biofluorescent worlds: biological fluorescence as a temporal
TRANSCRIPT
Biofluorescent Worlds:
Biological fluorescence as a temporal biosignature for flare star worlds
Jack O’Malley-JamesCarl Sagan Institute
Cornell University
Ithaca, NY 14853
USA
607-255-5869
Lisa KalteneggerCarl Sagan Institute
Cornell University
Ithaca, NY 14853
USA
Draft version August 23, 2016Preprint typeset using LATEX style AASTeX6 v. 1.0
BIOFLUORESCENT WORLDS:
BIOLOGICAL FLUORESCENCE AS A TEMPORAL BIOSIGNATURE FOR FLARE STAR WORLDS
Jack T. O’Malley-James and Lisa Kaltenegger
Carl Sagan Institute
Cornell University
Ithaca, NY 14853, USA
ABSTRACT
Habitability for planets orbiting active stars has been questioned. Especially, planets in the Habitable
Zone (HZ) of M-stars, like our closest star Proxima Centauri, experience temporal high-ultraviolet
(UV) radiation. The high fraction of M-stars (75%) within the solar neighborhood, the high occur-
rence rate of rocky planets around M-stars, and the favorable contrast ratio between the star and a
potentially habitable rocky planet, makes such planets interesting targets for upcoming observations.
During M-star flares, the UV flux on a HZ planet can increase by up to two orders of magnitude.
High UV radiation is harmful to life and can cause cell and DNA damage. Common UV protection
methods (e.g. living underground, or underwater) would make a biosphere harder to detect. How-
ever, photoprotective biofluorescence, “up-shifting” UV to longer, safer wavelengths (a proposed UV
protection mechanism for some corals), would increase the detectability of biota and even uncover
normally hidden biospheres during a flare. Such biofluorescence could be observable as a “temporal
biosignature” for planets around UV-active stars. We model temporal biofluorescence as a biosig-
nature for an exoplanet biosphere exposed to such conditions, based on planets in M-star HZs. We
use fluorescing coral proteins to model biofluorescence, comparing observable spectra, and colors, to
vegetation and fluorescent minerals. Our planetary models assume a present-day Earth atmosphere
and explore the effect of varying cloud coverage and land:ocean fractions. UV flare-induced biofluo-
rescence could be remotely detectable, comparable in strength to vegetation on Earth. On planets in
the HZ of M-stars, biofluorescence could be a temporary biosignature, distinguishable from fluorescing
minerals and vegetation.
Keywords: astrobiology — stars: activity — stars: flare
1. INTRODUCTION
M stars are the most common type of star in the
galaxy and make up 75% of the stars in the solar neigh-
borhood. They are also excellent candidates for habit-
able zone (HZ) terrestrial planet searches. Our closest
star, Proxima Centauri, is an active M6 star that cur-
rently experiences intense flares every 10-31 hours (Cin-
cunegui et al. 2007), with several teams searching for
a planet in the HZ around it. Another nearby M star,
the young M8 star TRAPPIST-1, has been shown to
have three Earth-sized planets close to its HZ (Gillon et
al. 2016). Recent observations suggest the inner plan-
ets, TRAPPIST-1 b and c, are likely to be terrestrial
(de Wit et al., 2016). Some possible orbits for the out-
ermost planet, TRAPPIST-1d, place it in the system’s
HZ (Gillon et al. 2016). Estimates of the HZ occur-
rence rate of Earth-sized (0.4-1 R⊕) planets around cool
dwarfs range between 15% and 66% (Traub 2011; Gai-
dos 2013; Dressing et al. 2013; Kopparapu 2013; Ari et
al. 2015). The upcoming TESS mission, scheduled for
launch in 2017, will survey nearby bright stars to iden-
tify transiting exoplanets, including terrestrial ones in
the HZ. It will be sensitive enough to identify HZ planet
candidates around nearby low mass stars (Teff ≤ 4000K;
late M and early K stars) for future ground and space-
based characterization (Ricker et al. 2014; Sullivan et
al. 2015). TESS is expected to find 100s of 1.25-2 R⊕planets and 10s of Earth-sized planets, with a handful
(< 20) of these planets in the HZ of their cool host stars
(Ricker et al. 2014; Sullivan et al. 2015; Barnes et al.
2015). This makes it likely that the first HZ planet that
can be characterized will be orbiting a nearby M star.
Planets that receive high doses of UV radiation are
generally considered to be less promising candidates in
the search for life (see e.g. Buccino et al. 2006). How-
ever, several teams have made the case that planets in
the HZ of M stars can remain habitable, despite periodic
high UV fluxes (see e.g. Rugheimer et al. 2015b, Scalo
2
et al. 2007; Buccino et al. 2007; Tarter et al. 2007;
Heath et al.1999).
Two recent studies (Rugheimer et al. 2015a, 2015b)
model the amount of radiation reaching the surface
of an Earth-like planet with a 1 bar surface pressure
through its geological evolution (following Kaltenegger
et al. 2007) for different star types. The study also
assessed the biological impact of that radiation for at-
mospheres corresponding to different times in the geo-
logical history of a planet, modeled on Earth. Gener-
ally, an Earth analog planet orbiting an inactive M star
would receive a lower UV flux than Earth. It’s biolog-
ically effective UV radiation dose is between 0.001-0.03
times that on the present Earth (Rugheimer et al, 2015a,
2015b). However, around active M stars, such planets
would be subject to periodic bursts of UV radiation,
because of the proximity of the HZ to the star. That
increases the surface UV flux on a HZ planet by up to
two orders of magnitude for up to several hours for the
most active M stars (Segura et al., 2010). M stars also
remain active for longer periods of time than the Sun
(see West et al., 2011). Figure 2 shows the change in
magnitude of the UV flux from AD Leo, an active M
star, before and during a flare, compared to the UV flux
at present-day Sun.
The close proximity of planets in the HZ of cool stars
can cause the planet’s magnetic field to be compressed
by stellar magnetic pressure, reducing the planet’s abil-
ity to resist atmospheric erosion by the stellar wind. X-
ray and EUV flare activity can occur up to 10-15 times
per day, and typically 2-10 times, for M dwarfs (Cuntz
& Guinan, 2016), which increases atmospheric erosion
on close-in planets. This results in higher fluxes of UV
radiation reaching the planet’s surface (Lammer et al.
2007; See et al., 2014) and, potentially, a less dense at-
mosphere. In addition, planets in the HZs of M stars
are subject to stellar particle fluxes orders of magnitude
stronger than those in the solar HZ (Cohen et al. 2014)
that could erode their protective ozone shield as well as
some of the atmosphere. Note that currently we can not
model the expected surface pressure on an exoplanet.
A decrease in surface pressure or atmosphere mass in-
creases the UV flux reaching the surface, assuming the
same atmospheric composition of a planet.
When UV radiation is absorbed by biological
molecules, especially nucleic acids, harmful effects, such
as mutation or inactivation can result, with shorter UV
wavelengths having the most damaging effects (see e.g.
Voet et al. 1963; Diffey 1991; Matsunaga et al. 1991;
Tevini 1993; Cockell 1998; Kerwin & Remmele 2007).
On the present-day Earth, the ozone layer prevents the
most damaging UV wavelengths, UV-C radiation (100 -
290 nm), from reaching the surface. However, on other
HZ planets, a protective ozone layer may not be present;
the early Earth, for example, lacked a significant ozone
layer (see e.g. Kaltenegger et al. 2007).
In the absence of an ozone layer, depending on the
atmospheric composition of a planet, other atmospheric
gases, such as sulphur compounds and CO2 can absorb
UV radiation (see e.g., Cockell et al. 2000; Rugheimer
et al. 2015a). The thinner the atmosphere of a planet
is, the more of the damaging radiation would reach the
planet’s surface. Hence, mechanisms that protect biota
from such radiation are a crucial part of maintaining
habitability, especially on planets with thin atmospheres
that, for example, are less massive and can therefore not
maintain a dense protective atmosphere.
On Earth, biological mechanisms such as protective
pigments and DNA repair pathways can prevent, or mit-
igate, radiation damage (see e.g. Cockell 1998; Heath
1999; Neale & Thomas 2016). Additionally, subsur-
face environments can reduce the intensity of radiation
reaching an organism (see e.g. Heath 1999; Wynn-
Williams et al., 2002; Ranjan & Sasselov 2016). Some
teams have suggested that life that is constrained to
habitats underwater, or beneath a planet’s surface, may
not be detectable remotely (see e.g. Cockell 2014),
making an inhabited planet appear uninhabited. How-
ever, photoprotective biofluorescence, in which protec-
tive proteins absorb harmful UV wavelengths and re-
emit them at longer, safer wavelengths, is a mecha-
nism that has not yet been considered. It could be a
strong temporal biosignature on planets orbiting active
M stars. On Earth, evidence suggests that some coral
species use such a mechanism to reduce the risk of dam-
age to symbiotic algae, which provide the coral with
energy (see e.g. Salih et al. 2000; Roth et al 2010): flu-
orescent proteins in corals absorb blue and UV photons
and re-emit them at longer wavelengths.
Here we outline the possibility of detecting a fluores-
cent biosphere responding to stellar UV flare events,
such as in the HZ of an M star. Section 2 describes
biofluorescence and modeled UV surface levels on exo-
planets, section 3 describes our models, section 4 shows
our results, and section 5 discusses open issues.
2. BIOFLUORESCENCE IN CORALS
Scleractinian (hard) coral reefs are comprised of
colonies of identical animals called polyps (small sac-
like animals with a set of tentacles surrounding a cen-
tral mouth) that secrete calcium carbonate to form a
hard exoskeleton. Excessive light can be harmful to
corals, either by damaging the algal photosystem, or by
increasing oxidative stress, via photochemical reactions
(Bhagooli & Hidaka 2004; Takahashi & Murata 2008;
Roth et al. 2010). Some species contain fluorescent pro-
teins with excitation spectra in the UV-A (315-400 nm)
and blue regions of the spectrum, which have emission
3
Figure 1. An example of coral fluorescence. Coral fluores-cent proteins absorb near-UV and blue light and re-emit itat longer wavelengths (see e.g. Mazel & Fuchs 2003). Imagemade available under Creative Commons CC0 1.0 UniversalPublic Domain Dedication.
Figure 2. The UV flux at the top of the atmosphere ofa planet at a 1 AU equivalent distance orbiting the activeM star AD Leo during quiescence (grey) and flaring (blue)(Segura et al. 2010). The solar UV spectrum is shown forcomparison.
maxima between 420-700 nm. UV absorption by these
proteins is a possible protection mechanism for symbi-
otic algae from harmful UV radiation, cf. Fig.1, Fig.2
(see e.g. Salih et al. 2000; Gorbunov et al. 2001)
The visible signature of a coral reef has two compo-
nents: elastic scatter (reflected light) and inelastic scat-
ter (fluorescent light). Fluorescence can be a significant
factor in the appearance of coral reefs (Fuchs 2001). The
magnitude of the increase in intensity at the emitting
wavelengths of fluorescent proteins varies depending on
the intensity of the radiation absorbed, the wavelengths
a protein absorbs/emits and the spectral overlaps be-
tween proteins with different excitation spectra within
a reef (Fuchs 2001). The four most commonly occurring
fluorescent proteins in coral have emission peaks at 486
nm, 515 nm, 575 nm and 685 nm (the 685 nm pigment
is associated with chlorophyll in symbiotic algae). The
properties of these pigments are summarised in Table
1. All except the 515 nm protein are excited by UV-A
radiation (Fuchs 2001).
Emission Excitation Fluorescent
peak (nm) range (nm) efficiency (%)
486 350-475 3-5
515 400-525 10-12
575 350-575 8-10
685 350-650 1-2
Table 1. The four most common fluorescent proteins incoral species. The quoted fluorescent efficiencies are fromcoral species selected for being highly fluorescent (Mazel &Fuchs 2003).
Fluorescent efficiency is the ratio of the number of
photons emitted to the number of photons absorbed. A
material is considered to be quite fluorescent if it has
an efficiency of approximately 10%. Coral fluorescence
becomes more significant under artificial lighting con-
ditions (see e.g. Mazel & Fuchs 2003; Hochberg et al.
2004), with increases in reflectance of up to an order of
magnitude; the increase depending on the spectral dis-
tribution of the illuminating light source (Mazel & Fuchs
2003). For example, laser-induced fluorescence enables
corals to be successfully identified and monitored from
the ocean surface, rather than in-situ (Myers et al. 1999;
Mumby et al. 2004). A more intense UV regime should
result in a stronger fluorescence effect.
2.1. UV surface fluxes during M-star flares
The proposed biofluorescent signature around an ac-
tive M star is induced by UV flares and corresponding
UV levels on the surface of a planet. Corals fluoresce
brightly under a range of light regimes on Earth. UV-A
wavelengths are the main UV excitation wavelengths for
coral fluorescence. Many fluorescent corals are found at
shallow ocean depths where the UV-A regime is similar
to the surface flux on Earth. However, bright coral flu-
orescence is observed down to depths of 50-60 m (Eyal
et al. 2015), where the UV-A flux can be as low as 12
Wm−2 (Eyal et al. 2015). The modeled UV flux reach-
ing the surface of a planet analog to the present-day
Earth, orbiting a quiescent M star at a 1 AU-equivalent
distance varies between 0.04-1.36 Wm−2, depending on
spectral type for a 1 bar surface pressure (Rugheimer et
al. 2015a). This is 1-2 orders of magnitude below the
UV flux on present-day Earth’s surface (32.3 Wm−2;
Rugheimer et al. 2015a).
For the active M star AD Leo, the modeled quiescent
UV surface flux on an Earth-analog planet with 1 bar
surface pressure in the HZ is 2.97 Wm−2 UV-A, 0.01
Wm−2 UV-B and 2.13×10−14 Wm−2 UV-C (Segura et
al. 2010), an order of magnitude lower than on present-
day Earth (consistent with results by Rugheimer et al.
2015b). However at the flare peak, UV-A and UV-B
levels reach 120.77 Wm−2 UV-A and 3.15 Wm−2 UV-B
4
Figure 3. (Left) Four sample coral species (A, B, C and D) spanning the color distribution of corals that we model asbiofluorescent surfaces. (Right) An example of modeled fluorescence using coral B and spectral data on fluorescence emissionof four common coral fluorescent proteins (Coral data from Roelfsema C. & Phinn S. 2006; Clark 2007).
(Segura et al. 2010), an order of magnitude higher than
on present-day Earth. Note that the modeled UV sur-
face values would increase when one accounts for the de-
formation of a planet’s compressed magnetic field due to
stellar magnetic pressure (Lammer et al. 2007). Lower
surface pressure would also increase the UV levels on a
planet’s surface. For different atmospheric compositions
- like a younger Earth (see Rugheimer et al. 2015a,b for
details) - the UV surface flux also increases.
Increased UV flux on a planet’s surface could occur
due to associated proton events with the flare, in which
charged particles from the star are accelerated in the di-
rection of a flare. If the planet is in the line of sight of
the charged particle stream, ionized particles can inter-
act with the planet’s atmosphere, forming odd hydro-
gen species and breaking up N2 to form NOx species,
both of which can destroy ozone (see e.g. Segura et
al. 2010; Grenfell et al. 2012). Accounting for a pro-
ton event associated with an M star flare, Segura et
al. (2010) calculated that approximately 70 days after
the flare, ozone depletion could cause surface UV fluxes
to increase for up to 2 years. Frequent flares and pro-
ton events could thus permanently weaken, or erode, a
planet’s ozone layer and overall atmosphere, allowing
more UV radiation to reach the planet’s surface if the
star’s activity cycle is shorter than the recovery time of
the atmosphere.
3. MODELS
3.1. Planetary Models
Our model initially explores the surface signal
strength if the surface is a global ocean inhabited by
biofluorescent life, or a hypothetical fully vegetation cov-
ered planet. This constitutes the ideal case with the
strongest signal. Next we add a clear atmosphere to
the planet (assuming an M3V host star) using EXO-
Prime (see e.g. Kaltenegger & Sasselov 2009); a coupled
1D radiative-convective atmosphere code developed for
rocky exoplanets. The code is based on iterations of a
1D climate model (Kasting & Ackerman 1986; Pavlov
et al. 2000; Haqq-Misra et al. 2008), a 1D photochem-
istry model (Pavlov & Kasting 2002; Segura et al. 2005,
2007), and a 1D radiative transfer model (Traub & Stier
1976; Kaltenegger & Traub 2009) that calculates the
model spectrum of a rocky exoplanet in the HZ. EXO-
Prime models exoplanet’s atmospheres and environment
depending on the stellar and planetary conditions, in-
cluding the UV radiation that reaches the surface and
the planet’s reflection, emission and transmission spec-
trum.
We then reduce the fraction of biofluorescent life on
the surface to test how long it can be detectable by
adding ocean (from the USGS Spectral Library) as an
additional surface to explore the effect of different frac-
tions of inhabited versus uninhabited surface on the
spectrum. We compare the biofluorescent signal to that
of vegetation (from the USGS Spectral Library) for de-
tectability.
Then we add clouds (from the USGS Spectral Library,
following Kaltenegger et al. 2007) to the model. We as-sume clouds block our view of any surface feature. We
model different cloud coverages up to 50% (an Earth-like
cloud fraction). Any surface feature signature is reduced
with increasing cloud cover because clouds are highly re-
flected and therefore strongly influence a planet’s spec-
trum (see e.g. Kaltenegger et al. 2007).
3.2. Biofluorescence Models
We use the efficiency limits of terrestrial fluorescent
proteins as a guide to the exploration of the magnitude
of our modeled biofluorescence. The first fluorescent
proteins studied were green fluorescent proteins (GFPs),
extracted from jellyfish (Shimomura et al. 1962; John-
son et al., 1962; Morin & Hastings 1971; Morise et al.
1974; Tsien 1998). Over time these have been adapted
and engineered for use in a variety of applications, from
fluorescent microscopy to transgenic pets (see e.g. Stew-
art (2006) and references therein). GFPs have been en-gineered in the lab to have a much higher fluorescence
5
efficiency by taking advantage of useful mutations. This
has resulted in proteins with high efficiencies of up to
100% (Ilagan et al. 2010; Goedhart et al. 2012). There-
fore, although this is not observed in nature on Earth, it
is feasible that, given the right evolutionary conditions,
highly efficient fluorescent proteins could evolve with up
to 100% efficiency (C. Mazel, private communication).
Furthermore, in a study on coral fluorescent proteins,
Roth et al. (2010) found that rapid changes in pro-
tein concentrations occurred in response to changes in
light intensity, and that pigment concentration strongly
correlates with fluorescence intensity. Hence, we postu-
late the possible evolution of dense fluorescent pigment
concentrations in a high-UV environment to explore the
idea of a biofluorescent biosignature. We assume bioflu-
orescence here that increases reflectance by 100%.
We simulated coral fluorescence by using models of
the fluorescence response of the four common fluores-
cent proteins in corals (shown in Tab. 1), which have
fluorescence peaks at 486 nm, 515 nm, 575 nm and 685
nm. The change in the reflectance spectrum caused by
simulated fluorescence is illustrated in Fig. 4 (left col-
umn).
3.3. Color-color diagrams as a diagnostic tool
We explore how to use a standard astronomy tool to
characterize stellar objects, a color-color diagram, to dis-
tinguish planets with and without biofluorescent biosig-
natures (following Hegde & Kaltenegger 2013). To de-
termine the difference between the reflectance, r, of two
different color bands, we use equation (1):
CAB = A−B = −2.5log10
(rArB
)(1)
where CAB is the difference between two arbitrarycolour bands, A and B.
We use standard Johnson-Cousins BVI broadband fil-
ters to define the color bands (0.4 µm < B < 0.5 µm;
0.5 µm < V < 0.7 µm; 0.7 µm < I < 0.9 µm). We used
color-color plots (Figs. 6 to 10) to explore color change
on model planets caused by flare-induced fluorescence
compared to vegetation for clear atmospheres, cloudy
conditions and different ocean fractions.
3.4. Investigating false positives
It is also possible for fluorescence to occur abiotically.
Some minerals (e.g. calcite, fluorite, opal, zircon) and
polycyclic aromatic hydrocarbons (PAHs; e.g. fluoroan-
thene, perylene, pyrene) are fluorescent at similar wave-
lengths to those of fluorescent corals; a result of metal
cation impurities in the case of minerals, and delocal-
ized electrons in aromatic molecule groups for the case of
PAHs (see e.g. McDougall 1952; Modreski 1987; Beltran
et al. 1998). Therefore, we explore how to distinguish
biofluorescence from fluorescent minerals.
Hydrocarbons and the metal inclusions within fluores-
cent minerals would not be subject to Darwinian evo-
lution, so we assume here that their fluorescent levels
would be comparable to those on Earth’s or slightly in-
creased, or decreased, scaled to the UV surface flux at
the exoplanet’s surface compared to Earth’s.
4. RESULTS
Figure 4 compares biofluorescent models with fluores-
cent mineral surfaces. The left column of Fig. 4 shows
reflectance spectra for simulated coral fluorescence for
four wavelengths compared to minerals that fluoresce at
similar wavelengths in the right-hand column for a clear
atmosphere. The spectra show that both have different
responses to UV flux. The signal of fluorescent minerals
has a different shape than the biofluorescent signal and,
under the assumption that biofluorescence evolves, the
mineral signature is also much weaker. The four mineral
species were chosen for their abilities to fluoresce at simi-
lar wavelengths to corals and represent the strongest flu-
orescent minerals (see e.g. Modreski 1987; fluorescence
information from: Luminescent Mineral Database1).
Figure 5 compares biofluorescent signal strength (left
column) to a commonly used biological surface feature,
vegetation (right column). We use a fluorescent coral
spectrum (coral B) and the 515 nm modeled biofluores-
cence as an example to compare them. All panels in-
clude a clear present-day atmosphere. In the top panel
of Fig.5 we model a planet that is completely covered by
the respective surface: biofluorescence (left) and vege-
tation (right). The middle panel of Fig.5 shows how an
increasing ocean fraction of 30% and 70% decreases the
detectable surface feature from either surface. The pan-
els on the bottom of Fig.5 show how cloud cover, assum-
ing fractions of 10% and 50%, decreases the detectable
biological surface features even more strongly than re-
ducing the surface cover of the biosignature, even for a
planet that is completely covered by the biological sur-
face. We call these two effects out separately, but they
of course combine, depending on surface fraction cov-
erage as well as cloud coverage (see discussion). Note
that the aim of Fig.5 is not to give a specific value but
to compare the strength and detectability of a biofluo-
rescent biosphere to the commonly used vegetation red
edge feature that is proposed as one of the biological
surface features for upcoming visible direct imaging mis-
sions (see e.g. Seager et al. 2005; Tinetti et al. 2006;
Schneider et al. 2010; Hu et al. 2012). Fig. 5 shows that
the proposed bioflourescence signature can be stronger
1 http://flomin.org
6
Figure 4. Reflectance spectra in the visible for coral (left column) and fluorescent minerals (right column). Fluorescence at eachof the common coral fluorescent protein emission wavelengths was simulated for increases in reflectance from 10% to 100%. Thefour mineral species shown were chosen for their abilities to fluoresce at similar wavelengths to corals and represent the strongestfluorescent minerals. (Non-fluorescent coral spectra from Roelfsema & Phinn 2006. Mineral spectra sources: Sources: USGSDigital Spectral Library (Clark 2007), ASTER spectral library, California Institute of Technology. Fluorescence was simulatedusing data from C. Mazel.)
7
than the surface biosignature of vegetation covering the
same surface area.
A standard astronomy tool that characterizes stellar
objects, a color-color diagram, can be used to distin-
guish planets with and without biofluorescent biosigna-
tures (following Hegde & Kaltenegger 2013). We used
these planet model spectra with and without clouds, and
with varying surface fractions of biological versus ocean
surface, to create standard Johnson colors from their flu-
orescent and non-fluorescent spectra for biofluorescence
and minerals. We add vegetation for comparison as well.
The minerals and coral fluorescence occupy separate
regions of the color-color diagram (Fig. 6), in both the
non-fluorescing and fluorescing cases. The curved shape
of the zircon reflectance spectrum in the visible region
causes it to occupy a slightly separate region in the color
space than the other minerals, which tend to have very
flat reflectance profiles in the visible spectrum. Fluores-
cence at 515 nm, 575 nm and 685 nm, moves the color of
the biofluorescent planet further away from fluorescent
minerals and vegetation. However fluorescence at 486
nm moves the biofluorescent color closer to the color of
minerals and vegetation. Coral D is closest in the color
space to vegetation. Mineral fluorescence causes a much
smaller change in color position, that is too small to
show on the scale in Fig. 6.
Figure 7 shows the color-color diagram for a planet
with different fractions of uninhabited ocean versus
biofluorescent surface (see Fig.5 for the corresponding
spectra) for two examples, 30% and 70%, the latter cor-
responding to the Earth’s ocean fraction, for each of
the four modeled biofluorescent pigments. The decrease
in overall biological surface area reduces the influence
of the biological surface on the overall planetary spec-
trum and color and therefore also the shift in colors due
to fluorescence. Still, biofluorescent signatures are well
distinguishable from minerals or vegetation in the color-
color diagram in both cases.
Figure 8 shows the effect of cloud coverage on the col-
ors of a planet (see Fig.5 for the corresponding spectra)
comparing a clear atmosphere to 10% and 50% cloud
coverage, the latter corresponding to Earth’s cloud cov-
erage. Increasing cloud cover moves the colors of all
planetary models closer together - whether the planet
is completely covered by minerals, vegetation or bioflu-
orescent corals. Note that a planet model with 100%
cloud coverage would not show any surface feature and
appear the same in spectra and color, no matter what
the underlying surface were covered in. The increase
in cloud coverage also reduces the magnitude of the ob-
servable shift in position during biofluorescence, because
only part of the surface is visible; the rest is blocked from
our view by clouds (see discussion).
Figures 9 and 10 compare the colors of biofluorescence
for the clear-sky case and partial cloud cover case, re-
spectively, with the colors of solar systems bodies (us-
ing spectra from Irvine et al. (1968) and Karkoschka
(1994)). Fluorescent coral colors are distinct from other
solar system bodies (with the exception of 515 nm fluo-
rescence of coral B, which moves close to the color po-
sition of Mars), as well as abiotic mineral fluorescence
(as shown before). Figures 6 to 10 show that all four
of the considered pigments produce notable changes in
position on the color-color diagram.
5. DISCUSSION
To produce a strong fluorescence effect during an M
star flare, the best fluorescent proteins will be those that
absorb most strongly in the UV part of the spectrum.
Alternatively to higher UV flux, an increased quantum
efficiency of fluorescent proteins could also enable de-
tectable biofluorescence, even for a lower UV surface
flux. A fluorescence quantum yield of 1 is possible, based
on lab-based improvements on the yield of the green flu-
orescent protein, discussed earlier, which would increase
the detectability of a biofluorescent signal. Addition-
ally, a denser concentration of fluorescent proteins may
be selected for under high UV conditions, enhancing the
strength of the fluorescent signal.
During a flare, models show that a planet in the HZ of
AD Leo would receive a UV-B surface flux 80% higher
than Earth’s (Segura et al. 2010), while the UV-C sur-
face flux would be several orders of magnitude higher
than Earth’s, if an ionized particle stream aligns with
the flare and depletes the planet’s ozone layer (Segura et
al. 2010). Hence, fluorescent proteins that downshift the
full UV flux (UV-A to -C) to benign visible wavelengths
(perhaps via a cascading chain of fluorescent proteins
with different excitation wavelengths; see, for example,
Gilmore et al. 2003) may be selected for under high UV
conditions, further strengthening a biofluorescent signal.
Most corals gain energy, carbohydrates and oxygen via
symbiotic relationships with algae, which in return feed
on CO2 and waste products from the coral. To maintain
this symbiosis, coral habitats are limited to the euphotic
zone of the oceans (to a maximum depth of up to 165
m on Earth; Kahng & Maragos 2006) to allow access
to photosynthetically active radiation (PAR). Water at-
tenuation would reduce coral fluorescent flux. Here we
make the simplifying assumption of shallow, transpar-
ent oceans. This is consistent with shallow-water coral
reef habitats on Earth, which are typically warm, clear
seawater environments, at depths as shallow as 0-3 m
(Kleypas et al. 1999; Nagelkerken et al. 2000). Further-
more, the lower flux of Earth-like PAR on an M star
planet would make shallow water habitats more prefer-
able for any aquatic photosynthesisers.
We do not argue for the evolution of an exact replica
8
Figure 5. Comparison of spectra for a planet with vegetation (left) and biofluorescent (right) surfaces. A present-day Earthatmosphere has been added to all models shown. We use a coral spectrum (coral B) with modeled fluorescence (515 nmfluorescence is used in this example). (Top) Surface biosignatures are assumed to cover 100% of the planet. (Middle) An oceanfraction of 30% and 70% is added, reducing the surface biosignature fraction to 70% and 30%, respectively. (Bottom) Cloudcover fractions of 10% and 50% are added to the model, assuming the surface of the planet is completely covered with thebiosignature (like the top panel) to show the effects of clouds separately from surface fraction coverage (middle).
of coral reefs on other worlds. In this hypothesis we use
corals as an example of an organism that exhibits fluo-
rescence and contains convergent evolutionary features
(i.e. features that have evolved independently in differ-
ent species that are not closely related). We argue for
the evolution of an organism that shares some of these
convergent traits with corals; specifically the ability to
form simple, collaborative colonies and the ability to
fluoresce.
Corals were one of the the earliest forms of animal
life to evolve on Earth, with an evolutionary history
spanning 500 Myr. Initially existing as simple, soli-
tary organisms, they later evolved into collective reefs.
Reef-building is an ancient trait for life on Earth. Many
marine species have independently evolved reef-building
abilities, but the trait can be traced back to colo-
nial groups of bacteria building stromatolites 3.5 Gya
(Kiessling 2009).
The fluorescent proteins in corals are descended from
green fluorescent proteins (Field et al. 2006). Green
fluorescent proteins are present in a variety of phyla,
suggesting an origin within an ancient common ancestor
in the very early metazoan (animal) life, over 500 Myr
ago (Chudakov et al. 2010).
Biofluorescence is widespread in life on Earth and is
thought to have evolved independently, multiple times
(Sparks et al., 2014; Gruber et al. 2015; Gruber &
Sparks, 2015), which strengthens a case for the evolution
of biofluorescence on other inhabited worlds, following
convergent evolution arguments.
M star HZs are subject to stellar particle fluxes or-
ders of magnitude stronger than those in the solar HZ
9
Figure 6. Color-color diagrams for four fluorescent corals and four fluorescent minerals, before (grey - labelled A, B, C, D) andduring (black - labelled A’, B’, C’, D’) fluorescence at each of the four common emission wavelengths. We assume 100% increasein reflectance over the wavelength range of the emission spectra for each coral fluorescent pigment during fluorescence. Notethat the change in position before and during mineral fluorescence is too small to plot on this scale.
(Cohen et al., 2014), which could trigger aurorae up to
103 times more intense than on Earth. For planets with
weaker magnetic fields than Earth, the auroral oval (the
amount of planetary area with open magnetic field lines)
could cover all latitudes on the planet (Vidotto et al.,
2013), changing the colors of planets during a flare, po-
tentially adding other absorption and emission features
in the atmosphere that were not taken into account in
our models. However, aurorae fluoresce at known wave-
lengths depending on the composition of the atmosphere
and therefore would be distinguishable from biofluores-
cence.
A planet will most likely have both clouds as well as
different surfaces, depending on how dense their atmo-
sphere is. With decreasing surface pressure, cloud cover-
age should also decrease, therefore we showed the effects
separately in Fig.5 to Fig.8, which allows insight into the
individual effects on detectability and characterization
of a planet in a spectrum, or in a color-color diagram.
For planets with non-complete cloud coverage, the ef-
fect of clouds can be distinguished from surface features
with many short, high signal-to-noise observations, be-
cause clouds should occupy all areas of the planet given
enough time. Thus one can separate them from sur-
face features that are bound to the rotation of a planet,
if the observations can be limited to about 1/20 of the
planet’s rotation period, or for the Earth, about an hour
(see Palle et al. 2008), requiring big future telescopes.
Even though removing the effect of clouds on planetary
spectra is not feasible for near-term observation with
telescopes like the E-ELT, it should be possible with
even more ambitious space and ground-based telescopes
that are being designed.
6. CONCLUSIONS
In this work, we explored the hypothesis and de-
tectability of exoplanets dominated by a biofluorescent
biosphere, which uses fluorescence as a UV damage-
mitigation strategy. Especially planets in the Habit-
able Zone (HZ) of M-stars would experience tempo-
ral high ultraviolet (UV) radiation. During an M-star
flare, the UV radiation flux on a HZ planet can in-
crease by up to two orders of magnitude. Photopro-
tective biofluorescence (the “up-shifting” of UV light to
longer, safer wavelengths, via absorption by fluorescent
proteins), a proposed UV protection mechanism of some
coral species, would increase the detectability of biota,
both in a spectrum, as well as in a color-color diagram.
Such biofluorescence could be observable as a “tempo-
ral biosignature” for planets around stars with changing
UV environments, like active flaring M stars, in both
their spectra as well as their color.
Using a standard astronomy tool to characterize stel-
lar objects, a color-color diagram, one can distinguish
planets with and without biofluorescent biosignatures.
The change in color caused by biofluorescence differs,
in position and magnitude, from that caused by abiotic
10
Figure 7. Color-color diagrams for a planet with different surface coverage of oceans versus biological surface, before (grey -labelled A, B, C, D) and during (black - labelled A’, B’, C’, D’) fluorescence at each of the four common emission wavelengthsfor 30% and 70% ocean coverage.
fluorescence, distinguishing both.
The TESS mission will be sensitive enough to iden-
tify rocky planets in the HZ of nearby M stars, providing
targets for testing this hypothesis with follow up obser-
vations. A planet in the HZ of Proxima Centauri will be
the best target to look for biofluorescence. Watching a
star during a flare event could uncover a biofluorescent
biosphere on a planet orbiting it. This suggests that
exoplanets in the HZ of active M stars are also interest-
ing targets in the search for signs of life beyond Earth.
Hence, high UV fluxes could actually make certain hid-
den biospheres detectable.
Acknowledgements The authors acknowledge helpful
discussions with Charles Mazel and David Gruber, Juan
Torres-Perez for providing coral spectral data, Sarah
Rugheimer for providing atmosphere spectra and the
USGS Digital Spectral Library and the ASTER Spec-
11
Figure 8. Color-color diagrams for planets with surfaces completely covered by biofluorescent corals, vegetation or minerals,for a clear atmosphere and for 10% and 50% cloud coverage. Non-fluorescing corals are marked with grey points, labelled A toD. Fluorescing corals are marked with black points labelled A’ to D’.
tral Library for reflectance spectral data for miner-
als, vegetation and ocean water. We would also like
to acknowledge funding from the Simons Foundation
(290357, Kaltenegger).
REFERENCES
[1]Ari, S., Gaidos, E., & Wu, Y. 2015, Astrophys. J., 799, 180
[2]Beltran, J. L., Ferrer, R., & Guiteras, J. 1998, Analytica
Chimica Acta, 373, 311
[3]Bhagooli, R., & Hidaka, M. 2004, Comparative Biochemistry and
Physiology Part A: Molecular & Integrative Physiology, 137,
547
[4]Buccino, A. P., Lemarchand, G. A., & Mauas, P. J. 2006, Icarus,
183, 491
[5]Buccino, A. P., Lemarchand, G. A., & Mauas, P. J. 2007, Icarus,
192, 582
[6]Chudakov, D. M., Matz, M. V., Lukyanov, S., & Lukyanov, K.
A. 2010, Physiological Reviews, 90, 1103
[7]Cincunegui, C., Dıaz, R. F., & Mauas, P. J. 2007, Astron.
Astrophys., 461, 1107
[8]Clark, R. N. 2007, USGS Digital Spectral Library splib06a, Data
Series 231.
[9]Cockell, C. S. 1998, J. theor. Biol., 193, 717
[10]Cockell, C. S. 2014, Phil. Tran. R. Soc. A, 372, 20130082
[11]Cohen, O., Drake, J. J., Glocer, A., et al. 2014, Astrophys. J.,
790, 57
[12]Cuntz, M., & Guinan, E. F. 2016. Astrophys. J., 827, 79
[13]de Wit, J., Wakeford, H. R., Gillon, M., et al. 2016, Nature,
doi:10.1038/nature18641
[14]Diffey, B. L. 1991, Physics and Medicine in Biology, 36, 299
[15]Dressing, C. D. & Charbonneau, D. 2013, Astrophys. J., 767, 95
[16]Eyal, G., Wiedenmann, J., Grinblat, M., et al. 2015, PloS one,
10, e0128697.
[17]Field, S. F., Bulina, M. Y., Kelmanson, I. V., Bielawski, J. P., &
Matz, M. V. 2006, Journal of Molecular Evolution, 62, 332
[18]Fuchs, E. 2001, Applied Optics, 40, 3614
[19]Gaidos, E. 2013, Astrophys. J., 770, 90
[20]Gillon, M., Jehin, E., Lederer, S. M., et al. 2016, Nature, 533,
221
[21]Gilmore, A. M., Larkum, A. W., Salih, A., et al. 2003,
Photochemistry and Photobiology 77, 515
12
Figure 9. Color-color diagrams of planet models with clear atmospheres, and a surface that is completely covered by bioflu-orescent corals, fluorescent minerals, or vegetation compared to the colors of planets in our own Solar System, before (grey -labelled A, B, C, D) and during (black - labelled A’, B’, C’, D’) fluorescence at each of the four common emission wavelengths.
Figure 10. Color-color diagrams of planet models with an atmosphere and 50% cloud cover, and a surface that is completelycovered by biofluorescent corals, fluorescent minerals, or vegetation compared to the colors of planets in our own Solar System,before (grey - labelled A, B, C, D) and during fluorescence at each of the four common emission wavelengths.
13
[22]Goedhart, J., von Stetten, D., Noirclerc-Savoye, M., et al. 2012,
Nat. Commun. 3, 751[23]Gorbunov, M. Y., Kolber, Z. S., Lesser, M. P., & Falkowski, P.
G. 2001, Limnology and Oceanography, 46, 75
[24]Grenfell, J. L., Grießmeier, J. M., von Paris, P., et al. 2012,Astrobiology, 12, 1109
[25]Gruber, D. F., & Sparks, J. S. 2015, American MuseumNovitates, 3845, 7pp
[26]Gruber, D. F., Gaffney, J. P., Mehr, S., et al. 2015, PloS one, 10,
e0140972.[27]Haqq-Misra, J. D., Domagal-Goldman, S. D., Kasting, P. J., &
Kasting, J. F. 2008, Astrobiology, 8, 1127
[28]Heath, M.J., Doyle, L. R., Joshi, M. M., & Haberle, R. M. 1999,Origins of Life and Evolution of the Biosphere, 29, 405
[29]Hegde, S., & Kaltenegger, L. 2013, Astrobiology, 13, 47
[30]Hochberg, E. J., Atkinson, M. J., Apprill, A., & Andrefouet, S.2004, Coral Reefs, 23, 84
[31]Hu, R., Ehlmann, B. L., & Seager, S. 2012, Astrophys. J. 752, 7
[32]Ilagan, R. P., Rhoades, E., Gruber, D. F., et al. 2010, FEBSJournal 277, 1967
[33]Irvine, W. M., Simon, T., Menzel, D. H., Pikoos, C., & Young,A. T. 1968, Astronom. J. 73, 807
[34]Johnson, F. H., Shimomura, O., Saiga, Y., et al. 1962, Journal of
Cellular and Comparative Physiology, 60, 85[35]Kahng, S. E., & Maragos, J. E. 2006, Coral Reefs 25, 254
[36]Kaltenegger, L., Traub, W. A., & Jucks, K. W. 2007, Astrophys.
J., 658, 598[37]Kaltenegger, L., & Sasselov, D. 2009, Astrophys J, 708, 1162
[38]Kaltenegger, L., & Traub, W. A. 2009, Astrophys J, 698, 519
[39]Karkoschka, E. 1994, Icarus, 111, 174[40]Kasting, J. F., & Ackerman, T. P. 1986, Science, 234, 1383
[41]Kerwin, B. A., & Remmele, R. L. 2007, Journal of
Pharmaceutical Sciences, 96, 1468[42]Kiessling, W. 2009, Annu. Rev. Ecol. Evol. Syst., 40, 173
[43]Kleypas, J. A., McManus, J. W., & Meez, L. A. 1999, Amer
Zool, 39, 146[44]Kopparapu, R. K. 2013, Astrophys. J. Lett. 767, L8
[45]Lammer, H., Lichtenegger, H. I., Kulikov, Y. N., et al. 2007,Astrobiology, 7, 185
[46]Matsunaga, T., Hieda, K., & Nikaido, O. 1991, Photochemistry
and Photobiology, 54, 403[47]Mazel, C. H., & Fuchs, E. 2003 Limnol. Oceanog., 48, 390
[48]McDougall, D. J. 1952, Am. Mineral, 37, 427
[49]Modreski, P. J. 1987, New Mexico Geology, 9, 25[50]Morin, J. G., & Hastings, J. W. 1971, Journal of cellular
physiology, 77, 313
[51]Morise, H., Shimomura, O., Johnson, F. H., & Winant, J. 1974,Biochemistry, 13, 2656
[52]Mumby, P. J., Skirving, W., Strong, A. E., et al. 2004, MarinePollution Bulletin 48, 219
[53]Myers, M. R., Hardy, J. T., Mazel, C. H., & Dustan, P. 1999,
Coral Reefs, 18, 179.[54]Nagelkerken, I., Van der Velde, G., Gorissen, M. W., et al. 2000,
Estuar Coast Shelf Sci, 51, 31
[55]Neale, P. J., & Thomas, B. C. 2016, Astrobiology (ahead ofprint), doi:10.1089/ast.2015.1360.
[56]Palle, E., Ford, E. B., Seager, S., Montanes-Rodrıguez, P., &
Vazquez, M. 2008, Astrophys. J., 676, 1319[57]Pavlov, A. A., Kasting, J. F., Brown, L. L., Rages, K. A., &
Freedman, R. 2000, Journal of Geophysical Research, 105, 981
[58]Pavlov, A. A., & Kasting, J. F. 2002, Astrobiology, 2, 27
[59]Ranjan, S., & Sasselov, D. D. 2016, Astrobiology 16, 68
[60]Ricker, G. R., Winn, J. N., Vanderspek, R., et al. 2014, Proc. of
SPIE 9143, 914320
[61]Roelfsema, C., & Phinn, S. 2006 Spectral reflectance library of
selected biotic and abiotic coral reef features In Heron Reef,
Bremerhaven, PANGAEA. http://epic.awi.de/31865/
[62]Roth, M. S., Latz, M. I., Goericke, R., & Deheyn, D. D. 2010 J.
Exp. Biology, 213, 3644
[63]Rugheimer, S., Segura, A., Kaltenegger, L., & Sasselov, D.
2015a, Astrophys. J., 806, 137.
[64]Rugheimer, S., Kaltenegger, L., Segura, A., Linsky, J., &
Mohanty, S. 2015b, Astrophys. J., 809, 57
[65]Salih, A., Larkum, A., Cox, G., Kuhl, M., & Hoegh-Guldberg,
O. 2000 Nature 408, 850
[66]Scalo, J., Kaltenegger, L., Segura, A., et al. 2007, Astrobiology,
7, 85
[67]Schneider, J., Leger, A., Fridlund, M., et al. 2010, Astrobiology,
10, 121
[68]Seager, S., Turner, E. L., Schafer, J., & Ford, E. B. 2005,
Astrobiology 5, 372
[69]See, V., Jardine, M., Vidotto, A. A., et al. 2014, Astron.
Astrophys. 570, A99
[70]Segura, A., Kasting, J. F., Meadows, V., et al. 2005,
Astrobiology, 5, 706
[71]Segura, A., Meadows, V. S., Kasting, J. F., Crisp, D., & Cohen,
M. 2007, Astron. Astrophys., 472, 665
[72]Segura, A., Walkowicz, L. M., Meadows, V., Kasting, J., &
Hawley, S. 2010, Astrobiology, 10, 751
[73]Shimomura, O., Johnson, F. H., & Saiga, Y. 1962, Journal of
Cellular and Comparative Physiology 59, 223
[74]Sparks, J. S., Schelly, R. C., Smith, W. L., et al. 2014, PLoS
ONE, 9, e83259
[75]Stewart, C. N. 2006, Trends in biotechnology, 24, 155
[76]Sullivan, P. W., Winn, J. N., Berta-Thompson, Z. K., et al.
2015, Astrophys. J. 809, 77
[77]Takahashi, S., & Murata, N. 2008, Trends in Plant Science, 13,
178
[78]Tarter, J. C., Backus, P. R., Mancinelli, R. L., et al. 2007,
Astrobiology, 7, 30
[79]Tevini, M. (ed.) 1993, in UV-B Radiation and Ozone Depletion:
Effects on Humans, Animals, Plants, Microorganisms, and
Materials (Boca Raton, Florida: Lewis Publishers)
[80]Tinetti, G., Meadows, V. S., Crisp, D., et al. 2006, Astrobiology,
6, 881
[81]Traub, W. A., & Stier, M. T. 1976, Applied Optics, 15, 364
[82]Traub, W. A. 2011, Astrophys. J., 745, 20
[83]Tsien, R. Y. 1998, Annual review of biochemistry, 67, 509
[84]Vidotto, A., Jardine M., Morin, J., et al. 2013, Astron.
Astrophys., 557, A67
[85]Voet, D., Gratzer, W. B., Cox, R. A., & Doty, P. 1963,
Biopolymers, 1, 193
[86]West, A. A., Morgan, D. P., Bochanski, J. J., et al. 2011
Astronom. J. 141, 97
[87]Wynn-Williams, D. D., Edwards, H. G. M., Newton, E. M., &
Holder, J. M. 2002, Int. J. Astrobiology, 1, 39