Eclipse effects on field crops and marine zooplankton: the 29 March 2006 Total Solar Eclipse

The e ﬀ ects in the biosphere from the Total Solar Eclipse of 29 March 2006 were investigated in ﬁeld crops and marine zooplankton. Taking into account the decisive role of light on the photoenergetic and photoregulatory plant processes, measurements of photosynthesis and stomatal behaviour were conducted on seven important ﬁeld- 5 grown cereal and leguminous crops. A drop in photosynthetic rates, by more than a factor of 5 in some cases, was observed, and the minimum values of photosynthetic rates ranged between 3.13 and 10.13 µ mol CO 2 m − 2 s − 1 . However, since solar irradiance attenuation has not at the same time induced stomatal closure thus not blocking CO 2 uptake by plants, it is probably other endogenous factors that has been responsible 10 for the observed fall in photosynthetic rates. Field studies addressing the migratory responses of marine zooplankton (micro-zooplankton (ciliates), and meso-zooplankton) due to the rapid changes in underwater light intensity were also performed. The light intensity attenuation was simulated with the use of accurate underwater radiative transfer modeling techniques. Ciliates, responded to the rapid decrease in light intensity during 15 the eclipse adopting night-time behaviour. From the meso-zooplankton assemblage, various vertical migratory behaviours were adopted by di ﬀ erent species.


Introduction
On 29 March 2006, millions of people in the Northern Hemisphere had the opportunity to observe a total solar eclipse. In the Greek territory, the phenomenon had a maximum 20 duration of 2 min and 58 s, at the small island of Kastelorizo in the east-southern part of the country.
The effects of solar eclipses on the behaviour of living organisms have been the subject of many observations. Plant behaviour was first studied by Deen and Bruner (1933) by following stomatal movements of the gray birch during the sun eclipse of tion experienced during the phenomenon. Although fluctuating light conditions are a common feature for natural habitats (e.g. through transitional light flecks in canopies, changing cloudiness, diurnal periods of light and dusk), the dynamics of the decline of radiation during an eclipse typically differ (Schulze and Hall, 1982;Kuppers et al., 1997). Moreover, observations have shown that a drop of sap flow in a number of plants 10 was related to solar eclipses (Ladefoged, 1963;Fernandez et al., 1996;Morecroft et al., 2000).
Developmental processes, such as transient aberrations in the chromosomal structure of root meristems and a delaying seed germination, have been also ascribed to the impact of an eclipse (Sathalah et al., 1984;Kumar et al., 1984). In addition, limited 15 short-term effects on photosynthesis and evapotranspiration of crop plants, such as pigeon pea and bread wheat, were reported by Singh et al. (1992). In mature forest trees of Picea abies, Fagus sylvatica and Quercus robur photosynthesis was reduced to an extent that allowed net CO 2 evolution from leaves during an eclipse (Häberle et. al., 2001 EGU aceans (e.g. Nannocalanus minor, Scolecithrix danae and Undinula vulgaris) in the Gulf of Mexico, responded to the noontime solar eclipse of March 1970 by migrating to the surface. Moreover, according to Bright et al. (1972), the response as reflected in the number of organisms captured at the surface, was larger than the respective response to a decrease in light intensity at night. Some studies of the movements of 5 scattering layers during solar eclipses showed that layers started ascending at the beginning of the eclipse (Backus et al., 1965;Tont and Wick, 1973;Kampa, 1975), while in other cases the scattering layers did not respond at all (Franceschini et al., 1970).
The diurnal vertical migration of many zooplankters is a well known behavior, with some of the organisms migrating several hundred meters in the water column (Mar- 10 shall and Orr, 1955;Longhurst, 1976). The factors controlling vertical migration in mesopelagic species, however, have not been intensively studied. It is generally accepted that down-welling irradiance plays an important role (Banse, 1964;Ringelberg, 1995), since most vertical migrations of plancton occur at sunrise and sunset. Although, most of the above studies concern mesozooplankton there is lack of informa- 15 tion concerning other zooplankton organisms.
This study aims at investigating the effects of solar eclipses on field crops and marine zooplankton in Greece, during the 29 March 2006 total solar eclipse. Since there is no available evidence on the mechanisms involved in the effects of solar eclipses on photosynthesis and stomatal behaviour, and taking also into account the decisive role Introduction EGU chamber, was found to be 1700±150 µmol photons m −2 s −1 . Pn and g s calculation was based on the equations of Leuning and Sands (1989)  irradiance spectra were measured during 9:00-13:00 UTC at a sampling frequency of 20 s and 8 s, respectively, using two diode-array spectrometers (Kazadzis et al., 2007;Blumthaler et al., 2006). 15 The water column structure was sampled before, during, and after the eclipse, at approximately 1 h intervals, during an 8 h period (Table 1). Vertical profiles of temperature and salinity were measured with a Seabird CTD System (911 plus) assembly. Phytoplankton fluorescence (Chl-a) was recorded with a Chelsea in situ fluorometer. Water samples were taken in the euphotic 0-100 m layer, using Niskin bottles with 20 teflon-coated springs and O-rings. For microzooplankton, 7 replicate samples were taken hourly, every 10 m in the water column, whereas for mesozooplankton, 6 replicates of samples were collected hourly at descrete layers (0-10 m, 10-20 m, 20-50 m, 50-100 m, 100-200 m Five hundred ml micro-zooplankton samples were preserved with borax-buffered formalin (final concentration 2% formaldehyde) and stored at 4 • C in the dark. Before examination, samples were left to settle in their bottles in the dark at 4 • C and after 48 h, the top 400 ml of the sample was slowly siphoned off. The bottom 100 ml of the 5 sample was transferred into sedimentation chambers, allowed to settle for 24 h (Utermohl, 1951) and was finally examined by means of an inverted microscope OLYMPUS IX 70.
Meso-zooplankton was sampled by vertical hauls of a WP-2 net (200 µm mesh size). Samples were fixed immediately after collection and preserved in a 4% borax buffered-10 formaldehyde seawater solution. In the laboratory the samples were split using a folsom splitter. The first half was used to get information on the taxonomic level of order or species and counted under an OLYMPUS stereoscope.

Underwater radiative transfer modeling
Measured changes in above water light intensity and underwater composition during 15 the solar eclipse, were used as input to the extensively validated Hydrolight Radiative Transfer Program (Mobley, 1988), to estimate changes in underwater radiation fields. In the absence of in situ underwater radiometric measurements, the model calculations allowed us to estimate the changes in light levels (e.g. PAR) at specific depths in the water column, where changes in micro-and meso-zooplankton concentrations were 20 observed during the solar eclipse event. Mobley (1994) has given a detailed description of the physical assumptions and mathematical calculations in the Hydrolight Model. Tzortziou et al. (2006) recently reported very good agreement (differences less than 10%) between in situ measurements of water leaving radiance and Hydrolight results, over a wide range of conditions in the bio-optically complex estuarine waters of the Introduction

EGU
Measurements of downward surface irradiance spectra, E s (λ), and chl-a vertical profiles were used as inputs to perform the model calculations. The water was modeled by three components: pure water, pigmented particles and covarying colored dissolved organic matter (CDOM). We assumed that mineral concentrations were negligible in the very clear, oligotrophic waters at the Kastelorizo site.

5
The Pope and Fry (1997) absorption values for pure water and the seawater scattering coefficients of Morel (1974) as retabulated by Smith and Baker (1981), were used in our model simulations. Particle absorption at depth z and wavelength λ, a p (z,λ), was estimated using the Morel (1991) model and the chlorophyll-specific absorption coefficients given in Prieur and Sathyentranath (1981). Absorption by 10 CDOM,a CDOM (z,λ), was assumed to covary with particle absorption according to a CDOM (z,λ)=0.2·a p (z,440)·exp[-0.014·(λ-440)] (Mobley and Sundman, 2000). Particulate scattering was estimated using the Gordon and Morel (1983) model, while CDOM was assumed to be non-scattering. Because measurements of particulate backscattering were not available, a backscattering fraction of 0.044 was used in our calcula- 15 tions based on previous studies for clear waters (e.g. Mobley, 1994;Stramski et al., 2004). Raman scattering and CDOM and chlorophyll-a fluorescence were included in all model runs using the Hydrolight default assumptions for fluorescence efficiency and wavelength redistribution functions for fluorescence by chlorophyll (Mobley, 1994) and CDOM (Hawes, 1992). 20 Water surface roughness was estimated assuming that the slope probability distribution of the capillary waves follows the Cox and Munk (1954) wind-direction independent distribution. A wind speed of 5 m s −1 was used in the simulations based on local meteorological observations. The water column was assumed to be infinitely deep below the greatest depth of interest (ca. 250 m). Model results included underwater vertical 25 profiles of photosynthetically active radiation, PAR (in µmol phot m −2 s −1 ), estimated according to Mobley (1994) from underwater scalar irradiance, E o (z,λ), and integrated over the spectral region 400-700 nm.  Fig. 1d). On the other hand, relative humidity increased by about 10% (Fig. 1c). The combination of light "switch off" and increased humidity, together with the decreased temperature during eclipses has been previously shown to have an impact on forest trees (e.g. Steppe et al, 2002). 15 Figure 2 illustrates the diurnal course of photosynthetic rate for the cereals and legumes examined. A drop in photosynthetic rates, by more than a factor of 5 in some cases, was observed as a response to the variation in PAR during the eclipse. The minimum values of observed photosynthetic rates ranged between 3.13 (faba beans) and 10.13 µmol CO 2 m −2 s −1 (bread wheat). The relative reduction of photosynthesis 20 during the eclipse varied among the different species (  Matyssek (1986) and Häberle et al. (2001). One-way ANOVA was applied, on three replicates for each plant, in order to test the statistical significance of the variance in stomatal conductance (g s ). Table 3 shows the diurnal changes of g s for the examined species. In all species except pea, g s showed a decline at 10:00 UTC and at earlier eclipse stages, occasionally significant 5 (e.g. in bread wheat, oats, and faba beans) in comparison with the preceding values. At 10:30 UTC however a time near the totality, g s returned to values before the beginning of the eclipse, maintained throughout the rest of the day. Thus, the drop in g s at 10:00 UTC is probably not eclipse induced and an additional important argument is that the timings of minimum g s and minimum photosynthetic rate do not coincide.

EGU
From the above it is evidenced that the diurnal course of g s followed the typical pattern expected for mesophytic crop species: higher values early in the day 6:00-7:00 UTC, steadily declining within the next two-three hours and remaining stable thereafter. It is also known that morning values of g s are higher than those observed in the afternoon for similar PAR (Rochette et al., 1991). A midday temporary decrease in g s 15 is also a common phenomenon in dry environments, like the one of southern Greece. Accordingly, the drop in g s observed at 10:00 UTC simply reflects normal, "midday stomatal closure" (Miller, 1938;Meidner and Mansfield, 1968).
Stomatal movements are greatly affected by environmental conditions. Light is reported to be amongst the most important factors determining the course of stomatal 20 behaviour (Meidner and Memsfield, 1968), although many interactions with other environmental (e.g. CO 2 -concentration, vapour pressure deficit, temperature, etc.) and plant factors (plant water status, endogenous rhythms) exist to a considerable extent. It is obvious from our results that the solar eclipse has influenced photosynthesis ( Fig. 2) but on the other hand it was not an important factor directly affecting the course of g s 25 (Table 3). Consequently, since solar irradiance attenuation has not induced stomatal closure thus not blocking CO 2 uptake by plants, it is probably other endogenous factors coupled with the absence of light that has been responsible for the observed fall in photosynthetic rates.

EGU
Photosynthesis is inherently dependent on light. The three functions of the light harvesting apparatus (light capturing, energy transfer and electron transfer) are controlled by the physical and chemical characteristics of the plant pigments. Absence of light disrupts electron transport, leading to a decline of photosynthesis (Schreiber et al., 1995). This sequence of events also occurs during a solar eclipse which, to some extent, re-5 sembles the daily down and dusk periods with their gradual changes in PAR. However, in contrast with normal diurnal courses, the time span of the decline in PAR and its subsequent recovery during a total solar eclipse last for only a few minutes each, which is much shorter than the duration of dawn and dusk in the temperate climates.
3.2 The effect of eclipse on marine zooplankton 10 The solar eclipse induced effects on ecosystems has been additionally investigated in the marine environment. These effects are mainly driven by the change of underwater irradiance that has been simulated via the Hydrolight Radiative Transfer Program for the case of Kastelorizo. Figure 3 shows the change with time in downward irradiance as measured above the water surface, E s , during the eclipse event. Light inten-15 sity dropped dramatically with the onset of the eclipse, with E s at 550 nm decreasing from ∼1.35 W m −2 nm −1 at 9:35 UTC (1st contact) to below the range of the instrument's sensitivity during totality 10:53 UTC, increasing again to ∼1.16 W m −2 nm −1 at 12:10 UTC (last contact). Model estimated photosynthetically active radiation (PAR) just below the water surface was ∼2000 µmol phot m −2 s −1 just before the 1st contact, 20 while the evolution of PAR at more depths (10 and 30 m) is also shown in Fig. 3.
The water column structure at the sampled station before, during and after the eclipse is described in Fig. 4. Temperature and salinity ranged from 16.2 to 17 • C and from 39 to 39.2 psu, respectively, in the upper 200 m. Due to spring warming, the surface layer (from the surface down to 20-30 m) is warmer and saltier than the underlying 25 waters, transported in the area by the Asia Minor current (POEM group, 1992 EGU remained almost constant during the eclipse event. Chl-a concentration varied from 0.14 to 0.19 µg l −1 , with a small chlorophyll maximum between 40 and 60 m. No significant changes in chl-a vertical distribution were observed during the eclipse event.
From the vertical abundance profiles (Fig. 5, K1 and K3) it was clear that before the eclipse ciliates showed a non-random distribution and tended to accumulate at 30 m depth, where they showed a distinct peak in cell numbers (up to 212 cells l −1 ). At this 10 depth, PAR was ∼11% of its surface value (cast K2 in Fig. 6).
However, at 10:20 UTC (cast K4), one hour after the 1st contact and 30 min before totality, we observed a vertical homogenous spreading of the ciliates in the water column. At this time, surface PAR had already decreased by ∼50%. It seemed like ciliates responded to the rapid decrease in light intensity during the eclipse, adopting a night-15 time behavior. At 11:20 UTC (cast K5), almost 30 min after totality, when surface PAR was ∼35% of its initial value before the eclipse, ciliates were found in greater numbers within the first 20 m from the surface. Two hours after totality (cast K6 at 12:50 UTC), the notable peak at 30 m was re-established.
Ciliate diel cycles have been surprisingly little studied before. Jonsson (1989) re- 20 ported that ciliates often accumulate around the pycnocline, mainly due to the effect of negative geotaxis. Chemokinetic and photokinetic responses have been shown to influence the vertical distributions in some ciliates in lake environments (Finlay et al., 1987). Interactions between sinking velocity, swimming velocity, tumbling rate and shape or density asymmetry of the cell have also been suggested to explain the vertical distribution of planktonic ciliates in natural waters (Jonsson, 1989). Perhaps the ciliate community tried to stay within its optimal photo-environment during the unusually rapid decrease in light intensity due to the eclipse. Vertical migration could also 1302 Introduction EGU be governed by other factors, such as predation. Ciliates could move downwards in order to avoid predation pressure from zooplankton, accumulated near the surface layers at night. Perez and Dolan (1995) found in a study at the Western Mediterranean sea, that chlorophyll containing ciliates were mainly concentrated at the chl-a maximum depth, however heterotrophic ciliates often migrated from 20-30 m depth during 5 the day to the surface at night or in the early morning. Measurements by Stoecker et al. (1989) across Georges Bank (Northwest Atlantic), revealed that oligotrichous ciliates with chloroplasts were usually located in the upper half of the euphotic zone during the day, showing diel changes in their distribution pattern. For example, Laboea strobila was concentrated near the surface just before sunrise and early in the morning, 10 but had a subsurface maximum at noon.

Mesozooplankton
The mesozooplankton assemblage was made up 39-74% by copepods, showing an increasing relative abundance with depth. Copepodites of Clausocalanus, Paracalanus, Oithona and Ctenocalanus dominated the copepod assemblage (up to 40, 42, 19 and 15 15% respectively). Less important were Clausocalanus paululus, Ctenocalanus vanus, Oithona plumifera and Oncea media (up to 8-10%). The non-copepod zooplankton made a significant contribution to the total abundance represented, by 5 taxa: Appendicullaria, Doliolidae, Chaetognatha, Medusae and Pteropoda. Figure 7 shows the results of copepod vertical distribution at the surface down to 20 100 m. Copepods exhibited the greatest density between 10-20 m depth. At K4 10:20 UTC when surface PAR was ∼50% (Fig. 6) of that before the eclipse, copepods showed higher abundance close to the surface which can be explained by a possible migrating behaviour. Calanus and Clausocalanus copepodites seemed to be synchronized with the exogenous changes of light due to the eclipse. Individuals of 25 these species showed a vertical migratory movement towards the surface. Among the other copepodites, Paracalanus mainly localized at 0-10m increased their abundance in this layer before totality. EGU tows appeared in greater abundance in the 0-10 m layer, as well as Oncea media concentration increased at the surface layers. The other copepodites, male and female copepods displayed no significant differences in distribution and apparent response to the eclipse. Similarly, the other zooplankton taxa were distributed high in the water column  20 m), whereas the fall of light intensity resulted in a migration upwards to the surface It seems that appendicullaria (Oikopleura dioica), meduses and doliolids responded to light changes and did move up in the water column towards the surface. Pteropods that maintained themselves mainly at 0-10m increased their abundance before totality, while chaetognaths did not respond to light changes during the eclipse. Avoidance of light by zooplankton has been demonstrated by many studies with a remarked preference for the surface layers during dawn and dusk (Conover et al., 1988;Hays et al., 1974). Earlier reports for Pseudocalanus and Calanus species (Runge and Ingram, 1991;Hattori and Saito, 1997), showed that medium to large calanoids exhibited a normal diel vertical migration moving upward into the surface layer sometime 15 between 19:00 h and midnight. Calanus is a fast swimmer moving in short bursts of about 15-66 m/h (Marshall and Orr 1955). In a study of the diurnal vertical movements of chaetognaths and appendicularia, Schmidt (1973) reported that these animal groups during daytime avoid the upper layers, although around sunset and sunrise they tend to accumulate above 25 m. Consistent with our results are Sherman and Honey (1972) 20 data, reporting that chaetognaths showed little change in vertical distribution during the eclipse. However the strong responses reported by Pepita (1955) for chaetognaths and decapod larvae differed from our results, probably due to species composition of the zooplankton. The diurnal course of photosynthetic rate for the seven important field-grown cereal and leguminous crops studied here, indicated a dramatic drop during the total phase of the eclipse. The minimum values of observed photosynthetic rates varied among the different species and ranged between 3.13 (faba beans) and 10.13 µmol CO 2 m −2 s −1 5 (bread wheat). Cereals were less influenced than legumes and in particular faba beans, oats and grass pea were the most intensely affected. The latter may enable their use as potential indices for future investigation of the effects of climate changes on field crops. The instantaneous changes observed in the diurnal cycle of stomatal conductance 10 (g s ) for the examined species could not be attributed to solar eclipse. Thus, since solar irradiance attenuation has not at the same time induced stomatal closure, thus not blocking CO 2 uptake by plants, it is probably other endogenous factors that has been responsible for the observed fall in photosynthetic rates. This should be taken under serious consideration for planning of future solar eclipse related experiments. 15 The solar eclipse induced effects on the marine environment were mainly driven by the change of underwater irradiance. Model estimated photosynthetically active radiation (PAR) just below the water surface was ∼2000 µmol phot m −2 s −1 just before the 1st contact. Temperature and salinity remained almost constant during the eclipse event. Chl-a concentration varied from 0.14 to 0.19 µg l −1 . No significant changes in 20 chl-a vertical distribution were observed during the eclipse event.
The behavior of marine populations has been investigated, separately for microzooplankton (ciliates) and meso-zooplankton. Ciliates, tending to accumulate at 30 m depth, responded to the rapid decrease in light intensity during the eclipse and adopting night-time behaviour showed a vertical homogenous spreading in the water column. EGU media increased their abundance in the surface layer before totality. The other copepodites, male and female copepods displayed no significant differences in distribution and apparent response to the eclipse. Similarly, the other zooplankton taxa such as appendicullaria (Oikopleura dioica), meduses and doliolids responded to light changes moving up towards the surface.

5
Overall, given the biodiversity of the Greek territory we seized the opportunity of the March 2006 Total Solar Eclipse to investigate the sensitivity of various species in the biosphere. The different responses ascertained between various species both in field crops and marine zoo-plankton, provide an early alert that future climate changes influencing the amount of radiation that reaches the earth's surface, may disturb the