The biomass distribution on Earth

Abstract

Significance
The composition of the biosphere is a fundamental question in biology, yet a global quantitative account of the biomass of each taxon is still lacking. We assemble a census of the biomass of all kingdoms of life. This analysis provides a holistic view of the composition of the biosphere and allows us to observe broad patterns over taxonomic categories, geographic locations, and trophic modes.

Full text

The biomass distribution on Earth
Yinon M. Bar-On
a
, Rob Phillips
b,c
, and Ron Milo
a,1
a
Department of Plant and Environmental Sciences, Weizmann Institute of Science, 76100 Rehovot, Israel;
b
Department of Physics, California Institute of
Technology, Pasadena, CA 91125; and
c
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA 91125
Edited by Paul G. Falkowski, Rutgers, The State University of New Jersey, New Brunswick, NJ, and approved April 13, 2018 (received for review July 3, 2017)
A census of the biomass on Earth is key for understanding the
structure and dynamics of the biosphere. However, a global,
quantitative view of how the biomass of different taxa compare
with one another is still lacking. Here, we assemble the overall
biomass composition of the biosphere, establishing a census of the
≈550 gigatons of carbon (Gt C) of biomass distributed among all of
the kingdoms of life. We find that the kingdoms of life concentrate
at different locations on the planet; plants (≈450 Gt C, the domi-
nant kingdom) are primarily terrestrial, whereas animals (≈2GtC)
are mainly marine, and bacteria (≈70 Gt C) and archaea (≈7GtC)
are predominantly located in deep subsurface environments. We
show that terrestrial biomass is about two orders of magnitude
higher than marine biomass and estimate a total of ≈6GtCof
marine biota, doubling the previous estimated quantity. Our anal-
ysis reveals that the global marine biomass pyramid contains more
consumers than producers, thus increasing the scope of previous
observations on inverse food pyramids. Finally, we highlight that
the mass of humans is an order of magnitude higher than that of
all wild mammals combined and report the historical impact of
humanity on the global biomass of prominent taxa, including
mammals, fish, and plants.
ecology
|
biomass
|
biosphere
|
quantitative biology
One of the most fundamental efforts in biology is to describe
the composition of the living world. Centuries of research
have yielded an increasingly detailed picture of the species that
inhabit our planet and their respective roles in global ecosystems.
In describing a complex system like the biosphere, it is critical to
quantify the abundance of individual components of the system
(i.e., species, broader taxonomic groups). A quantitative de-
scription of the distribution of biomass is essential for taking
stock of biosequestered carbon (1) and modeling global bio-
geochemical cycles (2), as well as for understanding the historical
effects and future impacts of human activities.
Earlier efforts to estimate global biomass have mostly focused
on plants (3–5). In parallel, a dominant role for prokaryotic
biomass has been advocated in a landmark paper by Whitman
et al. (6) entitled “Prokaryotes: The unseen majority.”New
sampling and detection techniques (7, 8) make it possible to re-
visit this claim. Likewise, for other taxa, such as fish, recent global
sampling campaigns (9) have resulted in updated estimates, often
differing by an order of magnitude or more from previous esti-
mates. For groups such as arthropods, global estimates are still
lacking (10, 11).
All of the above efforts are each focused on a single taxon. We
are aware of only two attempts at a comprehensive accounting of
all biomass components on Earth: Whittaker and Likens (12)
made a remarkable effort in the early 1970s, noting even then that
their study was “intended for early obsolescence.”It did not in-
clude, for example, bacterial or fungal biomass. The other at-
tempt, by Smil (13), was included as a subsection of a book
intended for a broad readership. His work details characteristic
values for the biomass of various taxa in many environments. Fi-
nally, Wikipedia serves as a highly effective platform for making
accessible a range of estimates on various taxa (https://en.wikipedia.
org/wiki/Biomass_(ecology)#Global_biomass) but currently falls
short of a comprehensive or integrated view.
In the past decade, several major technological and scientific
advances have facilitated an improved quantitative account of
the biomass on Earth. Next-generation sequencing has enabled a
more detailed and cultivation-independent view of the compo-
sition of natural communities based on the relative abundance of
genomes (14). Better remote sensing tools enable us to probe the
environment on a global scale with unprecedented resolution
and specificity. The Tara Oceans expedition (15) is among recent
efforts at global sampling that are expanding our view and cov-
erage. Continental counterpart efforts, such as the National
Ecological Observatory Network in North America, add more
finely resolved, continent-specific details, affording us more ro-
bust descriptions of natural habitats.
Here, we either assemble or generate estimates of the biomass
for each of the major taxonomic groups that contribute to the
global biomass distribution. Our analysis (described in detail in SI
Appendix) is based on hundreds of studies, including recent studies
that have overturned earlier estimates for many taxa (e.g., fish,
subsurface prokaryotes, marine eukaryotes, soil fauna).
Results
The Biomass Distribution of the Biosphere by Kingdom. In Fig. 1 and
Table 1, we report our best estimates for the biomass of each
taxon analyzed. We use biomass as a measure of abundance,
which allows us to compare taxa whose members are of very
different sizes. Biomass is also a useful metric for quantifying
stocks of elements sequestered in living organisms. We report
biomass using the mass of carbon, as this measure is independent
of water content and has been used extensively in the literature
(6, 16, 17). Alternative measures for biomass, such as dry weight,
are discussed in Materials and Methods. For ease of discussion,
we report biomass in gigatons of carbon, with 1 Gt C =10
15
gof
carbon. We supply additional estimates for the number of indi-
viduals of different taxa in SI Appendix, Table S1.
Significance
The composition of the biosphere is a fundamental question in
biology, yet a global quantitative account of the biomass of
each taxon is still lacking. We assemble a census of the biomass
of all kingdoms of life. This analysis provides a holistic view of
the composition of the biosphere and allows us to observe
broad patterns over taxonomic categories, geographic loca-
tions, and trophic modes.
Author contributions: Y.M.B.-O., R.P., and R.M. designed research; Y.M.B.-O. and R.M.
performed research; Y.M.B.-O. and R.M. an alyzed data; and Y.M.B.-O., R.P., and R.M.
wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This open access article is distributed under Creative Commons Attribution-NonCommercial-
NoDeriv atives L icense 4.0 ( CC BY-N C-ND).
Data deposition: All of the data used to generate our estimates, as well as the code used
for analysis, are available on GitHub at https://github.com/milo-lab/biomass_distribution.
1
To whom correspondence should be addressed. Email: ron.milo@weizmann.ac.il.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1711842115/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1711842115 PNAS Latest Articles
|
1of6
SYSTEMS BIOLOGY

The sum of the biomass across all taxa on Earth is ≈550 Gt C,
of which ≈80% (≈450 Gt C; SI Appendix, Table S2) are plants,
dominated by land plants (embryophytes). The second major
biomass component is bacteria (≈70 Gt C; SI Appendix, Tables
S3–S7), constituting ≈15% of the global biomass. Other groups,
in descending order, are fungi, archaea, protists, animals, and
viruses, which together account for the remaining <10%. Despite
the large uncertainty associated with the total biomass of bac-
teria, we estimate that plants are the dominant kingdom in terms
of biomass at an ≈90% probability (more details are provided in
the SI Appendix). Aboveground biomass (≈320 Gt C) represents
≈60% of global biomass, with belowground biomass composed
mainly of plant roots (≈130 Gt C) and microbes residing in the
soil and deep subsurface (≈100 Gt C). Plant biomass includes
≈70% stems and tree trunks, which are mostly woody, and thus
relatively metabolically inert. Bacteria include about 90% deep
subsurface biomass (mostly in aquifers and below the seafloor),
which have very slow metabolic activity and associated turnover
times of several months to thousands of years (18–22). Excluding
these contributions, global biomass is still dominated by plants
(SI Appendix, Fig. S1), mostly consisting of ≈150 Gt C of plant
roots and leaves and ≈9 Gt C of terrestrial and marine bacteria
whose contribution is on par with the ≈12 Gt C of fungi (SI
Appendix, Table S8).
Whereas groups like insects dominate in terms of species
richness [with about 1 million described species (23)], their
relative biomass fraction is miniscule. Some species contrib-
ute much more than entire families or even classes. For ex-
ample, the Antarctic krill species Euphausia superba contributes
≈0.05 Gt C to global biomass (24), similar to other prominent
species such as humans or cows. This value is comparable to
the contribution from termites (25), which contain many spe-
cies, and far surpasses the biomass of entire vertebrate classes
such as birds. In this way, the picture that arises from taking
a biomass perspective of the biosphere complements the fo-
cus on species richness that is commonly held (SI Appendix,
Fig. S3).
The Uncertainty Associated with Global Biomass Estimates. The
specific methods used for each taxon are highly diverse and are
given in detail in the SI Appendix, along with data sources.
Global biomass estimates vary in the amount of information they
are based on and, consequently, in their uncertainty. An estimate
of relatively high certainty is that of plants, which is based on
several independent sources. One of these is the Forest Re-
source Assessment, a survey on the state of world forests con-
ducted by the international Food and Agriculture Organization
(FAO). The assessment is based on a collection of country re-
ports that detail the area and biomass density of forests in each
country (26) using a standardized format and methodology. The
FAO also keeps a record of nonforest ecosystems, such as sa-
vannas and shrublands, in each country. Alternatively, remote
sensing data give high coverage of measurements that indicate
AB
Fig. 1. Graphical representation of the global biomass distribution by taxa. (A) Absolute biomasses of different taxa are represented using a Voronoi di-
agram, with the area of each cell being proportional to that taxa global biomass (the specific shape of each polygon carries no meaning). This type of vi-
sualization is similar to pie charts but has a much higher dynamic range (a comparison is shown in SI Appendix, Fig. S4). Values are based on the estimates
presented in Table 1 and detailed in the SI Appendix. A visual depiction without components with very slow metabolic activity, such as plant stems and tree
trunks, is shown in SI Appendix, Fig. S1.(B) Absolute biomass of different animal taxa. Related groups such as vertebrates are located next to each other. We
estimate that the contribution of reptiles and amphibians to the total animal biomass is negligible, as we discuss in the SI Appendix. Visualization performed
using the online tool at bionic-vis.biologie.uni-greifswald.de/.
Table 1. Summary of estimated total biomass for abundant
taxonomic groups
Taxon Mass (Gt C) Uncertainty (-fold)
Plants 450 1.2
Bacteria 70 10
Fungi 12 3
Archaea 7 13
Protists 4 4
Animals 2 5
Arthropods, terrestrial 0.2
Arthropods, marine 1
Chordates, fish 0.7
Chordates, livestock 0.1
Chordates, humans 0.06
Chordates, wild mammals 0.007
Chordates, wild birds 0.002
Annelids 0.2
Molluscs 0.2
Cnidarians 0.1
Nematodes 0.02
Viruses 0.2 20
Total 550 1.7
Values are based on an extensive literature survey and data integration as
detailed in the SI Appendix. Reported values have been rounded to reflect
the associated level of uncertainty. We report an uncertainty projection for
each kingdom as a fold-change factor from the mean, representing a range
akin to a 95% confidence interval of the estimate. The procedure for de-
riving these projections is documented in detail in Materials and Methods
and SI Appendix.
2of6
|
www.pnas.org/cgi/doi/10.1073/pnas.1711842115 Bar-On et al.

plant biomass (27–29). Remote sensing is used to measure, for
example, the height of trees or the number of tree stems per unit
area. Biomass is inferred by field measurements establishing a
connection between tree plant biomass and satellite-based re-
mote sensing measurements. Combining data from independent
sources such as these enables a robust assessment of the total
plant biomass (17).
A more characteristic case with larger uncertainties is exem-
plified by marine prokaryotes, where cell concentrations are
measured in various locations and binned based on depth. For
each depth range, the average cell concentration is calculated
and the total number of marine prokaryotes is estimated through
multiplication by the water volume in each depth range. The
total number of cells is converted to biomass by using the char-
acteristic carbon content per marine prokaryote. In cases where
there are fewer measurements (e.g., terrestrial arthropods, ter-
restrial protists), the possibility of systematic biases in the estimate
is greater and the uncertainty larger. To test the robustness of
our estimates, we used independent approaches and analyzed
the agreement between such independent estimates. Details
on the specific methodologies used for each taxon are provided in
the SI Appendix. Because most datasets used to estimate global
biomass rely on fragmentary sampling, we project large uncer-
tainties that will be reduced as additional data become available.
The Impact of Humanity on the Biosphere. Over the relatively short
span of human history, major innovations, such as the domesti-
cation of livestock, adoption of an agricultural lifestyle, and the
Industrial Revolution, have increased the human population
dramatically and have had radical ecological effects. Today, the
biomass of humans (≈0.06 Gt C; SI Appendix, Table S9) and the
biomass of livestock (≈0.1 Gt C, dominated by cattle and pigs; SI
Appendix, Table S10) far surpass that of wild mammals, which
has a mass of ≈0.007 Gt C (SI Appendix, Table S11). This is also
true for wild and domesticated birds, for which the biomass of
domesticated poultry (≈0.005 Gt C, dominated by chickens) is
about threefold higher than that of wild birds (≈0.002 Gt C; SI
Appendix, Table S12). In fact, humans and livestock outweigh all
vertebrates combined, with the exception of fish. Even though
humans and livestock dominate mammalian biomass, they are a
small fraction of the ≈2 Gt C of animal biomass, which primarily
comprises arthropods (≈1GtC;SI Appendix, Tables S13 and
S14), followed by fish (≈0.7 Gt C; SI Appendix, Table S15).
Comparison of current global biomass with prehuman values
(which are very difficult to estimate accurately) demonstrates the
impact of humans on the biosphere. Human activity contributed
to the Quaternary Megafauna Extinction between ≈50,000 and
≈3,000 y ago, which claimed around half of the large (>40 kg)
land mammal species (30). The biomass of wild land mammals
before this period of extinction was estimated by Barnosky (30)
at ≈0.02 Gt C. The present-day biomass of wild land mammals is
approximately sevenfold lower, at ≈0.003 Gt C (SI Appendix,Pre-
human Biomass and Chordates and Table S11). Intense whaling
and exploitation of other marine mammals have resulted in an
approximately fivefold decrease in marine mammal global bio-
mass [from ≈0.02 Gt C to ≈0.004 Gt C (31)]. While the total
biomass of wild mammals (both marine and terrestrial) de-
creased by a factor of ≈6, the total mass of mammals increased
approximately fourfold from ≈0.04 Gt C to ≈0.17 Gt C due to
the vast increase of the biomass of humanity and its associated
livestock. Human activity has also impacted global vertebrate
stocks, with a decrease of ≈0.1 Gt C in total fish biomass, an
amount similar to the remaining total biomass in fisheries and to
the gain in the total mammalian biomass due to livestock hus-
bandry (SI Appendix, Pre-human Biomass). The impact of human
civilization on global biomass has not been limited to mammals
but has also profoundly reshaped the total quantity of carbon
sequestered by plants. A worldwide census of the total number of
trees (32), as well as a comparison of actual and potential plant
biomass (17), has suggested that the total plant biomass (and, by
proxy, the total biomass on Earth) has declined approximately
twofold relative to its value before the start of human civilization.
The total biomass of crops cultivated by humans is estimated at
≈10 Gt C, which accounts for only ≈2% of the extant total plant
biomass (17).
The Distribution of Biomass Across Environments and Trophic Modes.
Examining global biomass in different environments exposes
stark differences between terrestrial and marine environments.
The ocean covers 71% of the Earth’s surface and occupies a
much larger volume than the terrestrial environment, yet land
biomass, at ≈470 Gt C, is about two orders of magnitude higher
than the ≈6 Gt C in marine biomass, as shown in Fig. 2A. Even
though there is a large difference in the biomass content of the
terrestrial and marine environments, the primary productivity of
the two environments is roughly equal (33). For plants, we find
that most biomass is concentrated in terrestrial environments
(plants have only a small fraction of marine biomass, <1GtC,in
the form of green algae and seagrass; Fig. 2B). For animals, most
biomass is concentrated in the marine environment, and for
bacteria and archaea, most biomass is concentrated in deep
subsurface environments. We note that several of the results in
Fig. 2Bshould be interpreted with caution due to the large un-
certainty associated with some of the estimates, mostly those of
total terrestrial protists, marine fungi, and contributions from
deep subsurface environments.
When analyzing trophic levels, the biomass of primary pro-
ducers on land is much larger than that of primary and secondary
consumers. In stark contrast, in the oceans, ≈1 Gt C of primary
producers supports ≈5 Gt C of consumer biomass, resulting in an
inverted standing biomass distribution as shown in Fig. 2C.Such
inverted biomass distributions can occur when primary producers
have a rapid turnover of biomass [on the order of days (34)], while
consumer biomass turns over much more slowly [a few years in the
case of mesopelagic fish (35)]. Thus, the standing stock of con-
sumers is larger, even though the productivity of producers is
necessarily higher. Previous reports have observed inverted bio-
mass pyramids in local marine environments (36, 37). An addi-
tional study noted an inverted consumer/producer ratio for the
global plankton biomass (16). Our analysis suggests that these
observations hold true when looking at the global biomass of all
producers and consumers in the marine environment.
Discussion
Our census of the distribution of biomass on Earth provides an
integrated global picture of the relative and absolute abundances
of all kingdoms of life. We find that the biomass of plants
dominates the biomass of the biosphere and is mostly located on
land. The marine environment is primarily occupied by microbes,
mainly bacteria and protists, which account for ≈70% of the total
marine biomass. The remaining ≈30% is mainly composed of
arthropods and fish. The deep subsurface holds ≈15% of the
total biomass in the biosphere. It is chiefly composed of bacteria
and archaea, which are mostly surface-attached and turn over
their biomass every several months to thousands of years (18–22).
In addition to summarizing current knowledge of the global
biomass distribution, our work highlights gaps in the current
understanding of the biosphere. Our knowledge of the biomass
composition of different taxa is mainly determined by our ability
to sample their biomass in the wild. For groups such as plants,
the use of multiple sources to estimate global biomass increases
our confidence in the validity of current estimates. However, for
other groups, such as terrestrial arthropods and protists, quan-
titative sampling of biomass is limited by technical constraints,
and comprehensive data are thus lacking. Beyond specific taxa,
there are entire environments for which our knowledge is very
Bar-On et al. PNAS Latest Articles
|
3of6
SYSTEMS BIOLOGY

limited, namely, the deep subsurface environments such as deep
aquifers and the ocean’s crust, which might hold the world
largest aquifer (38). Studies in these environments are scarce,
meaning that our estimates have particularly high uncertainty
ranges and unknown systematic biases. Main gaps in our knowl-
edge of these environments pertain to the distribution of biomass
between the aquifer fluids and the surrounding rocks and the
distribution of biomass between different microbial taxa, such
as bacteria, archaea, protists, and fungi. Scientists have closely
monitored the impact of humans on global biodiversity (39–41),
but less attention has been given to total biomass, resulting in
high uncertainty regarding the impact of humanity on the bio-
mass of vertebrates. Our estimates for the current and pre-
human biomasses of vertebrates are only a crude first step in
calculating these values (SI Appendix,Prehuman Biomass). The
biomass of amphibians, which are experiencing a dramatic
population decline (42), remains poorly characterized. Future
research could reduce the uncertainty of current estimates by
sampling more environments, which will better represent the
diverse biosphere on Earth. In the case of prokaryotes, some
major improvements were recently realized, with global esti-
mates of marine deep subsurface prokaryote biomass reduced
by about two orders of magnitude due to an increased diversity
of sampling locations (7).
Identifying gaps in our knowledge could indicate areas for
which further scientific exploration could have the biggest impact
on our understanding of the biosphere. As a concrete example,
we identify the ratio between attached to unattached cells in the
deep aquifers as a major contributor to the uncertainties asso-
ciated with our estimate of the biomass of bacteria, archaea, and
viruses. Improving our understanding of this specific parameter
could help us better constrain the global biomasses of entire
domains of life. In addition to improving our reported estimates,
future studies can achieve a finer categorization of taxa. For
example, the biomass of parasites, which is not resolved from
their hosts in this study, might be larger than the biomass of top
predators in some environments (43).
By providing a unified, updated, and accessible global view of
the biomass of different taxa, we also aim to disseminate knowl-
edge of the biosphere composition to a wide range of students and
researchers. Our survey puts into perspective claims regarding the
overarching dominance of groups such as termites and ants (44),
nematodes (45), and prokaryotes (6). For example, the biomass of
termites [≈0.05 Gt C (25)] is on par with that of humans but is still
around an order of magnitude smaller than that of other taxa,
such as fish (≈0.7 Gt C; SI Appendix, Table S15). Other groups,
such as nematodes, surpass any other animal species in terms of
number of individuals (SI Appendix, Fig. S2) but constitute only
about 1% of the total animal biomass.
The census of biomass distribution on Earth presented here is
comprehensive in scope and based on synthesis of data from the
recent scientific literature. The integrated dataset enables us to
draw basic conclusions concerning kingdoms that dominate the
biomass of the biosphere, the distribution of biomass of each
kingdom across different environments, and the opposite structures
of the global marine and terrestrial biomass pyramids. We identify
areas in which current knowledge is lacking and further research is
most required. Ideally, future research will include both temporal
and geographic resolution. We believe that the results described
in this study will provide students and researchers with a holistic
quantitative context for studying our biosphere.
Materials and Methods
Taxon-Specific Detailed Description of Data Sources and Procedures for
Estimating Biomass. The complete account of the data sources used for es-
timating the biomass of each taxon, procedures for estimating biomass, and
projections for the uncertainty associated with the estimate for the biomass of
each taxon are provided in the SI Appendix. To make the steps for estimating
the biomass of each taxon more accessible, we provide supplementary tables
that summarize the procedure as well as online notebooks for the calculation
of the biomass of each taxon (see data flow scheme in SI Appendix, Overview).
In Table 1, we detail the relevant supplementary table that summarizes the
steps for arriving at each estimate. All of the data used to generate our esti-
mates, as well as the code used for analysis, are open-sourced and available at
https://github.com/milo-lab/biomass_distribution.
Choice of Units for Measuring Biomass. Biomass is reported in gigatons of
carbon. Alternative options to represent biomass include, among others,
biovolume, wet mass, or dry weight. We chose to use carbon mass as the
measure of biomass because it is independent of water content and is used
extensively in the literature. Dry mass also has these features but is used less
frequently. All of our reported values can be transformed to dry weight to a
good approximation by multiplying by 2, the characteristic conversion factor
between carbon and total dry mass (46–48).
We report the significant digits for our values throughout the paper using
the following scheme: For values with an uncertainty projection that is higher
than twofold, we report a single significant digit. For values with an un-
certainty projection of less than twofold, we report two significant digits. In
cases when we report one significant digit, we do not consider a leading “1”
as a significant digit.
A
C
B
Fig. 2. Biomass distributions across different environments and trophic modes. (A) Absolute biomass is represented using a Voronoi diagram, with the area
of each cell being proportional to the global biomass at each environment. Values are based on SI Appendix, Table S23. We define deep subsurface as the
marine subseafloor sediment and the oceanic crust, as well as the terrestrial substratum deeper than 8 m, excluding soil (6). (B) Fraction of the biomass of each
kingdom concentrated in the terrestrial, marine, or deep subsurface environment. For fungi and protists, we did not estimate the biomass present in the deep
subsurface due to data scarcity. (C) Distribution of biomass between producers (autotrophs, mostly photosynthetic) and consumers (heterotrophs without
deep subsurface) in the terrestrial and marine environments. The size of the bars corresponds to the quantity of biomass of each trophic mode. Numbers are
in gigatons of carbon.
4of6
|
www.pnas.org/cgi/doi/10.1073/pnas.1711842115 Bar-On et al.

General Framework for Estimating Global Biomass. In achieving global esti-
mates, there is a constant challenge of how to move from a limited set of local
samples to a representative global value. How does one estimate global
biomass based on a limited set of local samples? For a crude estimate, the
average of all local values of biomass per unit area is multiplied by the total
global area. A more effective estimate can be made by correlating measured
values to environmental parameters that are known at a global scale (e.g.,
temperature, depth, distance from shore, primary productivity, biome type),
as shown in Fig. 3. This correlation is used to extrapolate the biomass of a
taxon at a specific location based on the known distribution of the envi-
ronmental parameter (e.g., the temperature at each location on the globe).
By integrating across the total surface of the world, a global estimate is
derived. We detail the specific extrapolation procedure used for each taxon
in both the SI Appendix and supplementary tables (SI Appendix, Tables S1–
S23). For most taxa, our best estimates are based on a geometric mean of
several independent estimates using different methodologies. The geo-
metric mean estimates the median value if the independent estimates are
log-normally distributed or, more generally, the distribution of estimates is
symmetrical in log space.
Uncertainty Estimation and Reporting. Global estimates such as those we use in
the present work are largely based on sampling from the distribution of
biomass worldwide and then extrapolating for areas in which samples are
missing. The sampling of biomass in each location can be based on direct
biomass measurements or conversion to biomass from other types of mea-
surement, such as number of individuals and their characteristic weight. Som e
of the main sources of uncertainty for the estimates we present are the result
of using such geographical extrapolations and conversion from number of
individuals to overall biomass. The certainty of the estimate is linked to the
amount of sampling on which the estimate is based. Notable locations in
which sampling is scarce are the deep ocean (usually deeper than 200 m) and
deep layers of soil (usually deeper than 1 m). For some organisms, such as
annelids and marine protists and arthropods, most estimates neglect these
environments, thus underestimating the actual biomass. Sampling can be
biased toward places that have high abundance and diversity of wildlife.
Relying on data with such sampling bias can cause overestimation of the
actual biomass of a taxon.
Another source of uncertainty comes from conversion to biomass. Conver-
sion from counts of individuals to biomass is based on either known average
weights per individual (e.g., 50 kg of wet weight for a human, which averages
over adults and children, or 10 mg of dry weight for a “characteristic”
earthworm) or empirical allometric equations that are organism-specific, such
as conversion from animal length to biomass. When using such conversion
methods, there is a risk of introducing biases and noise into the final estimate.
Neverthele ss, there is often no way around using such conversions. As such, we
must be aware that the data may contain such biases.
In addition to describing the procedures leading to the estimate of each
taxon, we quantitatively survey the main sources of uncertainty associated
with each estimate and calculate an uncertainty range for each of our bio-
mass estimates. We choose to report uncertainties as representing, to the best
of our ability given the many constraints, what is equivalent to a 95%
confidence interval for the estimate of the mean. Uncertainties reported in
our analysis are multiplicative (fold change from the mean) and not additive
(±change of the estimate). We chose to use multiplicative uncertainty as it is
more robust to large fluctuations in estimates, and because it is in accord
with the way we generate our best estimates, which is usually by using a
geometric mean of different independent estimates. Our uncertainty pro-
jections are focused on the main kingdoms of life: plants, bacteria, archaea,
fungi, protists, and animals.
The general framework for constructing our uncertainties (described in
detail for each taxon in the SI Appendix and in the online notebooks) takes
into account both intrastudy uncertainty and interstudy uncertainty. Intra-
study uncertainty refers to uncertainty estimates reported within a specific
study, whereas interstudy uncertainty refers to variation in estimates of a
certain quantity between different papers. In many cases, we use several
independent methodologies to estimate the same quantity. In these cases,
we can also use the variation between estimates from each methodology as
a measure of the uncertainty of our final estimate. We refer to this type of
uncertainty as intermethod uncertainty. The way we usually calculate un-
certainties is by taking the logarithm of the values reported either within
studies or from different studies. Taking the logarithm moves the values to
log-space, where the SE is calculated (by dividing the SD by the square root
of the number of values). We then multiply the SE by a factor of 1.96 (which
would give the 95% confidence interval if the transformed data were nor-
mally distributed). Finally, we exponentiate the result to get the multiplicative
factor in linear space that represents the confidence interval (akin to a 95%
confidence interval if the data were log-normally distributed).
Most of our estimates are constructed by combining several different
estimates (e.g., combining total number of individuals and characteristic
carbon content of a single organism). In these cases, we use intrastudy,
interstudy, or intermethod variation associated with each parameter that is
used to derive the final estimate and propagate these uncertainties to the
final estimate of biomass. The uncertainty analysis for each specific bio-
mass estimate incorporates different components of this general scheme,
depending on the amount of information that is available, as detailed on a
case-by-case basis in the SI Appendix.
In cases where information is ample, the procedure described above yields
several different uncertainty estimates for each parameter that we use to
derive the final estimate (e.g., intrastudy uncertainty, interstudy uncertainty).
We integrate these different uncertainties, usually by taking the highest
value as the best projection of uncertainty. In some cases, for example, when
information is scarce or some sources of uncertainty are hard to quantify, we
base our estimates on the uncertainty in analogous taxa and consultation
with relevant experts. We tend to round up our uncertainty projections when
data are especially limited.
Taxonomic Levels Used. Our census gives estimates for the global biomass at
various taxonomic levels. Our main results relate to the kingdom level: animals,
archaea, bacteria, fungi, plants, and protists. Although the division into
kingdoms is not the most contemporary taxonomic grouping that exists, we
chose to use it forthe current analysisas most of the data we rely upon doesnot
provide finer taxonomic details (e.g., the division of terrestrial protists is mainly
based on morphology and not on taxonomy). We supplement these kingdoms
of living organisms with an estimate for the global biomass of viruses, which
are not included in the current tree of life but play a key role in global bio-
geochemical cycles (49). For all kingdoms except animals, all taxa making up
the kingdom are considered together. For estimating the biomass of animals,
we use a bottom-up approach, which estimates the biomass of key phyla
constituting the animal kingdom. The sum of the biomass of these phyla
represents our estimate of the total biomass of animals. We give estimates for
most phyla and estimate bounds for the possible biomass contribution for the
remaining phyla (SI Appendix,Other Animal Phyla). Within chordates, we
provide estimates for key classes, such as fish, mammals, and birds. We esti-
mate that the contribution of reptiles and amphibians to the total chordate
biomass is negligible, as we discuss in the SI Appendix. We divide the class of
mammals into wild mammals and humans plus livestock (without a contri-
bution from poultry, which is negligible compared with cattle and pigs). Even
though livestock is not a valid taxonomic division, we use it to consider the
impact of humans on the total biomass of mammals.
Fig. 3. General framework for estimating global biomass. The procedure
begins with local samples of biomass across the globe. The more representa-
tive the samples are of the natural distribution of the taxon biomass, the more
accurate the estimate will be. To move from localsamples to a global estimate,
a correlation between local biomass densities and an environmental param-
eter (or parameters) is established. Based on this correlation, in addition to our
knowledge of the distribution of the environmental parameter, we extrapo-
late the biomass across the entire globe. The resolution of the resulting bio-
mass distribution map is dependent on the resolution at which we know the
environmental parameter. Integrating across the entire surface of the Earth,
we get a global estimate of the biomass of the taxon.
Bar-On et al. PNAS Latest Articles
|
5of6
SYSTEMS BIOLOGY

ACKNOWLEDGMENTS. We thank Shai Meiri for help with estimating the
biomass of wild mammals, birds, and reptiles and Arren Bar-Even, Oded Beja,
Jorg Bernhardt, Tristan Biard, Chris Bowler, Nuno Carvalhais, Otto Coredero,
Gidon Eshel, Ofer Feinerman, Noah Fierer, Daniel Fisher, Avi Flamholtz, Assaf
Gal, José Grünzweig, Marcel van der Heijden, Dina Hochhauser, Julie Huber,
Qusheng Jin, Bo Barker Jørgensen, Jens Kallmeyer, Tamir Klein, Christian
Koerner, Daniel Madar, Fabrice Not, Katherine O’Donnell, Gal Ofir, Victoria
Orphan, Noam Prywes, John Raven, Dave Savage, Einat Segev, Maya Shamir,
Izak Smit, Rotem Sorek, Ofer Steinitz, Miri Tsalyuk, Assaf Vardi, Colomban de
Vargas, Joshua Weitz, Yossi Yovel, Yonatan Zegman, and two anonymous
reviewers for productive feedback on this manuscript. This research was sup-
ported by the European Research Council (project NOVCARBFIX 646827), the
Israel Science Foundation (Grant 740/16), the ISF-NRF Singapore Joint Research
Program (Grant 76627 12), the B eck Canadian Center f or Alternative Energ y
Research, Dana and Yossie Hollander, the Ullmann Family Foundation, the
Helmsley Charitable Foundation, the Larson Charitable Foundation, the Wolf-
son Family Charitable Tru st, Charles Rothschild, and Selmo Nussenbaum. This
study was also supported by the NIH through Grant 1R35 GM118043-01
(MIRA). R.M. is the Charles and Louise Gartner Professional Chair.
1. Carvalhais N, et al. (2014) Global covariation of carbon turnover times with climate in
terrestrial ecosystems. Nature 514:213–217.
2. Quere CL, et al. (2005) Ecosystem dynamics based on plankton functional types for
global ocean biogeochemistry models. Glob Change Biol 11:2016–2040.
3. Vernadsky VI (1998) Photosynthetic living matter. The Biosphere (Copernicus, New
York, NY), pp 72–84.
4. Saugier B, Roy J, Mooney HA (2001) Estimations of global terrestrial productivity:
Converging toward a singlenumber. Terrestrial Global Productivity, eds Roy J, Saugier B,
Mooney HA (Academic, San Diego), pp 543–557.
5. Atjay GL, Ketner P, Duvigneaud P (1979) The Global Carbon Cycle, SCOPE Report 13,
eds Bolin B, Degens ET, Kempe S, Ketner P (Wiley, Chichester, UK), pp 129–181.
6. Whitman WB, Coleman DC, Wiebe WJ (1998) Prokaryotes: The unseen majority. Proc
Natl Acad Sci USA 95:6578–6583.
7. Kallmeyer J, Pockalny R, Adhikari RR, Smith DC, D’Hondt S (2012) Global distribution
of microbial abundance and biomass in subseafloor sediment. Proc Natl Acad Sci USA
109:16213–16216.
8. McMahon S, Parnell J (2014) Weighing the deep continental biosphere. FEMS
Microbiol Ecol 87:113–120.
9. Irigoien X, et al. (2014) Large mesopelagic fishes biomass and trophic efficiency in the
open ocean. Nat Commun 5:3271.
10. Williams CB (1960) The range and pattern of insect abundance. Am Nat 94:137–151.
11. Petersen H, Luxton M (1982) A comparative analysis of soil fauna populations and
their role in decomposition processes. Oikos 39:288–388.
12. Whittaker RH, Likens GE (1973) Carbon in the biota. Brookhaven Symp Biol 281–302.
13. Smil V (2013) Harvesting the Biosphere: What We Have Taken from Nature (MIT Press,
Cambridge, MA).
14. Schloss PD, Handelsman J (2006) Toward a census of bacteria in soil. PLoS Comput Biol
2:e92.
15. Bork P, et al. (2015) Tara Oceans. Tara Oceans studies plankton at planetary scale.
Introduction. Science 348:873.
16. Buitenhuis ET, et al. (2013) MAREDAT: Towards a world atlas of MARine ecosystem
DATa. Earth Syst Sci Data 5:227–239.
17. Erb K-H, et al. (2017) Unexpectedly large impact of forest management and grazing
on global vegetation biomass. Nature 553:73–76.
18. D’Hondt S, Rutherford S, Spivack AJ (2002) Metabolic activity of subsurface life in
deep-sea sediments. Science 295:2067–2070.
19. Jørgensen BB (2011) Deep subseafloor microbial cells on physiological standby. Proc
Natl Acad Sci USA 108:18193–18194.
20. Jørgensen BB, Boetius A (2007) Feast and famine–Microbial life in the deep-sea bed.
Nat Rev Microbiol 5:770–781.
21. Braun S, et al. (2017) Microbial turnover times in the deep seabed studied by amino
acid racemization modelling. Sci Rep 7:5680.
22. Trembath-Reichert E, et al. (2017) Methyl-compound use and slow growth charac-
terize microbial life in 2-km-deep subseafloor coal and shale beds. Proc Natl Acad Sci
USA 114:E9206–E9215.
23. Chapman AD (2009) Numbers of Living Species in Australia and the World (Australian
Biological Resources Study, Canberra, Australia).
24. Atkinson A, Siegel V, Pakhomov EA, Jessopp MJ, Loeb V (2009) A re-appraisal of the
total biomass and annual production of Antarctic krill. Deep Sea Res Part I Oceanogr
Res Pap 56:727–740.
25. Sanderson MG (1996) Biomass of termites and their emissions of methane and carbon
dioxide: A global database. Global Biogeochem Cycles 10:543–557.
26. MacDicken KG (2015) Global Forest Resources Assessment 2015: What, why and how?
For Ecol Manage 352:3–8.
27. Saatchi SS, et al. (2011) Benchmark map of forest carbon stocks in tropical regions
across three continents. Proc Natl Acad Sci USA 108:9899–9904.
28. Thurner M, Beer C (2014) Carbon stock and density of northern boreal and temperate
forests. Glob Ecol Biogeogr 23:297–310.
29. Baccini A, et al. (2012) Estimated carbon dioxide emissions from tropical deforestation
improved by carbon-density maps. Nat Clim Change 2:182–185.
30. Barnosky AD (2008) Colloquium paper: Megafauna biomass tradeoff as a driver of
Quaternary and future extinctions. Proc Natl Acad Sci USA 105:11543–11548.
31. Christensen LB (2006) Marine mammal populations: Reconstructing historical abun-
dances at the global scale. Fisheries Centre Research Reports 14.
32. Crowther TW, et al. (2015) Mapping tree density at a global scale. Nature 525:
201–205.
33. Field CB, Behrenfeld MJ, Randerson JT, Falkowski P (1998) Primary production of the
biosphere: Integrating terrestrial and oceanic components. Science 281:237–240.
34. Zubkov MV (2014) Faster growth of the major prokaryotic versus eukaryotic
CO2 fixers in the oligotrophic ocean. Nat Commun 5:3776.
35. Catul V, Gauns M, Karuppasamy PK (2011) A review on mesopelagic fishes belonging
to family Myctophidae. Rev Fish Biol Fish 21:339–354.
36. Dortch Q, Packard TT (1989) Differences in biomass structure between oligotrophic
and eutrophic marine ecosystems. Deep Sea Res A 36:223–240.
37. Gasol JM, del Giorgio PA, Duarte CM (1997) Biomass distribution in marine planktonic
communities. Limnol Oceanogr 42:1353–1363.
38. Johnson HP, Pruis MJ (2003) Fluxes of fluid and heat from the oceanic crustal reser-
voir. Earth Planet Sci Lett 216:565–574.
39. Newbold T, et al. (2015) Global effects of land use on local terrestrial biodiversity.
Nature 520:45–50.
40. Ceballos G, Ehrlich PR, Dirzo R (2017) Biological annihilation via the ongoing sixth
mass extinction signaled by vertebrate population losses and declines. Proc Natl Acad
Sci USA 114:E6089–E6096.
41. Isbell F, et al. (2017) Linking the influence and dependence of people on biodiversity
across scales. Nature 546:65–72.
42. Wake DB, Vredenburg VT (2008) Colloquium paper: Are we in the midst of the sixth
mass extinction? A view from the world of amphibians. Proc Natl Acad Sci USA 105:
11466–11473.
43. Kuris AM, et al. (2008) Ecosystem energetic implications of parasite and free-living
biomass in three estuaries. Nature 454:515–518.
44. Schultz TR (2000) In search of ant ancestors. Proc Natl Acad Sci USA 97:14028–14029.
45. Wilson EO (2003) The encyclopedia of life. Trends Ecol Evol 18:77–80.
46. Elser JJ, et al. (2000) Nutritional constraints in terrestrial and freshwater food webs.
Nature 408:578–580.
47. Ferris H, Venette RC, Lau SS (1997) Population energetics of bacterial-feeding nem-
atodes: Carbon and nitrogen budgets. Soil Biol Biochem 29:1183–1194.
48. Sohlenius B, Bjorn S (1979) A carbon budget for nematodes, rotifers and tardigrades
in a Swedish coniferous forest soil. Ecography 2:30–40.
49. Jover LF, Effler TC, Buchan A, Wilhelm SW, Weitz JS (2014) The elemental composition
of virus particles: Implications for marine biogeochemical cycles. Nat Rev Microbiol 12:
519–528.
6of6
|
www.pnas.org/cgi/doi/10.1073/pnas.1711842115 Bar-On et al.