Distinguishing multicellular life on exoplanets by testing Earth as an exoplanet
Christopher E. Doughty, Andrew Abraham, James Windsor, Michael Mommert, Michael Gowenlock, Tyler Robinson, David Trilling
(cid:105)(cid:105) (cid:105) “astro” — 2020/9/3 — 1:02 — page 1 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105)
Submitted to InternationalJournal of Astrobiology
Key words: multicellular, exoplanet, biosignature,BRDF
Author for correspondence:
Christopher E. Doughty, Email:[email protected]
Distinguishing multicellular life on exoplanetsby testing Earth as an exoplanet
Christopher E. Doughty , Andrew Abraham , James Windsor ,Michael Mommert , Michael Gowanlock , Tyler Robinson , and David Trilling School of Informatics, Computing, and Cyber Systems, Northern Arizona University, Flagstaff, AZ.86011, USA, Department of Astronomy and Planetary Science, Northern Arizona University, Flagstaff,AZ. 86011, USA and University of St. Gallen, Institute of Computer Science, Rosenbergstrasse 30, 9000St. Gallen, Switzerland
Abstract
Can multicellular life be distinguished from single cellular life on an exoplanet? We hypoth-esize that abundant upright photosynthetic multicellular life (trees) will cast shadows at highsun angles that will distinguish them from single cellular life and test this using Earth as anexoplanet. We first test the concept using Unmanned Arial Vehicles (UAVs) at a replica moonlanding site near Flagstaff, Arizona and show trees have both a distinctive reflectance signa-ture (red edge) and geometric signature (shadows at high sun angles) that can distinguish themfrom replica moon craters. Next, we calculate reflectance signatures for Earth at several phaseangles with POLDER (Polarization and Directionality of Earth’s reflectance) satellite direc-tional reflectance measurements and then reduce Earth to a single pixel. We compare Earth toother planetary bodies (Mars, the Moon, Venus, and Uranus) and hypothesize that Earthâ ˘A ´Zsdirectional reflectance will be between strongly backscattering rocky bodies with no weather-ing (like Mars and the Moon) and cloudy bodies with more isotropic scattering (like Venus andUranus). Our modelling results put Earth in line with strongly backscattering Mars, while ourempirical results put Earth in line with more isotropic scattering Venus. We identify potentialweaknesses in both the modelled and empirical results and suggest additional steps to determinewhether this technique could distinguish upright multicellular life on exoplanets.
1. INTRODUCTION
Recently, a 1.3 Earth mass planet only ∼ a r X i v : . [ a s t r o - ph . E P ] S e p (cid:105) “astro” — 2020/9/3 — 1:02 — page 2 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105) Earth has more than 3 trillion trees [Crowther et al., 2015], eachwith a vertical structure that casts shadows differently than objectson a lifeless planet with weather and climate. Almost all trees areat a 90 ◦ angle to the ground while less than 1 percent of the surfaceof the Earth has with a slope greater than 45 ◦ [Hall et al., 2005].This is simply because weather and climate, which are thoughtto be necessary on any planet capable of sustaining multicellu-lar life [Kasting et al., 2003] will erode much abiotic topographyover time. For instance, one study suggested a lifeless planet withweather will be very similar to Earth topologically [Dietrich andPerron, 2006]. Therefore, shadows at certain sun angles may beindicative of multicellular life, but could we detect them on anexoplanet?Earth Scientists know a great deal about tree shadows becauseto accurately estimate terrestrial reflectance (with, for example,Landsat or MODIS (Moderate Resolution Imaging Spectrora-diometer) satellite data) shadows at different sun angles must beremoved. Therefore, a great deal of effort has been put into devel-oping a quantitative framework to predict shadows at different sunangles. This framework, called the bidirectional reflectance dis-tribution function (BRDF), is the change in observed reflectancewith changing view angle or illumination direction [Schaepman-Strub et al., 2006]. Forests seen from different sensor sun angleshave predictable differences in reflectance [Bréon et al., 2002,Bréon and Henriot, 2006, Li and Strahler, 1992, Wolf et al., 2010].Previous work used a semi-empirical BRDF model [Bacour andBréon, 2005, Maignan et al., 2004] at the global scale to explorewhether, in theory, Earth with vegetation would have differentalbedo at different sensor sun angles versus an Earth without veg-etation [Doughty and Wolf, 2010]. They found that even if theentire planetary albedo were rendered to a single pixel, the rate ofincrease of albedo as a planet approaches full illumination wouldbe comparatively greater on a vegetated planet than on a non-vegetated planet. It was hypothesized that the technique wouldwork at 4 light years (and greater depending on knowledge oncloud abundance and a coronagraph design) meaning it could betested on the recently discovered planet in the habitable zone ofProxima Centauri.The method was then tested empirically [Doughty and Wolf,2016] using the Galileo space probe data and first principles, ina similar methodology as Sagan et al. [1993]. Sagan et al. [1993]detected multiple stages of life on Earth, but they did not have atechnique to distinguish between single and (non-technological)multicellular life on Earth. Doughty and Wolf [2016] used theGalileo space probe data but because the Galileo dataset had onlya small change ( < ◦ ) in phase angle (sun-satellite position), theobserved anisotropy signal was small, and they could not detectmulticellular life on Earth. In contrast, in this paper, we proposeto use to the POLDER satellite (Polarization and Directional-ity of Earth’s reflectance) data to test this question. This datasetgives global reflectance, directionality (BRDF), and polarizationmeasurements at 20km resolution and phase angles of > ◦ [Bicheron and Leroy, 2000]. Therefore, we can create a view ofEarth at different phase angles and determine empirically if, evenscaling to a single pixel, we could distinguish between single andmulticellular life on Earth.However, could the BRDF technique distinguish betweenabundant vertical structures like moon craters and abundant veg-etation on an exoplanet? Most such craters would in theory be eroded on a lifeless planet with weather and climate. However,we test the BRDF of craters on Earth to understand how they castshadows at different sun angles. We took advantage of moon-likecraters near our university that were created by the USGS in 1967to help Apollo astronauts train by simulating different-sized lunarimpact craters. A total of 497 craters were made within two sitescomprising 2,000 square feet. We fly a UAV above a cratered land-scape at different sun angles meant to replicate the moon landingsite.We can also use detection of the red edge as corroborating evi-dence for the existence of vegetation. Our goal is to compare thereflectance properties at the red edge of plants with the BRDFor geometric optics, for example, the shape and arrangement ofobjects within a pixel that transmit or block light [Torrance andSparrow, 1967], using Earth as an Exoplanet at various scales(Figure 1). We propose to test this at the following scales: at thereplica moon landing crater field, at the Amazon basin and theSahara Desert, on all of Earthâ ˘A ´Zs cloud free continental terres-trial surface and for the Earth as a whole. We will then comparethe phase function of the Earth as a single pixel to phase func-tions of other planets in the solar system. We will compare Earthempirically (with POLDER data) and for Earth modelled with andwithout vegetation with a BRDF model [Maignan et al., 2004,Bacour and Bréon, 2005] (Figure 1).
2. METHODS2.1. Site information
To test NDVI and BRDF as biosignatures, we took advantage ofan “extraterrestrial landscape” near our university that we call thereplica moon landing crater site (35.30594920 lon, -111.50617530lat). Moon-like craters were created by the USGS in 1967 by dig-ging holes and filling them with various amounts of explosives,which were detonated to simulate different-sized lunar impactcraters. The human-made craters range in size from 1.5–12 metersin diameter. This area was chosen for the craters because of thebasaltic cinders from an eruption of the Sunset Crater Volcano 950years ago. After the explosions, the excavated lighter clay materialspread out from the blast craters and across the fields, like ejectafrom actual meteorite impacts. A total of 497 craters were madewithin two sites comprising 2,000 square feet (Figure 2).
We flew the Parrot Bluegrass (Parrot) UAV with 4 wavelengths(green 550 nm (40nm bandwidth (bw)), red â ˘A ¸S 660 nm(40nmbw), red edge 735nm (10 nm bw), and NIR 790nm (40 nm bw))above the replica moon landing crater site described above. Weflew at various times to get different sun instrument angles (5:30,7:30, 9:00 and 11:30 am) comparing three landscape types (bareground, craters, and ponderosa pine trees). The Parrot takes ∼ ∼
300 m (88 by 338m or 6ha) witha resolution of 4.7cm/pixel. We use the program Pix4DCaptureto plan the flight paths and Pix4Dmapper to orthomosaic the rawimages into reflectance values (WGS 84 coordinate system). Thisprogram created geotiffs for each band which we uploaded into (cid:105) “astro” — 2020/9/3 — 1:02 — page 3 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105) Distinguishing multicellular life on exoplanets by testing Earth as an exoplanet Fig. 1.
Our conceptual design of a distant observer monitoring Earth and the change in backscattering as it revolves around the sun. Θ is the azimuth angle, Ω i is thesolar zenith angle, Ω v is the view angle, and Ψ is the phase angle. Fig. 2. (A) The Apollo Astronaut training ground as originally photographed in1967 (from USGS archives). (B) An example of UAV flyover measuring NDVI at5am in 2018 with a current google earth image as a background image. (C)Closeups of two example regions of interest (tree and crater) at three differenttimes of the day in the NIR (790 nm) band. Note, the crater shadows visible at 5am but not at later times while tree shadows are visible at all three times. the Google Earth Engine. We used matlab (Mathworks) to furtheranalyze this data.
POLDER (Polarization and directionality of Earth’s reflectance)gives global reflectance, directionality (BRDF), and polarizationmeasurements [Bacour and Bréon, 2005, Bicheron and Leroy,2000]. The ground size or resolution of a POLDER-measuredpixel is 6x7 km at nadir. 12 directional radiance measurementsat each spectral band are taken for each point on Earth. We down-loaded data that capture the period from October 30, 1996 toFebruary 28, 1997. During that period, we chose 21 days inter-spersed within this broader period and aggregated data from thosedays (Specifically â ˘A ¸S Oct 30,31, Nov 1–6 and Dec 30–31 1996and Jan 8, 9, 10, 11, 12, 14, 16, 17, 22, 23, 27 1997). We also col-lected solar zenith angle (which is relative to the local zenith andmay vary between 0 ◦ (sun at zenith) and approximately 80 ◦ ) andview zenith angle, (which is relative to the local zenith and mayvary between 0 ◦ (POLDER at zenith) and approximately 75 ◦ ) (see Figure 1 for an example of the geometries). For each day, we sub-tracted the view zenith angle from the solar zenith angle (but wedid not control for azimuth angle) to estimate phase angle for thewavelengths 565 nm (20 nm bandwidth) and 763 nm (10 nm band-width). POLDER also has bands 670 nm and 865 nm, which arecloser to traditional NDVI bands [Masek and Lim, 2006] and havebeen used previously to characterize vegetation cover and BRDFresponses [Bacour and Bréon, 2005]. However, these bands arealso not ideal as they use polarized filters which are unlikely to beon future space telescopes. Therefore, we use bands 670 nm and865 nm in Figs S1-2 and table S1, but use 565 nm and 763 nmin the rest of the manuscript. These two wavelengths were thenused to create NDVI (Normalized difference vegetation index)according to the following equation: N DV I = (763 nm − nm ) / (763 nm + 565 nm ) We then created separate data maps for < ◦ phase angleranges, then 1–3 ◦ , then 3–6 ◦ , 6–20 ◦ , and 20–30 ◦ . We aggregatedall available data for these five different phase angles and createdcloud free land images of the Amazon basin, the Sahara Desertregion, and all regions combined together. We averaged thesemaps as a single pixel at the different phase angles to replicatewhat Earth might look like to a distant observer as it circles thesun at different phase angles. (cid:105) “astro” — 2020/9/3 — 1:02 — page 4 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105) To compare the how Earth would look circling the sun at a dis-tance to other planetary bodies, we digitized data from Sudarskyet al. [Sudarsky et al., 2005] where they aggregated data for opti-cal phase functions for Mars, Venus, the moon, and Uranus alongwith a Lambert model where radiation is scattered isotropicallyoff a surface regardless of its angle of incidence [Sudarsky et al.,2005]. A classical phase function normalizes planetary albedo to1 at a phase angle of 0 ◦ . Data for Mars is originally from Thorpe[1977], for the Moon from Lane and Irvine [1973], for Uranusfrom Pollack et al. [1986] and Sudarsky et al. [2005] does notstate where the Venus data is originally from.We normalized all the datasets (Earth-POLDER, Earth no veg-etation, Earth with vegetation, Mars, Venus, Uranus, and themoon) so that the albedo at phase angle of 0 ◦ was one. We thensubtracted these from a Lambert curve to highlight the impact ofdirectional scattering from each of these bodies.
3. RESULTS
The Apollo astronaut training ground offers a unique opportunityto compare NDVI and BRDF in an “extraterrestrial landscape”with trees. In 1967, a flyover of the area early in the morningshows large shadows for both the craters and the local ponderosapine trees (Figure 2(a)). It is therefore conceivable that craterscould replicate the shadows and BRDF is not a good multicel-lular life biosignature. However, our UAV demonstrates why atlater times of the day (at lower phase angles) the story changes.Figure 2(b) shows our UAV NDVI image for the region at 5am.The trees clearly have a higher NDVI and the craters still haveshadows. However, Figure 2(c) shows strong shadows with thecraters at 5:30am but not at 9am and 11am. In contrast, the treesshow clear shadows at all times even towards noon (at lower phaseangles). This effect will change slightly with latitude [Doughtyand Wolf, 2010].We can quantify these qualitative observations with our UAVcollected reflectance data. Figure 3 shows the reflectance his-tograms for trees and craters in the NIR (790 nm) at differenttimes of the day. Because the UAV flew overhead, the daytimescorrespond with high (5am), medium (9am) and low phase angles(11am). In Figure 3(a), at 5:30 am the histogram of the cratershows a strong shadow peak at ∼ ∼ ∼ ∼ ∼ ◦ phase angle and reduced brightness at higher phaseangles is our hypothesized “multicellular life biosignature”.NDVI showed different reflectance peaks for trees than for bareground and craters. The “tree” NDVI signal included shadows andbare ground which reduced the overall NDVI signal. However,even with the mixed signal, NDVI also showed a clear signal thatcould distinguish between the three areas with NDVIâ ˘A ´Zs medianhistogram of 0.06 for the trees and ∼ Fig. 3.
Histograms of NIR reflectance (790 nm) for (top) craters and (middle)trees at different times of the day (5am and 9am for craters, 9am and 11am fortrees). For clarity, we aggregate all tree reflectance pixels greater than 0.15 to0.15. (bottom) NDVI for trees (green), craters (black), and bare ground (blue) at11am. ground (Figure 3(c)). Was the NDVI or BRDF signal greater? Forexample, a typical region of interest with 50 percent tree cover, 50percent ground at 9am might have 25 percent of the ground cov-ered in shadow. At 9am, our scene might have an NIR reflectanceof 0.09 (0.15*0.5 (tree)+0.01*0.25 (shadow)+0.05*0.25 (ground))while at noon, as the shadows are masked, it would change to0.10 (0.15*0.5 (tree)+0.05*0.50 (ground)). This is a relativelysmall change of 0.01. We have shown that moon craters wouldnot show this change and the 0.01 signal is the “multicellularlife biosignature”. However, the NDVI signal of ∼ < ◦ phase anglecontained many regional blank areas, especially tropical regionswith great cloud cover, and we did not include it in our final anal-ysis. We discuss this more in the discussion section. Therefore,we focused on the phase angle ranges of 1–3 ◦ , 3–6 ◦ , 6–20 ◦ , and20–30 ◦ . Averaging over 21 days gave sufficient cloud free imagesto create maps for most of the planet. There were still gaps in ourcoverage, both at high latitudes, where POLDER did not cover,and in parts of the tropics where clouds were very abundant. Table 1.
Absolute change of reflectance (between 1–3 ◦ phase angle and 20–30 ◦ phase angle) for band 763 nm, NDVI and the percent change for band 763nm for the Amazon, Sahara, all land and the world. AmazonSaharaAll landworld nm .
016 0 .
007 0 .
015 0 . N DV I .
055 0 .
009 0 .
043 0 . per nm . . . . These cloud free images allowed us to compare two multicel-lular life endmembers â ˘A ¸S the Amazon basin, with abundant tree (cid:105) “astro” — 2020/9/3 — 1:02 — page 5 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105)
Distinguishing multicellular life on exoplanets by testing Earth as an exoplanet cover, and the Sahara Desert, with very few trees. In Figure 7(a),we show the average reflectance for these two regions at both 565and 763nm at several different phase angles. The changes weresmaller than we had hypothesized with our BRDF model possiblybecause we missed the large change between 0–1 ◦ phase angle.At 763nm between phase angle 1–3 ◦ and 20–30 ◦ there was a dif-ference of 0.016 reflectance units or ∼ ∼ Fig. 4.
Cloud free terrestrial reflectance at 763nm from POLDER at the phaseangle (pa) ranges (a) 1–3 ◦ , (b) 3–6 ◦ , (c) 6–20 ◦ , and (d) 20–30 ◦ aggregated andaveraged from the 21-day period described in the methods. We next created a global view of Earth (including land,clouds and oceans) at the different phase angles (Figure 5) anda NDVI of the entire Earth at different phase angles (Figure 6). InFigure 7(b) and (c), we average Figure 4, 5, and 6 as a single pixelat the different phase angles. As a single pixel, at 565nm, there areonly minor reflectance changes between phase angle 1–3 ◦ and 20–30 ◦ . However, at 763 nm, the land only had reflectance changes ∼ ∼
12 percent and the whole world had a slightly smallerchange of 0.011 or ∼ Fig. 5.
All pixels (including ocean and clouds) reflectance at 763nm fromPOLDER at the phase angles (a) 1–3 ◦ , (b) 3–6 ◦ , (c) 6–20 ◦ , and (d) 20–30 ◦ aggregated and averaged from the 21-day period described in the methods. reflectance) and modelled Earth with and without vegetation[Doughty and Wolf, 2010]. All planetary bodies have very dif-ferent albedos, but for comparison purposes, we standardized theaverage albedo to 1 at a phase angle of 0. We initially hypothesizedthat Earth would have a phase function between Mars and Venus(with both POLDER and the vegetation model in agreement). Inother words, Earth might be a partially cloudy planet with somedirectional reflectance. However, our modeled estimates of Earth,with and without vegetation showed similar directional reflectanceto Mars but our empirical results using POLDER data showedEarth was more similar to Venus (Figure 8). Fig. 6.
All (including ocean and clouds) NDVI pixels from POLDER at the phaseangles (a) 1–3 ◦ , (b) 3–6 ◦ , (c) 6–20 ◦ , and (d) 20–30 ◦ aggregated and averagedfrom the 21 day period described in the methods.
4. DISCUSSION
Why there was a large divergence between our modelled resultsof Earth at different phase angles and our empirical ones? To (cid:105) “astro” — 2020/9/3 — 1:02 — page 6 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105) review, modelled Earthâ ˘A ´Zs reflectance at different phase anglesis similar to Mars while empirical POLDER data of Earth-â ˘A ´Zs reflectance at different phase angles are similar to Venus(Figure 8). We hypothesize that both the model and empirical datahave issues that make them not align. For instance, our model usesthe best vegetation BRDF model, but it did not have a good BRDFmodel for other components of the Earth, such as oceans, cloudsand atmosphere. Therefore, it likely missed key components ofatmospheric scattering and cloud directional reflectance. In con-trast, we hypothesize that there were also issues with the empiricaldata because by excluding our phase angle data of < ◦ degree inour empirical analysis, we missed the largest change in BRDF.Our BRDF model suggests the largest change in reflectance fromvegetation will be between phase angles of 0-1 ◦ and 1–3 ◦ . There-fore, by missing this peak, and showing little change < ◦ , ourphase curve is more like an isotropic body like Venus. Fig. 7. (top) Averaged reflectance at different phase angles at differentwavelengths (565 nm and 763 nm) for the Amazon region and the Sahararegion. (middle) Averaged reflectance at different phase angles for all Earth andall terrestrial land at different wavelengths (565 nm and 763 nm). (bottom)Averaged NDVI at different phase angles for a cloud covered Earth (red), allterrestrial land (black), the Amazon region (green), and the Sahara region (blue).
Mars and the moon both have greater backscattering than Earth.For solid bodies with thin atmospheres like Mars, previous workhas shown that backscattering can be significant [Thorpe, 1977].This is because Mars (currently) has no liquid water to erode andsmooth its rough edges. Our phase curve (Figure 8), shows thatthe moon has even stronger backscattering than Mars, which isinitially surprising [Lane and Irvine, 1973]. However, this is dueto a phenomenon called coherent backscatter which occurs onvery dry soils where particles have a diameter that is similar tothe wavelength of the photon used to view them [Hapke et al.,1993]. A planet with climate like Earth does not exhibit coherentbackscatter, even in dry areas, such as deserts, because the par-ticle sizes are too big (generally between 0.05 to 2mm) at 800nm or less [Tarbuck and Lutgens, 2008]. Therefore, Earth showsless backscattering than Mars or the moon because of the pres-ence of abundant isotropic clouds. The presence of craters on themoon and Mars also affects backscattering. At low phase angles the BRDF of craters is substantially different than that of trees(Figures 2 and 3). Earth has few craters due to abundant erosioncaused by climate. It is interesting to note the large amount oferosion of the craters at the replica moon landing site that hasalready occurred due to weather and climate in the 50 years sincethe craters were first formed.In contrast, Venus and Uranus have scattering more similarto Lambert scattering where radiation is scattered isotropicallyoff a surface. Lambert scattering is a good approximation forobjects such as Uranus [Pollack et al., 1986], and to a lesserextent Venus [Sudarsky et al., 2005]. Surprisingly, our empiricallyderived phase function for Earth was less steep than either Venusor Uranus (Figure 8). This is surprising because Earth has manystrong backscattering surfaces like trees. We hypothesize that thisis due to excluding our phase angle data of < ◦ in our empiricalanalysis.To improve our future empirical analysis, we need to bettercapture low phase angles. With the POLDER data, averagingfor phase angles of 1 degree or less was inherently more patchybecause it was averaging over a smaller dataset. Key regions, likeAmazonia were missing because of high cloud cover. In fact, thecloudier terrestrial areas, and the regions less represented at < ◦ phase angle, were those most likely to have abundant tree cover(like Amazonia). For this reason, we were not confident includingour maps of < ◦ phase angle. POLDER was only available fora few months during 1996-1997 and it is currently the only satel-lite of its kind to capture the Earth at all phase angles. Capturingplanets at low phase angles will also be a problem with any view-ing of an exoplanet because it could be washed out by the light ofits star, even with the most advanced coronagraph design [Guyonet al., 2006]. However, in theory, we could observe the planet dur-ing continuous rotation cycles which could increase the amount ofdata available to analyze the exoplanet for vegetation structure.To improve our modelling analysis, we need to better modelthe BRDF of non-vegetated surfaces. We used a state of the artBRDF model for vegetation [Bicheron and Leroy, 2000, Bacourand Bréon, 2005], but only averaged BRDF values for clouds,atmosphere and oceans. With this improved model, how do weenvision using the model in the future to distinguish a planet withmulticellular life versus just single cellular life? We could createa model of an exoplanet based on the exoplanetâ ˘A ´Zs size, den-sity, cloud cover, distance to star, and the starâ ˘A ´Zs irradiance.For instance, let us imagine we had the proper technology andcoronagraph to observe the 1.3 Earth mass planet only ∼ (cid:105) “astro” — 2020/9/3 — 1:02 — page 7 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105) Distinguishing multicellular life on exoplanets by testing Earth as an exoplanet exciting target were to be discovered, more telescope time couldincrease this to ∼
5. CONCLUSIONS AND FUTURE DIRECTIONS
Overall, in theory, BRDF could distinguish between multicellu-lar and single cellular life on exoplanets, but we have recognizedissues with both our models and our empirical observations thatmust be improved before this technique could be used with confi-dence. The easiest short-term step is to improve the modelling bycombining the various BRDF models. Further empirical validationwill be more challenging as POLDER is a unique satellite. Herewe demonstrate that BRDF is challenging to detect and will bea smaller signal than NDVI, which has already proven to be chal-lenging to detect with Earth as an exoplanet [Montañés-Rodríguezet al., 2006]. Should this line of research therefore be abandoned?Theoretically, it could still work and since we are not aware ofother techniques to distinguish an exoplanet with multicellularlife, we believe further work should still continue.
Fig. 8. (Top) The phase function for several solar system objects (from Sudarskyet al. [2005]), Earth with and without vegetation structure (from Doughty andWolf [2010]) and empirically calculated for a cloud covered Earth with POLDERdata from this paper. The phase function normalizes for albedo by forcing albedoto one at a phase angle of 0 ◦ . We also show a lambert model from Sudarskyet al. [2005] which assumes an object that scatters light perfectly isotopically. Inthe bottom figure, we show the same data but subtract the Lambert curve tomore clearly show backscattering differences. Acknowledgement.
This project was funded by NASAâ ˘A ´Zs HabitableWorldâ ˘A ´Zs program with the project name: “Testing methods to detect 3Dvegetation structure on exoplanets” (16-HW16-2-0025).
References
Guillem Anglada-Escudé, Pedro J Amado, John Barnes, Zaira MBerdiñas, R Paul Butler, Gavin AL Coleman, Ignaciode La Cueva, Stefan Dreizler, Michael Endl, Benjamin Giesers,et al. A terrestrial planet candidate in a temperate orbit aroundproxima centauri.
Nature , 536(7617):437–440, 2016. doi:10.1038/nature19106.Cédric Bacour and François-Marie Bréon. Variability of biomereflectance directional signatures as seen by polder.
RemoteSensing of Environment , 98(1):80–95, 2005. ISSN 00344257.doi: 10.1016/j.rse.2005.06.008.Patrice Bicheron and Marc Leroy. Bidirectional reflectance dis-tribution function signatures of major biomes observed fromspace.
Journal of Geophysical Research: Atmospheres , 105(D21):26669–26681, 2000. ISSN 01480227. doi: 10.1029/2000JD900380.FM Bréon and N Henriot. Spaceborne observations of ocean glintreflectance and modeling of wave slope distributions.
Jour-nal of Geophysical Research: Oceans , 111(C6), 2006. ISSN21699291. doi: 10.1029/2005JC003343. (cid:105) “astro” — 2020/9/3 — 1:02 — page 8 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105)
Francois-Marie Bréon, Fabienne Maignan, Marc Leroy, and IanGrant. Analysis of hot spot directional signatures measuredfrom space.
Journal of Geophysical Research: Atmospheres ,107(D16):AAC–1, 2002. ISSN 01480227. doi: 10.1029/2001JD001094.James H. Brown.
Scaling in biology . Oxford University Press,2000.D.E. Brownlee and D Ward.
Rare Earth . Nicolaus Copernicus,2000.William D Collins, Philip J Rasch, Byron A Boville, James JHack, James R McCaa, David L Williamson, Bruce P Briegleb,Cecilia M Bitz, Shian-Jiann Lin, and Minghua Zhang. The for-mulation and atmospheric simulation of the community atmo-sphere model version 3 (cam3).
Journal of Climate , 19(11):2144–2161, 2006. ISSN 08948755. doi: 10.1175/JCLI3760.1.Thomas W Crowther, Henry B Glick, Kristofer R Covey, CharlieBettigole, Daniel S Maynard, Stephen M Thomas, Jeffrey RSmith, G Hintler, Marlyse C Duguid, Giuseppe Amatulli, et al.Mapping tree density at a global scale.
Nature , 525(7568):201–205, 2015. ISSN 14764687. doi: 10.1038/nature14967.William E Dietrich and J Taylor Perron. The search for a topo-graphic signature of life.
Nature , 439(7075):411–418, 2006.ISSN 14764687. doi: 10.1038/nature04452.Michael J Donoghue. Key innovations, convergence, and success:macroevolutionary lessons from plant phylogeny.
Paleobiol-ogy , 31(sp5):77–93, 2005. ISSN 0094-8373. doi: 10.1666/0094-8373(2005)031[0077:kicasm]2.0.co;2.Christopher E Doughty and Adam Wolf. Detecting tree-like multi-cellular life on extrasolar planets.
Astrobiology , 10(9):869–879,2010. ISSN 15311074. doi: 10.1089/ast.2010.0495.Christopher E Doughty and Adam Wolf. Detecting 3d vege-tation structure with the galileo space probe: Can a distantprobe detect vegetation structure on earth?
PloS one , 11(12):e0167188, 2016. ISSN 19326203. doi: 10.1371/journal.pone.0167188.Yuka Fujii, Daniel Angerhausen, Russell Deitrick, ShawnDomagal-Goldman, John Lee Grenfell, Yasunori Hori,Stephen R Kane, Enric Pallé, Heike Rauer, Nicholas Siegler,Karl Stapelfeldt, and Kevin B Stevenson. Exoplanet Biosigna-tures: Observational Prospects.
Astrobiology , 18(6):739–778,6 2018. ISSN 1531-1074. doi: 10.1089/ast.2017.1733. URLhttps://doi.org/10.1089/ast.2017.1733.Linda E Graham, Martha E Cook, and James S Busse. Theorigin of plants: body plan changes contributing to a majorevolutionary radiation.
Proceedings of the National Academyof Sciences , 97(9):4535–4540, 2000. ISSN 00278424. doi:10.1073/pnas.97.9.4535.Olivier Guyon, EA Pluzhnik, MJ Kuchner, B Collins, andST Ridgway. Theoretical limits on extrasolar terrestrial planetdetection with coronagraphs.
The Astrophysical Journal Sup-plement Series , 167(1):81, 2006. ISSN 0067-0049. doi:10.1086/507630.Ola Hall, Giacomo Falorni, and Rafael L Bras. Characterizationand quantification of data voids in the shuttle radar topographymission data.
IEEE Geoscience and Remote Sensing Letters , 2(2):177–181, 2005. ISSN 1545598X. doi: 10.1109/LGRS.2004.842447.Bruce W Hapke, Robert M Nelson, and William D Smythe.The opposition effect of the moon: the contribution of coher-ent backscatter.
Science , 260(5107):509–511, 1993. ISSN00368075. doi: 10.1126/science.260.5107.509.James S Jenkins, Joseph Harrington, Ryan C Challener, Nicolás TKurtovic, Ricardo Ramirez, Jose Peña, Kathleen J McIntyre,Michael D Himes, Eloy Rodríguez, Guillem Anglada-Escudé,et al. Proxima centauri b is not a transiting exoplanet.
MonthlyNotices of the Royal Astronomical Society , 487(1):268–274,2019. doi: 10.1093/mnras/stz1268.James F Kasting, David Catling, et al. Evolution of a habitableplanet.
Annual Review of Astronomy and Astrophysics , 41(1):429–463, 2003. ISSN 0066-4146. doi: 10.1146/annurev.astro.41.071601.170049.Ray Kenny and L Paul Knauth. Stable isotope variations inthe neoproterozoic beck spring dolomite and mesoprotero-zoic mescal limestone paleokarst: Implications for life on landin the precambrian.
Geological Society of America Bul-letin , 113(5):650–658, 2001. ISSN 00167606. doi: 10.1130/0016-7606(2001)113<0650:SIVITN>2.0.CO;2.Centrella Kouveliotou, E Agol, N Batalha, J Bean, M Bentz,N Cornish, A Dressler, E Figueroa-Feliciano, S Gaudi,O Guyon, et al. Enduring quests-daring visions (nasaastrophysics in the next three decades). arXiv preprintarXiv:1401.3741 , 2014.Adair P Lane and William M Irvine. Monochromatic phase curvesand albedos for the lunar disk.
The Astronomical Journal , 78:267, 1973. ISSN 00046256. doi: 10.1086/111414.Xiaowen Li and Alan H Strahler. Geometric-optical bidirectionalreflectance modeling of the discrete crown vegetation canopy:Effect of crown shape and mutual shadowing.
IEEE trans-actions on Geoscience and Remote Sensing , 30(2):276–292,1992. ISSN 15580644. doi: 10.1109/36.134078.Timothy A Livengood, L Drake Deming, Michael F A’Hearn,David Charbonneau, Tilak Hewagama, Carey M Lisse, Lucy AMcFadden, Victoria S Meadows, Tyler D Robinson, Sara Sea-ger, et al. Properties of an earth-like planet orbiting a sun-likestar: Earth observed by the epoxi mission.
Astrobiology , 11(9):907–930, 2011. doi: 10.1089/ast.2011.0614.Wolfgang Lucht, Crystal Barker Schaaf, and Alan H Strahler. Analgorithm for the retrieval of albedo from space using semiem-pirical brdf models.
IEEE Transactions on Geoscience andRemote sensing , 38(2):977–998, 2000. ISSN 01962892. doi:10.1109/36.841980.F Maignan, F.-M Bréon, and R Lacaze. Bidirectional reflectanceof Earth targets: evaluation of analytical models using a largeset of spaceborne measurements with emphasis on the HotSpot.
Remote Sensing of Environment
IEEE Geoscience and (cid:105) “astro” — 2020/9/3 — 1:02 — page 9 — (cid:105)(cid:105)(cid:105) (cid:105) (cid:105)(cid:105)
Distinguishing multicellular life on exoplanets by testing Earth as an exoplanet Remote Sensing Letters , 3(5), 2006. doi: 10.1109/LGRS.2005.857030.Pilar Montañés-Rodríguez, E Pallé, PR Goode, and FJ Martín-Torres. Vegetation signature in the observed globally integratedspectrum of earth considering simultaneous cloud data: appli-cations for extrasolar planets.
The Astrophysical Journal , 651(1):544, 2006. ISSN 0004-637X. doi: 10.1086/507694.James B Pollack, Kathy Rages, Kevin H Baines, Jay T Bergstralh,Daniel Wenkert, and G Edward Danielson. Estimates of thebolometric albedos and radiation balance of uranus and nep-tune.
Icarus , 65(2-3):442–466, 1986. ISSN 10902643. doi:10.1016/0019-1035(86)90147-8.Carl Sagan, W Reid Thompson, Robert Carlson, Donald Gurnett,and Charles Hord. A search for life on earth from the galileospacecraft.
Nature , 365(6448):715–721, 1993. doi: 10.1038/365715a0.Gabriela Schaepman-Strub, Michael E Schaepman, Thomas HPainter, Stefan Dangel, and John V Martonchik. Reflectancequantities in optical remote sensingâ ˘AˇTdefinitions and casestudies.
Remote sensing of environment , 103(1):27–42, 2006.ISSN 00344257. doi: 10.1016/j.rse.2006.03.002.Edward W. Schwieterman, Nancy Y. Kiang, Mary N. Parenteau,Chester E. Harman, Shiladitya Dassarma, Theresa M. Fisher,Giada N. Arney, Hilairy E. Hartnett, Christopher T. Reinhard,Stephanie L. Olson, Victoria S. Meadows, Charles S. Cock-ell, Sara I. Walker, John Lee Grenfell, Siddharth Hegde, SarahRugheimer, Renyu Hu, and Timothy W. Lyons. ExoplanetBiosignatures: A Review of Remotely Detectable Signs of Life,2018a. ISSN 15311074.Edward W Schwieterman, Nancy Y Kiang, Mary N Parenteau,Chester E Harman, Shiladitya DasSarma, Theresa M Fisher,Giada N Arney, Hilairy E Hartnett, Christopher T Reinhard,Stephanie L Olson, et al. Exoplanet biosignatures: a review ofremotely detectable signs of life.
Astrobiology , 18(6):663–708,2018b. ISSN 15311074. doi: 10.1089/ast.2017.1737.David Sudarsky, Adam Burrows, Ivan Hubeny, and Aigen Li.Phase functions and light curves of wide-separation extraso-lar giant planets.
The Astrophysical Journal , 627(1):520, 2005.ISSN 0004-637X. doi: 10.1086/430206.EJ Tarbuck and F.K. Lutgens.
Earth, An introduction to physicalgeology . Prentice Hall, 2008.Thomas E Thorpe. Viking orbiter photometric observations of themars phase function july through november 1976.
Journal ofGeophysical Research , 82(28):4161–4165, 1977. ISSN 0148-0227. doi: 10.1029/js082i028p04161.Giovanna Tinetti, Victoria S Meadows, David Crisp, WilliamFong, Evan Fishbein, Margaret Turnbull, and Jean-Pierre Bib-ring. Detectability of planetary characteristics in disk-averagedspectra. i: The earth model.
Astrobiology , 6(1):34–47, 2006.ISSN 15311074. doi: 10.1089/ast.2006.6.34.K. E. Torrance and E. M. Sparrow. Theory for off-specular reflec-tion from roughened surfaces.
Journal of the Optical Societyof America (1917-1983) , 57(9):1105, September 1967. ISSN0030-3941. doi: 10.1364/josa.57.001105. Angelos Tsiaras, Ingo P. Waldmann, Giovanna Tinetti, JonathanTennyson, and Sergey N. Yurchenko. Water vapour in theatmosphere of the habitable-zone eight-Earth-mass planet K2-18 b.
Nature Astronomy , 3:1086–1091, September 2019. doi:10.1038/s41550-019-0878-9.S. G. Turyshev. Direct Multipixel Images of an Exo-Earth witha Solar Gravitational Lens Telescope.
Journal of the BritishInterplanetary Society , 71:361–368, January 2018.MR Walter, R Buick, and JSR Dunlop. Stromatolites 3,400–3,500myr old from the north pole area, western australia.
Nature , 284(5755):443–445, 1980. doi: 10.1038/284443a0.Geoffrey B West, James H Brown, and Brian J Enquist. A gen-eral model for the origin of allometric scaling laws in biology.
Science , 276(5309):122–126, 1997.Adam Wolf, Joseph A Berry, and Gregory P Asner. Allometricconstraints on sources of variability in multi-angle reflectancemeasurements.