12  Manipulating UV, VIS and NIR radiation

Light sources and light modifiers

First edition authors

Pedro J. Aphalo, Andreas Albert, Andy R. McLeod, Anu R. Heikkilä, Iván Gómez, Félix López Figueroa, T. Matthew Robson, Åke Strid

12.1 Safety considerations

12.1.1 Risks related to sunlight exposure

When working in field experiments, workers and researchers are exposed to sunlight as in any other outdoor activity. To avoid health hazards adequate protection should be used, either in the form of clothing or sun-blocking lotions. In some cases, for example when there is strong reflection of radiation by water, sand or snow-covered surfaces, eye protection in the form of sunglasses or goggles should be used.

A common and simple parameter of the UV radiation level outdoors is the index, which was developed by the World Health Organization (WHO) including recommended protection (WHO 2002). The index is used for example in weather forecasts in integer numbers from 1 to 11+, representing the exposure from low to extreme. [^14]

12.1.2 Risks related to the use of UV lamps

Lamps emitting and create a higher risk of eye damage than sunlight because when radiation is not accompanied by strong visible light the eye pupils remain wide open. Also many lamps emit , which is not present in sunlight at ground level and is more damaging than . It is important to use goggles to protect the eyes not only from radiation from the front but also from the sides of the face. It is important to ensure that the goggles used are designed to protect from radiation, rather than just from the impact of flying particles. The skin should also be well protected, because in the same way as for plants, the effect of -B on humans is enhanced under low visible light irradiance by the low rate of photoreactivation (light-driven repair of DNA damage).

Locations where lamps are in use should have visibly located warning signs. Although there is no standardized symbol for radiation hazard, the warning symbol for optical radiation (Figure [fig:warning:optical]) accompanied by the text ‘UV radiation’ can be used. Additional text, ‘protect eyes and skin’ or ‘do not look directly at light source’ can be added. Access to outdoor experiments should be restricted by fences with locked gates, and by locking access doors to greenhouses, controlled environment rooms and cabinets. This is to prevent exposure of people who are unaware of the risks involved. Everybody with access to the area should be informed of risks, and trained to use the protection and work procedures required to mitigate them.

Lamps also pose risks unrelated to radiation. Fluorescent tubes contain mercury and should disposed of as hazardous waste. The glass envelope of lamps is fragile, and can cause injuries if it breaks. Xenon-arc lamps also present a relatively high risk of explosion, goggles and other face and body protection should be used. This risk is caused by the high pressure inside this type of lamp and is present irrespective of the xenon-arc lamps being on or off.

Lasers, can easily produce injuries, and should be handled with great care. Even the very low power laser pointers used during lectures are capable of injuring the eye if pointed towards the audience. As with light sources, access to any type of laser should be restricted to trained users, and eye protection should be worn to match the wavelength emitted by each particular laser. Recommendations relating to warning signs, given above, apply equally to lasers and UV lamps.

12.1.3 Risks related to electrical power

Using mains power for either instruments or lamps, involves a risk of electric shock and this is especially true in humid conditions and outdoors. All components, like connectors, enclosures, and switches, should have an IP rating[^15] and any other codes required for their use in a particular environment. When using mains power in a room where there is water, it is usually a requirement to have grounded mains sockets protected with residual current devices (RCD). These sense the current imbalance in the power line caused by a current leak to ground, for example caused by the accidental flow of current through the body of an operator. When this residual current is sensed, the device trips and disconnects power in a few milliseconds so preventing serious injury. RCDs and earth grounding should be regularly tested. Never use an electrical device that has a grounded plug in a mains socket lacking ground contacts because this will create a risk of electrocution and in addition may negatively affect the functioning of measuring instruments.

12.1.4 Safety regulations and recommendations

The EU directive 2006/25/EC about protecting workers from optical radiation can be found at http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2006:114:0038:0059:EN:PDF. This directive does not apply to solar radiation. It gives maximum allowed levels of exposure from artificial optical radiation sources and other requirements on how to achieve these and the monitoring of workers health. Britain’s Health and Safety Executive (HSE) has produced a guide on how to apply this directive (not approved yet by EU), that can be found at http://www.hse.gov.uk/radiation/nonionising/optical.htm. At this address, there is also information on sun exposure indicating that ‘UV radiation should be considered an occupational hazard for people who work outdoors’. The World Health Organization (WHO) has issued general recommendations about protection from the sun, which can be found at http://www.who.int/uv/sun_protection/en/.

12.2 Artificial sources of UV radiation

12.2.1 Lamps

12.2.1.1 Fluorescent lamps and tubes

Fluorescent lamps and tubes are low pressure mercury vapour lamps. The mercury vapour emits radiation at specific spectral lines, mostly in the UV region of the spectrum. Except for the case of germicidal lamps the inside of the glass tube is coated with a layer of ‘phosphor’[^16] which absorbs UV radiation and re-emits the energy as fluorescence, at longer wavelengths, either as , , or visible radiation. There are even some fluorescent tubes producing far-red radiation around 750 nm. For environmental studies we are interested in and emitting lamps. These lamps emit across a rather broad wavelength band, and have several minor secondary peaks in the and visible regions. Figure [fig:TL12:UVB313:UVA340] shows the emission spectra of some UV lamps.

image

It should be remembered when designing experiments that lamps should be always filtered with cellulose diacetate film or some other material to remove , otherwise the effects observed will not be only due to but also to radiation. The emission of radiation might look small in energy terms but, being very effective in eliciting biological effects, it is capable of completely distorting the apparent response of plants and other organisms to radiation in an experiment. In the case of lamps, both and radiation should be removed by filtration.

Ultraviolet emitting fluorescent lamps also emit a small amount of visible radiation, even in the red and orange regions of the spectrum. This should not be forgotten when using these lamps as the only source of radiation. For example the very small amount of red-orange light can be enough to enhance germination of silver birch seeds in Petri dishes when irradiated with lamps (Pedro J. Aphalo, unpublished data).

The most commonly used and lamps are 1200 mm-long tubes, rated at 40 W of electrical power. These lamps are sold for materials testing (Q-Panel UVB-313 and UVB-340, Q-Labs Inc., Farnworth, England) and for medical use (Philips TL-40 W12, and TL-40 W 01, Philips, Amsterdam, The Netherlands). The second of these lamp types from Philips emits over a very narrow peak, and are not suited for simulating ozone depletion, but might be useful when the aim is not to simulate solar radiation. There are also other lamp types available, for example small-sized compact fluorescent lamps such as PL-S 9W 01 from Philips.

Using fluorescent lamps requires some additional equipment (Figure [fig:fluorescent:tube]). The main component is a “ballast” which limits the current through the lamp. Traditional ballasts are electromagnetic, comprising a coil wound on a ferrous core, and require one more component, a ‘starter’ for turning on the lamps. Lamps driven by such ballasts run at power-line frequency, 50 Hz in Europe and 60 Hz in USA, causing emission to flicker in time. This flicker is not visible to humans, but is visible to some insects. The modern alternatives are electronic ballasts driving the lamps at high frequency (in the order of 50–100 kHz) which together with the latency of the phosphor yields an almost constant radiation output from the lamps. Consequently, we recommend the use of high frequency ballasts to avoid artifacts. Electronic ballasts do not require starters.

image

Some types of electronic ballasts are dimming ballasts , allowing the adjustment of lamp output from full power down to a ballast-type-specific low value, usually somewhere between 1 and 10% of full power. Dimming of lamp output can also be achieved using electromagnetic ballasts and phase-angle dimming controllers but these enhance the visible flicker of the lamps. Some ballasts are designed for newer T8 (thin, 26 mm diameter) tubes rather than the old-fashioned T12 (thick, 38 mm diameter) tubes used for most lamps, which may cause some problems with dimming, and affect the lowest power level achievable. Dimming ballasts are controlled digitally (for example DSI system for Tridonic PCA 1/38 T8 ECO, Tridonic GmbH & Co KG, Dornbirn, Austria) or by a direct current voltage signal (for example 1–10 V for Quicktronic HF 1X36/230-240 DIM UNV1, Osram, Munich, Germany). Dimming ballasts can be useful for adjusting doses for different treatments, or when building modulated supplementation systems (see section 8.2.6 on page ).

Radiation output from fluorescent lamps varies markedly with ambient temperature as shown in Figure [fig:UV:lamp:temperature]. The data in this figure were obtained by varying the temperature in a growth chamber where the lamps were located, and measuring the spectral irradiance with a spectroradiometer maintained at near constant temperature outside the chamber. Only the optical fibre entered the chamber through an instrumentation port. In non-modulated systems, and in experiments with treatments at different temperatures this should be taken into account and the irradiance from lamps should always be measured at the same temperature as during use of the lamps in the experiments. When the ambient temperature fluctuates, the lamp output should be measured continuously, or at least at a range of temperatures and irradiances estimated from a continuous record of temperature.

image

irradiance should never be adjusted by using different plant to lamp distances in different treatments. Changing the distance modifies the shading effect of lamp frames, and even small differences in can create spurious differences between treatments (see Flint et al. 2009 for details). Furthermore, it is essential to have controls both under frames with unenergised lamps and also under frames with energized lamps but filtered with polyester film to remove the , as the only difference between and non-treatments should be the irradiance of (see Newsham et al. 1996). If we want to assess the effect of shading by lamp frames alone an additional control with no lamp frames could be added to an experiment.

Fluorescent tubes are normally held on metal frames, and each frame supports several lamps. Because these tubes are long and narrow, using single tubes would provide a very uneven irradiance field. Care should be taken when choosing the spacing of the tubes, and the vertical distance between the tubes and plants. The model of Björn and Teramura (1993) can be used to simulate the spatial distribution of irradiance or fluence rate under such an array of lamps (for a reimplementation of the model as an R package see section 8.8 on page ). Figure [fig:position:tubes] shows the results of six such simulations, three for tubes evenly distributed along the frame (‘equidistant’) and three for tubes arranged following the projection on a plane of what would be equidistant distribution around the perimeter of a half circle (‘cosine’ pattern). The closer together the contour lines are, the steeper the change in irradiance. The examples in Figure [fig:position:tubes] are for irradiance, and these results do not apply to fluence rate. For each size and number of lamps and size of frames, it should be possible to optimize either the evenness of irradiance or of fluence rate by means of the model (Figure [fig:position:tubes:square]). The unevenness of the radiation field should be taken into account when deciding where to the place plants, and also when measuring exposures under the frames.

When using UV lamps to supplement solar radiation, shading by the lamps and their supporting structures can become a problem. This is discussed in section 9.10.3 on page in relation to effective doses and enhancement errors.

12.2.1.2 Xenon arc lamps

Xenon arc lamps are specialised light sources that produce intense visible and UV radiation from an electric discharge in high pressure xenon gas. The lamps comprise tungsten electrodes inside a quartz envelope and produce an intense plasma ball at the cathode at 6000–6500 K temperature generating a spectral output throughout the visible and UV (Figure [fig:xenon:lamp:spectrum]). Xenon arc lamps must be operated in a fully enclosed housing (Figure [fig:xenon:lamp:photo]) that usually provides mirrors and focussing adjustments, electrical ignition and ventilation to cool the lamp. Xenon arcs provide intense point sources but with careful operation and adjustment are sometimes useful for experiments in photobiology. They are extremely strong sources and have a very high heat output in the infra-red waveband. This heat output typically requires their use with a water filter and/or a dichroic mirror reflector to reduce heating if used in experiments with plant material. A dichroic mirror transmits infra-red onto a heat sink and reflects only the required waveband (e.g. UV radiation) onto the target. After removal of excessive heat using a water filter or dichroic mirror they can be used with narrow band-pass filters for the determination of effects at different wavelengths such as action spectra. Care must be taken not to focus the source onto the mirror or filter which can be readily damaged by excessive heat.

image

image

In addition to the essential safety considerations when using any UV sources (detailed in section 8.1), important additional safety precautions are necessary when using xenon arc lamps. The high pressure inside the lamps creates a risk of explosion during installation and during operation, which increases with lamp age. An impact resistant face shield and heavy protective gloves should be worn when handling a lamp and it must never be touched with the fingers as grease marks increase the risk of failure. Lamps are supplied in a protective cover which should remain in place until installation and they should only be used inside a specialised housing that provides explosion protection to contain flying glass should an explosion occur. A specialised lamp housing is normally provided with a fan for cooling the lamp which gets extremely hot and ventilation should continue for at least five minutes after the lamp is switched off. This is usually achieved in a specialised housing with temperature sensors or timers. However, some lamps (sometimes called ‘UV-enhanced’) produce short UV wavelengths (\(<\)250 nm) that generate large quantities of ozone which is toxic and can cause respiratory problems and asthma attacks. The ozone must be removed from the room and building by additional ventilation systems or absorbed and destroyed by special filters. Alternatively, ‘ozone-free’ lamps are most often used and have a quartz envelope that absorbs below 250 nm and so prevents excessive ozone formation. A xenon arc lamp is extremely bright and the intense visible radiation can damage the eyes in addition to UV effects. The arc or its intense image must never be viewed directly as it can cause permanent eye damage. When using the equipment, the illuminated areas may also be extremely bright and dark coloured goggles should be worn, such as those used in welding that also provide UV protection, ensuring that the eyes are also protected from side illumination as well as directly. It is essential to follow the manufacturers’ instructions with respect to lamp life and operating current in order to maximise their useful life and minimise the risk of failure due to explosion. Further information on xenon arc sources and other high intensity lamp sources sometimes used in photobiology is provided by Holmes (1984).

12.2.2 Deuterium lamps

High intensity water-cooled deuterium lamps (150 W) have a fairly flat radiant intensity curve in the UV-B region (approximately 1.5–3.0 between 280 and 300 nm at a distance of 30 cm) that is appropriate for mechanistic plant UV photobiology studies. This type of radiation sources (e.g. produced by Hamamatsu Photonics) requires a dedicated control box for operation but is still an affordable alternative to lasers, starting at about 2,500€ for a setup. High intensity deuterium lamps have been used for wavelength dependence measurements both in photobiology (Kalbin et al. 2005; Kalbina et al. 2008) and photochemistry (Kalbin et al. 2005) when fitted with the appropriate filters. The lamp has to be mounted horizontally to allow efficient water cooling but using spreading lenses and a mirror mounted in a 45\(^\circ\) angle to the incident radiation a full rosette can easily be irradiated (Kalbina et al. 2008) and, the UV irradiation can be supplemented with PAR from external sources as desired. A drawback of this type of equipment is the short life of the radiation source (a few hundred hours). Spare lamps cost approximately 1 500 €.

12.2.3 LEDs

The principle of a light emitting diode (LED) is totally different to conventional incandescent light bulbs. It is a semiconductor consisting of two types of layer: one layer releasing electrons as a direct current to fill holes in another. Energy is released as a photon at a wavelength (colour) that depends on the band gap energy of the materials comprising each of the two layers (Figure [fig:LED:junction]).

image

Due to this emission of photons at a very precise wavelength, the typical spectrum of a LED shows one distinct peak. The emitted wavelength range can be increased by adding a fluorescing material to the LED. This is shown in Figure [fig:LED] together with a measurement of the solar irradiance at the Helmholtz Zentrum München in Neuherberg, Germany. The maximum spectral irradiance of the LED peak (measurement at 20 cm distance) is more than one order of magnitude lower than that of the solar irradiance (which is the reason a semilogarithmic scale is used). All measurements were made using a double-monochromator spectroradiometer (Bentham, Reading, UK).

image

image

If the correct semiconductor material is chosen, almost any wavelength can be produced even in the UV range. The intensity of LEDs (Figure [fig:UVLED]) is much lower than that of LEDs emitting visible light, even at a shorter distance than in Figure [fig:LED]. Normally, the optical power of a LED is around 1 W or less for visible radiation and only around 0.5 mW for UV radiation. There are LEDs available with higher power (many of them are arrays of tens of chips on a single package) but they require an additional cooling device. These arrays can emit up to 50 mW of radiation, however, they are very expensive. This is one of the most important reasons why LEDs alone are not useful at the moment for exposing plants to radiation. In addition, many different LEDs or specially designed LEDs are necessary to approximate the wavelength distribution of the solar spectrum. On the other hand, LEDs have been in common use for photosynthesis research for a long time (Tennessen, Singsaas, and Sharkey 1994) and many gas-exchange instruments use LEDs as a light source.

LEDs are low voltage direct current (DC) devices. The current through a LED must be limited. The simplest current limiter is a resistor connected in series with the LED. The calculator at http://ledcalculator.net/ can be used to determine the value of the required resistor. There is also a handy pair of Android applications, ‘elektor LED Resistor Calculator’ and ‘elektor Resistor Color Code’, available for free from Google Play (https://play.google.com/).

12.2.4 Spectrographs

A spectrograph composed of a light source and a monochromator may be used in applications requiring spectrally-resolved UV radiation exposure of biological specimens. This may be the case when searching for wavelength dependencies in plant responses or investigating specific effects known a priori to depend on wavelength.

The selection of a light source is application driven and depends on the requirements imposed by the study. The main requirements concern the intensity and the spectral distribution of the radiant output of the lamp. The geometry of the setup, including the source-target-distance and the area of exposure, sets certain limits not only on the light source but also on the characteristics of the monochromator.

A schematic representation of a typical single monochromator is shown in Figure [fig:spectrograph]. The light emitted by the light source is directed onto the entrance slit of the monochromator. The collimating mirror reflects the light onto the grating that diffracts the light into its spectral components. The diffracted light is reflected by the focusing mirror onto the exit slit. Two single monochromators may be aligned in a way that the exit slit of the first monochromator serves as the entrance slit of the second monochromator. This arrangement makes up a double monochromator which provides better stray:light rejection than single monochromators. Commercially available monochromators of both types exist.

image

In the setup of a single monochromator shown in Figure [fig:spectrograph], photons of only one wavelength would ideally exit the exit slit. However, as the slits have finite dimensions, the actual output would be a narrow line of radiation, peaking around a certain nominal central wavelength. Two approaches are commonly adopted to achieve the production of multiple spatially-separated lines. The grating may be equipped with a drive arm rotating the grating. Ideally, for each position of the grating, only radiation of a certain wavelength is reflected and focused onto the exit slit one at a time. Alternatively, the focusing mirror may be omitted and replaced by a sample exposure plane. Ideally, in each position on the plane, only radiation of a certain wavelength would be present (Figure [fig:UVEMA]).

image

The main characteristics of a monochromator that should be considered include wavelength region, wavelength accuracy, bandwidth of the lines, and the amount of stray light. These characteristics set guidelines for design of the layout of the optical components, selection of the slit widths, and formulation of the specifications for the grating. An unavoidable compromise must be made between the wavelength resolution and the intensity. Finer separation of the wavelengths comes with lower intensities and vice versa. Stray light rejection may be improved by placing suitably-designed baffles inside the enclosure of the spectrograph. For applications requiring superior stray light rejection, a double monochromator should be used.

The main characteristics of the grating include groove spacing and a ruled area. For a spectrograph operating at UV radiation wavelengths, a holographic grating is a more feasible choice than a ruled grating because it can achieve a higher resolution. The main difference between ruled and holographic gratings is the process of their manufacture. A ruled grating has grooves scribed by a diamond on a ruling machine, whereas holographic gratings are produced by a photolithographic process using lasers which are able to achieve a higher groove density. Should the focal plane be flat, a concave grating must be used. The size of the grating has certain practical limits set by the manufacturers. The maximum size of the exposure area is essentially limited by the size of the grating.

Selection of the light source is defined by the desired wavelengths, the intensity levels, and the spectral distribution of the radiation. For experiments imitating natural exposure, light sources exhibiting radiative characteristics resembling those of natural sunlight, such as xenon-arc lamps, should be used. The heat tolerances of the substrates of the gratings are limited. As a consequence, the light entering the monochromator may have to be filtered to remove excessive infra-red radiation which is often achieved using a water filter. (see subsubsection 8.2.1.2 on page )

In addition to the small spectrographs described above, there are also much larger instruments. Many of these dispense with the exit slit, and project the dispersed spectrum onto a ‘stage’. Watanabe et al. (1982) describe a large spectrograph designed for irradiating biological samples. Over a curved 10 m-long focal plane covering wavelengths of 250 to 1000 nm, many samples can be irradiated simultaneously. There are few such spectrographs in the world, but they are important for measuring action and response spectra (e.g. Saitou et al. 1993).

12.2.5 Lasers

Laser stands for ‘Light Amplification by Stimulated Emission of Radiation’. Stimulated emission can be induced in many different materials, and when light is confined in a cavity between mirrors the stimulated emission is amplified yielding a narrow beam of spatially- and temporarily coherent, and collimated light. Lasers usually produce very narrow and intense beams of monochromatic light, although there are exceptions. Some lasers are tunable, meaning that the wavelength of the emitted light beam can be varied. One of the many ways of achieving this is to replace the mirror at one end of the cavity with a movable prism or grating. The lasing medium can be a solid crystal, a gas or a dye in solution. Lasers are said to be ‘pumped’ by a source of light, for example a lamp external to the lasing cavity.

Laser diodes are semiconductor lasers. They are solid state devices producing a laser beam based on the same principle as other lasers. They are pumped by the light emitted when electrons and ‘holes’ interact in the semiconductor junction. They are used to excite other types of lasers, in printers and CD/DVD/BD players. Blu-ray disc players use laser diodes which emit blue-violet light. There are emitting laser diodes of low power, for example, emitting 20 mW at 375 nm. The semiconductor material used in this case is . In addition the catalogue of Roithner Lasertechnik GmbH (Vienna, Austria) lists ‘diode pumped solid state lasers’ emitting at 266 nm and at 355 nm. Non solid-state lasers that emit in the region are also available. For example argon ion lasers continuously emit at wavelengths of 334 and 351 nm.

The power of lasers varies from 1–5 mW for laser pointers to 100 kW for lasers under development for military use. The coherent and concentrated beam can cause serious damage to eyesight, and safety precautions should be taken during their use, unless they are of very low power (see section 8.1 on page ).

Lasers can be used for example in characterizing the slit function of spectrometers (see section 9.7.2.2 on page ), for aligning optics, and in the ubiquitous barcode readers used in supermarkets and laser pointers used by speakers during lectures. They are used in instruments like confocal microscopes and different analytical procedures in chemistry and biology. They can also be used to excite photochemical or photobiological systems if the target is small.

For the purpose of UV photobiology, tunable optical parametric oscillator (OPO) pump lasers (pump wavelength 355 nm) are especially useful in mechanistic studies such as in accurate wavelength dependence determinations or action spectroscopy (O’Hara, Strid and Jenkins, in preparation). Placing a cuvette holder in the beam, a plant extract or a protein solution can be accurately irradiated using 50 or 100 cuvettes. By using the appropriate optics (lenses, mirrors, etc.; Figure [fig:tunable:laser]), the geometry of the beam can be manipulated so that a larger area or sample (several cm\(^2\)) can be illuminated. This is large enough for exposing for instance part of an rosette, or for simultaneously irradiating two detached leaves. Tuneable lasers can also be used for studies of complex mechanistic interactions between several plant photoreceptors, since they can be used within a short time interval (seconds) at different wavelengths, e.g. for irradiation first in the , then in the blue or red.

image

When using lasers for photobiological purposes, a number of circumstances need to be taken into account. Lasers emit their radiation in the form of pulses and this should be kept in mind when designing the experiments. Usually, both the energy of the pulses and the pulse rate can be varied within limits, depending on the sophistication (and price!) of the equipment. Pulse rates between 1 to 20 Hz in the more affordable machines up to 1000 Hz in the more expensive ones can be obtained. The pulses are typically 5–10 ns in length and the linewidth is usually around 0.5 nm (Figure [fig:laser:output]), making this type of instrumentation ideal for studies of the initial UV signalling events on the biochemical and cell biological levels. Also, due to the physical principles on which a tunable laser in the UV region relies, some residual radiation with double the desired wavelength may still be present in the beam exiting the apparatus. This has to be taken care of by using a filter in the light-path blocking the longer wavelengths. For instance, when irradiating a plant with a wavelength of 300 nm, care must be taken so that any 600 nm component will be blocked out to be sure that the biological effect is due to the UV-B alone and not the red light. Of course, the laser UV-B radiation can be supplemented with PAR from other light sources as required.

image

Laser radiation is measured as the energy emitted (in J) in each pulse or in a train of pulses using a pulse energy meter. This then has to be related to the area actually irradiated. The maximal output energy for a flash at 280 nm typically ranges between 40 and 200 . The spectral bandwidth and the accuracy of the wavelength setting should be checked prior to use with a calibrated diode array spectroradiometer (Figure [fig:laser:output]). As might be expected, the major drawback of using lasers, in addition to the small area that usually can be irradiated, is the cost. Lasers useful for UV-B photobiological purposes at present start at approximately 50 000 €.

12.2.6 Modulated UV-B supplementation systems

There are two types of supplementation systems for use outdoors: square-wave and modulated. Square-wave supplementation systems work by simply switching the lamps on and off. The only control on the dose of is the length of time that the lamps are energized each day. Modulated systems work by continuously adjusting lamp -radiation output, which is usually called dimming. Furthermore, in modulated systems, the dimming of the lamps is adjusted by means of a feedback control system. radiation is measured both with supplement and without and the control system is programmed so that the treatment is a fixed percent increase in above the control condition, or in some cases, above true ambient conditions. McLeod (1997) reviews many of the early modulated and square-wave systems, and discusses many of the compromises and limitations involved in their use and design. Figure [fig:botania] shows one lamp frame from a modulated system, and Figure [fig:botania:new] shows an overhead view.

image

image

A modulated system automatically adjusts the supplement following changes in natural irradiance with time of day and also compensates for ageing of cellulose diacetate filters, and changes in lamp output with ambient temperature.

Caldwell et al. (1983) describe a modulated system based on custom built electronics, which works even at very low ambient temperatures. Aphalo, Tegelberg, and Julkunen-Tiitto (1999) describe a system built from off-the-shelf components and controlled by a datalogger, which does not work at temperatures below 0. Figure [fig:UV:modulated] shows, for this system, the daily course of irradiance under near-ambient control and enhancement frames throughout two days with different cloud conditions. Occasionally, modulated systems have also been used inside greenhouses (e.g. Hunt and McNeil 1998). Another approach is to use a PID algorithm in an industrial micro-controller module to control the dimming of the lamps. Such a modulated systems has been built based on Gantner intelligent modules (Gantner Instruments Test & Measurement GmbH, Darmstadt, Germany, Matti Savinainen, pers. comm.). It is also possible to use a personal computer and a PID control algorithm implemented in a graphical instrument control and measurement language like LabVIEW (National Instruments Corporation, Austin, TX, USA) or FlowStone (DPS Robotics, UK).

image

Modulated systems, like any other supplementation system, require frequent checks and replacement of burnt lamps and aged filters. It is recommended to replace acetate filters regularly. For lamps (except for the odd ‘early deaths’) it is recommended that they are replaced following a fixed schedule as lamp output degrades with lamp age. This is especially important in those systems not having separate feedback control for each lamp frame, as having lamps of different ages in monitored and slave frames would lead to inaccuracies in the level of supplementation between frames.

Modulated systems are preferable to square-wave systems as they avoid excessive supplementation during periods of low and low natural irradiance. They also compensate for fluctuations in lamp output. However, they are more difficult to design and build, and consequently more expensive. The drawbacks of square-wave systems can be moderated by completely switching off the system on rainy and cloudy days, or by automatically switching off supplementation when irradiance goes below a certain threshold. The duration of square-wave exposure may also be limited to fixed hours either side of solar noon. Musil et al. (2002) evaluated the errors involved in the use of square-wave systems and concluded that they are not very large, whereas Díaz et al. (2006) concluded that modulated systems are preferable. However, it is difficult to assess the effect of the unrealistic treatments in square-wave systems on the outcome of experiments as comparisons have been solely based on differences in irradiation regimes rather than on the responses of plants.

Enhancement errors in the calculation of effective irradiance are discussed in section 9.10.3 on page .

12.2.7 Exposure chambers and sun simulators

Two main types of sun simulators exist: small systems based, for example, on xenon-arc lamps (see 8.2.1.2 on page ), and large systems built by combining several types of lamps and filters to achieve a simulation of the solar spectrum; so called “sun simulators” and “(walk-in-size) exposure chambers”. The latter are sometimes called a ‘phytotron’ (Bickford and Dunn 1972). The principal concept of how to design plant growth chambers can be found for example in Langhans and Tibbitts (1997).

Here, a description is given in more detail of the phytotron built at the Research Unit Environmental Simulation of the Helmholtz Zentrum München, Neuherberg, Germany[^17]. This phytotron facility consists of a set of seven closed chambers (length \(\cdot\) width \(\cdot\) height):

• four walk-in-size chambers (3.4 m \(\times\) 2.8 m \(\times\) 2.5 m),

• two medium-size sun simulators (1.4 m \(\times\) 1.4 m \(\times\) 1.0 m),

• one small sun simulator (1.2 m \(\times\) 1.2 m \(\times\) 0.4 m).

In order to extrapolate response of plants from experiments to those in their natural habitat, these experiments need to be performed under realistic and reproducible conditions. The radiation provided must be as realistic as possible. Not only the quantity but also the spectral quality of radiation has to match the seasonal and diurnal variations occurring in nature (Caldwell and Flint 1994). This includes the steep absorption characteristics of radiation that result from the filtering of solar irradiance by stratospheric ozone as well as the balance between the , , and . As no single artificial light source can simulate both spectral quality and spectral quantity of global irradiance, a combination of metal halide lamps (400 W HQI Daylight, Osram, Germany), quartz halogen lamps (500 W Halostar, Osram, Germany), and blue fluorescent tubes (40 W TLD 18, Philips, the Netherlands) are used in order to simulate the spectrum from the to the near IR wavelengths. Excess infra-red radiation is removed by a layer of water. Underneath this water filter, additional quartz halogen lamps (300 W Halostar, Osram, Germany) are installed to adjust the mid and far IR radiation. The missing irradiance is supplemented by fluorescent tubes (40 W TL 12, Philips, the Netherlands). The radiation output of these fluorescent tubes, however, extends to well below 280 nm. This portion must be blocked very efficiently. Selected borosilicate and lime glass filters as well as acrylic ‘glass’ filters and plastic sheets provide a sufficiently steep cut-off at the desired wavelength. Different combinations of these glasses and films allow the cut-off wavelength to be altered, thus enabling a simulation various scenarios as shown for example in figure [fig:Poly:Spectra] in section 7.8 on page . The diurnal variations of the irradiance are achieved by switching appropriate groups of lamps on and off. The optimised lamp configuration of the ceiling is presented in Figure [fig:SoSi:Lamps]. A more detailed description is given by Seckmeyer and Payer (1993), Döhring et al. (1996) and Thiel et al. (1996). Figure [fig:sunsimulator] shows a picture of a sun simulator and the schematic outline.

image

The lamps are mounted far above the cultivation area in order to obtain a homogeneous spatial distribution of the radiation. Deviations from homogeneity, indicated by the ratio of spectral irradiance at any position to that at the central position, do not exceed 20% for all wavelengths but, due to non-symmetric lamp mounting, can depend on wavelength. The horizontal distribution of in the cultivation area is shown in figure [fig:SoSi:horizontal]. The horizontal distribution of and radiation is similar and not shown here.

image

Figure [fig:SoSi:sun] shows a comparison of the spectral irradiance of the small sun simulator, provided with a double layer of Tempax glass (Schott, Germany), and a typical outdoor spectrum measured at the field station of the Helmholtz Zentrum München, Neuherberg, Germany. Measurements in the sun simulator show that the steep, realistic shape of the edge and the ratio can be simulated very close to nature. The ratio of the sun simulator is \(1:23:194\) and matches the natural conditions of \(1:25:206\) very well.

Typical irradiance data from the phytotron compared to a field measurement (on 17 April 1996 at the Helmholtz Zentrum München, Neuherberg, Germany, solar zenith angle \(\theta_s = 38\degree\)) are listed in table [tab:SoSi:values].

	Small	Large	Outdoor
	1.53	1.20	1.10
	54.9	22.7	47.2
	430	250	390

Besides the lighting, temperature, and humidity, the atmosphere in the chamber is also controlled. Typical gaseous pollutants such as ozone, nitric oxides, and combustion residuals can be introduced. The effects of carbon dioxide and hydrocarbons on plants can also be studied. Modern control technology with central monitoring ensures a safe and well-defined operation.

12.3 Filters

In UV research, optical filters are used in different contexts, in measuring instruments and in UV radiation sources. In the case of instruments, they are used in most broadband sensors to achieve the desired spectral response. They are also used in some spectrometers to improve stray light performance and remove second order artifacts.

They are used in lamp sources like laboratory solar simulators based on xenon arc lamps. In experiments with UV-B fluorescent lamps they are used to absorb short wavelength radiation in the UV-B treatment and to achieve comparable no UV-B controls. These uses are discussed in section 8.2.1 on page . In the current section, in addition to introducing the optical properties of filters, we will mainly discuss the use of filters to manipulate the spectrum of sunlight in field experiments.

12.3.1 Optical properties of filters

There are different types of filters. Band-pass filters transmit radiation in a given range of wavelengths or band, and absorb or reflect light outside this range. The band can be either wide or narrow. Cut-off filters can be either long-pass, or short-pass: respectively transmitting all radiation of longer wavelengths than a cutoff wavelength and transmitting all radiation of wavelengths shorter than a cutoff (less common). The transition is a slope that can be steep or gradual depending on the filter.

Just like any object, filters reflect, absorb or transmit radiation. These are the only three possible fates of incident radiation. reflectance (), absorptance () and transmittance () are the corresponding fractions of the incident radiation such that \(\refl + \abst + \trans = 1\) (or if expressed as percentages they add up to 100%). Absorptance should not the confused with absorbance () which is a different quantity given by \(\absb = \log \frac{1}{\trans}\). All these quantities can be defined either for a broad band, or for very narrow band (e.g. where \(\lambda\) is the wavelength). In the latter case, we can measure across wavelengths obtaining a spectrum.

When describing filters, the main property of interest is spectral transmittance. Most filters work by absorbing radiation of the unwanted wavelengths. However, some special filters work by reflecting the unwanted radiation. This is important when working with high power light sources: reflecting filters warm up much less that absorbing filters. An example of reflecting filters are the so called heat mirrors, which reflect infra-red radiation.

For light absorbing filters, like many of those made from plastic or optical glass, the optical characteristics depend not only on the material but also on its thickness. For this reason it is important to always indicate both type and thickness when describing a filter. In many cases one can take advantage of this phenomenon when planning experiments as the cutoff wavelength depends on the thickness of the filter.

Another important property of filters is whether they scatter radiation or not. When light is scattered the diffuse radiation component increases. For example, normal glass does not scatter visible radiation while opaline (white glass) does. As the proportion of diffuse radiation in PAR affects canopy photosynthesis (Okerblom, Lahti, and Smolander 1992; Urban et al. 2007, 2012; Markvart et al. 2010) one should use filters with similar scattering properties for all treatments.

Filters can be made from optical glass, gelatine, and many different synthetic organic compounds. In many cases, coloured substances are used as additives to the whole thickness of the material but sometimes a layer is merely applied to the surface or between two layers of transparent material. Glass filters tend to be very stable, but many plastics and the additives they contain react when illuminated, and deteriorate gradually when exposed to visible- and especially UV radiation. Filters made from optical glass are a more expensive than those made from plastics and tend to be readily available only in small sizes. Plastic films used as filters tend to be available in large sheets or rolls. Sometimes, liquid filters, usually aqueous solutions of chemicals, can also be useful (see Chapter 11 in Montalti et al. 2006 for spectral transmittance for several liquid filters). Sampath-Wiley and Jahnke (2011) describe a new type of liquid filter which makes it possible to obtain a realistic simulation of the solar spectrum under laboratory conditions using normal fluorescent lamps.

The engineering quality of atmospheric UV absorption by glass filter techniques is sufficient for most plant experiments. However, this method has its limitations. Strong UV exposure causes rapid ageing of borosilicate glass filters due to a physico-chemical effect known as solarisation. This originates in UV-induced changes in the oxidation state of iron contaminants present in the glass matrix. The oxidation of Fe\(^{2+}\) to Fe\(^{3+}\) is accompanied by an absorption shift to longer wavelengths within the UV range. A decrease in the transmittance of fresh glass filters occurs within their first few hours of use, followed by a more-gradual long-term decline. Hence, changes in transmission during an experiment can be largely avoided by the pre-ageing of new filters. In addition, there is a gradient of ageing with the depth of the glass. Soda-lime glass exhibits very much reduced ageing compared to borosilicate glass. The solarisation of the soda-lime glass ceases after a few hours of UV treatment and this may be due to its lower content of iron contaminants (Döhring et al. 1996).

12.3.2 Manipulating UV radiation in sunlight

When using plastic films in the field they need to be supported in a way such that they remain in place even under windy conditions and also so that rainwater does not accumulate on top of them. It is important to be aware that plastic filters modify the microclimate in several ways. If they absorb infra-red radiation, even if transparent at other wavelengths, they will alter the energy balance of plants and soil under them. This produces a greenhouse effect that increases temperatures. If they absorb they will affect photosynthesis, and even small differences in transmittance () between treatments could be important. Plants are partially protected from wind and completely shielded from rain by the filters. In some cases even plasticizer additives (different phthalates) added during plastic manufacture have been implicated in artifacts caused by the use of cellulose diacetate film in exclusion experiments, particularly when ventilation is restricted (Krizek and Mirecki 2004).

As filters have so many side effects, it is not surprising that a comparison against a no-filter control usually yields large differences, demonstrating that comparisons examining the effects of UV attenuation should always be done against controls under UV transparent films. Of course, we may want to know how similar the conditions under control filters are to natural conditions. In this case at least two types of controls are needed: a control with a UV transparent filter and a control without any filter. Figures [fig:filters:photo:field], [fig:filters:photo:field1], [fig:filters:photo:one:frame] and [fig:filters:photo:branch] show some typical setups for potted plants and branches respectively. Filters on tree branches have been used by Rousseaux et al. (2004), Kotilainen et al. (2008) and others. Filters have also been used to cover patches of natural vegetation by Phoenix et al. (2003), Robson et al. (2004) and others. Filters have been used in experiments with potted seedlings or plants by Hunt (1997), Kotilainen et al. (2009), Morales et al. (2010), and others.

image

image

image

image

The spectral transmittance of several commonly used films is shown in Figure [fig:filter:T:spectra], and the resulting filtered sunlight spectral irradiance in Figure [fig:filter:sunlight:spectra]. For near-ambient-UV controls cellulose diacetate, polythene or polytetrafluoroethylene (PTFE) are normally used. Cellulose diacetate is not a good option as its optical properties deteriorate fast, and after deterioration it absorbs more UV-B radiation (Figure [fig:filter:T:spectra:aging]), tears easily, and is affected by water. Cellulose diacetate is used for removing from the radiation emitted by lamps (see section 8.2.1.1 on page ). In contrast as there is no in solar radiation at ground level, other materials that are more durable can be used for near-ambient controls in attenuation experiments. Polythene film (types without UV absorbing additives only) is very stable, and cheap, but usually scatters light slightly more than the films used for UV-B attenuation. PTFE and related polymers (brand names Teflon, Hostaflon, etc.) are extremely stable, transmit UV radiation very well, and some types produce little scattering[^18]. Also films made from polychlorotrifluoroethylene (PCTFE) (sold under the trade name ACLAR) can be used. For UV-B attenuation experiments, polyester film (e.g. brand names Mylar, Melinex, Autostat) is the filter material most frequently used for attenuating with only a small effect on irradiance. Sometimes soda glass (normal window glass) or special acrylic plates are used instead of polyester. Polyester film produces little scattering and glass and acrylic panes almost none. For attenuation of the whole UV band, UV-absorbing theatrical gels can be used. They are called gels for historical reasons but nowadays either polyester or polycarbonate is used as base material. The most useful type is the one with code #226, available with similar specifications from several manufacturers (Rosco, Lee filters, Formatt filters, see the Appendix on page for addresses). Rosco #0 is an UV-A and UV-B transparent theatrical gel with spectral transmittance rather similar to thin cellulose diacetate. Some polyester films can also remove all UV wavelengths (e.g. brand name ‘Courtgard’). Some experimental greenhouse cladding films have cut-off wavelengths in the middle of the UV-A band and provide additional possibilities. Theatrical gels are available only in one standard thickness, but cellulose diacetate, polyester, PTFE and polythene films are available in several different thicknesses. Thickness affects both mechanical and optical properties (Figure [fig:filter:T:spectra:thickness]). The most common thicknesses used for cellulose diacetate and polyester are 100 to 125 . For polythene, thinner films can be used to minimize scattering if mechanical strength is not limiting (e.g. 50 ). As PTFE is a strong material, thin films can also be used, but one should be careful to match PAR transmittance between all treatments.

image

image

image

image

Filters can be mounted on wooden, metal or wire frames, or sometimes, especially when covering branches, on chicken wire net. It is important to carefully chose the fastening method. On wooden frames staples or thumb pins can be used, but the films tend to tear where they have been perforated so reinforcing the edges with transparent or duct tape may be necessary. It is important to use tapes that produce no toxic fumes, and to use the same amount and type of tape for all treatments (e.g. not use more tape for the films that break more easily). In harsh environments, the propensity for filters to tear or fracture can be greatly reduced by stretching them taut over the filtered plots to keep them still. This can be achieved by unrolling filters firmly attached to two cylindrical rods/poles clamped under tension to a structure anchored in the ground (Figure [fig:filters:photo:field1] on page ), . This approach also has the benefit over wooden frames of enabling a large number of filters to be carried to the field at once undamaged, so can be of use to researchers working at inaccessible field sites. When mounting the filters good ventilation should be maintained, to avoid warming up of the plants and soil under them. Careful consideration is needed when making the compromise between the UV reduction achieved by adding plastic curtains to increase the filtration of diffuse:rad (or direct:rad when the solar elevation angle is small at high latitudes), and the exacerbation of warming and further reduction in ventilation that this will cause. In long-term filtration experiments, the growth of dense vegetation under and around the filters is a common problem which can lead to large unwanted changes in the microclimate in experimental plots.

The impracticality of filtering entire trees means that branch filters are often the most acceptable substitutes when assessing the effects of solar UV radiation on trees. However their installation and maintenance involves several additional considerations beyond those attached to fixed frames. Branch filters should be located only on the top and sides of branches, while the underside should be left open for ventilation and access. A structure of chicken wire and aluminium cables can maintain the shape of the filter while minimising contact between the filter and the treated part of the branch, however the weight of the structure should be kept to a minimum to reduce shading and the risk of mechanical damage to the tree. Filters mounted on branches behave like flags in the wind, so branches on which filters are installed need to be tethered to the ground or their movement stabilised using wooden canes attached to a different large branch or limb of the tree, or preferably a combination of both these restraints, so that they do not move un-naturally or break in windy weather. By necessity in such experiments, the filtered area is relatively small and requires frequent maintenance to keep the shoots and leaves from growing into the filter or growing too far away from the filter into areas where they receive too much unfiltered and diffuse UV radiation. Particular attention should be given to the orientation of branch filters on a tree. For instance, branches and filters orientated to the north receive a very different dose of visible and UV radiation from those orientated to the south (Rousseaux et al. 2004). Comparing similar branches of the same tree under different filtration treatments is an obvious way to reduce the random variability among experimental units, but it is important to be aware that signals could potentially pass between different filtered branches and unfiltered branches which may dilute the response of a particular branch to its UV treatment. One method of controlling for communication between branches involves excising the phloem of filtered branches. However, this isolation brings other problems as the realism of the experiment is reduced, so the need to adopt this approach will depend on the aims of each individual experiment. Maintaining branch filters in situ during the autumn and winter is not always practically feasible, however when using species that have determinate growth, their environment, particularly the radiation they receive, during bud set and the formation of leaf primordia can influence subsequent growth and leaf traits during the following growing season.

Sometimes the design of an experiment requires intermediate levels of attenuation, and this can be achieved by mounting strips of one type of plastic film on a base of another type (Figure [fig:filters:photo:one:frame]). In such a system the strips are usually fastened at the edges of the wooden or wire frame on which the whole filter is attached. Unless the frames are small (less than 30 to 50 cm long sides) it may also be necessary to attach the strips in the middle with tape. In such a case the tape used must be clear to both PAR and UV radiation. The cheapest old-fashioned stationery tapes use a cellulose acetate substrate and are good for this purpose (e.g. Scotch Crystal Clear). Strips are usually 10 to 15 mm wide and if the desired effect is 50% attenuation, they are attached with a separation equal to their own width. Cutting the strips with scissors is a big task, but printers’ shops can usually cut them efficiently for a small charge. The UV irradiance under filters with strips is heterogeneous but the patches move with solar transit through the sky. This is more of a problem for measurement of irradiances than for the application of the treatments as reciprocity between irradiation time and irradiance apparently holds for such experiments (Rosa et al. 2001).

Even if we use filters that are totally opaque to UV-B, the treatment will not lead to total UV-B exclusion. That is why we use the phrase ‘UV-B attenuation’ for such treatments (Figure [fig:filter:maps]). As discussed in section 7.4 on page , solar UV radiation includes a large proportion of diffuse radiation coming from the sky, even under non-cloudy conditions. Consequently, some UV radiation penetrates under the edges of filters. It is therefore necessary to avoid locating experimental plants near the edges of the filters, and we recommend to have at least one row of border plants, forming a so-called ‘guard row’, from which no data are collected. Furthermore, it is possible to use curtains of the same filter material on the sides of the frames, but one should not close all sides as air movement is needed to avoid elevated temperatures and unnaturally still air. The filters should not be too high above the plants and irradiance should be measured at all the different positions where measured plants are located, rather than just under the centre of the frame.

12.3.3 Measuring the spectral transmittance of filters

The recommended instrument for measuring filter spectral transmittance is a spectrometer or spectrophotometer equipped with an integrating sphere. Using an integrating sphere ensures that all transmitted radiation, both non-scattered and scattered is taken into account in the measurement. For non-scattering filters, one can obtain a reasonable approximation of the total spectral transmittance with a regular spectrophotometer. In most cases it is adequate to cut a piece of the film to the size of a cuvette and insert it carefully in the cuvette holder. In some models with an open light path (e.g. HP 8453, Hewlett Packard Gmbh, Waldbronn, Germany, now Agilent) it is possible to put a larger piece of filter material in the light path outside the cuvette holder but close to it. For measuring in the UV-B band one needs an instrument using a deuterium lamp as radiation source. The spectral resolution is important, and ideally should be 1 or 2 nm.

The optical properties of all filters change with time, even if the material from which they are made is stable, because filters get scratched and dirty. Dust and pollen stick to the filter surfaces and particles attach more tenaciously to some materials than others so that the filters in some treatments may become dirtier than in others. In addition to cleaning and replacing the filters regularly, it is necessary to measure the spectral transmittance of both new and used filters. We also recommend that samples of the filters are taken both at the start and end of their use and are stored protected from light and heat at least until the results of the experiments have been published, but preferably for longer.

12.4 Manipulating UV-B in the aquatic environment

12.4.1 Incubations in the field

In contrast to the situation in terrestrial environments, field experimentation with aquatic organisms is constrained not only by the underwater light conditions but also by a suite of physical perturbations (e.g. wave action, tidal fluctuations, currents, sedimentation, etc.). In the case of benthic macrophytes, an important factor that has to be considered is their vertical distribution. Deep or constantly submerged environments impose different difficulties compared to e.g. intertidal locations, which are subjected to changing light conditions due to the action of tides. In lakes and rivers, differences in the concentration of suspended absorbing substances between the surface and deeper habitats can be important. An experimental design for the study of seaweeds normally includes different filters cutting-off and (see section 8.3 for details).

Algae can be exposed to these different radiation conditions by using self-made incubators constructed with plastic frames and nets with different diameters and forms depending of the size of the macrophyte (Figure [fig:marine:incubation]). Both the cylindrical and sheet Plexiglass incubators can be covered by appropriate UV cut-off films and located at different water depths by using ropes or buoys. E.g. as used by Gómez et al. (2005) with the red alga in the Quempillen Estuary (Chile). The cylindrical incubators can be also used in intertidal pools, allowing them to float or locating them at the bottom of a pool. In general, experiments dealing with effects of natural radiation on aquatic organisms are carried out in a range of depths between 0 and 10 m to represent the penetration depth in coastal waters.

image

The diurnal variability of solar radiation is one of the most important factors to consider when working with aquatic macrophytes in the field. In general, at temperate and cold-temperate locations, daily changes in the irradiance conditions are also exacerbated by depth, which affects the number of hours during which algae are exposed to radiation. It is well known that photosynthetic physiology changes during the day with maximum rates of photoinhibition (a mechanism of dissipation of excess energy) occurring during the highest solar irradiance at midday, while a recovery in the photosynthetic capacity is found in the afternoon (Häder et al. 1996; Figueroa et al. 1997; Gómez et al. 2004). This pattern is well generalized in macrophytes and strongly dependent on the growth depth: individuals growing at shallow depths show a more efficient photoinhibition than their deeper counterparts (Franklin and Forster 1997). Thus, experimental designs employing transplantation or incubations at different depths along the water column should consider these potential confounding effects.

Tides are also a relevant factor affecting radiation in shallow marine waters, especially for intertidal species. Fluctuations in the height of the water column caused by tidal regimes can reduce or increase considerably the available UV radiation. For example, in a study comparing two coastal systems with different tidal range, Huovinen and Gómez (2011) demonstrated that algae from fjords, with tidal ranges close to 7 m, exhibited less difference in photosynthetic light demand and susceptibility to radiation than algae from an open coast where maximal tidal variation was 2 m. Probably, this is a major factor influencing experimentation with benthic intertidal macrophytes, especially in highly dynamic systems that may also be characterized by strong wave action.

12.4.2 Incubation under artificial lamps

The effect of radiation on aquatic plants can also be studied using incubations under controlled illumination conditions (using lamps) in the laboratory (Figure [fig:lab:incubation]). Normally the sources are similar to those used for incubations of terrestrial plants (see section 8.2.1).

image

In general, experimentation on aquatic plants using small vessels under laboratory conditions is designed to test for effects of in isolated cells or small sized macrophytes. In the case of large seaweeds (e.g. kelps), space limitation restricts incubations to pieces or sample discs. In the case of seaweeds, due to their simple morphological organization, the use of cut sections still allows a realistic extrapolation of the situation in the whole thallus.

The optical characteristics and shape of the vessels containing the samples strongly determine the orientation of the sample to the UV source and the experimental setup. Often the vessels used in routine incubations are opaque to radiation (e.g. Pyrex, and many types of glass ware and polycarbonate), and thus lamps should be located above them. However, there are various transparent materials, e.g. methacrylate (Type GS-2458) or quartz glass, which permits irradiation from different sides and mimics better the natural underwater conditions. In experimental setups which cover vessels with cut-off filters, bubbles from aeration can affect the penetration and diffraction patterns within the vessels. Thus, cut-off and neutral filters should be kept at sufficient distance from the water.

Due to the heat emission of lamps, temperature within the incubation chamber can vary considerably, which should be taken into account, similar as with incubations of terrestrial plants. This factor is especially relevant for algae subject to small changes in temperature in their habitat, e.g. some deep water and polar algae. Thermoregulated water baths can minimize this problem (Figure [fig:lab:incubation]).

The use of artificial sources presents various advantages. Firstly, the manipulation of the :ratio can permit the examination of some processes that normally are masked by the prevailing high irradiances in the field, as well as simulations of different scenarios. Secondly, it is possible to standardize conditions for physiology, a situation not normally possible in the natural habitat, where unpredictable environment conditions (e.g. water motion, scattered light field) make in situ experimentation difficult.

12.5 Suitable treatments and controls

In every experiment one must include a suitable control in addition to treatments. What is a suitable control will depend on the aims of the study. In order to assess the effect of radiation, then the only difference between an treatment and the control treatment should be the irradiance of . This may seem obvious, but it is important to think about the consequence of this for the design of experiments. Many sources emit in addition to radiation, , -A and visible radiation. Common sources emit in addition to , , and visible radiation. As most fluorescent and some other UV sources emit small amounts of radiation, which is absent in sunlight, radiation must be removed by filtration. See Box [ex:lamps:controls] for descriptions of some typical experimental designs and their advantages and pitfalls.

Case

We use Q-Panel UVB-313 tubes, or Philips TL12 fluorescent tubes outdoors. These lamps are sold as broadband lamps, but they emit radiation of other wavelengths (, , and visible) in addition to (see Figure [fig:TL12:UVB313:UVA340] on page ). Our objective is to study the effect of radiation.

Design 1

We remove from the treatment by filtering the lamps with cellulose diacetate (see Figure [fig:filter:T:spectra] on page ). We add (and a very small amount of visible radiation) to the controls by having identical lamps as in the treatments but filtered with polyester film, which absorbs and . This also ensures that any effect of the lamps on the temperature of plants is similar in treated and control plants, and shading of sunlight is similar. This type of control is usually called ‘control’.

Design 2

As above, but in addition we add a second control with unenergized lamps. We can compare this control to the control to assess whether there is any side effect related to the functioning of the lamps but unrelated to radiation.

Design 3

We add a third control with no lamps or frames. By comparing this control with the control with unenergized lamps we can test for the effect of shading by lamps and supporting structures.

Design 4

We use cellulose diacetate-filtered lamps for treatments and unenergized lamps for controls. In this case the effect of is confounded with other effects of the lamps, except for shading.

Design 5

We use cellulose diacetate-filtered lamps for treatments and no lamps or frames for the controls. In this case the effect of is confounded with all other effects of using lamps.

Caveat

When discussing very small effects of treatments and the different controls there is always the risk of misinterpretation as the films used as filters are far from perfect, there is a transition zone between high absorptance and high transmittance spanning tenths of nanometres in wavelength (see Figure [fig:filter:T:spectra]). Consequently, controls will not be exposed to exactly the same irradiance of as treated plants. Usually they will be exposed to a slightly lower irradiance of plus a trace of .

Comparison

Design 1 is the simplest design that can be used to test for the effects of . Design 2, is preferable as it allows an assessment of the possible secondary effects of lamps, except for shading. Design 3, is rarely used, but allows an assessment all side-effects of lamps in addition to the effect of . Is useful in an ecological context where we are interested in assessing how much the experimental setup disrupts natural conditions. Design 4 should be avoided as it is unsuitable for testing the effects of radiation as all the different effects of the lamps are confounded with the effect of enhancement, except for shade. Design 5 is the worst possible, and should never be used.

When using filters, one should take into account the side-effects of using them. In addition to absorbing different amounts and wavelengths of radiation, filters may have effects on visible and infra-red irradiance. No filter transmits 100% of visible light. In addition filters block precipitation and wind, and may increase the temperature of the air, vegetation and soil below them. So, one should always use, as the control for assessing effects of attenuation, a transparent filter rather than no filter. See Box [ex:filters:controls] for some examples of good and bad experimental setups.

Case

We use polyester (‘Mylar’) film outdoors to attenuate radiation (see Figure [fig:filter:sunlight:spectra] on page ). Our objective is to study the effect of solar radiation.

Design 1

We attenuate radiation (but not radiation) in the \(-\)treatment by filtering sunlight with a polyester film. We use as a control a film that transmits and . This ensures that most of the effect of the films on precipitation, wind and temperature is similar for treated and control plants. Also shading of sunlight by supporting structures is similar. This type of control is usually called ‘near ambient control’.

Design 2

As above, but in addition we add a second treatment with a filter absorbing both - and radiation. We can compare this treatment to the attenuation treatment to assess whether there is an effect of radiation.

Design 3

We add a second control without filters or frames (‘ambient control’). By comparing this control with the near-ambient control we can test for the effect of filters and supporting structures.

Design 4

We attenuate radiation in our \(-\)treatment by filtering sunlight with a polyester film. We have as controls plots with no filters or supporting structures. In this case the effect of is confounded with other effects of the filters.

Caveat

When discussing very small effects of treatments and the different controls there is always the risk of misinterpretation as the films used as filters are far from perfect, there is a transition zone between high absorptance and high transmittance spanning tenths of nanometres in wavelength (see Figure [fig:filter:T:spectra] on page ). Consequently, near ambient controls will not be exposed to exactly the same irradiance of as plants under the -attenuation treated plants. Usually they will be exposed to a slightly lower irradiance of plus a trace of long wavelength .

Comparison

Design 1 is the simplest design that can be used to test for the effects of . Design 2, is preferable as it allows us to also assess the effects of solar radiation. Design 3, is not always used, but allows to assess all side-effects of filters in addition to the effect of radiation. This design is useful in an ecological context where we are interested in assessing how much our experimental setup disrupts natural conditions. Design 4 should be avoided as it is unsuitable for testing the effects of radiation as all the different effects of the filters are confounded with the effect of attenuation.

In the case of controlled environments, various combinations of lamps and filters are used to create different radiation treatments. As discussed above, if in order to test for effects of radiation, exposure should be the only difference between treatment and control conditions. For example, one can have a sun simulator, or at least a combination of lamps providing both - and visible radiation and then have different filters between the lamps and plants. The same principles as discussed in Examples [ex:lamps:controls] and [ex:filters:controls] apply to experiments in controlled experiments.

12.6

12.6.1 Recommendations for outdoor experiments

In this section we list recommendations related to the manipulation of radiation in outdoor experiments, and in section 8.6.2 list some additional recommendations on radiation manipulations inside greenhouses and controlled environments. See section 9.15 for recommendations on how to quantify radiation, section 10.9 for recommendations about growing conditions, and section 11.12 for recommendations about statistical design of experiments.

1. Make sure that the only difference between the treatments you compare is what you want to test. If possible make measurements of the environmental conditions in the different treatments and report them in your publications.

   1. In experiments with lamps avoid differences in irradiance among treatments. Such confounding differences can be caused by shade from lamps and the frames supporting them if they are kept at different heights. All treatments must have lamps that are switched on or off, covered with different filters or dimmed to different radiance values.

   2. In experiments with filters avoid difference in irradiance and scattering. Filters used to block radiation and the frames supporting them have some attenuating and scattering effect on . All treatments must have similar supporting structures, be positioned at the same height, and the filters themselves must have as similar as possible transmittance in regions of the spectrum outside the region. All filters must have as similar as possible light scattering properties, as even if irradiance is the same, differences in the proportion of diffuse light affects the growth of plants.

   3. In experiments with lamps avoid differences in temperature among treatments. Keep the distance between lamps and the top of the plants at least 0.4–0.5 m. If enclosures like open-top chambers are used, provide enough ventilation or cooling to remove the heat generated by the lamps.

   4. In experiments with filters avoid differences in temperature among treatments. The main factor to consider in this case is ventilation and distance from the film to the top of the plants. An additional factor is the transmittance of the different films to longwave infra-red radiation. The filters can affect the temperatures both during the day and at night. If possible, measure the temperature of the plants and soil, in addition to the temperature of the air.

   5. In experiments with filters avoid artifacts caused by plasticisers used in some plastic films used as filters. Ensure good ventilation and if possible avoid using cellulose diacetate films for near-ambient controls. When not using lamps there is no need to remove radiation and consequently the more stable and less toxic PTFE (Teflon) or polythene films should be used instead.

2. In experiments with lamps do not use unfiltered lamps. Most lamps also emit some that must be filtered with a cellulose diacetate filter in treated plots. Most lamps also emit , so an additional control should be included in all experiments. This control is achieved by filtering out both and radiation by means of polyester film. The smallest well designed experiment should include three treatments:

3. If the difference between near-ambient controls and true ambient conditions is of interest, then an additional true ambient control should be included to test how big is the effect of the small differences in air, soil and leaf temperatures, irradiance, air humidity and any other side-effect of the manipulations. In any case, when designing an experiment we should strive to minimize shading, alterations in temperature and ventilation.

4. In experiments with lamps, if at all possible, use modulated systems that avoid unrealistic ratios between - and irradiances. If the use of a modulated system is impossible, keep the lamps on only a few hours centred on solar noon and switch them off during cloudy weather.

5. In experiments with lamps use high frequency electronic ballasts rather than electromagnetic ballasts to drive the lamps, so as to avoid flicker in the radiation output, which can affect insect behaviour.

6. In all experiments using filters check periodically whether the spectral characteristics of the filters have changed and replace them when needed. Cellulose diacetate degrades particularly fast and, for example when used at about 0.3 m from lamps it should be replace after about 50 h of lamp irradiation. If wrapped on lamps, it should be replaced even more frequently. These times are approximate and can also be extended when using modulated systems.

7. In experiments using filters to attenuate in sunlight, even if the filters are not yet degraded or if they are made of stable materials like glass or PFTE (Teflon) the transmission characteristics will change by the accumulation of dust, dirt and pollen. If this happens, the filters should be cleaned or replaced.

12.6.2 Recommendations for experiments in greenhouses and controlled environments

Many of the recommendations in section 8.6.1 can be adapted to apply to indoor experiments. Here we list additional recommendations applicable to experiments in greenhouses and controlled experiments.

1. Keep the balance between irradiance, irradiance and as similar as possible to that in sunlight and/or vegetation canopies. Many growth chambers and growth rooms achieve relatively low irradiances of . Avoid using high doses of in such cabinets. Some chambers and rooms have sheets of polycarbonate (PC) separating the lamps from the plants, in such chambers levels are negligible. In chambers using bare lamps, irradiance is unrealistically low, but usually not equal to zero. A visible- and radiation spectral composition truly matching sunlight can be achieved only with sun simulators.

2. Unless you are specifically studying the effect of step changes in irradiance, for annual plants we recommend that the treatment starts at the time of seed germination or earlier, and for perennial plants before budburst.

3. In greenhouses, depending on the cladding material used, the and irradiances will differ from those outdoors. Irradiances of and are always somewhat lower than outdoors as the cladding materials and the structure of the greenhouse absorb part of the radiation.

4. Be aware that even though the temperature may be adequate for growing plants in a greenhouse during the winter, light and especially irradiances are much lower than in the summer, even when high pressure sodium or metal halide lamps are used to increase levels. For this reason, in experiments with lamps the balance between and can become very different to that under natural conditions.

5. Of course, if you are not interested in what happens under natural conditions but are researching the management of crops under cover, then you only have to make sure that your treatments match what can be achieved in commercial production systems.

12.7 Further reading

The following links are to directives and recommendations concerning protection from exposure.

http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2006:114:0038:0059:EN:PDF (EU directive 2006/25/EC about protecting workers from optical radiation)
http://www.hse.gov.uk/radiation/nonionising/optical.htm (guide on how to apply this directive (not approved yet by EU)
http://www.who.int/uv/sun_protection/en/ (general recommendations by WHO)

12.8 Appendix: Calculating the radiation field under an array of lamps

This section includes code for calculating the radiation field under different arrays of lamps (Figures [fig:position:tubes] and [fig:position:tubes:square]). The algorithm is that in the BASIC program in Björn and Teramura (1993) with a few changes related to the density of grid points used, and the area of the grid for which the light field is calculated. The grid area is expanded as it is also important to assess the distance at which neighbouring arrays can be located. Please, see Björn and Teramura (1993) for the details of the algorithm. The algorithm assumes that there are no reflecting surfaces near the array, so it is better suited for simulating the light field under arrays of lamps located outdoors, than for those located in controlled environments.

The functions are in the R package ‘lamps’ which is available from the handbook web site. The following example also uses package ggplot2, available from CRAN (comprehensive R archive network, at http://cran.r-project.org/ and many mirror sites). The first example (Box [box:light:field:R]) is for calculating and plotting the radiation field under and array of fluorescent tubes, using mostly the defaults. See the package documentation for the many function arguments available. The second example shows how to draw the positions of the fluorescent tubes (Box [box:tube:positions:R]).

12.9 Appendix: Suppliers of light sources and filters

In this section we provide names and web addresses to some suppliers of light sources and filters. This is certainly an incomplete list and exclusion reflects only our ignorance.

UV lamps:
http://www.lighting.philips.com/ (and many other types of lamps)
http://www.q-lab.com/ (‘Q-Panel’ and lamps)
Ballasts for fluorescent lamps
http://www.osram.com/ (‘Quicktronic’ electronic dimming ballasts)
http://www.tridonic.com/ (electronic dimming ballasts)
LEDs:
http://www.osram.com/
http://www.lighting.philips.com/
http://www.roithner-laser.com/
http://www.valoya.com/
Xenon arc lamps:
http://www.newport.com/oriel/
http://www.muller-elektronik-optik.de/
Lasers:
http://www.opotek.com (Tunnable UV lasers)
http://www.roithner-laser.com/
Filters:
http://www.formatt.co.uk/ (theatrical ‘gels’)
http://www.leefilters.com/ (theatrical ‘gels’)
http://www.macdermidautotype.com/ (‘Autostat’ polyester film)
http://www.nordbergstekniska.se (cellulose diacetate, Mylar and other films)
http://www.rosco.com/ (theatrical ‘gels’)
http://www.schott.com/ (glass filters, and special glasses)
http://www.thermoplast.fi/ (‘Autostat’, ‘Aclar’, and many other types of films.

Aphalo, P. J., R. Tegelberg, and R. Julkunen-Tiitto. 1999. “The Modulated UV-B Irradiation System at the University of Joensuu.” Biotronics 28: 109–20. http://ci.nii.ac.jp/naid/110006175827/en.

Bickford, E. D., and S. Dunn. 1972. Lighting for Plant Growth. Ohio, USA: Kent State University Press.

Björn, L. O., and A. H. Teramura. 1993. “Simulation of Daylight Ultraviolet Radiation and Effects of Ozone Depletion.” In Environmental UV Photobiology, edited by A. R. Young, L. O. Björn, J. Moan, and W. Nultsch, 41–71. New York: Plenum Press.

Caldwell, M. M., and S. D. Flint. 1994. “Lighting Considerations in Controlled Environments for Nonphotosynthetic Plant Responses to Blue and Ultraviolet Radiation.” In Proceedings of the International Lighting in Controlled Environments Workshop, NASA-CP-95-3309:113–24.

Caldwell, M. M., W. G. Gold, G. Harris, and C. W. Ashurst. 1983. “A Modulated Lamp System for Solar UV-B (280-320 Nm) Supplementation Studies in the Field.” Photochemistry and Photobiology 37: 479–85. https://doi.org/10.1111/j.1751-1097.1983.tb04503.x.

Díaz, S., C. Camilión, J. Escobar, G. Deferrari, S. Roy, K. Lacoste, S. Demers, et al. 2006. “Simulation of Ozone Depletion Using Ambient Irradiance Supplemented with UV Lamps.” Photochemistry and Photobiology 82 (4): 857–64. https://doi.org/10.1562/2005-09-28-RA-700.

Döhring, T., M. Köfferlein, S. Thiel, and H. K. Seidlitz. 1996. “Spectral Shaping of Artificial UV-B Irradiation for Vegetation Stress Research.” Journal of Plant Physiology 148: 115–19. https://doi.org/10.1016/S0176-1617(96)80302-6.

Figueroa, F. L., S. Salles, J. Aguilera, C. Jiménez, J. Mercado, B. Viñegla, A. Flores-Moya, and M. Altamirano. 1997. “Effects of Solar Radiation on Photoinhibition and Pigmentation in the Red Alga .” Marine Ecology Progress Series 151: 81–90. https://doi.org/10.3354/meps151081.

Flint, Stephan D, Ronald J Ryel, Timothy J Hudelson, and Martyn M Caldwell. 2009. “Serious Complications in Experiments in Which UV Doses Are Effected by Using Different Lamp Heights.” Journal of Photochemistry and Photobiology, B 97 (1): 48–53. https://doi.org/10.1016/j.jphotobiol.2009.07.010.

Franklin, L. A., and R. M. Forster. 1997. “The Changing Irradiance Environment: Consequences for Marine Macrophyte Physiology, Productivity and Ecology.” European Journal of Phycology 32: 207–32. https://doi.org/10.1080/09670269710001737149.

Gómez, I., F. L. Figueroa, P. Huovinen, N. Ulloa, and V. Morales. 2005. “Photosynthesis of the Red Alga Under Natural Solar Radiation in an Estuary in Southern Chile.” Aquaculture 244: 369–82. https://doi.org/h10.1016/j.aquaculture.2004.11.037.

Gómez, I., F. L. Figueroa, N. Ulloa, V. Morales, C. Lovengreen, P. Huovinen, and S. Hess. 2004. “Photosynthesis in 18 Intertidal Macroalgae from Southern Chile.” Marine Ecology Progress Series 270: 103–16. https://doi.org/10.3354/meps270103.

Häder, D.-P., M. Lebert, A. Flores, C. Jiménez, J. Mercado, S. Salles, J. Aguilera, and F. L. Figueroa. 1996. “Photosynthetic Oxygen Production and PAM Fluorescence in the Brown Alga (Linnaeus) Lamouroux Measured in the Field Under Solar Radiation.” Marine Biology 127: 61–66. https://doi.org/10.1007/BF00993644.

Holmes, M. G. 1984. “Light Sources.” In Techniques in Photomorphogenesis, edited by H. Smith and M. G. Holmes, 43–79. Academic press.

Hunt, J. E. 1997. “Ultraviolet-B Radiation and Its Effects on New Zealand Trees.” Ph.D. Dissertation, Canterbury, New Zealand: Lincoln University.

Hunt, J. E., and D. L. McNeil. 1998. “Nitrogen Status Affects UV-B Sensitivity of Cucumber.” Australian Journal of Plant Physiology 25 (1): 79–86. https://doi.org/10.1071/PP97102.

Huovinen, P., and I. Gómez. 2011. “Spectral Attenuation of Solar Radiation in Patagonian Fjords and Coastal Waters and Implications for Algal Photobiology.” Continental Shelf Research 31: 254–59. https://doi.org/10.1016/j.csr.2010.09.004.

Kalbin, Georgi, Shaoshan Li, Helena Olsman, Mikael Pettersson, Magnus Engwall, and Åke Strid. 2005. “Effects of UV-b in Biological and Chemical Systems: Equipment for Wavelength Dependence Determination.” Journal of Biochemical and Biophysical Methods 65: 1–12. https://doi.org/10.1016/j.jbbm.2005.09.001.

Kalbina, I., S. Li, G. Kalbin, L. O. Björn, and Å Strid. 2008. “Two Separate UV-b Radiation Wavelength Regions Control Expression of Different Molecular Markers in .” Functional Plant Biology 35 (3): 222–27. https://doi.org/10.1071/FP07197.

Kotilainen, Titta, Riitta Tegelberg, Riitta Julkunen-Tiitto, Anders Lindfors, and Pedro J. Aphalo. 2008. “Metabolite Specific Effects of Solar UV-A and UV-B on Alder and Birch Leaf Phenolics.” Global Change Biology 14: 1294–304. https://doi.org/10.1111/j.1365-2486.2008.01569.x.

Kotilainen, Titta, Tuulia Venäläinen, Riitta Tegelberg, Anders Lindfors, Riitta Julkunen-Tiitto, Sirkka Sutinen, Robert B. O’Hara, and Pedro J. Aphalo. 2009. “Assessment of UV Biological Spectral Weighting Functions for Phenolic Metabolites and Growth Responses in Silver Birch Seedlings.” Photochemistry and Photobiology 85: 1346–55. https://doi.org/10.1111/j.1751-1097.2009.00597.x.

Krizek, D T, and R M Mirecki. 2004. “Evidence for Phytotoxic Effects of Cellulose Acetate in UV Exclusion Studies.” Environmental and Experimental Botany 51: 33–43. https://doi.org/10.1016/S0098-8472(03)00058-3.

Langhans, R. W., and T. W. Tibbitts, eds. 1997. Plant Growth Chamber Handbook. Vol. SR–99. North Central Regional Research Publication 340. Iowa Agriculture; Home Economics Experiment Station. http://www.controlledenvironments.org/Growth_Chamber_Handbook/Plant_Growth_Chamber_Handbook.htm.

Markvart, J., E. Rosenqvist, J. M. Aaslyng, and C. -O. Ottosen. 2010. “How Is Canopy Photosynthesis and Growth of Chrysanthemums Affected by Diffuse and Direct Light?” European Journal of Horticultural Science 75 (6): 253–58.

McLeod, A. R. 1997. “Outdoor Supplementation Systems for Studies of the Effects of Increased Uv-b Radiation.” Plant Ecology 128: 78–92. https://doi.org/10.1023/A:1009794427697.

Montalti, Marco, Alberto Credi, Luca Prodi, and M. Teresa Gandolfi. 2006. Handbook of Photochemistry. 3rd ed. Boca Raton, FL, USA: CRC Press.

Morales, Luis O, Riitta Tegelberg, Mikael Brosché, Markku Keinänen, Anders Lindfors, and Pedro J Aphalo. 2010. “Effects of Solar UV-A and UV-B Radiation on Gene Expression and Phenolic Accumulation in Betula Pendula Leaves.” Tree Physiol 30: 923–34. https://doi.org/10.1093/treephys/tpq051.

Musil, C. F., L. O. Björn, M. W. J. Scourfield, and G. E. Bodeker. 2002. “How Substantial Are Ultraviolet-b Supplementation Inaccuracies in Experimental Square-Wave Delivery Systems?” Environmental and Experimental Botany 47 (1): 25–38. https://doi.org/DOI: 10.1016/S0098-8472(01)00108-3.

Newsham, K. K., A. R. McLeod, P. D. Greenslade, and B. A. Emmett. 1996. “Appropriate Controls in Outdoor UV-B Supplementation Experiments.” Global Change Biology 2: 319–24. https://doi.org/10.1111/j.1365-2486.1996.tb00083.x.

Okerblom, P., T. Lahti, and H. Smolander. 1992. “Photosynthesis of a Scots Pine Shoot - A Comparison of 2 Models of Shoot Photosynthesis in Direct and Diffuse Radiation Fields.” Tree Physiology 10 (2): 111–25. https://doi.org/10.1093/treephys/10.2.111.

Phoenix, G. K., D. Gwynn-Jones, J. A. Lee, and T. V. Callaghan. 2003. “Ecological Importance of Ambient Solar Ultraviolet Radiation to a Sub-Arctic Heath Community.” Plant Ecology 165: 263–73. https://doi.org/10.1023/A:1022276831900.

Robson, T. Matthew, Verónica A Pancotto, Carlos L Ballaré, Osvaldo E Sala, Ana L Scopel, and Martyn M Caldwell. 2004. “Reduction of Solar UV-B Mediates Changes in the Microenvironment and the Peatland Microfungal Community.” Oecologia 140: 480–90. https://doi.org/10.1007/s00442-004-1600-9.

Rosa, T. M. de la, R. Julkunen-Tiitto, T. Lehto, and P. J. Aphalo. 2001. “Secondary Metabolites and Nutrient Concentrations in Silver Birch Seedlings Under Five Levels of Daily UV-B Exposure and Two Relative Nutrient Addition Rates.” New Phytologist 150: 121–31. https://doi.org/10.1046/j.1469-8137.2001.00079.x.

Rousseaux, M. C., R. Julkunen-Tiitto, P. S. Searles, A. L. Scopel, P. J. Aphalo, and C. L. Ballaré. 2004. “Solar UV-B Radiation Affects Leaf Quality and Insect Herbivory in the Southern Beech Tree Nothofagus Antarctica.” Oecologia 138: 505–12. https://doi.org/10.1007/s00442-003-1471-5.

Saitou, Tsutomu, Yoshinobu Tachikawa, Hiroshi Kamada, Masakatsu Watanabe, and Hiroshi Harada. 1993. “Action Spectrum for Light-Induced Formation of Adventitious Shoots in Hairy Roots of Horseradish.” Planta 189: 590–92. https://doi.org/10.1007/BF00198224.

Sampath-Wiley, Priya, and Leland S Jahnke. 2011. “A New Filter That Accurately Mimics the Solar UV-B Spectrum Using Standard UV Lamps: The Photochemical Properties, Stabilization and Use of the Urate Anion Liquid Filter.” Plant Cell Environ 34: 261–69. https://doi.org/10.1111/j.1365-3040.2010.02240.x.

Seckmeyer, G., and H.- D. Payer. 1993. “A New Sunlight Simulator for Ecological Research on Plants.” Journal of Photochemistry and Photobiology B: Biology 21: 175–81. https://doi.org/10.1016/1011-1344(93)80180-H.

Tennessen, D. J., E. L. Singsaas, and T. D. Sharkey. 1994. “Light-Emitting Diodes as a Light Source for Photosynthesis Research.” Photosynthesis Research 39: 85–92. https://doi.org/10.1007/BF00027146.

Thiel, Stephan, Thorsten Döhring, Matthias Köfferlein, Andre Kosak, Peter Martin, and Harald K. Seidlitz. 1996. “A Phytotron for Plant Stress Research: How Far Can Artificial Lighting Compare to Natural Sunlight?” Journal of Plant Physiology 148 (3–4): 456–63. https://doi.org/10.1016/S0176-1617(96)80279-3.

Urban, Otmar, Dalibor Janous, Manuel Acosta, Radek Czerny, Irena Markova, Martin Navratil, Marian Pavelka, et al. 2007. “Ecophysiological Controls over the Net Ecosystem Exchange of Mountain Spruce Stand. Comparison of the Response in Direct Vs. Diffuse Solar Radiation.” Global Change Biology 13: 157–68. https://doi.org/10.1111/j.1365-2486.2006.01265.x.

Urban, Otmar, Karel Klem, Alexander Ac, Katerina Havránková, Petra Holisová, Martin Navrátil, Martina Zitová, et al. 2012. “Impact of Clear and Cloudy Sky Conditions on the Vertical Distribution of Photosynthetic Uptake Within a Spruce Canopy.” Functional Ecology 26: 46–55. https://doi.org/10.1111/j.1365-2435.2011.01934.x.

Watanabe, Masakatsu, Masaki Furuya, Yasuhiro Miyoshi, Yasunori Inoue, Isao Iwahashi, and Koichi Matsumoto. 1982. “Design and Performance of the Okazaki Large Spectrograph for Photobiological Research.” Photochemistry and Photobiology 36: 491–98. https://doi.org/10.1111/j.1751-1097.1982.tb04407.x.

WHO. 2002. “Global Solar UV Index: A Practical Guide.” World Health Organization. http://www.unep.org/PDF/Solar_Index_Guide.pdf.