unlikely to sustain revenue if warming exceeds 1 C. With limited water resources to supply irrigation to peanut crops in the future, solutions must be found to support production and maintain economic sustainability. Recently, progress has been made toward identification of cultivated and diploid peanut accessions that exhibit phenotypic expression of traits associated with drought tolerance that can be used in breeding for resistance (Balota & Isleib, 2020; Bennett et al., 2022; Kumar et al., 2022; Kumar et al., 2024; Massa et al., 2024). Risk management tools have been, also, deployed along with management practices to reduce the effect of drought (Martins et al., 2023; Nboyine et al., 2023), but more research is needed to achieve sustaining yields and nutritional quality of peanuts under increasingly harsh conditions.

A decrease in gas exchange is one of the first physiological responses to drought in peanut, as evident by decreases in transpiration, stomatal conductance, and photosynthesis (Balota et al., 2024; Mohandass et al., 1989; Reddy et al., 2003; Pilon et al., 2018; Zhang et al., 2022; Soba et al., 2024). Ultimately, decreased growth and leaf area, as well as lower photosynthetic efficiency, leads to a decline in net canopy carbon assimilation, directly impacting yield (Reddy et al., 2003; Soba et al., 2024). Thus, of particular interest is the identification of the limiting factors and susceptibility of photosynthetic mechanisms to drought which reduce yield. These factors include stomatal and non-stomatal photosynthetic mechanisms that respond differently to abiotic stress depending on species and severity of the stress (Rucker et al., 1995; Pilon et al., 2018). Additionally, morphological factors such as root architecture, rooting depth, and plant acclimation capacity can contribute directly to plant water status and photosynthetic/source strength (Dang et al., 2024; Puppala et al., 2023; Rowland et al., 2012a; Rowland et al., 2012b). Pilon et al. (2018) addressed stomatal and non-stomatal limitations to carbon assimilation in peanut and showed that while declines in net CO₂ assimilation were driven by decreased stomatal conductance, the primary mechanisms contributing to low photosynthetic rates were non-stomatal (Soba et al., 2024). These non-stomatal limitations included both disruption of electron transport rates through photosystem (PS) II, RuBP regeneration, rubisco carboxylation, and diffusional limitations in the chloroplast (Pilon et al., 2018; Soba et al., 2024). Presumably both the stomatal and non-stomatal mechanisms are affected by and even damaged by low leaf water potential status.

From a plant breeding perspective, research aimed at the discovery and modeling of targeted photosynthetic traits to improve yield and stress tolerance in peanut and other crops; these traits included early biomass accumulation and leaf area index (LAI), ‘chlorophyll stability’, and net assimilation rate (A) (Balota et al., 2024; Bennett et al., 2022; Dang et al., 2024; Zhang et al., 2022; Rucker et al., 1995; Fischer et al., 1998; Kiniry et al., 2005; Nigam and Aruna, 2007; Arunyanark et al., 2008; Balota et al., 2008; Songsri et al., 2008). A comprehensive analysis of traits proposed to improve drought and heat resistance in peanut has been recently published by Puppala et al. (2023). Earlier research showed that LAI and early biomass accumulation had significant effects on peanut yield under stress conditions through improved light interception, radiation use efficiency, higher A, and extended photosynthetic duration (Collino et al., 2001; Kiniry et al., 2005; Arunyanark et al., 2008). This research also emphasized that substantial genetic variation for these traits exists in peanut (Kiniry et al., 2005; Nigam and Aruna, 2007; Arunyanark et al., 2008). In particular, A seemed to be a key trait as its cumulative rate over the growing season is the main driver of biomass and yield. For example, experimental and theoretical analyses suggested that improvement of assimilation in C₃ plants is feasible and can be achieved through increasing photosynthetic rate per unit leaf area (Zhang et al., 2022). Using carbon isotope composition and discrimination (Δ¹³C) chemistry terms, Farquhar and colleagues (1989) developed the theory linking Δ¹³C to A, transpiration efficiency (TE, the ratio of A to transpiration rate), and water use efficiency (WUE, the ratio of biomass to water used to produce that biomass). Since then, numerous studies used Δ¹³C to estimate water use of C₃ plants, including peanut, but this analysis is expensive and low throughput for regular use in breeding (Nageswara Rao and Wright, 1994; Rowland and Lamb, 2005; Zhang et al., 2022). Limited-transpiration (LT) rate at elevated vapor pressure deficit (VPD) and transpiration decrease (TD) early in the soil drying were also proposed as two physiological mechanisms by which soil water could be conserved and used by plants for a longer time in between rain events (Devi et al., 2010).

A fundamental strategy that plants have evolved for regulating water use is altering hydraulic conductance, ultimately influencing transpiration flux. Alterations in transpiration flux can have significant impacts on crop yield because of its impact on CO₂ assimilation for photosynthesis (de Wit, 1958; Passioura, 1996; Blum, 2009). For warm-season C₃ plants like peanut, the general pattern of water use for non-stressed plants is a gradual increase in water use during morning hours peaking at mid-day and declining in the afternoon as leaf water potential decreases (Mohandass et al., 1989). Declining leaf water potential has also been closely related to declines in stomatal conductance ultimately reducing water use (Brodribb and Holbrook, 2003).

Relative humidity (RH), dew, and rain vary greatly across the peanut producing regions of the U.S. For example, RH, the ratio of actual vapor pressure to saturation vapor pressure, depends greatly on temperature. In general, the western hot-semi-arid regions of the U.S. have lower vapor pressures than the hot-humid climate of the southeastern and eastern U.S. Vapor pressure deficit (VPD), i.e. the difference between the amount of actual moisture and how much moisture air can hold, is a measure of how close air is to saturation. For a given environment, VPD is easily derived from RH and temperature, and has spatial and temporal variation through the crop growth cycle. When soil water is unlimited, because the sub-stomatal cavity of the leaf is saturated with water vapor, water diffuses through open stomata outside the leaf under high atmospheric VPD (less moisture in the air). This may result in leaf dehydration, faster soil water depletion, and on-set of water deficit stress.

Sinclair et al. (2017) and others (Turner et al., 1985; Devi et al., 2009; Shekoofa et al., 2014, 2015a, b; Pradhan et al., 2019) capitalized on the response of stomatal conductance to diurnal VPD variation and proposed that LT response to increased midday VPD could be a mechanism of dehydration avoidance. Through this mechanism, plants can conserve soil water at early growth stages for use later in the season to sustain seed filling. The predominant physiological mechanism most likely influencing the LT trait is stomatal regulation. Shekoofa et al. (2015a) documented the LT trait threshold occurring across peanut genotypes was not influenced by relative water content. This indicates that bulk leaf water status or active metabolic processes (abscisic acid transport) is not influencing the LT trait (McAdam and Brodribb, 2014; Sinclair et al., 2017). The current evidence is that LT trait could be a result of particular genotypes having relatively lower hydraulic conductivity (passive metabolic process) (Sinclair et al., 2008; Choudhary et al., 2013; Choudhary et al., 2014; Ocheltree et al., 2014). The general concept is that low hydraulic conducting genotypes under higher VPD have insufficient water flux into leaves to match water loss via transpiration, thus reducing stomatal aperture (stomatal limitation) balancing the water flux in and out of the leaf (Sinclair et al., 2017). Therefore, the presence of LT trait may be associated with resistance factors limiting leaf water transport between xylem and guard cells (Sinclair et al., 2008).

When day/night temperature was maintained at 27/21 C in a growth chamber, Devi et al. (2009) reported the LT trait was present in 9 of the 17 peanut genotypes, i.e. 9 genotypes showed a break of linearity between transpiration and VPD when VPD was greater than 2.2 kPa. Other controlled chamber studies evaluating the impact of temperature on the genotypic LT response concluded that increasing temperature could cause a lack of expression of this trait (Shekoofa et al., 2015a). When the chamber temperature was 31 C, three of the 6 genotypes expressed the LT trait, but at a temperature of 36 C the LT trait was diminished in all genotypes. The lack of response of this trait at warmer temperatures was also observed in a field study (Shekoofa et al., 2015a). The impact of temperature on the expression of this trait seems to be more influential on peanut than C₄ determinate crops of sorghum and maize (Shekoofa et al., 2013; Shekoofa et al., 2015b). Transpiration decrease (TD) in response to soil drying expressed as the fraction transpirable soil water (FTSW) was also investigated as a peanut strategy to conserve soil water at early stages for later use. Devi et al. (2009) studied 17 peanut genotypes and found large variation in the FTSW among the genotypes, i.e. FTSW ranged from 0.22 to 0.71. However, less variability of this trait has been reported for the U.S. commercial peanut cultivars (Devi et al., 2011; Shekoofa et al., 2013). Stomatal sensitivity to soil drying was observed in a field study when comparing genotypes of subspecies fastigiata and hypogaea (Zurweller et al., 2022).

Crop production loss from water deficit stress primarily occurs because of reduction in photosynthesis from stomatal closure limiting CO₂ assimilation (de Witt, 1958; Pilon et al., 2018; Tanner and Sinclair, 1983). Therefore, mechanisms that limit transpiration such as LT and TD must, to some extent, also reduce CO₂ assimilation and may result in counterproductive effects on yield. For peanut, LT and TD were documented with certainty under growth chamber conditions, but evidence of the traits’ expression in the field is limited and their direct contribution to soil water conservation and yield advantage under drought is still to be determined.

Diurnal monitoring of air and canopy temperature allows measurement of canopy temperature depression (CTD), a surrogate for stomatal conductance and transpiration of plants in the field. This technique was used with success for decades in plant breeding for drought and heat tolerance (Ehrler et al., 1978; Blum et al., 1982; Reynolds et al., 1994; Amani et al., 1996; Fischer et al., 1998; Balota et al., 2007). For field grown peanut in a sub humid environment, diurnal CTD variation was monitored with wireless infrared thermometers (Balota et al., 2012a, b). As temperature and VPD increased, transpiration rate estimated from CTD increased from 8:00 to 10:00 EST, i.e., VPD went from 0 to 1.5 kPa during this time; CTD decreased at midday, i.e. from 11:00 to14:00 when VPD went from 2 to 2.5 kPa; then CTD increased again in the afternoon until 15:00 when it started to decline, one hour after VPD began to decline (Fig. 9). A negative linear relationship between transpiration rate and VPD from 11:00 to 14:00 EST for field grown peanut was confirmed by directly measuring stomatal conductance on 8 peanut genotypes with a LI-6400 portable photosynthesis meter (LI-COR Biosciences Inc.) (Hawkins, 2014). This agrees with the classical model that stomata closes when the water content in the sub-stomatal cells declines from intensified water diffusion outside the leaf and reopens when the water content is replenished by hydraulic conductivity (Taiz et al., 2014; Brodribb and McAdam, 2011). Unlike results noted in growth chamber tests, when LT and TD to soil drying were expressed in a handful of peanut lines when temperature was 32 C and even in fewer lines when temperature was 36 C (Shekoofa et al., 2015a; Pradhan et al, 2019), in tests in the field by Balota et al. (2012b) on 30 peanut genotypes, all genotypes partially closed stomata in response to midday VPD increase, regardless of temperature and soil moisture content (Fig. 10). However, by this mechanism of LT, yield variation among genotypes from 4992 to 6758 kg ha^-1 was not explained (Balota, 2010).

Similarly, there is little evidence that growth chamber selected peanut genotypes for LT response to increased VPD or TD with soil drying have yield advantage when grown in the field. For example, in growth chamber testing, Sinclair et al. (2018) found line N12006ol with decreased transpiration rate to soil drying in terms of FTSW, i.e. high FTSW, and speculated that TD may have caused higher yield of N12006ol in low-yield environments in the 2014 PVQE testing. Indeed, in 2014 PVQE trials, line N12006ol had an average yield of 5426 kg ha^-1 across five locations, like the popular cultivar Bailey, but its sister line, N12008olFCSmT, produced 15% more yield than Bailey and N12006ol when grown under the same environmental conditions (Balota et al., 2015). N12008olFCSmT had significantly lower FTSW than N12006ol (Sinclair et al., 2018) but was released as Bailey II in 2017 for its superior yield.

There is no evidence that LT and TD traits expressed early in the season can conserve water for use at later physiological growth stages of seed development under field conditions. In general, growth chamber studies not only provide conditions different than in the field and expressed phenotypes may be different in controlled conditions than in the field, but they disregard the soil-plant-atmosphere relationships, which are the driving force for crop production. For example, in the VC region, the soil is sandy and shallow, and drought stress can develop within 10 to 20 days of no rainfall. Also, at early growth stages, evaporation is more intense than later in the season when plants cover the ground. Under these conditions, field crops in Suffolk, VA, may become more drought stressed than crops in TX in mid-July (Fig. 11). However, some phenotypes can withstand severe field dehydration. The peanut genotype that remains green and turgid plants in Fig. 11, while the entire field is water-stressed, is GP-NC WS 17 (Tallury et al., 2014). This line is an exotic derived line from interspecific hybridization with Arachis cardenasii (Krapov. & W.C. Gregory) with drought and heat tolerance (Tallury et al., 2014). This genotype was screened for LT in growth chambers by Shekoofa et al. (2015a) and did not show the LT trait. However, GP-NC WS 17 showed the greatest stomatal conductance recovery after water deficit stress, greater chlorophyll fluorescence estimated as the ratio of variable to maximum fluorescence (Fv/Fm), and the least membrane injury, when compared with other genotypes in the field (Rosas-Anderson et al., 2014; Singh et al., 2014). Pilon et al. (2018) also documented that non-stomatal factors such as electron transport are important for maintaining relatively high levels of CO₂ assimilation under drought. Similarly, under field conditions when drought was imposed by coverage with rainout shelters, Zhang et al., (2022) showed that some peanut genotypes, i.e. PI 502120 and AU-NPL 17, maintained high photosynthetic and stomatal conductance rates, and produced high yields under drought. The authors called these genotypes “water spenders”, like cultivars of wheat (Triticum aestivum L.) and common bean (Phaseolus vulgaris L.) previously identified with similar responses to drought (Blum et al., 2009; Sanz-Saez et al., 2019). Other cultivars, however, developed an opposite strategy to cope with drought; they closed the stomata early during the onset of drought. These cultivars, i.e. Line-8 and AU16-28, called “water savers” showed lower stomatal conductance but similar yields with “water spenders” (Zhang et al., 2022). The authors concluded that, indeed, water savers did not have a yield advantage over water spenders under field-imposed drought.

Drought tolerance is a complex of morphological and physiological adaptive plant responses including escape, avoidance, and resistance (Blum, 1988). Considering that drought is usually accompanied by high temperatures, the complexity may become overwhelming and unrealistic to address with single traits identifiable through low-throughput techniques in controlled conditions. For agricultural production, adaptive responses must have a yield benefit under drought, while biotic resistance needs not to be forgotten; this makes drought tolerance even more complicated from a breeding perspective. The single trait selection approach has limited physiological applications in plant breeding and constitutes a major point of controversy among breeders and physiologists for decades. Hopefully, the next generation strategy of genomic breeding coupled with advances in phenomics, and machine learning will allow for increasingly more effective inclusion of physiology in breeding and accelerated development of peanut cultivars with improved drought tolerance. Because these strategies have the power to combine a diversity of plant responsive traits, i.e. traits associated with yield, biotic and abiotic tolerance, the mechanisms mostly contributing to increased yield under various “drought” conditions will be more likely to be successfully identified and further used to develop drought tolerant peanut cultivars.

Temperature influences peanut growth and development throughout the season. For optimal productivity, air temperatures for vegetative growth should be from 30 to 35 C, whereas for reproductive development, the optimal range is from 22 to 30 C (Wood, 1968; Cox, 1979; Williams and Boote, 1995).

Multiple physiological processes are affected by sub- or supra-optimal temperatures. Soil temperatures below 18 C slow the germination rate and result in a high percentage of seedlings with thick hypocotyls and underdeveloped, curled root system (Kvien et al., 2022). Conversely, a rise in soil temperatures above 35 C decreases the overall germination percentage and necrotic areas are likely to develop on hypocotyls and roots of seedlings due to disruption in cell membranes.

The balance between photosynthesis and respiration mediates cell growth and expansion in peanut leaves; thus, photosynthesis and respiration are essential physiological processes associated with tissue growth and development. The air temperature for optimal photosynthesis in peanut plants is around 30 C, but photosynthesis occurs with variable rates over an extended temperature range (10 to 35 C). Cold stress is more likely to occur over the first 30 DAP in spring-sown peanut fields, when temperatures are not consistently warm enough to sustain plant growth. Low temperatures can considerably impair early season growth. The lowest critical average temperature at which growth is interrupted, is near 12 C (Cox and Martin, 1974). At temperatures below 20 C, peanut seedlings produce fewer smaller leaves and shorter stems when compared to seedlings grown at 30 C. Faster development of the first peanut true leaves contributes to overall plant growth at early season (Virk et al., 2019). The reduced growth of leaves and stems caused by low temperatures is primarily associated with decreased photosynthetic rate. The underlying factors causing decreased photosynthetic rate under low temperatures include reduced quantum efficiency of PS II and poor performance of the PS II and PS I intersystem, which reduces the overall energy produced during the light-dependent reactions with further inhibition of photosynthetic performance (Virk et al., 2019). Temperatures below 22 C resulted in delayed flower appearance, decreased number of flowers, and impaired pollination (Zhang et al., 2019a). Shifting planting to a later date can provide higher temperatures at the seedling stage. Moreover, temperatures lower than 10 C or higher than 35 C may promote irreversible damage to the photosynthetic system (Berry and Björkman, 1980).

High air temperatures generally accelerate the progress of reproductive growth stages in peanut. Plants initiate and complete flowering earlier and pods mature more rapidly under heat stress (Akbar et al., 2017). In a study with 62 peanut genotypes, plants reached 75% flowering between 35 and 50 days under temperatures ranging from 28 to 30 C. When the genotypes were exposed to temperatures from 33 to 39 C, 75% flowering was achieved in 29 to 37 days (Akbar et al., 2017). Pod maturity also occurred sooner under hot environments. Although early pod set and maturity may benefit planting and harvest logistics, overall yield is considerably decreased by warmer temperatures, partly because of shorter growth cycle and less resource capture (Prasad et al., 2003).

The reproductive stages of peanut are more susceptible to the negative effects of high temperatures than the vegetative stages. Flower fertility is highly sensitive to temperatures above 35 C, resulting in low fruit set (Prasad et al., 2000). Flower bud temperature is typically similar to air temperature up to 30 C. However, at air temperatures above 35 C, the cooling effects are insufficient to prevent the flower buds’ heat injury. Fruit-set may be reduced due to impaired micro- and mega-sporogenesis, and reduced fertilization of flowers present on the plant (Prasad et al., 2003). Reduced flower fertilization under high temperatures occurs due to poor pollen production and viability, and low pollen tube growth (Prasad et al., 2003). Since self-pollination of peanut flowers occurs early morning and the flower opens only after fertilization, the most critical period for heat is the first hours of the day. Optimum temperatures for flower production are between 25 and 30 C, whereas plants tolerate temperatures up to 34 C for pod development. Pollen viability is above 90% at temperatures of approximately 32 C, decreasing to 68% at 40 C with a further decline to 0% at 44 C (Prasad et al., 2003).

High air temperatures also negatively impair the photosynthetic process in peanut leaves. Temperatures above 40 C during flowering and pod development (74 DAP) decreased net photosynthesis and stomatal conductance by 35% on average for seven genotypes (Pilon et al., unpublished). However, nighttime respiration was not impacted by midday high temperatures. The lower photosynthetic rate resulted in less pod dry weight only for two of the seven genotypes, suggesting that five peanut genotypes exhibited greater heat tolerance. These genotypes decreased photosynthesis during heat stress but subsequently increased their photosynthetic rates after the stress period, allocating a greater proportion of photoassimilates to pod development (Pilon et al., unpublished).

Soil temperature is important in regulating pod initiation, pod growth rate, and mature seed mass. Soil temperatures near 20 C in the fruiting zone improved pod initiation rate, resulting in a greater number of pods per plant; but 20 C did not result in optimal pod growth rate, thus seed mass at maturity was low at this temperature (Golombek and Johansen, 1997). Temperatures between 30 and 32 C are suitable for greater pod growth rate and mature seed mass (Golombek and Johansen, 1997). Yet, soil temperature increases to 35 C may reduce peg and pod development by 33% (Ketring, 1984), and seed mass by 45%, resulting in lower overall pod yield (Akbar et al., 2017).

Growth of vegetative tissues (leaves and stems) is generally accelerated by high temperatures up to 35 C (Prasad et al., 2003). In contrast, reproductive units (pegs and pods) require slightly lower temperatures (near 30 C) for optimal growth. Under heat stress, crop growth rate (CGR) is high, suggesting great accumulation of photo-assimilates; however, pod growth rate (PGR) is generally low, thus indicating insufficient partitioning of accumulated photoassimilates (Craufurd et al., 2003), which is caused by poor pod-set (Prasad et al., 2003). The ultimate consequence of low pod-set and poor partitioning of photosynthates from the leaves to pods under heat stress results in significant reduction in pod yield, regardless of vigorous aboveground growth.

The partitioning factor (ratio of PGR to CGR) has been used as a criterion for heat tolerance selection in peanuts (Ntare et al., 2001; Craufurd et al., 2003). Other physiological assessments employed as screening techniques for temperature tolerance included cell membrane thermostability, chlorophyll fluorescence, and flower to fruit-set ratio (Srinivasan et al., 1996; Talwar et al., 1999). Comprehensive knowledge of the physiological mechanisms underlying tolerance to cold and heat would contribute to identifying cultivars adapted to locations with predominant extreme temperatures.

Phenotype corresponds to plant characteristics that can be quantitatively measured, and which depend on the interaction between genotype and environment (G × E). Plant physiologists and breeders have collected phenotypic data for decades to understand how plants respond to biotic and abiotic stressors, and how to breed for the most suitable phenotypes under specific environmental conditions. Historically, phenotypes of interest include but are not limited to plant height, biomass, LAI, photosynthesis, root angle, root length density, seed number, and hundred seed weight (Balota et al., 2024; Kolhar and Jagtap, 2023; Sarkar et al., 2020; Sarkar et al., 2021). Currently, phenotypes can be taken using simple and more technologically complex methodologies that, although appropriate and accurate, are limited by the cost and time to collect the data. As breeding programs analyze hundreds of genotypes in multiple spatial and temporal conditions, deployment of high throughput methods is needed.

The development of DNA sequencing technologies has generated the opportunity to analyze plant genomes, providing scientists with tools to understand the genetic regulation of specific phenotypes in quick and affordable ways (Jiao and Schneeberger, 2017; Yang et al., 2017). Development of recombinant inbred lines (RIL), near isogenic lines, and diversity panels of plant populations, makes it possible to map Quantitative Trait Loci (QTL) and thus code quantitative phenotypes needed to improve crop performance under different environments (Adachi et al., 2011; Dhanaphal et al., 2015; Oakley et al., 2018; Kassie et al., 2023). In peanut, QTLs have been identified for plant height, leaf length, yield components, and yield (Khedikar et al., 2018; Lv et al., 2018; Mondal et al., 2019); seed composition traits such as oil content (Hu et al., 2018; Liu et al., 2024); seed dormancy (Wand et al., 2022); mineral nutrient contents (Zhang et al., 2019b); disease tolerance (for a review on this theme see Desmae et al., 2018; Choudhary et al., 2019; Liang et al., 2021); and yield components related with drought tolerance (Faye et al., 2015; Desmae et al., 2018; Kumar et al., 2024). However, to produce strong correlations between phenotypes and their genomic regions, thousands of genotypes have to be phenotyped across environments, growth stages, and years; this is difficult to accomplish with traditional phenotyping methods. Thus, phenotyping has become the bottleneck for genomic discovery and the breeding efforts (Ninomiya, 2022).

In the past 30 years, improvements in digital image acquisition, automation, and robotics have made high-throughput phenotyping (HTP) possible (Furbank and Tester 2011; Fiorani and Schurr, 2013). HTP methods were initially developed for greenhouses and growth chamber applications using plants in early stages of growth; thus, they were unreliable for field screening and not valid for breeding (Araus et al., 2014; Ninomiya, 2022). After 2010, field HTP platforms started to be developed (Araus and Cairns, 2014; Araus et al., 2018a). Field HTP combines remote sensing via satellite and aerial sensing and uses terrestrial and unmanned aerial vehicle (UAV) platforms equipped with a variety of sensors to assess physiological phenotypes related with yield and stress adaptation (Araus and Cairns, 2014; York et al., 2015).

Remote sensing technologies are becoming increasingly more affordable to the end-users, breeders, and farmers. These technologies collect multiple plant characteristics in a relatively short time, repeatedly and consistently over the season with increased accuracy. Each of these technologies requires two main components. First, they require a vehicle or a platform to bring the sensors on site. Platforms are of different types, and they are complex and costly. The second part is the technology itself in the form of one or multiple sensors for data collection.

Phenotyping platforms

Several types of systems are currently available, with different levels of adaptability and upgradability, mobile or stationery, and with a drastic price range from hundreds to millions of dollars (Fig. 12) (Yuan et al., 2023). Platforms can be immobile, using single or multiple cameras and other sensors mounted on single or multiple poles to take time-lapse images (Brocks et al., 2016; Zhou et al., 2017). Others are stationary field scanner systems on a movable gantry (Virlet et al., 2017; Burnette et al., 2018). The Field Phenotyping Facility at the University of Nebraska Lincoln is one acre in size, and includes sensors and cameras mounted on an automatic cable suspended carrier system (Spidercam) attached to four giant poles in each corner of the field (Kirchgessner et al., 2017; Bai et al., 2019). Significant limitations of the stationary platforms are the high maintenance cost and the limited number of experiments that can be accommodated, e.g. use of any chemicals as treatments may result in negative effects on the subsequent experiments.

Alternatives to the stationary platforms are mobile, manned or unmanned, platforms. Mobile remote sensing systems include phenomobiles, aircraft, and satellites. The phenomobiles are ground-based vehicles such as unmanned robots or manned diesel-powered tractors and pushcarts (Pittman et al., 2015; Yuan et al., 2018; Antonucci et al., 2023). While phenomobiles allow freedom of use in multiple environments over stationary platforms, they are also limited by the geography of the land through which they can be driven. For example, their use can be limited for tall canopies, wide-spread ground cover crops, and uneven ground. For tractor-based platforms, a significant limitation is the area where they can be safely and time-effectively moved to collect data.

Manned and unmanned aircraft bypass the terrain limitations of the terrestrial platforms. Aerial imaging allows faster and broader coverage of a field than ground-based vehicles. Manned aircraft are well-suited for large landscapes consisting of several hundred to thousands of acres, but their high cost, flight restrictions to certain zones, and the requirement and availability of pilots make their use limited. Drones or unmanned aircraft (UAVs) with either fixed wings or multi-rotors are better options for small farms and research areas. A disadvantage of UAVs, however, is that battery capacity limits the flight mission and payload.

Satellite imagery collected from the Sentinels of the European Space Agency and Landsat from the US is used for phenotyping, mainly for large landscape areas. Satellite imagery is based on vegetation-derived reflectance at multiple wavelengths, from which numerous vegetative indices have been developed for different applications (Reed et al., 1994; Labus et al., 2002; Johnson et al., 2003). The major disadvantage of satellites is their lower spatial resolution in comparison with aircraft. Image resolution is 30 m² or more for Landsat 8, 3 m² for Planetscope, and 1 m² for SkySat, which eliminates small plot comparisons, and intra-plot and plant-level phenotyping. The passage of the satellite above the experiment area as well as the presence of clouds can also limit their use. A general disadvantage of all terrestrial and aerial HTP platforms is that they require expertise at multiple levels to use and maintain them.

Sensors used in field phenotyping

The sensors used in remote sensing are diverse. Laser and ultrasonic sensors are mainly used on ground-based vehicles to estimate plant height. These systems emit a laser or an ultrasonic pulse at a recurrent frequency, which hit the ground or a plant, then reflect back to the sensor. The longest distance measured by the emitted pulse is the bare ground and the smaller is the plant. The subtraction of both distances gives the height of the plant. In this way plant height of maize, bermudagrass, wheat, and alfalfa (Fumiki and Omasa, 2009; Ehlert et al., 2010; Selbeck et al., 2010; Pittman et al., 2015; Cazenave et al., 2019a, b) was successfully estimated.

Light Detection and Ranging (LiDAR) technology has been used for several decades in meteorological applications. More recently, LiDAR is being used to phenotype crops, including peanut, by creating 3D images, either from terrestrial or aerial platforms (Yuan et al., 2019; Sun et al., 2017; Saeed et al., 2023). LiDAR has been used to successfully estimate plant height and, in combination with ground-based platforms, estimate canopy architecture and its components, i.e. stem height, length of lateral branches, and canopy shape and density (Fig. 13).

Photogrammetry uses high resolution cameras mounted on either terrestrial or aerial platforms to sense reflected light in the visible red, green, and blue (RGB) bands of the electromagnetic spectrum. In peanut, seedling vigor, main stem height, wilting, and yield under drought and late leaf spot disease [caused by the fungus Northopassarola personata (Berk and Curt)] were RGB-estimated with great accuracy (Balota and Oakes, 2016, 2017; Sarkar et al., 2020; Sarkar et al., 2021a; Sarkar et al., 2021b; Kassim et al., 2022; Sie et al., 2022; Oteng-Frimpong et al., 2023; Balota et al., 2024; Chapu et al., 2022; Chapu et al., 2024).

Near infra-red and thermal multispectral sensors are used to assess crop color, temperature as result of transpiration, and biomass accumulation from reflectance in the near infra-red and other spectra. Reflectance-derived vegetation indices such as the Normalized Difference Vegetation Index (NDVI), Normalized Difference Red Edge (NDRE), LAI, and canopy temperature in peanut (Balota and Oakes, 2017; Yuan et al., 2018; Sarkar et al., 2020) and other crops (Freeman et al., 2007; Erdle et al., 2011; Gnyp et al., 2014; Pittman et al., 2015; Cazenave et al., 2019b) have been recently developed.

Phenotyping complex physiological processes and hidden plant organs, e.g. roots and peanut pods, is challenging. For these, HTP and remote sensing can only be partially implemented. Therefore, phenotyping mainly relies on low throughput methods (LTP). However, recently Gimode et al. (2023) used Ground Penetrating Radar (GPR) to analyze the below-ground peanut pods non-destructively.

Gas exchange phenotyping

Photosynthesis is a complex process that includes different complex reactions that, for simplification, can be divided into light reactions driven by electrons passing the different photosystems, and CO₂ fixation. Measurement of CO₂ fixation parameters uses infrared gas analyzers (IRGA). Depending on the type of photosynthetic parameter, measurements can take between one min, i.e. midday CO₂ fixation (A), to 90 min, i.e. the maximum rate of Rubisco carboxylation (V_cmax) and electron transport (J_max), and A response to internal CO₂ (Ci) (A-Ci) curves (Farquhar et al., 1980; Bernacchi et al., 2003). Nonetheless, IRGA measurements only give periodic information that may not be representative of the whole plant performance over the growing season. Farquhar et al. (1989) used carbon isotope discrimination (Δ¹³C), as it offers information on photosynthesis and WUE in a more integrated way over the life span of the plant.

WUE phenotyping through carbon isotope discrimination

In peanut, WUE has been demonstrated to be an important trait to select to improve drought stress tolerance (Vadez and Ratnakumar, 2016; Zhang et al., 2022). Water use efficiency can be calculated as the amount of biomass fixed per amount of water used by the plant. The most accurate way to calculate the amount of water used by a plant is gravimetrically, i.e. weighing potted plants every day, re-watering to pre-calculated weights to achieve desired water levels and calculating the amount water lost through transpiration (Vadez and Ratnakumar, 2016; Zhang et al., 2022). However, this method is time consuming, and few facilities can afford to perform such big scale, controlled experiments. Farquhar et al. (1989) reported a strong relationship between WUE calculated gravimetrically and Δ¹³C. This relationship has been demonstrated in several crop species such as common bean (Ehleringer et al., 1991), soybean (Dhanapal et al., 2015), wheat (Del Pozo et al., 2016), and peanut (Wright et al., 1993), allowing Δ¹³C to be used as surrogate phenotyping for WUE. In peanut, Δ¹³C is highly inheritable (Hubick et al., 1988), therefore allowing identification of the genomic regions responsible for changes in Δ¹³C and then WUE. However, the Δ¹³C method is expensive, i.e. mass spectroscopy analysis costs around $8 per sample.; Due to the cost and complexity of the peanut genome, Δ¹³C method has not yet been connected successfully to QTLs in peanut breeding; it is more successfully used in soybean and wheat (Dhanapal et al., 2015; Del Pozo et al., 2016). In soybean and wheat, GWAS and QTL mapping have been performed and genomic regions codifying Δ¹³C are now available (Dhanapal et al., 2015; Mora et al., 2015). Despite its success, Δ¹³C is far from being considered an HTP technique because of the time required for biomass sampling and preparation, and the subsequent spectroscopy analysis. Near infrared (NIR) technology is based on the unique reflectance patterns that different leaf and seed chemical compounds produce when they are exposed to light with a known light spectrum. Ferrio et al. (2001) used NIR technology to accurately predict Δ¹³C, and thus WUE. These authors predicted Δ¹³C from NIR reflectance with an accuracy ranging from 82 to 86%.

Hyperspectral reflectance to predict photosynthetic parameters.

Hyperspectral sensors capture electromagnetic radiation reflected from vegetation in the visible (VIS, 400–700 nm), near-infrared (NIR, 700–1,300 nm), and short-wavelength infrared (SWIR, 1,400–3,000 nm) regions of the electromagnetic spectrum, which contain information about leaf physiological status including pigments, structural constituents of biomass, and water content (Curran, 1989; Penuelas and Filella, 1998). Variation of foliar reflectance at different wavelengths is specific to the variations in chemical and structural characteristics of the leaf and photosynthetic apparatus (Fig. 14a, b, c) (Serbin et al., 2012). Hyperspectral reflectance spectroscopy as an HTP tool is becoming recognized as promising in agricultural research (Serbin et al., 2012; Weber et al., 2012; Araus and Cairns, 2014; Silva-Perez et al., 2017; Yendrek et al., 2017). Hyperspectral hand-held sensors such as the Field Spec Hi-Res-4, have been used to predict photosynthetic parameters including maximum rate of rubisco carboxylation Vc_max (Fig. 14) and midday CO₂ assimilation of peanut plants (Buchaillot et al., 2022). However, this technology is low throughput as individual plants need to be measured one by one. To reach high throughput phenotyping, researchers are currently using hyperspectral cameras mounted on drones to predict these parameters with similarly accurate results predicted with the hand-held equipment (Bagherian et al., 2023). The combination of drone and hyperspectral technologies will allow breeding programs to phenotype plant photosynthetic parameters in a high throughput manner in the future.

Phenotyping roots - the hidden resource.

Roots provide water and nutrient uptake. Different root traits enable plants to respond, adapt and tolerate different environmental stresses. For example, peanut cultivars with drought tolerance showed greater root length at deeper depths than drought-sensitive cultivars (Thangthong et al., 2017). As the root system is underground, historically, it has been difficult to phenotype (York et al., 2015; Atkinson et al., 2019). However, recently, technological improvements in image acquisition and data analysis have allowed great advancements in root and pod phenotyping, as mentioned earlier in this section (Gimode et al., 2023). Regardless of various obstacles, e.g. soil type, heterogeneity and moisture, GPR technology can accurately detect below-ground plant parts significantly contributing to the understanding of plant response mechanisms to tolerate stress (Dobreva et al., 2021; Gimode et al., 2023).

Other non-destructive field measurements rely on rhizotrons and rhizo-boxes systems with different automation capabilities and, more recently, on X-ray computed tomography and GPR (Garbout et al., 2012; Nagel et al., 2012; Rogers et al., 2016; Thangthong et al., 2016; Liu et al., 2018). For example, for peanut, Thangthong et al. (2016) developed a rhizotron that can be placed in the field to monitor root development in quadrants of 10 cm² (Fig. 15a), from the beginning of the season until harvest. Although this type of system is advanced and more representative of “normal” root development in the field, soil particles can interrupt the visibility of the roots causing missed data (Atkinson et al., 2019). Soil and soilless systems (Fig. 15b, c) used in greenhouse applications are less informative as root phenotypes can be drastically different from the phenotypes in the field (Kuijken et al., 2015).

In summary, research achievements in peanut physiology have been numerous in recent years. These achievements are predominant in the areas of crop requirements for sustainable production, plant growth and development, mechanisms of adaptation to biotic and abiotic factors, and, more recently, the development of high-throughput tools for phenotyping. Along with the achievements in genome sequencing and genetic mapping technologies, plant physiology discoveries and advancements in HTP technologies pave the way for the implementation of new strategies in peanut breeding, i.e. molecular breeding and genomic selection. These strategies appear more robust than traditional breeding for predicting valuable traits for increased yield and adaptation to environmental conditions; and allow pyramiding these traits more effectively in future peanut cultivars to meet production requirements for the specific geographical regions in which cultivars will be grown.