SEARCH

SEARCH BY CITATION

Keywords:

  • canopy temperature;
  • climate change;
  • drought;
  • energy balance;
  • global warming;
  • heat wave;
  • infrared heating;
  • plant–climate interactions;
  • plants;
  • temperature manipulation

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References

1. The infrared (IR) heating technique used in terrestrial vegetation warming experiments has been continuously improved over the past two decades, but the temperature control methods applied thus far inadequately reflect natural conditions.

2. Current control methods base the IR radiation administered on measured canopy or air temperatures. However, these temperatures are influenced by stomatal responses of the vegetation, which depend on moisture conditions and therefore also on warming. In nature, drought-stressed vegetation warms up more than well-watered vegetation, leading to potential differences in heat stress. Current control methods preclude such differences from developing.

3. We propose an alternative approach to render temperature control independent of plant responses. Theoretical canopy temperatures associated with given (target) air temperatures are calculated, based on reference canopy conductance determined from controls and fluctuating meteorological conditions. The IR radiation needed to attain the theoretical temperature is subsequently applied. Actual canopy and air temperatures are free to deviate from that temperature, as measured temperatures no longer control the heaters’ radiation output.

4.Synthesis. We devised an alternative control method for IR heaters that is independent of plant responses to heating. This is especially critical in extreme event studies, where differences in plant water status, and therefore canopy temperatures, are likely to be exacerbated.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References

Experimentally subjecting plants to a warmer environment while striving for natural conditions is by no means an easy endeavour. Tools such as indoor growth chambers struggle to provide a natural light environment, while outdoor greenhouses are usually associated with lower light levels, reduced wind speed and changes in humidity. The use of heating cables also grants full temperature control, but this technique warms only the soil and seems unrealistic given the important and established connection between above- and below-ground plant parts. Passive warming methods include open top chambers (Marion et al. 1997) and the use of retractable infrared (IR) reflective covers (Beier et al. 2004), but these systems, which do not allow temperature control, depend very much on the prevailing meteorological conditions and are hence variable.

A technique that has been gaining ground in climate manipulation studies is the use of IR lamps (Aronson & McNulty 2009). Harte et al. (1995) were presumably the first to use such overhead heaters. In their field experiment, which started in 1991, Harte et al. suspended commercially available IR lamps above montane meadow plots. The lamps emitted a constant flux, which warmed vegetation and soil year-round. Later, Nijs et al. (1996) improved this technique by adding a modulator of the IR flux, which made it possible to accurately uphold a constant difference between the surface temperatures of warmed and control plots. Without such modulation, fluctuations in wind speed, which bring cooler ambient air to the warmed plots, enhance surface temperature variability together with increasing the mean. One of the benefits of using IR heaters is that this approach does not require enclosing the plants, so wind speed and light are hardly influenced. Furthermore, the warming is direct: the IR lamps heat the canopy surface without having to overcome a boundary layer resistance, which makes the technique very responsive (Nijs et al. 1996; Kimball et al. 2008), although this does imply that the air is only warmed indirectly and generally to a lesser extent than the surface. Finally, the whole canopy and the soil are warmed (Kimball 2005), which makes it a highly inclusive technique. A drawback in IR irradiation experiments is that the air underneath the heaters is relatively drier (lower relative humidity), creating unrealistic vapour pressure deficits. Kimball (2005) therefore introduced a corrective method through irrigation, mimicking a constant relative humidity, in line with climate projections. De Boeck et al. (2010) noted that such a correction is not required in heat wave studies, as relative humidity is naturally reduced during such extreme events. The fact that the advantages seem to outweigh the disadvantages is prompting more and more researchers to use IR heaters for their warming experiments. These include longer term experiments (Morin et al. 2010), extreme event simulations (Van Peer et al. 2004; Smith 2011), studies focused on soils (Xia, Chen & Wan 2010) and night-time-only warming (Mohammed & Tarpley 2009).

Controlling canopy or air temperatures?

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References

Despite the apparent growing popularity of IR heating in ecological research, an issue that is still unresolved is whether to control for air or for canopy temperature. In most studies, temperature control is based on canopy (surface) temperatures (e.g. Braun, Buchner & Neuner 2002; Marchand et al. 2004; Morin et al. 2010). This seems a logical choice, as this is the variable that is directly affected by the applied heating. Moreover, the temperature that ultimately determines metabolic plant processes is that of the plant itself, not that of the air. Indeed, it is well documented that species from both cold and hot ecosystems are morphologically adapted to force plant temperatures to be closer to the metabolic optimum, in spite of adverse air temperatures (Larcher 2003). For example, alpine plants often form cushions, which increase the effective characteristic dimension (acting as a big leaf) and decrease the wind speed, leading to leaf temperatures substantially above those of the surrounding air. In other studies, however, temperature control is based on air rather than canopy temperatures (e.g. Wan et al. 2002; Sherry et al. 2008; Mohammed & Tarpley 2009). Dissipation of sensible heat of the canopy under IR heaters indeed increases air temperatures as well. As both historical records and climate change projections contain only information on air temperatures, and as these sources are commonly the basis for deciding which temperature regime to impose, using this same variable in the experiment seems rational. So, which temperature to control for? The problem is that increases in canopy temperatures are not necessarily equivalent to (natural) rises in air temperatures. A regression between canopy temperatures measured at experimental plots and air temperatures recorded at a nearby meteorological station in the study of Marchand et al. (2006) proved significant but showed marked variation (R2 of 0.21). This is unsurprising as canopy temperatures depend not only on air temperatures, but also on wind speed, canopy conductance, air humidity and the general radiative environment. These variables exert an influence on air temperature as well, but less directly.

Kimball (2005) stressed the sensitivity of canopy temperatures under IR heaters to canopy conductance, noting that the power needed to warm the canopy by 1 °C is drastically reduced at night and in water-stressed canopies (i.e. when stomatal conductance is low). Marchand et al. (2006) also stated that the reduced soil water contents observed in their heat wave experiment probably reinforced the temperature increment, and that a heating treatment with the same soil moisture as in the ambient plots would therefore experience a less intense heat wave. A schematic based on standard energy balance equations (see eqn 1) illustrates this (Fig. 1). At the same air temperature, leaf temperature can differ substantially depending on stomatal behaviour. The fact that the plant water status affects canopy temperature is problematic, as most IR heating experiments do not allow for canopy temperatures to vary freely; therefore, the natural plant response is restricted. The problem is not resolved by using air temperature as the control variable, because the warmer air temperatures result from the surface (canopy) warming generated by the IR heaters. Indeed, the fact that the plant response and the level of warming influence each other under IR heaters is an unavoidable consequence of controlling the rate of warming based on a variable that is directly or indirectly a measure of the plant’s response.

image

Figure 1.  Schematic overview of the deviation of leaf temperatures from air temperatures (1:1 line depicted), and how these differences depend on the stomatal behaviour of plants. Figure based on standard energy balance equations (see text for details), with fixed dew point temperature, wind speed and radiation.

Download figure to PowerPoint

The application of a constant energy flux is one way in which warming can be applied independent of plant responses (e.g. Saleska et al. 2002). However, the resulting warming (i) increases temperature variability due to the dependence on fluctuating environmental conditions (mainly air temperature, wind and radiation), (ii) is uncontrolled – no target temperatures can be set. Also, the amount of energy that needs to be applied is unclear, as for example, merely adding the extra radiative forcing from climate projections to simulate a future climate will hardly affect temperatures (Kimball 2005). The use of a constant energy flux therefore seems an inadequate solution for many experiments, especially those testing specific scenarios of average temperature increases or extreme deviations from normal (heat waves).

Excluding plant responses

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References

What is needed is an adjustable amount of warming that avoids the artefacts associated with the plant response and does not rely on the uncontrolled constant flux approach. To achieve this, the amount of heat that is added should be independent of plant responses. We propose to accomplish this by (i) calculating theoretical canopy temperatures associated with given (target) air temperatures, under fluctuating meteorological conditions, and (ii) simultaneously computing the energy output of the IR heaters required to achieve said theoretical canopy temperature. Step (i) can be done with a standard energy balance equation in the form of:

  • image

where Rn is net radiation, H is sensible heat flux and λE is latent heat loss. This equation can be resolved (see common textbooks on environmental physics, e.g. Campbell & Norman 1998) for canopy temperature (Tcanopy) from variables that can be measured with common devices that measure air temperature (Ta), relative humidity, wind speed and radiation, which are installed in most climate manipulation experiments:

  • image(eqn 1)

where Rabs is absorbed radiation, εs the surface emissivity (which can be regarded as a constant), σ the Stefan–Boltzman constant, λ the latent heat of vaporization for water (= −42.575 Ta + 44994, in kJ mol−1), G the heat flux from or into the soil (measurable with heat flux plates), g the canopy conductance (see later), D the vapour deficit of the air (which can be derived from relative humidity and air temperature), pa the atmospheric pressure, cp the specific heat of air (a constant), gr the radiative conductance (= 4εsσTa3/cp), gHa the canopy (boundary layer) conductance (which is dependent on wind speed) and finally s the slope of the saturation mole fraction = Δ/pa with Δ = bces(T)/(c + T)2 and es(T ) = a exp(bT/(T+c)), where a, b and c are constants.

However, the input air temperature should be the target air temperature: preferably a predefined increment above the fluctuating ambient (measured) air temperature (e.g. + 3 °C), as this reflects the natural daily temperature course (see next). In other words, the formula calculates what the future canopy temperature would be under the prevailing conditions of irradiation, wind speed, humidity, etc., given a prescribed warming scenario. With the target canopy temperature computed, the control device of the IR heaters should then modulate their output (e.g. by altering the voltage, see Nijs et al. 1996) to generate this temperature, whilst calculating the amount of radiation (or power) needed to achieve this (cf. Kimball 2005). Existing modulators could fairly easily be adapted to achieve the control proposed here. The resulting algorithm (power as a function of environmental factors) should then be used throughout the experiment. A crucial input parameter is canopy conductance, which has to reflect reference conditions. Canopy conductance should therefore be based on the study’s controls by entering the current environmental conditions measured in those plots and solving the energy balance equation for stomatal conductance. This needs to be done throughout the experiment, as canopy conductance depends on growth and plant age, which are bound to change during the course of the experiment. The fact that growth can differ between controls and treatments is no issue as such differences are treatment-induced and, because the purpose of this approach is to render the temperature control independent of plant responses, growth should therefore not be allowed to affect temperature control. Indeed, it is important that – during the treatment – the only temperature input is the target air temperature, with the actually measured canopy and air temperatures being the result of the applied energy addition in the form of IR radiation as well as of the plant and whole-system response to this. As such, the deviation of the actual canopy temperature from the theoretically computed temperature is indicative of the stomatal conductance: if the deviation is positive, conductance is lower than normal (signalling potential drought stress); if the deviation is negative, conductance is higher than normal. This method allows plant responses to freely affect temperature levels, without these responses in turn affecting the applied warming, hence creating artefacts.

Kimball (2005) showed that only approximately half the energy is required to warm water-stressed vegetation by 1 °C compared with well-watered vegetation. Under the control methods based on canopy temperature, this would mean that the IR heater power output would indeed be halved in such water-stressed vegetation. Using our method, the power output would remain identical to that under reference conditions. If these reference conditions reflected stomatal conductance under field capacity soil moisture (well watered), dried vegetation would, in the example of Kimball (2005), end up being warmed by ∼ 2 °C instead of 1 °C. We argue that this is more true to natural conditions because, by heat dissipation through transpiration, well-watered plant communities can avoid high temperatures potentially causing heat stress, while drought-affected communities cannot (cf. Fig. 1). By ignoring the difference between well-watered and drought-affected plants, the previously used methods may have created a bias in favour of drought-stressed plants by limiting the potential natural increase in leaf temperatures. The proposed new way of controlling IR heaters avoids such bias, thereby further strengthening the appeal of this technique for use in any climate manipulation study. Note that our approach would not only improve the accuracy of the warming simulation under drought, but also under other abnormal water conditions. For example, Sherry et al. (2008) combined IR warming with a precipitation addition treatment (double the ambient amounts), aiming for a 4 °C increase in air temperatures in both precipitation treatments. Higher soil moisture levels (caused by the 100% increase in precipitation) could lead to higher stomatal conductance and therefore increased transpirational cooling. The warming of the double precipitation treatment could therefore result in lower actual temperature increases than in the warmed ambient precipitation treatment, at least using a method where the radiation administered is independent of the plant responses (such as in our approach). This would be a realistic scenario, but would potentially affect the study’s outcome compared to using a predefined increase in air temperature (here: 4 °C) for both warming treatments.

Use in extreme event simulation

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References

As an example, we show how heat wave experiments could benefit from using the new heater control presented here. First note that IR heating is currently the only technique capable of generating the large temperature increases needed to simulate heat waves in free air. In our own IR heater set-up, we are able to impose increases in excess of 10 °C above ambient air temperatures. It has been shown numerous times that warming experiments often induce soil drying (e.g. Saleska, Harte & Torn 1999; De Boeck et al. 2008). As the rate of warming needed to create heat waves is (logically) much higher than the temperature increases in so-called trend experiments, this drying effect is expected to be exacerbated. To separate direct (high temperatures) from indirect (soil drying) heat wave effects, two treatments should hence be imposed: one in which soil drying is avoided altogether, and one in which it is allowed (e.g. Smith 2011). The stomatal response to the different water status would probably create substantially deviating canopy temperatures in both heat wave treatments. For example, at an air temperature of 35 °C, a drought-stressed canopy can be more than 10 °C hotter than a well-watered canopy (Campbell & Norman 1998).

In Fig. 2, we conceptually illustrate which canopy temperature trends the three main control methods would induce in the case of a heat wave with developing drought. The traditional control directly based on canopy or air temperature (Method 1) would not allow a (natural) drought-induced increase in canopy temperature to develop, ignoring the difference between a heat wave with and without drought (i.e. both heat wave treatments would have the same canopy temperature). The second method administers a constant radiative flux (fixed IR heater power), in this hypothetical case with a priori calculation of the fixed power level required to attain a certain (target) maximum canopy temperature under average conditions (temperature, wind speed, canopy conductance, etc.). As the IR radiation supplied is constant, canopy temperature will rise as drought-induced stomatal responses start to take effect (cf. Fig. 1). A heat wave treatment with or without drought would therefore yield different canopy temperatures using this method, a realistic scenario. However, as the true ambient conditions during the treatment are bound to deviate from the long-term averages of temperature, wind speed, etc., used for the a priori calculation of the IR heater output, the ultimate canopy temperatures will be subject to increased and uncontrolled variability unrelated to the heat wave treatment. Finally, our approach combines the independence from plant responses of Method 2 with the controllability of Method 1.

image

Figure 2.  Conceptual comparison of how a heat wave with developing drought imposed by infrared (IR) heating affects canopy temperatures using: (i) a typical control that maintains a fixed difference of canopy or air temperatures between the treatment and the reference plots (black squares); (ii) a predefined constant radiative flux emitted by IR heaters (grey triangles); (iii) the proposed alternative approach that maintains radiative output of infrared heaters independent of plant responses by calculating a (theoretical) canopy temperature based on variable environmental conditions, a target air temperature and a reference canopy conductance (white circles).

Download figure to PowerPoint

As daytime temperatures during heat waves are often increased more than night-time temperatures compared with average (De Boeck et al. 2010), we advise to incorporate such differences into the target temperatures. Target minimum and maximum air temperatures could then be calculated from meteorological records to quantify the intensity of the heat wave by means of its return time, something that would facilitate comparison between extreme event studies. To avoid the unnatural situation of one constant day and night-time temperature (a method sometimes adopted in growth chamber experiments), we propose to add a fixed temperature difference (ΔT) to fluctuating ambient air temperatures instead, based on weather predictions. Specifically, if temperature minima and maxima of 12 and 25 °C are predicted, and the target heat wave minima and maxima are 17 and 33 °C, then the control system should be set with a ΔT of 5 °C during night time and 8 °C during daytime. These settings should be adjusted daily. This approach entails: (i) custom-adjustable target temperatures (from projections or historical records), (ii) unconstrained vegetation responses, which can alleviate or aggravate heat stress, and (iii) natural daily temperature progression. This method can likewise be used to provide reliable temperature increases during drought, precipitation intensity manipulation, and other extreme event studies.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References

Since researchers first started using IR heaters for warming vegetation two decades ago, the technique has been continuously improved upon. A hitherto unresolved issue has been the manner in which to control the temperature increase. We argue that neither controlling for canopy nor for air temperature is correct, as these both depend on the stomatal response, which is in turn indirectly affected by warming. The alternative control method presented here operates independently of the plant response. As such, it should improve the accuracy and naturalness of the applied warming, and it further strengthens the arguments in favour of using the IR heating technique as an appropriate tool in warming studies. We illustrated this by discussing the potential use of the new approach in extreme climate event research.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References

H.J.D.B. is a post-doctoral research associate of the Fund for Scientific Research – Flanders. We thank Toon De Groote and three referees for useful advice.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Controlling canopy or air temperatures?
  5. Excluding plant responses
  6. Use in extreme event simulation
  7. Conclusions
  8. Acknowledgements
  9. References
  • Aronson, E.L. & McNulty, S.G. (2009) Appropriate experimental ecosystem warming methods by ecosystem, objective, and practicality. Agricultural and Forest Meteorology, 149, 17911799.
  • Beier, C., Emmett, B., Gundersen, P., Tietema, A., Penuelas, J., Estiarte, M., Gordon, C., Gorissen, A., Llorens, L., Roda, F. & Williams, D. (2004) Novel approaches to study climate change effects on terrestrial ecosystems in the field: drought and passive nighttime warming. Ecosystems, 7, 583597.
  • Braun, V., Buchner, O. & Neuner, G. (2002) Thermotolerance of photosystem 2 of three alpine plant species under field conditions. Photosynthetica, 40, 587595.
  • Campbell, G.S. & Norman, J.M. (1998) An Introduction to Environmental Biophysics, 2nd edn. Springer-Verlag, New York, USA.
  • De Boeck, H.J., Lemmens, C.M.H.M., Zavalloni, C., Gielen, B., Bossuyt, H., Malchair, S., Carnol, M., Merckx, R., Van den Berge, J., Ceulemans, R. & Nijs, I. (2008) Biomass production in experimental grasslands of different species richness during three years of climate warming. Biogeosciences, 5, 585594.
  • De Boeck, H.J., Dreesen, F.E., Janssens, I.A. & Nijs, I. (2010) Climatic characteristics of heat waves and their simulation in plant experiments. Global Change Biology, 16, 19922000.
  • Harte, J., Torn, M.S., Chang, F.R., Feifarek, B., Kinzig, A.P., Shaw, R. & Shen, K. (1995) Global warming and soil microclimate – results from a meadow-warming experiment. Ecological Applications, 5, 132150.
  • Kimball, B.A. (2005) Theory and performance of an infrared heater for ecosystem warming. Global Change Biology, 11, 20412056.
  • Kimball, B.A., Conley, M.M., Wang, S., Lin, X., Luo, C., Morgan, J. & Smith, D. (2008) Infrared heater arrays for warming ecosystem field plots. Global Change Biology, 14, 309320.
  • Larcher, W. (2003) Physiological Plant Ecology, 4th edn. Springer-Verlag, Berlin, Germany.
  • Marchand, F.L., Nijs, I., De Boeck, H.J., Kockelbergh, F., Mertens, S. & Beyens, L. (2004) Increased turnover but little change in the carbon balance of High-Arctic tundra exposed to whole growing season warming. Arctic Antarctic and Alpine research, 36, 298307.
  • Marchand, F.L., Verlinden, M., Kockelbergh, F., Graae, B.J., Beyens, L. & Nijs, I. (2006) Disentangling effects of an experimentally imposed extreme temperature event and naturally associated desiccation on Arctic tundra. Functional Ecology, 20, 917928.
  • Marion, G.M., Henry, G.H.R., Freckman, D.W., Johnstone, J., Jones, G., Jones, M.H., Levesque, E., Molau, U., Molgaard, P., Parsons, A.N., Svoboda, J. & Virginia, R.A. (1997) Open-top designs for manipulating field temperature in high-latitude ecosystems. Global Change Biology, 3, 2032.
  • Mohammed, A.R. & Tarpley, L. (2009) Instrumentation enabling study of plant physiological response to elevated night temperature. Plant Methods, 5, 7.
  • Morin, X., Roy, J., Sonié, L. & Chuine, I. (2010) Changes in leaf phenology of three European oak species in response to experimental climate change. New Phytologist, 186, 900910.
  • Nijs, I., Kockelbergh, F., Teughels, H., Blum, H., Hendrey, G. & Impens, I. (1996) Free air temperature increase (FATI): a new tool to study global warming effects on plants in the field. Plant, Cell and Environment, 19, 495502.
  • Saleska, S.R., Harte, J. & Torn, M.S. (1999) The effect of experimental ecosystem warming on CO2 fluxes in a montane meadow. Global Change Biology, 5, 125141.
  • Saleska, S.R., Shaw, R.M., Fischer, M.L., Dunne, J.A., Still, C.J., Holman, M.L. & Harte, J. (2002) Plant community composition mediates both large transient decline and predicted long-term recovery of soil carbon under climate warming. Global Biochemical Cycles, 16, 1055.
  • Sherry, R.A., Weng, E.S., Arnone, J.A., Johnson, D.W., Schimel, D.S., Verburg, P.S., Wallace, L.L. & Luo, Y.Q. (2008) Lagged effects of experimental warming and doubled precipitation on annual and seasonal aboveground biomass production in a tallgrass prairie. Global Change Biology, 14, 29232936.
  • Smith, M.D. (2011) An ecological perspective on extreme climatic events: a synthetic definition and framework to guide future research. Journal of Ecology, 99, 656663.
  • Van Peer, L., Nijs, I., Reheul, D. & De Cauwer, B. (2004) Species richness and susceptibility to heat and drought extremes in synthesized grassland ecosystems: compositional vs physiological effects. Functional Ecology, 18, 769778.
  • Wan, S., Yuan, T., Bowdish, S., Wallace, L., Russell, S.D. & Luo, Y.Q. (2002) Changes in microclimate induced by experimental warming and clipping in tallgrass prairie. Global Change Biology, 8, 754768.
  • Xia, J., Chen, S. & Wan, S. (2010) Impacts of day versus night warming on soil microclimate: results from a semiarid temperate steppe. Science of the total environment, 408, 28072816.