Short- and long-term warming events on photosynthetic physiology, growth, and yields of field grown crops

Global temperatures are rising from increasing concentrations of greenhouse gases in the atmosphere associated with anthropogenic activities. Global warming includes a warmer shift in mean temperatures as well as increases in the probability of extreme heating events, termed heat waves. Despite the ability of plants to cope with temporal variations in temperature, global warming is increasingly presenting challenges to agroecosystems. The impact of warming on crop species has direct consequences on food security, therefore understanding impacts and opportunities to adapt crops to global warming necessitates experimentation that allows for modification of growth environments to represent global warming scenarios. Published studies addressing crop responses to warming are extensive, however, in-field studies where growth temperature is manipulated to mimic global warming are limited. Here, we provide an overview of in-field heating techniques employed to understand crop responses to warmer growth environments. We then focus on key results associated with season-long warming, as expected with rising global mean temperatures, and with heat waves, as a consequence of increasing temperature variability and rising global mean temperatures. We then discuss the role of rising temperatures on atmospheric water vapor pressure deficit and potential implications for crop photosynthesis and productivity. Finally, we review strategies by which crop photosynthetic processes might be optimized to adapt crops to the increasing temperatures and frequencies of heat waves. Key findings from this review are that higher temperatures consistently reduce photosynthesis and yields of crops even as atmospheric carbon dioxide increases, yet potential strategies to minimize losses from high-temperature exist.


Introduction
Emissions of greenhouse gases to the atmosphere are rising steadily associated with anthropogenic activities [1]. The impacts of rising greenhouse gases on the global climate system are far-reaching, with the most direct impact being an increase in global mean temperatures [2]. To date, anthropogenic greenhouse gas emissions have resulted in a ∼1°C increase in global temperatures although the increase is not uniform across the planet with terrestrial areas warming more than marine [2] . Concurrent with increasing mean temperatures is the increase in temperature variability [3]. While warmer mean temperature and greater temperature variance, independently, result in more extreme temperatures, combined warming, and increased temperature variability have a multiplicative impact on the frequency of extreme, high-temperature events. This is consistent with the trend toward more common and increasing intensity of heat waves [4].
Air temperatures fluctuate widely throughout multiple timescales ranging from within-to between days, over weeks, months, and seasons. As sessile organisms, plants experience this wide range of temperatures and as a result have developed myriad strategies to sense [5] and cope [6] with temperature changes and, to a limited extent, regulate their temperature [7,8]. Despite the adaptations that allow plants to thrive over a range of growth conditions, temperature is known to have a direct impact on all facets of biochemical and biophysical activities at the cellular, organ, and whole plant scale. Plant temperatures frequently deviate from air temperature as a consequence of absorption of radiation, transpiration, and a variety of other factors. Yet rising air temperatures and plant canopy temperatures are coupled through the energy budget which leads to direct increases in plant temperatures as air temperatures rise. While plants can thermally acclimate key metabolic processes when changes in growing conditions occur [9], these adaptations are likely intended for plant survival and not necessarily to maintain optimal physiological activity leading to maximum growth or fecundity. Projections suggest that plant temperatures that exceed 30°C result in potential yield losses for major crops [10], a threshold that is often exceeded [11], more so as the globe warms. Because of this, current and future projected global warming and increasing occurrences of extreme heat events are both likely to drive significant acclimation responses of plants and impact the physiological, morphological, growth, and reproductive success of plants growing in the terrestrial biosphere. Consequently, the reality of anthropogenically induced changes in atmospheric composition and warming portends future challenges to agriculture and global food security.
Photosynthesis accounts for the entry point for carbon into the terrestrial biosphere. As a result, any factors that impact photosynthetic rates are certain to influence the growth and yield of crops. Understanding how photosynthesis can be adapted to higher temperatures is pressing. Global temperatures have already exceeded optimal temperatures for photosynthesis in many locations and continued warming could lead to a 40% decrease in terrestrial biosphere productivity [12]. The general response of photosynthesis to short-term fluctuations in temperature has been characterized extensively for both C3 and C4 species [13] and the underlying parameterization necessary to predict photosynthesis over a wide range of temperatures has been successfully modeled [14][15][16][17][18]. While the importance of thermal acclimation of photosynthesis has been well established [9,[19][20][21], actual adjustments in photosynthetic physiology that will occur with plants grown under higher temperatures over prolonged periods, consistent with global warming, are likely to alter growth and yield of crops under future climate conditions.
In this review, we outline the crop response of photosynthesis, growth, and yield to elevated temperature. Reproductive structures are known to be sensitive to elevated temperatures, however, in the context of this review, we will not focus on the mechanisms accounting for the temperature sensitivity of reproduction and instead recommend previous reviews on this topic (e.g. [22]). We will focus exclusively on field-based manipulative heating experiments, and, when appropriate, we will include interacting global change factors in addition to heating. The focus of the research presented in the manuscript is on canopy-scale heating of plants grown under field conditions as it is this scale that determines crop responses to the environment, that micrometeorological factors feedback on plant physiology and growth, and that global changes are likely to impact crop yields. Being a field-technique based review on major crops, we will focus on warming restricted to what are considered extreme temperature for major crop growing regions of the planet, where critical thresholds for heat stress are projected to occur at ∼30°C [10] and set an upper threshold of ca. 40°C to prevent extreme stress conditions that lead potentially to mortality. First, we will provide a summary of in-field heating technologies used to increase growth temperature for crops over long-term ( prolonged heating) and short-term (heat wave) scenarios. This will be followed by a summary of responses observed for field crops grown under hightemperature environments. We will briefly outline confounding factors associated with in-field heating, particularly the impact on leaf-to-air vapor pressure deficit (VPD). Finally, we will discuss the potential for adapting crops to high-temperature growth environments.
In-field heating technology to manipulate the growth environment for plants A variety of heating methods have been used to increase canopy temperatures, but when studying global climate change, it is important to consider methods that most closely mimic future climate scenarios and what effects the treatment will have on data interpretation [23,24]. This section outlines the various methods used to increase the temperature of field-grown crops. Each technique listed will provide an overview of the technique, how it's been applied, and considerations/challenges associated with the technique.
Generally, methods that heat air can be categorized as either passive or active. The advantages and disadvantages of these methods have been reviewed and debated [25][26][27][28][29]. Passive techniques include field-installed greenhouses and passive open-top chambers. These structures trap long-wave radiation by partially or completely covering plots and increasing the boundary layer resistance which can more effectively trap heat. Despite a limited ability to control the amount of temperature change, it is a relatively low-cost, straightforward, and flexible method [30][31][32]. Because they are adaptable systems that require no power, passive structures are the preferred methods of temperature manipulation in remote ecosystems [27,[33][34][35]. Furthermore, these passive systems can store humidity through the same mechanisms that trap heat, therefore potentially increasing the plant canopy temperature without increasing the VPD around the vegetation.
Active air heating techniques include open-top chambers, soil warming, and infrared (IR) heating elements that require powered infrastructure to modulate heating to warm temperatures above ambient. Open-top chambers typically use heating elements to warm air that is blown into the chamber to raise plot temperatures [36,37]. A benefit of open-top chambers includes the ability to alter many variables, e.g. raising CO 2 , tropospheric ozone, and/or temperature to study multiple global climate change factors with a single delivery system [38]. However, chamber structure and associated blowers have a significant effect on the microclimate beyond the intended treatments [39,40]. Resistance cables buried underground have been applied to studies focusing on understanding soil ecology impacts of global warming [41,42], however, they are not a practical method for increasing canopy temperatures.
IR heating has become a widely used technique for increasing canopy temperatures in a range of ecosystems without the use of an enclosure , which is shown to heat somewhat consistently throughout the plant canopy profile (e.g. [64]). IR heating systems have undergone numerous advances since their inception for in-field heating purposes [47]. Some experiments set IR heater output to a constant level which causes small temperature differences during the day and large differences at night [59]. More commonly, IR heating arrays consist of several components (Figure 1), which work in unison to increase the temperature of a plant canopy to a setpoint above a reference plot that can be maintained or adjusted based on the research questions being investigated.
The key components needed to operate an IR heating array system are relatively straightforward (Figure 1). While there are a variety of techniques employed, the most common setup involves using IR radiometers to measure the temperature of a reference plot and a plot that is intended to be heated. These are non-contact sensors that are used to measure the temperature of a surface, at which they are pointed. The IR sensors are wired into a measurement and control peripheral that determines the reference plot temperature and is programmed to control IR heater output to maintain a user-defined heated treatment. IR heater control is usually accomplished by using an industrial dimmer with scaled voltage control interfaced with a voltage output module that is part of the measurement and control system. Specific details coupled with examples of each peripheral are provided in Figure 1 and elsewhere [49,[53][54][55].
There are multiple challenges associated with construction and operation beyond the array components. First, high-powered utility service is required for system operation [54,56]. The heating arrays associated with the newest generation of heating systems range from 6 kW (e.g. [65]) to 24 kW (e.g. [55]) per array at maximum output, although actual output to maintain a setpoint will generally be lower than maximal output. Achieving maximum IR heater output requires nearby utility power and potentially dedicated transformers, which limits the potential locations where this technique can be applied. In rare cases when transformer availability was limited, a 37.5 kW diesel generator has been used by the authors as a stopgap at the Soybean Free Air CO 2 Enrichment (SoyFACE; http://soyface.illinois.edu/) IR heating experiment. The second consideration is the cost of electricity consumption. Analysis of one plot resulted in an average seasonal mean electricity demand of 3.8 kW hours per day to achieve a 3.2°C temperature increase in the heated plot relative to the reference plot. However, the energy required to maintain heating is likely to vary substantially based on many factors including wind speed, target temperature increase, crop type, water availability, solar radiation, etc. Within the context of these costs is the need for replication. For example, a full-factorial elevated CO 2 by warming experiment with four replicate blocks (n = 4) requires eight total arrays (e.g. [55]). Third, safety is a critical consideration when operating an IR heating experiment. Dangerously high electricity requirements coupled with extremely hot IR heating elements necessitate proper training, multiple fuse or circuit protection safety points, routine inspection of wires, professional contractors for key steps during setup, and personal protective equipment. Finally, as with most other in-field heating techniques, the use of IR heating arrays invariably increases the leaf-to-air VPD of the plants located within the heated treatment [28,49,66]. This results in the confounding influences of VPD and moisture feedbacks in addition to the intended warming [67], which, like all other active heating methods, challenges the assumption of heating as the main driver for physiological and growth responses [68,69].

Crop photosynthetic responses to rising temperature
Photosynthesis is the primary pathway for carbon assimilation in plants, thus how photosynthesis responds to changes in temperature has a major impact on growth and development. The primary impact of rising temperatures on photosynthesis involves Ribulose-1,5-bisphosphate carboxylase-oxygenase (Rubisco), the primary enzyme to fix atmospheric CO 2 in C3 plants. In addition to fixing CO 2 to ribulose 1,5-bisphosphate (RuBP) to drive photosynthesis, C3 plants suffer from Rubisco oxygenation of RuBP, which initiates the energetically wasteful photorespiratory pathway [70]. Photorespiration has potential consequences for lower crop growth and (A) An aerial representation of a heated and reference plot representing one experimental block. The goal is to increase the heated plot temperature by a setpoint increase (°C) above the reference plot. The heated plot is surrounded by six heaters, each with four 1000 W heating elements, hexagonally arranged. An infrared thermometer (B) is mounted above both the heated and reference plot and wired into a datalogger (C, top) which is used to measure the surface temperature differences between the two plots. If the temperature difference between the plots differs from the target temperature setpoint increase for the heated plot, the datalogger uses a PID algorithm to adjust the programmable voltage output module (C, bottom) to a value between 0 and 10 V. The scalable output range (0-10 V) is wired into a heavy-duty industrial dimmer (D) to scale energy output (0 V = 0% output; 10 V = 100% output) to the heaters. Not shown are multiple circuit breakers and fused junctions, grounding, telecommunications, and enclosure lock-out power safety switches.
yield. The carboxylation efficiency of Rubisco, defined as carbon fixation per photon of photosynthetically active radiation absorbed by the plants, is determined by many factors including the specificity of Rubisco for carboxylation vs. oxygenation (S C/O ), Rubisco activation state, the concentrations of oxygen and carbon dioxide surrounding Rubisco, light intensity, and the rate of RuBP regeneration.
Decreased discrimination by Rubisco against oxygen, defined as a decline in S C/O , drive a decline in the rate of carboxylation (V c ) relative to oxygenation (V o ) as temperature increases [13, [71][72][73][74][75][76]. For example, S C/O in Glycine max (soybean) declines from ∼100 at 25°C to ∼90 at 30°C [77]. As a result, most C3 plants have a thermal optimum between 20 and 30°C, with losses in photosynthetic potential above this optimum being driven by a lower V c /V o . Higher temperatures carry additional consequences related to declining concentration of CO 2 relative to O 2 in the chloroplast. As temperatures rise, the solubility of both O 2 and CO 2 declines, however, this decline is greater for CO 2 than for O 2 . Additionally, mesophyll conductance in some species is shown to rise with temperature and remain relatively constant in others [16,78]. Even when g m increases it is insufficient to offset the impact of solubility and the decline in S C/O with temperature leading to higher limitations photosynthetic carbon assimilation (e.g. [16]). In addition to the impact of higher temperatures on the supply of CO 2 to Rubisco, the activation state of Rubisco is highly regulated by Rubisco activase [72,79]. Rubisco activase is shown to be highly temperature sensitive [74,80] and is shown to lower photosynthetic carboxylation efficiency through a decline in Rubisco activation state [81].
The C4 photosynthetic pathway effectively minimizes photorespiration by actively concentrating CO 2 in the bundle sheath where Rubisco carboxylation occurs [82][83][84][85][86][87]. The exceptionally high concentration of CO 2 achieved under most circumstances by C4 plants outcompetes oxygenation at biologically relevant temperatures, which leads to a higher temperature optimum for photosynthesis than C3 plants [13]. Furthermore, evidence suggests that Rubisco activase has a higher thermal stability in C4 plants than in most C3 crop species [88]. Despite these differences in photosynthetic responses of C3 and C4 species to temperature, photosynthetic acclimation to differing growth conditions can lead to plasticity in measured relative to theoretical responses [21,89]. This raises the importance of experimental approaches that manipulate growth temperatures for fieldgrown C3 and C4 crops both over the long-term (warming throughout the growing season) and short-term (high-intensity, shorter duration heat waves).
Soybean responses to elevated temperature at 3.5°C above background, with and without elevated CO 2 , have been investigated over five growing seasons [55,91,97]. Leaf photosynthetic rates varied in response to the heated treatments based on cultivar, back-ground meteorological conditions, and whether measured in situ midday [97] or integrated diurnally [55]. For most years, photosynthesis did not vary between control and elevated temperature treatments, except for one year when heating was applied to an already exceptionally warm year [55]. For this year the maximum canopy temperature in the heated treatment, set to 3.5°C above control plots, was the hottest of any year with daily maximum canopy temperatures ranging from 32.5°C to 41.5°C. This treatment saw reductions in photosynthesis by 15% [55]. These responses are supported by studies suggesting that the impact of warming on a region is dependent on the baseline conditions [98]. Despite the lack of differences among the treatments, observed maximum rates of electron transport (J max ) were shown to be consistently lower for soybean grown in elevated temperatures [97,99]. Despite the variable responses associated with photosynthesis, above-ground biomass was consistently lower in elevated temperature relative to control for all years (Figure 2), which suggests that the leaf-level photosynthetic measurements may not be representative of whole-season responses and/or of canopy photosynthesis. Seed yield responses to elevated temperature were similar to those observed for above-ground biomass (Figure 2).
Soybean experiments have also evaluated the combined impact of elevated CO 2 with warming [55,91,97,99]. Photosynthetic rates for the combined warming and elevated CO 2 treatment were only slightly lower than observed for elevated CO 2 alone ( Figure 2) but rates for the combined treatment were much lower than the predicted synergistic effects based on modeling analysis [20,89]. This result is likely associated with both a reduction in stomatal conductance (gs; [55]) and a down-regulation of photosynthetic parameters [99] associated with both elevated CO 2 and warmer temperatures. Above-ground biomass declined in the combined treatment to a greater extent than observed for photosynthesis, again suggesting that canopy-integrated photosynthetic responses may be greater than the responses observed at the leaf scale. Yield was reduced further than biomass in the combined treatment, likely due to the impacts of warming on photosynthesis, g s , and on reproductive development [90,91,96]. Furthermore, soybean grown in elevated CO 2 shows consistently warmer canopy temperatures as a result of lower g s [100], thus increases in temperature associated with IR heating arrays and elevated CO 2 -induced declines in g s raise canopy temperatures further beyond the thermal optimum [55].
Wheat responses to growth in elevated temperature are shown to be highly variable relative to soybean. A sensitivity analysis of wheat yield responses to in-field heating of ∼1.5°C showed that winter wheat yield does not necessarily decrease with warmer temperatures in regions with cooler growing season temperatures or when the water supply is not limited [93]. Over a five-year experiment where temperatures were consistently lower than 25°C, wheat yield was shown to increase by ∼16% despite a temperature-induced shortening of the overall growing season. The yield increase was related to prolonged grain-filling developmental stages, relative to the non-heated plots, despite the overall shortened growing season [92,93]. In another experiment, plots of spring wheat were planted in staggered intervals over time from September to May (Fall to Spring) with supplemental warming of ∼1.3°C during the day and ∼2.7°C at night [61]. This experimental design allowed for variation in growth temperature associated with the staggered planting throughout the season coupled with the imposed heating conditions from the IR heaters. Results showed no yield impacts from supplemental heating when temperatures were near the optimum, and large yield increases with supplemental heating when background temperatures were below the optimum [61]. However, spring wheat yield was reduced in treatments where temperatures rose above the thermal optimum for wheat yield of ∼15°C [61].
Rice grown under in-field heating of ∼+1.5°C showed a reduced leaf area index, dry biomass, and yield [95]. Interestingly, these biomass and yield reductions were observed despite increases in the maximum rate of photosynthesis due to higher V cmax and J max in the heated treatment relative to control [95]. These results, like the ones observed in soybean, show that increases in photosynthesis are not always translated to increases in yield under a global warming scenario, and this can be related to other physiological processes, like grain development and grain growth. Consequently, warming accelerated rice development by reducing the days to heading and the days to get the maximum tiller number [94,101], which decreases spikelet density and consequently yield [101]. Photosynthesis data represents the mean in situ rates of photosynthesis collected midday over multiple measurement days during each season and averaged throughout the growing season. The above-ground biomass and seed yield data were collected at the end of each growing season. Target temperature increases were 3.5°C above background levels measured according to canopy temperature measurement. Full details for each measurement are described previously [55,97]. The bars represent the mean relative difference to the control (above 1 corresponds to an increase and below one a decrease in observed values relative to control), and error bars represent one standard error of the mean. Each symbol represents mean values for each growing season.
Despite being a C4 crop, maize grown under in-field heating resulted in reduced photosynthetic rates, again attributed to photosynthetic acclimation to higher temperatures [96]. No changes in g s and in intrinsic water use efficiency (iWUE) were observed for maize grown under high temperatures, relative to the control, which may be attributed to inherently lower g s associated with C4 species. Despite these photosynthetic responses, there was no observed difference in above-ground biomass for maize among the treatments, yet seed yield was lower in the heated treatments relative to the control, and this response was observed with and without elevated CO 2 [96] . The lower yields in elevated temperature can potentially be attributed to lower photosynthetic rates, but higher temperatures led to more rapid progression through key developmental stages [91] and led to fewer kernels per cob [96] compared with the non-heated control, both of which are likely factors in the yield decline. The lower kernel count can be related to factors such as changes in silk receptivity, e.g. due to desiccation [102,103] and ear or kernel development, including kernel abortion [91,[104][105][106][107], however, mechanistic understanding of these responses under in-field heating is still needed. While high temperature is likely to impact pollen viability [103,108], heated plots in these experiments are generally surrounded by non-heated vegetation, likely leading to an abundance of non-heat-treated and therefore fully viable pollen.

Heat waves
Extreme heat events will become more intense and more frequent [4, [109][110][111][112]. The definition of a heat wave is broad and because they are a major cause of weather-based mortality, definitions are often based on the amount of critical physiological stress posed to humans [113,114]. Studies on the effects of heat waves on plant growth and productivity are less common. Generally, there are two ways to study the effects of heat waves on plants. First one can use weather and historical yield data to model the effects of past extreme heat events on yield. Second, the heating technologies reviewed above can be used to impose heat waves in the field to help explore physiological breaking points.
Heat waves have significant effects on managed and unmanaged ecosystems [115]. A 2013 heat wave in southern China caused the largest crop yield anomaly in fifty years and reduced terrestrial carbon uptake in the region by ∼40-50% [116]. Similar heat waves have plagued other regions of the planet with a similar effect [117][118][119]. A useful framework has been provided for future heat wave research and outlines the role of compounding stresses and severe heat waves on crop failure and ecosystem die-off [120]. For crop loss in particular, it is clear that heat waves have the greatest impact on crop yield when they occur at a critical point of the growing season [121][122][123] and extreme temperatures play a major role in the projected yield reductions associated with climate change [124][125][126]. Given that temperature affects every biological process in the plant, it is important to simulate heat waves in the field as close to future projected conditions as possible during critical growth stages to identify processes that reduce yield.
Experimentally imposed heat waves in the field have large immediate effects on photosynthesis which tend to recover after the heat wave treatment. Heat waves that increased canopy temperatures 6°C above background temperatures for a three-day duration in soybean [123] and maize [122] consistently reduced photosynthetic rates during the heat waves, but photosynthesis recovered to control plot levels when the heatwave ended. From these same experiments, heat waves imposed during vegetative growth had no lasting impacts on yield, but heat waves imposed during sensitive reproductive stages in corn and soy caused significant decreases in yield [122,123]. An imposed three-day heat wave that increased the temperature to 38°C in wheat significantly reduced photosynthesis, but plants recovered completely within 30 days after the treatment ended [127]. In Lens culinaris (lentils), a three-day heat wave of 38-40°C caused reductions in ΦPSII and electron transport rate, proxies for reduced photosynthesis [128]. The heat wave reduced lentil biomass by 6%, but yield was reduced by 30%, which suggests the sensitivity of reproductive structures to high-intensity heating.
There have been several FACE studies examining the interaction between elevated CO 2 and heat waves on plant growth. In a study using lentils, elevated CO 2 stimulated yield by 34%, but this yield increase disappeared when plants were grown in combined elevated CO 2 and heat wave treatment [128]. Alternatively, the effects of elevated CO 2 on moderate (+5°C) and extreme (+9°C) heat waves for soybean showed yield was maintained in the +5°C but not the +9°C treatment [129]. A low-intensity heat wave imposed on rice during reproductive development negatively impacted yields, and offset the potential benefit of elevated CO 2 [63]. Ultimately, the impact of imposed heat waves on photosynthesis appears to be relatively limited to the duration of the heating event and has relatively minor impacts on total biomass, but appears to have a relatively consistent impact on yields when they are imposed during reproductive development. However, there are too few studies addressing the role of imposed heat waves on crops, necessitating more research to fully elucidate the underlying mechanisms and thresholds (duration and heating intensity) for responses.
A literature search did not yield any studies on a three-way interaction between elevated CO 2 , elevated seasonal temperature, and heat waves. From what we know about the acclimatory effects of elevated CO 2 and season-long warming on C3 photosynthetic metabolism, heat waves may have an even greater effect on photosynthesis than has been measured. The photosynthetic acclimation to warming and elevated CO 2 would lower photosynthetic rates at high temperatures and lower g s [55,97,130], two things that are likely to happen during a heat wave. Alternatively, evidence exists that crops experiencing heat stress early in a growing season have a potential 'priming' effect that may provide protection later in the growing season [11]. Additional research on the effects of heat waves and the critical process that they affect are needed to understand, and potentially mitigate, scenarios that will lead to crop failure.
In-field heating effects on crop water use efficiency through higher VPD Atmospheric humidity is determined by the actual amount of water vapor in the atmosphere (actual water vapor pressure; e a ) relative to the total amount of water vapor that the atmosphere can hold (saturation water vapor pressure; e s ). The difference between e s and e a is termed VPD and is used as a metric for the atmospheric evaporative demand. Saturation vapor pressure increases exponentially as a function of temperature while e a is determined by water available to the atmosphere and is therefore highly variable across spatial and temporal scales. Rising global temperatures are resulting in higher e s [69,[131][132][133][134] whereas the impact of global warming is having a lesser effect on e a [132]. Even in locations where e a is increasing, it is insufficient to offset the rise in e s leading to increases in VPD [135].
Disentangling the effects of higher temperatures and higher VPD on crops is essential given that differential increases in warming vs. atmospheric drying may elicit variable responses among crop types [136]. Plants respond to higher VPD by decreasing stomatal conductance [137]. The lower stomatal conductance conserves water loss associated with the higher evaporative demand but also can lower the CO 2 flux into the leaves, lowers evaporative cooling, and, for C3 species and under some circumstances for C4 species, increase stomatal limitation to photosynthesis [138]. Given the tight linkage between temperature and VPD, separating physiological responses of plants to these variables, particularly under field conditions, is difficult. As a result, fieldbased understanding of how vegetation responds to heating vs. higher VPD is scarce in the literature. Despite major experimental challenges required to isolate physiological and mechanistic responses to both temperature and VPD independently, an improved understanding of the mechanisms of warming vs. drying would carry the potential for improved food security and, more accurate ecosystem modeling.
Rising VPD caused by global warming can have several impacts on plant growth and water use efficiency. High VPD can increase transpiration rates, which decreases water use efficiency and potentially can lead to increased water stress [68]. To avoid excess water loss and conserve water, g s generally decreases which may limit CO 2 uptake for photosynthesis, also driving lower water use efficiency. Optimal VPD levels vary between plant species, but in general, a VPD range of 0.5-1.5 kPa is considered suitable for most plants. Overall, changes to VPD caused by global warming can have significant impacts on the growth, functionality, and water use efficiency of plants, which can have cascading effects on ecosystems and food production.
Experiments that address ecosystem responses to VPD suggest widespread reductions in productivity over a wide range of ecosystems based on a range of techniques and over many different scales but are generally limited to analysis of historic data or modeling output [134,[139][140][141]. Rising VPD is shown to have significant impacts on crop physiology, primarily on g s and the subsequent impact of changes in g s on canopy and soil processes [68]. While much of this previous work shows the role of VPD on the growth and yield of crops, methods are primarily based on statistical analysis or models with a limited mechanistic representation of crop responses to VPD vs. temperature. Recently, a field-based humidifying system was developed and deployed for a northern European forest over whole growing seasons, showing significant impacts of decreasing VPD on tree species [142]. With the in-field IR heating arrays developed to test high-temperature impacts on crops, the addition of a similar in-field humidifying system could potentially lead to the ability to disentangle the impacts of warming temperature vs. rising VPD on crop growth, physiology, and yield. While more complicated than irrigating to offset increased evaporative demand, humidifying the air potentially removes the physiological responses to higher VPD associated with heating treatments. Carrying out an in-field humidification experiment poses some challenges much like the installation for the IR heating arrays. Adding humidity requires an infrastructure investment including a supply of water, a high-pressure misting system, a feedback control system, and electricity to pump the water. Furthermore, moist warm environments may increase the possibility of pathogens being present, so the systems require periodic inspections to treat for diseases before they impact crop functioning. Nevertheless, incorporating humidifying techniques into IR heating experiments can help understand the relative influences of heating plants without the consequent drying of the air around the plants.

Adapting crops to high temperatures
Improving photosynthetic efficiency is essential to meet growing demands for agricultural products [143] yet, as outlined in this review, rising global temperatures are placing additional constraints on photosynthesis. The impact of warming temperature on photosynthesis defines a central target for adaptation as it is an important determinant of crop yield [144]. As the entry point for carbon into an ecosystem, photosynthesis is included in the list of likely targets for adapting crops to global warming [145,146]. In this section we review current strategies being investigated that can either minimize carbon losses or ideally increase carbon gains at higher temperatures.
Field research reviewed above points to the heat sensitivity of biochemical processes of carbon assimilation as the major limitation for both C3 and C4 photosynthesis. Overcoming the 10% reduction in S c/o observed as plants are warmed from 25 to 30°C [77] would require a compensating increase in [CO 2 ]/[O 2 ] at the active site of Rubisco on the order of 10%. Although the predicted increase in atmospheric [CO 2 ] to ∼550 ppm by the year 2050 will lower the impact of increased temperature on S c/o , it will not fully offset it [147]. The [CO 2 ] in chloroplasts of C3 crops where carboxylation occurs will be affected by changes in stomatal and mesophyll conductances along with the larger reduction in the solubility of CO 2 as compared with O 2 with warmer temperatures. Increases in stomatal and mesophyll conductance are expected to increase net carboxylation, however, increased stomatal, but not mesophyll, conductance will also drive significantly more water loss. This suggests that increasing mesophyll conductance can be a potential target for crop adaption to higher temperatures as it will increase the CO 2 in the chloroplast without increases in water loss. The temperature response of mesophyll conductance has been shown to increase or not change with increasing temperature depending upon species [16,78]. In the herbaceous species Nicotiana tabacum, Gossypium hirsutum, and Glycine max mesophyll conductance increased two-to threefold between 15 and 40°C, which is adaptive to the higher temperature. Increasing mesophyll conductance by exploiting natural variation or by engineering are thus potential strategies for adapting crops to warming.
The decrease in S c/o that accompanies increased temperature results, at least in C3 plants, in a decrease in net photosynthesis due to increased photorespiration. Photorespiration is energetically very expensive, consuming as much as 30% of a plant's metabolic energy [147] to recycle the glycolate produced by Rubisco oxygenation. Less energetically expensive alternative transgenic pathways to recycle glycolate have been demonstrated in plants (e.g. [148][149][150]) for which theory predicts they should have their greatest benefit at high temperature where V o and the consequent energetic costs are the highest. This prediction was confirmed with transgenic plants in which the native photorespiratory pathway had been bypassed, and glycolate flux was directed to a synthetic glycolate metabolism pathway installed in chloroplasts [151]. In canopy warming field experiments, it was shown that the substantial inhibition of growth by a season-long 5°C warming was largely prevented in plants containing the synthetic photorespiratory bypass. Another manipulation of photorespiration that may be adaptive to increasing growing season temperatures is the overexpression of the H-protein of glycine decarboxylase that has been shown to enhanced enhance net photosynthesis and growth of Arabidopsis thaliana [152] and Nicotiana tabacum [153]. While the mechanism by which the overexpression of this photorespiratory protein subunit stimulates photosynthesis is uncertain, there is again an expectation that it may have an increasing impact with rising Rubisco oxygenation rates with temperature.
The rate of regeneration of the CO 2 /O 2 acceptor RuBP is also sensitive to temperature and variable with growth conditions and across species [154,155]. At current CO 2 levels, light-saturated photosynthesis in C3 crop species operates at the intersection of Rubisco-limited and RuBP-limited photosynthesis, necessitating that photosynthesis will become increasingly limited by RuBP regeneration capacity as CO 2 increases [156]. Additionally, RuBP regeneration limitation increases at higher temperatures [13], and RuBP-limited photosynthesis benefits from lower photorespiration at elevated CO 2 , implying that increases in the RuBP regeneration rate should raise the temperature optimum of photosynthesis at elevated CO 2 . Although there are multiple carbon reduction cycle enzymes along with the full suite of thylakoid membrane electron transport and ATP synthesis reactions involved in RuBP regeneration, modeling has consistently identified sedoheptulose-1,7-bisphosphatase, aldolase, and transketolase as exerting the greatest control [157,158]. There are numerous demonstrations confirming that the overexpression of sedoheptulose-1,7-bisphosphatase stimulates photosynthesis, implying a RuBP regeneration limitation [159], and protects against the inhibition of photosynthesis by moderate heat stress [160]. Moreover, transgenic expression of the cyanobacterial bifunctional fructose-1,6/ sedoheptulose-1,7-bisphosphatase was shown to enhanced carbon assimilation and seed yield in soybean and resulted in significantly higher V c,max and J max under elevated CO 2 and at elevated temperature protected against the seed yield loss that was seen in wild-type plants [97].
Rubisco activase (Rca), a nuclear-encoded chloroplast localized protein, regulates the proportion of Rubisco that is catalytically active by remodeling Rubisco to release inhibitory sugar phosphates from Rubisco active sites. The catalytic events that generate the compounds which inactivate Rubisco increase with temperature, causing progressive inhibition of Rubisco-limited photosynthesis during mild heat stress [161]. Paradoxically, Rca itself is sensitive to moderate heat stress in many species. Thermal stability of Rca was shown to be improved through gene shuffling to increase rates of photosynthesis and growth in genetically transformed Arabidopsis exposed to heat stress [162]. Nearly all plant species express two forms of Rca, the β form and the longer redox regulated α form. These isoforms come either from alternative splicing of the transcript of a single gene or from one or more copies of separate α and β isoform genes. For example, wheat has three Rca isoforms arising from one α and two β isoform genes. Heat induces the expression of one of the β genes suggesting that the heat stress response in wheat is in part mediated by this Rca β isoform [163]. In vitro studies confirmed that this β isoform is a heat-stable variant of Rca for wheat whereas the other β isoform is the most heat sensitive. The specific amino acids able to confer Rca thermostability to the wheat β isoform were identified [164]. It was further shown that substitution of a methionine residue with isoleucine in the heat-sensitive wheat Rca β isoform improved the thermal optimum without affecting the efficiency of Rubisco activation [165]. C4 grass species, such as Sorghum bicolor (sorghum), also have separate α and β genes, but in these grass species the Rca-α form is expressed only at temperatures above 42°C. Induction rate of Rca-α was shown to match the recovery rate of photosynthesis and Rubisco activation from the 42°C heat shock [88]. This association between Rca-α isoform expression and maintenance of Rubisco activation at high temperature suggests that Rca-α has a functional thermo-protective role in carbon fixation in C4 grasses by sustaining Rubisco activation at high temperature. These discoveries suggest that engineering Rca and/or manipulating its regulation could be possible strategies to improve Rubisco carboxylation efficiency in crop plants in a warming world but also reveal that due to the diversity in Rca regulation across species, the specific solutions for improving thermotolerance of photosynthesis via Rca engineering will differ across species.

Conclusion
Whether the focus is on inherent intra-and inter-annual variability in growing temperature or the accelerating rate of global warming, understanding how crops respond to warming is critical to predict food security and to identify strategies to adapt crops to climate change. While there is still much more to understand about how warming can drive reductions in plant productivity, in-field heating technology is helping to provide some insights into crop responses and adaptation strategies to warming. Bridging field-based warming experiments with high-throughput phenotyping techniques, such as leaf spectral reflectance [166] and sun-induced chlorophyll fluorescence (SIF; [167]), are likely to accelerate understanding of crop photosynthetic physiological effects to warming over a range of temporal and spatial scales. These techniques are likely to create opportunities for future research on regional and global scale monitoring of crop responses to warming.
While our focus here is on the impact of increasing temperatures on photosynthesis, almost every biological process in a plant, from phenology to reproduction to water relations, is affected by high-temperature stress. Thus, while the warming climate does profoundly affect photosynthesis and photorespiration by various mechanisms, the diverse impacts of higher temperatures on other metabolic and physiological processes will unavoidably feedback on photosynthesis in ways that must be acknowledged and considered.