Characterizing the purple Earth: Modelling the globally-integrated spectral variability of the Archean Earth
E. Sanromá, E. Pallé, M. N. Parenteau, N. Y. Kiang, A. M. Gutiérrez-Navarro, R. López, P. Montañés-Rodríguez
aa r X i v : . [ a s t r o - ph . E P ] N ov Not to appear in Nonlearned J., 45.
Characterizing the purple Earth: Modelling theglobally-integrated spectral variability of the Archean Earth
E. Sanrom´a , , E. Pall´e , , M. N. Parenteau , , N. Y. Kiang , A. M. Guti´errez-Navarro , R.L´opez , and P. Monta˜n´es-Rodr´ıguez , Instituto de Astrof´ısica de Canarias (IAC), V´ıa L´actea s/n 38200, La Laguna, Spain [email protected]
ABSTRACT
The ongoing searches for exoplanetary systems have revealed a wealth of plan-ets with diverse physical properties. Planets even smaller than the Earth havealready been detected, and the efforts of future missions are placed on the discov-ery, and perhaps characterization, of small rocky exoplanets within the habitablezone of their stars. Clearly what we know about our planet will be our guide-line for the characterization of such planets. But the Earth has been inhabitedfor at least 3.8 Ga, and its appearance has changed with time. Here, we havestudied the Earth during the Archean eon, 3.0 Ga ago. At that time one of themore widespread life forms on the planet were purple bacteria. These bacteriaare photosynthetic microorganisms and can inhabit both aquatic and terrestrialenvironments. Here, we used a radiative transfer model to simulate the visi-ble and near-IR radiation reflected by our planet, taking into account severalscenarios regarding the possible distribution of purple bacteria over continentsand oceans. We find that purple bacteria have a reflectance spectrum which hasa strong reflectivity increase, similar to the red edge of leafy plants, althoughshifted redwards. This feature produces a detectable signal in the disk-averaged Departamento de Astrof´ısica, Universidad de La Laguna, Spain NASA Ames Research Center, Exobiology Branch, Mountain View, California 94035, USA SETI Institute, Mountain View, California 94035, USA NASA Goddard Institute for Space Studies, New York, NY 10025, USA Department of Microbiology, Faculty of Biology, University of La Laguna, Spain
Subject headings:
Astrobiology — Earth — Planets and satellites: atmospheres,surfaces — Radiative transfer
1. INTRODUCTION
Since the discovery of the first exoplanet orbiting a main sequence star in 1995 (Mayor & Queloz1995), nearly 950 extrasolar planets have been detected, and more than 2000 potential planetcandidates from the Kepler mission are waiting to be confirmed (Batalha et al. 2013). In thelast few years, we have been able to discover several planets in the super-Earth mass range(e.g. Udry et al. 2007; Charbonneau et al. 2009; Pepe et al. 2011; Borucki et al. 2012), someof them lying within, or close to, the habitable zone of their stars (e.g. Borucki et al. 2012;Barclay et al. 2013; Anglada-Escud´e et al. 2013). Even some Earth and Moon-sized plan-ets have been recently announced (Fressin et al. 2012; Muirhead et al. 2012; Gilliland et al.2013; Borucki et al. 2013), and this number is expected to increase in the future. In fact,early statistics have pointed out that around 62% of the Milky Way’s stars might host asuper-Earth (Cassan et al. 2012), while studies from NASA’s Kepler mission indicate thatabout 16.5% of stars have at least one Earth-size planet with orbital periods up to 85 days(Fressin et al. 2013). To be prepared for the characterization of future discovered exoearths,first we must take a look to our own solar system and its planets.Without a doubt, the possibility of finding life will drive the characterization of rockyexoplanets over the coming decades. Earth is the only planet where life is known to exist;thus observations of our planet will be a key instrument for characterization and the searchfor life elsewhere. However, even if we discovered a second Earth, it is very unlikely thatit would present a stage of evolution similar to the present-day Earth. The Earth has beenfar from static since its formation about 4.5 Ga ago. On the contrary, during this time, ithas undergone multiple changes in its atmospheric composition, its temperature structure,its continental distribution, and even changes in the forms of life that inhabit it. All thesechanges have affected the global properties of Earth as seen from an astronomical distance.Thus, it is of interest not only to characterize the observables of the Earth as it is today, butalso at different epochs (Kaltenegger et al. 2007; Sanrom´a & Pall´e 2012). 3 –Aiming at determining how Earth would look like to a hypothetical distant observer,several studies have been carried out over the last years. Earthshine observations have beenone of the observational approaches used for this purpose, providing a tool to study thespectrum of Earth in the visible (e.g. Goode et al. 2001; Woolf et al. 2002; Qiu et al. 2003;Pall´e et al. 2003, 2004), and also in the near-infrared (Turnbull et al. 2006; Pall´e et al. 2009)and in the near-UV (Hamdani et al. 2006). Sterzik et al. (2012) studied the use of the linearpolarization content of the earthshine to detect clouds and biosignatures.Another possible approach is through analysis of Earth’s observations obtained fromremote-sensing platforms (e.g., Tinetti et al. 2006a; Cowan et al. 2011; Robinson et al. 2011;Fujii et al. 2013). Cowan et al. (2009) performed principal components analysis in order todetermine if it was possible to identify surface features such as oceans and continents fromthe EPOXI data. They were able to reconstruct a longitudinally averaged map of the Earth’surface. Crow et al. (2011) were able to categorize Earth among the planets of the solarsystem by using visible colors.Also using EPOXI data, Kawahara & Fujii (2010, 2011) and Fujii & Kawahara (2012)proposed an inversion technique which allowed them to sketch two-dimensional planetaryalbedo maps from annual variations of the disk-integrated scattered light.Some authors have attempted to detect the vegetation red edge through earthshine mea-surements (Arnold et al. 2002; Woolf et al. 2002; Seager et al. 2005; Monta˜n´es-Rodr´ıguez et al.2006; Hamdani et al. 2006), and also using simulations (Tinetti et al. 2006a,b; Monta˜n´es-Rodr´ıguez et al.2006). The red edge is characterized by strong absorption in the visible part of the spectrumdue to the presence of chlorophyll, which contrasts with a sharp increase in reflectance inthe NIR due to scattering from the refractive index difference between cell walls and thesurrounding media. This particular signature of vegetation has been proposed as a possiblebiomarker in Earth-like planets (e.g., Seager et al. 2005; Monta˜n´es-Rodr´ıguez et al. 2006;Kiang et al. 2007a). The possibility of detecting hypothetical alien vegetation on terrestrialplanets has also been studied. Tinetti et al. 2006c explored the detectability of exovegetationin a planet orbiting an M star, on which vegetation photosynthetic pigments might show ashifted red edge signature. Kiang et al. 2007b conjectured further about rules for pigmentsadaptations to other stellar types.In this paper we concentrate on the Archean eon (3.8-2.5 Ga ago), particularly on theEarth at 3.0 Ga ago when the Sun was about 80% as bright as it is today (Gough 1981;Bahcall et al. 2001), and the atmospheric composition of our planet was completely differentto that of present day. At this time, the Earth’s atmosphere was likely dominated by N , CO , and water vapor (e.g., Walker 1977; Pinto et al. 1980; Kasting 1993; Kasting & Brown1998), with little or no free oxygen. Methane might have also been present as well, help- 4 –ing in the compensation for the reduced solar luminosity (e.g., Kiehl & Dickinson 1987;Pavlov et al. 2000; Haqq-Misra et al. 2008).While controversial, the first evidence of life is at 3.8 Ga in isotopically light graphiteinclusions in apatite from Greenland (Mojzsis et al. 1996), and most likely it was non-photosynthetic, although this is still a subject of debate. The earliest photosynthetic lifewas probably anoxygenic bacteria like purple bacteria (Xiong et al. 2000; Olson 2006), uti-lizing reductants such as H or H S instead of water. The Archean biosphere has beenproposed to be a mix of anoxygenic phototrophs and chemotrophs such as sulfate-reducingbacteria, methanogens, and other anaerobes (Kharecha et al. 2005). The former performphotosynthesis requiring a band gap energy smaller than that needed to split water, suchthat the photosynthetically active radiation relevant for anoxygenic photosynthetic bacteriacan extend into the near-infrared to as long as ∼
2. MODEL DESCRIPTION
For our calculations, we make use of a line-by-line radiative transfer algorithm, based onthe DISORT (Discrete Ordinates Radiative Transfer Program for a Multi-Layered Plane-Parallel Medium) code (Stammnes et al. 1988), in order to derive disk-integrated spectra ofthe early Earth. This radiative transfer model (RTM) utilizes spectral albedo of differentsurface types, profiles of atmospheric composition and temperature, cloudiness information,and viewing and illumination angles as input data for the calculations. Only a single angleof incidence and ten angles of reflection can be used for each model run. The RTM usedhere is basically an extension of the RTM for transits described in Garc´ıa Mu˜noz & Pall´e(2011) and Garc´ıa Mu˜noz et al. (2012) to a viewing geometry for which the light reachingthe observer has been reflected at the planet. With this RTM, we have generated a databaseof about 160 one-dimensional synthetic spectra that cover a wide range of illumination and ftp://climate1.gsfc.nasa.gov , Temperature and atmospheric composition profiles for the Archean were calculated by R.Ramirez (private communication). A 1-D radiative-convective climate model, first developedby Kasting et al. (1984) and recently substantially updated by Kopparapu et al. (2013) andRamirez et al. (2013), was used to calculate the atmospheric properties.These atmospheric profiles consist of 1% CO , 0.2% CH , according to Kaltenegger et al.(2007), being the remaining gas N . For the relative humidity, a Manabe-Wetherald profilewas used (Manabe & Wetherald 1967). For the calculation of these profiles the Sun wasassumed to have ∼
79% of its present-day luminosity, as we aimed to simulate the Earth 3.0Ga ago. The temperature and mixing ratios profiles of these species are shown in Figure 1.In our model, we divided the atmosphere into 33 uneven layers, which go from theboundary layer to 100 Km height, with the spacing between layers of 1 km near the bottomof the atmosphere, and 5 km or more above 25 km height. As the original atmospheric profileswere prescribed in layers up to 70 Km, we assumed the same constant values between 70and 100 Km.
To perform the disk-averaged spectra of the ancient Earth, we have considered fourdifferent planetary surfaces: water, desert, water with purple bacteria in suspension, andpurple bacteria in microbial mats. There is some discussion in the literature whether purplenon-sulphur bacteria could have colonized extended areas of soil, and whether such a signalwould be remotely detectable. Here we have assumed as the most likely scenario, that these 6 –microbial mats are located in marine intertidal environments. The wavelength-dependentsurface reflectivity of the two first surface types were derived from the ASTER SpectralLibrary and the USGS Digital Spectral Library . Figure 2 shows the spectral albedo ofthese surface types. The wavelength-dependent albedo of surfaces involving purple bacteriawere obtained as described in Section 3 (Figure 3). The movement of Earth’s tectonic plates has caused the formation and the break up ofcontinents over Earth’s history, including the formation of supercontinents. The orientationof Earth’s earliest continents is still unknown, although it is believed that the fraction ofthe surface covered by continents during the Archean was smaller than present day (e.g.Goodwin 1981; Belousova et al. 2010; Dhuime et al. 2012). Earth might have been almostentirely covered by water with some small continents. Hence, due to the complete lack ofinformation about the continental distribution of the Earth 3.0 Ga ago, we decided to usethat of the Earth during the Late Cambrian (500Ma ago, see Figure 4) and at present day,as two possible characteristics examples. The continental distribution of these two epochshas been taken from Ron Blakey’s website , where these surface maps are available online.The Earth geologic information has been regridded into the 64x32 pixel grid used by ourmodel.We have also defined the coastal areas of these maps. These coastal zones were deter-mined as those land grid cells that have adjacent ocean grid cells and those ocean cells thathave adjacent land grid cells. For these simulations we have taken into account three different cloud layers: low (1000-680 mb), mid (680-440 mb) and high cloud (440-30 mb). The optical properties of eachcloud type, wavelength-dependent scattering and absorption coefficients, and the asymmetryparameter, were taken from the Optical Properties of Aerosols and Clouds (OPAC) database (Hess et al. 1998). We have considered physical cloud thicknesses of 1 km, and we have http://speclib.jpl.nasa.gov http://speclab.cr.usgs.gov/spectral-lib.html http://jan.ucc.nau.edu
3. PURPLE BACTERIA’S REFLECTANCE SPECTRA
To obtain the reflectance spectra of purple bacteria, we used pure cultures of
Rhodobactersphaeroides
ATCC 49419, a purple non-sulfur bacterium growing as a suspension of cells inliquid media. This type of phototroph exhibits diverse metabolic abilities, allowing survival ina wide range of dynamic environmental conditions. Purple non-sulfurs can grow aerobicallyin the dark as a chemoheterotroph, and anaerobically in the light using hydrogen and organiccompounds as electron donors for photosynthesis.We made two different measurements to retrieve the reflectance spectrum of these bacte-ria. In the first one, the reflectance of a liquid pure culture of purple bacteria was measuredusing a UV/VIS/NIR spectrophotometer (VARIAN, CARY 5E). The measurements weretaken from 0.3 to 2.5 µm .In the second experiment, we used a LiCor LI1800 spectroradiometer with a remotecosine receptor that was positioned 5 cm above the culture. These measurements cover the0.35-1.1 µm spectral range. This culture of purple bacteria was in a petri dish sitting on topof a piece of white paper. In order to calculate the spectral albedo, the spectral irradianceof the Sun was also measured.Both sets of retrieved reflectances spectra were merged into one, covering from thevisible to the near-infrared (Figure 3), using as a reference the second experiment, whichhad a higher signal to noise ratio. In order to absolutely calibrate the reflectance spectra 8 –of the bacteria, we also measured the reflectance spectra of a set of known leaves. Thecomparison of our measured leaf reflectance spectra with those tabulated in the ASTERlibrary gave us a measure of our reflected flux, which we then applied to the reflectancespectra of the purple bacteria. This way we transformed the ratio scale into a reflectancescale.Figure 3 shows the reflectance spectrum of the purple non-sulfur bacterium Rhodobactersphaeroides . The photosynthetic pigments of these bacteria are bacteriochlorophyll a ester-ified with phytol, and carotenoids of the spirilloxanthin series. Due to the combination ofthese two pigments, living cells of this species show absorption features at 375, 468, 493,520-545, 589, 802, 860-875 nm (Imhoff 2005). Some of these features can be detected, andare marked, in Figure 3. Moreover, as the purple non-sulfur bacteria cultures used were red,the reflectance increase between ∼
600 and 700 nm is due to the reddish light reflected backfrom the cultures’ cells. The most noticeable feature of this spectrum is the sharp increasein reflectivity from approximately 0.9 µm to 1.1 µm , and the equally strong decrease from1.3 µm to 1.4 µm . Information about the physical nature of the absorption features at λ > µ m is not found in the literature. Starting at 1.4 µm and redwards, the spectrum doesnot show any measurable features, and the overall albedo value is probably that of water,made slightly more reflective due to the presence of bacteria in suspension that lower itstransmissivity. The overly-featureless variability is probably due to the low sensitivity of theinstrument used at these wavelengths.The bacteria concentration in our sample culture was very high ( ∼ cells/ml), proba-bly much more than the typical concentrations that would be found in seawater. Thus, whenwe modeled purple bacteria in open oceans, we used a combination of pure seawater andour bacteria(+water) spectra, weighted in varying percentages, to simulate different bacte-ria concentrations. Throughout the rest of the paper we refer to percentage dissolution (forexample a dissolution of 10% means 9 parts of seawater and 1 part of our culture, equivalentto concentrations of the order of ∼ cells/ml). Note that as we used liquid cultures tomeasure the reflectance of purple bacteria, the effect of the transmittance of water is alreadyincluded in the spectrum.
4. THE SPECTRA OF THE EARLY EARTH
In order to determine if it would be possible to discern the presence of life forms such aspurple bacteria in the spectra of an extrasolar planet, we have simulated the disk-integratedspectra of the ancient Earth, taking into account different continental distribution, cloudcoverage, and several abundance scenarios which go from a planet where purple bacteria 9 –have colonized both oceans and continents, to a planet where purple bacteria are only foundin oceanic coastal areas in low concentrations.In all the cases studied in this manuscript, both the observer and the Sun are locatedover the planet’s equator in such a way that the observer is looking at a half-illuminatedplanetary disk, i.e., at a phase angle of 90 ◦ . This is the most relevant geometry for studyingexoplanets, since the maximum angular separation of an extrasolar planet from its parentstar along its orbit, takes place at phase 90 ◦ , as defined from the observer’s position.In order to see the effect of considering different atmospheric compositions in the disk-averaged spectra of a lifeless planet, Figure 5 compares the spectrum of a planet with aCO -CH rich atmosphere and no oxygen (black), with the spectrum of a planet with Earth’scurrent atmospheric composition (blue). In both cases, the continental distribution is that ofEarth 500 Ma ago and continents are assumed to be completely covered by deserts. Figure5 (black lines) show strong absorption in the NIR part of the spectrum due to the increasedlevels of CO and CH , while in the visible region, the most noticeable difference with Earth’scurrent atmosphere is the lack of the absorption features typical of O and O . The visibility of surface inhomogeneities, such as continents or surface types, on a planetis naturally very dependent on the frequency of cloud formation. The top panels of Figure6 shows synthetic disk-averaged spectra of the ancient Earth over a course of a day, onespectrum every two hours, covering the spectral range between 0.4 and 2.5 µm , for both acloud free (left) and a cloudy atmosphere (right). Cloud cover is assumed to be 50%. Inthe top panels, the continental distribution corresponds to that of Earth 500Ma ago, andcontinents are dry lands (deserts), while the coastal points closest to land are completelycovered by purple bacteria, and the coastal points closest to ocean are a mixture of 10%purple bacteria and 90% of water. We have chosen this particular scenario since bacteria areexpected to be found where nutrients are more abundant, like in shallow waters or coastalzones.It is expected that as continents come in and out of the field of view, the light reflectedback by the Earth changes, with these changes more drastic for the cloud-free cases. Thereflectance is higher when continents occupy most of the observable half-disk (at 8:00-10:00UT, when the percentage of continental surface in the sunlit area of the planet visible fromour observer’s location is ∼ ∼ The land-mass distribution of the Earth 500 Ma ago consisted of a large continentalmass, mostly located in the southern hemisphere. There is also a further group of threelarge islands also in the southern hemisphere. On the other hand, the present-day Earth hastwo major continental land masses spreading over the north and southern hemispheres. Inorder to estimate the effect of considering different land-mass distributions, Figure 6 bottompanels show the same spectra as the top panels, but here the input continental distributioncorresponds to that of the present-day Earth. Although these two continental distributionsare considerably different, the disk-integrated spectra of both cases are quite similar. Therotational variability is also similar in both cases, with a comparable amplitude, only slightlysmaller for the present day Earth when clouds are considered.
Figure 7 (left) shows disk-integrated spectra obtained for the early Earth where conti-nents are completely covered by purple bacterial mats, and oceans are a mixture of waterand bacteria, 90% and 10%, respectively. Both a cloud-free (red) and a 50% cloudy case(black) are shown. The continental distribution used here is that of Earth 500Ma ago.Whether purple bacteria would have been able to colonize the continental surfaces duringthe Achean is still unsolved. Due to the lack of O in the early Earth’s atmosphere, andtherefore the lack of a ozone layer, harmful radiation did probably reach the Earth’s surface,and purple bacteria might have suffered from DNA damage. Studies of modern microbes,however, suggest that their photoprotective pigments, that absorb in the blue and UV (e.g.,carotenoids), are sufficient to have allowed for their survival in terrestrial and shallow waterenvironments on the early Earth (Cockell 1998). Moreover, some purple bacteria have beenshown to use reduced iron (Fe(II)) for photosynthesis, and Pierson et al. (1993) pointed outthat the oxidized iron products of this type of photosynthesis could have provided substantialprotection from UV radiation for surface-dwelling phototrophs prior to the development of 11 –an ozone shield. Thus, the presence of bacterial mats in continental areas, while not beingour most likely scenario, cannot be ruled out.In Figure 7 (left) we only show the spectra at the time when the continental presence isat maximum, i.e., at 08:00 UT, when the percentage of land over ocean that is illuminatedand visible at the same time is about 50%. The same is shown in the right panel of thisFigure, but here considering that purple bacteria do not exist over continents, and are onlyfound in the oceans. For comparison, Figure 5 (black lines) shows the spectra of a planetwithout bacteria in either water or land.When the amount of purple bacteria is high, 100% over continents and 10% diluted inthe water, the presence of purple bacteria on the early Earth produces a strong feature, asteep increase in reflectance around 1.0 µm in its disk-integrated spectrum (Figure 7 left).When clouds are included in the model, this signature is naturally significantly diluted.However, it is still easily detectable by simple inspection of the disk-average spectra.When considering a more realistic case where purple bacteria are only found in coastalareas (Figure 6), the increase in reflectivity around 1.0 µm due to these bacteria is readilyseen in the cloud-free case and is still detectable in the cloudy case.For a planet with marine bacterial life only, and bare continental surfaces, this spectralfeature produced by purple bacteria becomes harder to discriminate in the spectra (Figure 7right), being practically undetectable in the cloudy case. Here we have considered a mixtureof 10% bacteria and 90% water. In fact, if one compares the cloudy spectrum of this case witha case where there are no bacteria, neither over continents nor in the water (Figure 5, blacklines) it is impossible to distinguish between them by simple exploration of the spectra. Thus,to estimate the marine-only purple bacteria detectability, we have considered several bacteriaconcentrations in oceans: 20, 30, 40, and 50%. Table 1 shows the slope of the straight linethat connects the averaged planetary radiance in the 0.745-0.770 µm and 1.010-1.034 µm spectral intervals. The data are given as a function of bacteria concentration in water for botha cloud-free and a cloudy atmosphere. The slope between these two spectral regions, free ofatmospheric absorption features, is mainly influenced by the contribution of purple bacteriato the globally-integrated spectrum of Earth. Thus, this slope can be used as a measure ofthe strength of the purple bacteria signal, similar to what has been previously done withto quantify the vegetation’s red edge (Monta˜n´es-Rodr´ıguez et al. 2006). Table 1 shows howincreasing the bacteria concentration in water from 10% to 50% monotically increases thisslope. Although not shown here, the identification of the presence of purple bacteria bysimple inspection of these spectra in the cloudy atmosphere case is almost impossible forbacteria concentrations in oceans lower than 30%. 12 –Finally, although it is a very improbable scenario, we have run a comparison test toestimate the purple bacteria detectability. Figure 8 shows disk-averaged spectra of Earthconsidering the present-day continental and cloud distribution, and the early Earth atmo-sphere. Cloud distribution was taken as the 1984-2006 climatology of ISCCP cloudinessdata. Here we have assumed that continents are completely deserts and we have used the2012 annual mean ocean chlorophyll a content map from the SeaWiFS project as a proxyfor the distribution of purple bacteria. Thus, we have considered that the ocean latitudinalrange 90 ◦ (cid:21) ◦ , both Nord and South, and the -10 ◦ (cid:21) ◦ latitudinal range are a mixture of90% bacteria and 10% of water. Coastal zones are also populated with purple bacteria, andthe rest of the ocean is a mixture of 10% bacteria and 90% water. The figure shows thespectra at 8:00 UT (when oceans dominate the field of view; blue lines) and at 18:00 UT(when continents dominate the field of view; black lines). As in previous cases in a cloud-freeatmosphere purple bacteria are readily detectable. For a present-day cloud amount (roughly60%) the detectability is not so obvious, and a high signal-to-noise ratio spectra of the planetwould be needed.
5. Photometric Light-Curves
In the case of a terrestrial planet in the habitable zone of a G, F or K star, obtainingin the near-future even a low resolution spectra might be a difficult task to perform (e.g.,Pall´e et al. 2011; Rugheimer et al. 2013; Hedelt et al. 2013). Photometric observations on afew filters might actually be a more realistic possibility, even if the data are obtained viaspectroscopic observations, as is the case of Hubble and Spitzer observations nowadays (e.g.,Tinetti et al. 2007; Swain et al. 2008; D´esert et al. 2011; Pont et al. 2013). The fact that aplanet shows photometric variability along one rotation already speaks about the presenceof inhomogeneities in its surface or atmosphere (Ford et al. 2001). With sufficiently accuratetime series it might be possible to distinguish the presence of cloudiness and continents basedon this variability (Pall´e et al. 2008). Moreover, the rotational photometric variability as afunction of wavelength can reveal information about the major wide-spread composition ofcontinental surfaces.Thus, following Sanrom´a et al. (2013), we have convolved our modeled disk-integratedspectra against standard astronomical filters, both visible and near-infrared, namely B, V,R, I, z, J, H, and K. Figure 9 shows the photometric daily variations in each photometric http://isccp.giss.nasa.gov http://oceancolor.gsfc.nasa.gov/
13 –filter of the disk-averaged reflected light, for a cloud-free atmosphere (red) and for a cloudyone (black).For these simulations, we have used a conservative scenario: the continental distributionassumed is that of Earth 500 Ma ago, continents are bare desert, coastal land areas arecovered by purple bacteria mats, oceanic coastal areas are a mixture of 10% purple bacteriaand 90% of water, and open ocean areas are only water.The light curves shown in Figure 9 all have a similar shape; the peak in brightness takesplace in each of the photometric filters when continents dominate the field of view, between8:00 and 10:00 UT, while minimum values of reflectance occur when oceans occupy most ofthe view, approximately at 4:00 and at 16:00 UT.The cloud-free case shows a larger variability in the diurnal light curves in each filterand a considerable rise in brightness that increases monotically redwards. A similar resultis found when clouds are added to the model, although it shows much less variability owingto the effect of clouds on the spectra.Figure 10 (top) shows the amplitude of albedo variations of the two scenarios shown inFigure 9 as a function of the different photometric filters. As found before, the variabilityof the light reflected back by the planet in both the cloud-free and cloudy cases increasestoward the red, with this increase much more dramatic in the cloud-free case than in thecloudy one. The amplitude of these variations goes from a few percents to around 180% and50%, for a cloud-free and a cloudy atmosphere, respectively.Figure 10 (bottom) shows the expected photometric variability for a modern Earth withdifferent surface compositions (data from Sanrom´a et al. 2013), together with the ArcheanEarth results for the cloudy case (from top panel). The different color lines show the am-plitude of the albedo variability for a cloudy atmosphere with an atmospheric compositionsimilar to that of present Earth, for a planet totally desert (black line), completely coveredby vegetation (green line), by microbial mats (red and dark blue lines), and for a planetwhere continents are a mixture of microbial mats and deserts (yellow line), and microbialmats and vegetation (cyan line). In all cases the atmospheric composition in the modelsused to generate the data is that of the modern Earth. The purple line shows the albedovariability of a cloudy planet with an atmospheric composition similar to that of the earlyEarth (3.0 Ga ago) which continents are covered by deserts and coasts are a mixture ofpurple bacteria and water.In Sanrom´a et al. (2013) we concluded that it would be possible to discriminate be-tween vegetated continents, large extension of microbial mats and bare continental surfacesby comparing the amplitude of the albedo change, along the course of a day, taken in different 14 –photometric filters. Here, the purple curve suggests that it would also be possible to dis-criminate a purple Archean-like planet among the other scenarios studied in Sanrom´a et al.(2013). In contrast to the other scenarios, in the visible portion of the spectrum, the am-plitude of albedo change of the Archean Earth increases monotically from the B filter tothe z filter, an effect due to the combination of the lack of oxygen in the atmosphere andthe reflectance spectra of the purple bacteria. When one moves towards the red part of thespectrum, this increase is sharper and the curve peaks at the H filter to decrease redwardsin K. However, if one only takes into account the near-IR filters, it would be difficult todiscriminate between a desert planet and the purple scenario.
6. CONCLUSIONS
In this paper we have presented disk-integrated spectra of the Earth during the Archeaneon with the aim of studying how a different atmospheric composition and the presence ofearly life forms, such as purple bacteria, may have affected the way our planet looked fromafar.As one of the inputs of our models is the reflectance spectrum of different surface types,we carried out two different experiments in order to retrieve the spectral albedo of thesebacteria in the spectral rage 0.3-2.5 µ m. We found that purple bacteria show a reflectancespectrum with a sharp increase in reflectivity similar to the red edge of leafy plants, butshifted redwards.In order to determine if it would be possible to detect such a biomarker, we haveconsidered three different scenarios: one where purple bacteria have colonized the wholeplanet, both water and continents, other where purple bacteria are only found in oceans,and finally a scenario where these bacteria are found only in coastal regions. We have takeninto account the effect of clouds in our models finding that the inclusion of clouds resultsin the increase of the reflectivity of the planet, reduces drastically the albedo variabilityover the course of a day, and makes more difficult the identification of surface types underclouds. Changing the continental distribution does not seem to have a high impact in theglobally-averaged spectral variations.We find that when the amount of purple bacteria is high, they can be readily detectedin disk-averaged spectra, both in cloud-free and in cloudy atmospheres. While if purplebacteria are only found in oceans, their spectral feature becomes nearly undetectable in thecloudy case. When considering a more realistic scenario where purple bacteria are found incoasts, their presence can be detectable in the cloud-free case, and even in the cloudy case, 15 –although the signal is smaller.Finally, by convolving these simulated spectra against standard astronomical filters, weconclude that using photometric observations in different filters might allow us to discrimi-nate between a present day Earth with continental surface covered by deserts, vegetation ormicrobial mats, from an Archean Earth where purple bacteria have colonized large extensionsof the planet. Acknowledgments
We would like to thank Ant´ıgona Segura and Ramses Ramirez for kindly providing uswith the atmospheric profiles of the early Earth.
REFERENCES
Anglada-Escud´e, G., Tuomi, M., Gerlach, E., et al. 2013, A&A, 556, A126Arnold, L., Gillet, S., Lardi`ere, O., Riaud, P., & Schneider, J. 2002, A&A, 392, 231Bahcall, J. N., Pinsonneault, M. H., & Basu, S. 2001, ApJ, 555, 990Barclay, T., Burke, C. J., Howell, S. B., et al. 2013, ApJ, 768, 101Batalha, N. M., Rowe, J. F., Bryson, S. T., et al. 2013, ApJS, 204, 24Belousova, E. A., Kostitsyn, Y. A., Griffin, W. L., et al. 2010, Lithos, 119, 457Borucki, W. J., Agol, E., Fressin, F., et al. 2013, Science, 340, 587Borucki, W. J., Koch, D. G., Batalha, N., et al. 2012, ApJ, 745, 120Cassan, A., Kubas, D., Beaulieu, J.-P., et al. 2012, Nature, 481, 167Charbonneau, D., Berta, Z. K., Irwin, J., et al. 2009, Nature, 462, 891Cockell, C. S. 1998, J Theor Biol, 193, 717Cowan, N. B., Agol, E., Meadows, V. S., et al. 2009, ApJ, 700, 915Cowan, N. B., Robinson, T., Livengood, T. A., et al. 2011, ApJ, 731, 76Crow, C. A., McFadden, L. A., Robinson, T., et al. 2011, ApJ, 729, 130 16 –D´esert, J.-M., Bean, J., Miller-Ricci Kempton, E., et al. 2011, ApJ, 731, L40Dhuime, B., Hawkesworth, C. J., Cawood, P. A., & Storey, C. D. 2012, Science, 335, 1334Ford, E. B., Seager, S., & Turner, E. L. 2001, Nature, 412, 885Fressin, F., Torres, G., Charbonneau, D., et al. 2013, ApJ, 766, 81Fressin, F., Torres, G., Rowe, J. F., et al. 2012, Nature, 482, 195Fujii, Y. & Kawahara, H. 2012, ApJ, 755, 101Fujii, Y., Turner, E. L., & Suto, Y. 2013, ApJ, 765, 76Garc´ıa Mu˜noz, A. & Pall´e, E. 2011, J. Quant. Spec. Radiat. Transf., 112, 1609Garc´ıa Mu˜noz, A., Zapatero Osorio, M. R., Barrena, R., et al. 2012, ApJ, 755, 103Gilliland, R. L., Marcy, G. W., Rowe, J. F., et al. 2013, ApJ, 766, 40Goode, P. R., Qiu, J., Yurchyshyn, V., et al. 2001, Geophys. Res. Lett., 28, 1671Goodwin, A. M. 1981, Science, 213, 55Gough, D. O. 1981, Sol. Phys., 74, 21Hamdani, S., Arnold, L., Foellmi, C., et al. 2006, A&A, 460, 617Haqq-Misra, J. D., Domagal-Goldman, S. D., Kasting, P. J., & Kasting, J. F. 2008, Astro-biology, 8, 1127Hart, M. H. 1978, Icarus, 33, 23Hedelt, P., von Paris, P., Godolt, M., et al. 2013, A&A, 553, A9Hess, M., Koepke, P., & Schult, I. 1998, Bull. Am. Met. Soc., 79, 831Imhoff, J. 2005, Genus IX. Rhodopseudomonas. In: Bergey’s Manual of Systematic Bac-teriology, vol 2, 2nd Edition (Brenner, DJ., Krieg, N.R. Staley, J.T., Springer. NewYork)Kaltenegger, L., Traub, W. A., & Jucks, K. W. 2007, ApJ, 658, 598Kasting, J. F. 1993, Science, 259, 920 17 –Kasting, J. F. & Brown, L. L. 1998, The early atmosphere as a source of biogenic compounds.In the molecular origin of lifeKasting, J. F., Pollack, J. B., & Ackerman, T. P. 1984, Icarus, 57, 335Kasting, J. F. & Siefert, J. L. 2002, Science, 296, 1066Kawahara, H. & Fujii, Y. 2010, ApJ, 720, 1333Kawahara, H. & Fujii, Y. 2011, ApJ, 739, L62Kharecha, P., J., K., & Siefert, J. 2005, Geobiology, 3, 53Kiang, N. Y., Segura, A., Tinetti, G., et al. 2007a, Astrobiology, 7, 252Kiang, N. Y., Siefert, J., Govindjee, & Blankenship, R. E. 2007b, Astrobiology, 7, 222Kiehl, J. T. & Dickinson, R. E. 1987, J. Geophys. Res., 92, 2991Kopparapu, R. K., Ramirez, R., Kasting, J. F., et al. 2013, ApJ, 765, 131Manabe, S. & Wetherald, R. T. 1967, Journal of Atmospheric Sciences, 24, 241Mayor, M. & Queloz, D. 1995, Nature, 378, 355Mojzsis, S. J., Arrhenius, G., McKeegan, K. D., et al. 1996, Nature, 384, 55Monta˜n´es-Rodr´ıguez, P., Pall´e, E., Goode, P. R., & Mart´ın-Torres, F. J. 2006, ApJ, 651, 544Muirhead, P. S., Johnson, J. A., Apps, K., et al. 2012, ApJ, 747, 144Olson, J. 2006, Photosynth Res, 88, 109Pall´e, E., Ford, E. B., Seager, S., Monta˜n´es-Rodr´ıguez, P., & Vazquez, M. 2008, ApJ, 676,1319Pall´e, E., Goode, P. R., Monta˜n´es-Rodr´ıguez, P., & Koonin, S. E. 2004, Science, 304, 1299Pall´e, E., Goode, P. R., Yurchyshyn, V., et al. 2003, J. Geophys. Res. (Atmos.), 108, 4710Pall´e, E., Zapatero Osorio, M. R., Barrena, R., Monta˜n´es-Rodr´ıguez, P., & Mart´ın, E. L.2009, Nature, 459, 814Pall´e, E., Zapatero Osorio, M. R., & Garc´ıa Mu˜noz, A. 2011, ApJ, 728, 19 18 –Pavlov, A. A., Kasting, J. F., Brown, L. L., Rages, K. A., & Freedman, R. 2000, J. Geo-phys. Res., 105, 11981Pepe, F., Lovis, C., S´egransan, D., et al. 2011, A&A, 534, A58Pierson, B. K., Mitchell, H. K., & Ruff-Roberts, A. L. 1993, Origins of Life and Evolutionof the Biosphere, 23, 243Pinto, J. P., Gladstone, G. R., & Yung, Y. L. 1980, Science, 210, 183Pont, F., Sing, D. K., Gibson, N. P., et al. 2013, MNRAS, 432, 2917Qiu, J., Goode, P. R., Pall´e, E., et al. 2003, J. Geophys. Res. (Atmos.), 108, 4709Ramirez, R., Kopparapu, R., Zugger, M. E., et al. 2013, Nat.Geosc. (in review)Robinson, T. D., Meadows, V. S., Crisp, D., et al. 2011, Astrobiology, 11, 393Rugheimer, S., Kaltenegger, L., Zsom, A., Segura, A., & Sasselov, D. 2013, Astrobiology,13, 251Sanrom´a, E. & Pall´e, E. 2012, ApJ, 744, 188Sanrom´a, E., Pall´e, E., & Garc´ıa Mun˜oz, A. 2013, ApJ, 766, 133Scheer, H. 2003, The pigments. In Advances in Photosynthesis and Respiration, Vol. 13:Light-Harvesting Antennas in Photosynthesis (B.R. Green and W.W. Parson, Dor-drecht, The Netherlands, Kluwer Academic Publishers, pp. 29-81)Seager, S., Turner, E. L., Schafer, J., & Ford, E. B. 2005, Astrobiology, 5, 372Stammnes, K., Tsay, S. C., Wiscombre, W., & Jayaweera, K. 1988, Appl. Opt., 27, 2502Sterzik, M. F., Bagnulo, S., & Palle, E. 2012, Nature, 483, 64Swain, M. R., Vasisht, G., & Tinetti, G. 2008, Nature, 452, 329Tinetti, G., Meadows, V. S., Crisp, D., et al. 2006a, Astrobiology, 6, 34Tinetti, G., Meadows, V. S., Crisp, D., et al. 2006b, Astrobiology, 6, 881Tinetti, G., Rashby, S., & Yung, Y. L. 2006c, ApJ, 644, L129Tinetti, G., Vidal-Madjar, A., Liang, M.-C., et al. 2007, Nature, 448, 169Turnbull, M. C., Traub, W. A., Jucks, K. W., et al. 2006, ApJ, 644, 551 19 –Udry, S., Bonfils, X., Delfosse, X., et al. 2007, A&A, 469, L43Walker, J. C. G. 1977, Evolution of the atmosphereWoolf, N. J., Smith, P. S., Traub, W. A., & Jucks, K. W. 2002, ApJ, 574, 430Xiong, J., Fischer, W., Inoue, K., Nakahara, M., & Bauer, C. 2000, Science, 289, 1724
This preprint was prepared with the AAS L A TEX macros v5.2.
20 –Fig. 1.— Atmospheric composition and temperature profiles of the Earth 3.0 Ga ago.Fig. 2.— Spectral reflectance of water, vegetation, desert, and snow. Data taken from theASTER Spectral Library and the USGS Digital Spectral Library. 21 –Fig. 3.— Wavelength-dependent albedo obtained for purple bacteria in the VIS-NIR spectralrange where major absorption features of carotenoids and bacteriochlorophyll a are labeled.Table 1: Strength of the purple bacteria signal as a function of bacteria percentage in oceans.Bact. in Water No Clouds With Clouds10% 0.0295 0.008820% 0.0400 0.014030% 0.0505 0.019340% 0.0610 0.024550% 0.0715 0.0298
Note. — Slope between the intensity in the 0.745-0.770 µm and the 1.010-1.034 µm range.
22 –Fig. 4.— Left: A map of the Earth’s continental distribution during the Late Cambrian (500Ma ago). Image credit, Ron Blakey. Right: Same as in left panel but here oceans, continentsand coastal zones are indicated in blue, yellow and purple, respectively. Data are plottedwith a geographical resolution of 64x32 grid cells (longitude by latitude), the same that ourmodels use. Coastal areas constitute 14% of the total grid cells when using this continentaldistribution and geographic resolution. In our simulations the 00:00 hours UT, correspondto the subsolar point crossing the image center.Fig. 5.— Visible and near-infrared disk-averaged spectra of a planet with a continentaldistribution of the Earth 500 Ma ago, with an atmospheric composition similar to that ofthe Earth 3.0 Ga ago (black), and with present-day composition (blue). Here, continentsare totally desert and we have considered both, a cloud-free atmosphere (left), and a cloudyatmosphere (right). We have assumed that clouds cover 50% of the surface. The spectrahave been smoothed with a 100 point running mean for display purposes. 23 –Fig. 6.— Visible and near-infrared Earth’s reflectance spectra 3.0 Ga ago, taken as ππ