of 26
Descriptor
Anthroponumbers.org: A quantitative database of
human impacts on Planet Earth
Graphical abstract
Highlights
d
We present a holistic view of the many ways humans alter
Earth at a global scale
d
We consider how these global quantities vary across
geography
d
We further explore the time- and population-dependent
dynamics of these impacts
d
We enumerate and describe key properties associated with
each entry in the database
Authors
Griffin Chure, Rachel A. Banks,
Avi I. Flamholz, ..., Yinon Bar-On,
Ron Milo, Rob Phillips
Correspondence
griffinchure@gmail.com (G.C.),
phillips@pboc.caltech.edu (R.P.)
In brief
The environmental impacts of human
action on Earth are being felt on many
fronts. Despite our deep knowledge of
these impacts, finding reliable
quantitative information is burdensome,
often requiring domain expertise and
programmatic acumen. We present the
Human Impacts Database, which houses
a diverse array of quantities regarding
human impacts, making them easily
accessible and searchable. We use this
database to present a broad view of the
Anthropocene, exploring the global
magnitudes, spatial dependence, and
temporal dynamics of human impacts.
Chure et al., 2022, Patterns
3
, 100552
September 9, 2022
ª
2022 The Author(s).
https://doi.org/10.1016/j.patter.2022.100552
ll
Descriptor
Anthroponumbers.org: A quantitative database
of human impacts on Planet Earth
Griffin Chure,
1
,
2
,
10
,
11
,
*
Rachel A. Banks,
3
,
4
,
5
,
10
Avi I. Flamholz,
3
,
4
Nicholas S. Sarai,
6
Mason Kamb,
5
Ignacio Lopez-Gomez,
4
,
7
Yinon Bar-On,
8
Ron Milo,
8
and Rob Phillips
3
,
5
,
9
,
*
1
Department of Biology, Stanford University, Stanford, CA, USA
2
Department of Applied Physics, California Institute of Technology, Pasadena, CA, USA
3
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA, USA
4
Resnick Sustainibility Institute, California Institute of Technology, Pasadena, CA, USA
5
Chan-Zuckerberg BioHub, San Francisco, CA, USA
6
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA, USA
7
Department of Environmental Science and Engineering, California Institute of Technology, Pasadena, CA, USA
8
Department of Plant and Environmental Sciences, Weizmann Institute of Science, Rehovot, Israel
9
Department of Physics, California Institute of Technology, Pasadena, CA, USA
10
These authors contributed equally
11
Lead contact
*Correspondence:
griffinchure@gmail.com
(G.C.),
phillips@pboc.caltech.edu
(R.P.)
https://doi.org/10.1016/j.patter.2022.100552
SUMMARY
The Human Impacts Database (www.anthroponumbers.org) is a curated, searchable resource housing quan-
titative data relating to the diverse anthropogenic impacts on our planet, with topics ranging from sea-level
rise to livestock populations, greenhouse gas emissions, fertilizer use, and beyond. Each entry in the data-
base reports a quantitative value (or a time series of values) along with clear referencing of the primary
source, the method of measurement or estimation, an assessment of uncertainty, and links to the underlying
data, as well as a permanent identifier called a Human Impacts ID (HuID). While there are other databases that
house some of these values, they are typically focused on a single topic area, like energy usage or green-
house gas emissions. The Human Impacts Database facilitates access to carefully curated data, acting as
a quantitative resource pertaining to the myriad ways in whic h humans have an impact on the Earth, for prac-
ticing scientists, the general public, and those involved in education for sustainable development alike. We
outline the structure of the database, describe our curation procedures, and use this database to generate a
graphical summary of the current state of human impacts on the Earth, illustrating both their numerical values
and their intimate interconnections.
THE BIGGER PICTURE
Over the last 10,000 years, human activities have transformed Earth through
farming, forestry, mining, and industry. The complex results of these activities are now observed and quan-
tified as ‘‘human impacts’’ on Earth’s atmosphere, oceans, biosphere, and geochemistry. While myriad
studies have explored facets of human impacts on the planet, they are necessarily technical and often high-
ly focused. Thus, finding reliable quantitative information requires a significant investment of time to
assess each quantity and associated uncertainty. We present the Human Impacts Database (www.
anthroponumbers.org), which houses a diverse array of such quantities. We review a subset of these values
and how they help build intuition for understanding the Earth-human system. While collation alone does not
tell us how to best ameliorate human impacts, we contend that any future plans should be made in light of a
quantitative understanding of the interconnected ways in which humans influence the planet.
Production:
Data science output is well understood
and (nearly) universally adopted
Patterns
3
, 100552, September 9, 2022
ª
2022 The Author(s).
1
This is an open access article under the CC BY license (
http://creativecommons.org/licenses/by/4.0/
).
ll
OPEN ACCESS
INTRODUCTION
One of the most important scientific developments of the last
two centuries is the realization that the evolution of Earth is
deeply intertwined with the evolution of life. Perhaps the most
famous example of this intimate relationship is the large-scale
oxygenation of Earth’s atmosphere following the emergence
of photosynthesis.
1
This dramatic change in the composition
of the atmosphere is believed to have caused a massive extinc-
tion, as the biosphere was not adapted to an oxygenated atmo-
sphere.
2–4
Over the past 10,000 years, humans have likewise
transformed the planet, directly affecting the rise and fall of
ecosystems,
5–13
the pH and surface temperature of the
oceans,
14
,
15
the composition of terrestrial biological and hu-
man-made mass,
16
,
17
the planetary albedo and ice cover,
18–27
and the chemistry of the atmosphere,
28–33
to name just a few
examples. The breadth of human impacts on the planet is so
diverse that it touches on nearly every facet of the Earth system
and every scientific discipline.
Technological advances in remote sensing, precision mea-
surement, and computational power have made it possible to
measure these anthropogenic impacts with unprecedented
depth and resolution. However, as scientists with different
training use distinct methods for measurement and analysis,
report data in different units and formats, and use nomenclature
differently, these studies can be very challenging to understand
and relate to one another. Even seemingly simple questions such
as ‘‘how much water do humans use?’’ can be difficult to answer
when search engines are not optimized for finding numeric data,
and a search of the scientific literature yields an array of compli-
cated analyses with different units, varying definitions about
what constitutes water use, and distinct approaches to quanti-
fying flows. This problem persists beyond the primary scientific
literature, as governmental, intergovernmental, and industry
datasets can be similarly tricky and laborious to interpret.
Writing from California, as several of the authors are, where we
now have a ‘‘wildfire season’’ and a multi-decadal drought,
34
,
35
we wanted to develop a deeper understanding of the ways in
which human activities might have produced such dramatic
and consequential changes in our local and global environment.
In pursuit of basic understanding, we asked many questions, like
‘‘how much water and land do humans use?’’ and ‘‘how much
methane is emitted annually?’’ In our search for answers, even
when the question is well defined (as is the case for methane
emissions), we often encountered the same challenges: dispa-
rate technical studies written for expert audiences must be un-
derstood, evaluated, and synthesized just to answer simple
questions. It seemed to us that a referenced compendium of
‘‘things we already know,’’ akin to the
CRC Handbook of Chem-
istry and Physics
, would be very useful for us and others.
In building the Human Impacts Database, we took
inspiration from our previous experience building and using
the BioNumbers Database
36
(
https://bionumbers.hms.harvard.
edu
), a compendium of quantitative values relating to cell and
organismal biology. Over the past decade, the BioNumbers
Database has become a widely accessed resource that serves
not only as an index of biological numbers, but also as a means
of finding relevant primary literature, learning about methods of
measurement, and teaching basic concepts in cell biology.
37
We believe that a centralized, searchable database for quantita-
tive data encompassing the breadth of human impacts on Earth
would be similarly transformative for researchers, students, and
the interested public. While reading an entry in the Human Im-
pacts Database is not a replacement for reading the primary liter-
ature, the database serves as a resource to expedite the process
of finding quantitative data and exploring their interconnection.
Importantly, we do not put forward projected scenarios or spe-
cific policy proposals for combating anthropogenic effects on
Earth. However, we are convinced that such proposals should
be evaluated in the light of a comprehensive and quantitative
understanding of the Earth-human system.
RESULTS
Finding and compiling numbers from scientific
literature, governmental and non-governmental reports,
and industrial datasets
We have established the Human Impacts Database (
http://
anthroponumbers.org
) as a repository for the rapid discovery
of quantities describing the Earth-human system. We here pro-
vide a more complete description of the database structure,
the values it holds, and the stories it tells us about how humans
affect the Earth. As of this writing, the database holds > 300
unique and manually curated entries covering a breadth of
data sources, including primary scientific literature, govern-
mental and non-governmental reports, and industrial communi-
ques. Before it is added to the database and made public, each
entry is vetted extensively by the administrators (see
Note S1
for
detailed curation procedures). Included in each entry is a sum-
mary of the method by which it was determined, an assessment
of the corresponding uncertainty, and an explicit statement of
any known caveats important for interpretation of the data. While
these
z
300 entries include those we consider to be essential for
a quantitative understanding of human impacts on Earth, it is not
an exhaustive list. This database will continue to grow and evolve
as more data become publicly released, our understanding of
the human-Earth system improves, and members of the scienti-
fic community suggest values to be added.
Figure 1
shows the Human Impacts Database Entry for
perhaps the most emblematic anthropogenic impact: the stand-
ing atmospheric CO
2
concentration. The first two components of
an entry are the quantity title and its assigned category and sub-
category (
Figures 1
A and 1B). Primary categorization falls into
one of five classes: ‘‘land,’’ ‘‘water,’’ ‘‘energy,’’ ‘‘flora & fauna,’’
and ‘‘atmospheric & biogeochemical cycles.’’ Of course, these
categories are broad, and entries can be associated with several
categories. For this reason, each entry is also assigned a nar-
rower ‘‘subcategory,’’ such as ‘‘agriculture,’’ ‘‘urbanization,’’ or
‘‘carbon dioxide.’’ While this categorization is not meant to be
exhaustive, and many other schemes could be implemented,
we found that this organization allowed us to quickly browse
and identify quantities of interest.
Following the title and categorization, we report the measured
atmospheric CO
2
concentration. This corresponds to the most
recent reported measurement, which is, as of this writing, roughly
416 parts per million (ppm) in 2021 (
Figure 1
C). Importantly, we
report an approximate value for the CO
2
concentration rather
than a precise value to many significant digits. While the most
ll
OPEN ACCESS
Descriptor
2
Patterns
3
, 100552, September 9, 2022
Figure 1. A representative entry in the Human Impacts Database
(A–I) The entry page for HuID 81043, ‘‘Atmospheric CO
2
concentration,’’ is diagrammed with important features highlighted. Each entry in the Human Impacts
Database has (A) a name, (B) a primary and secondary categorization, (C) the numerical value with other units when appropriate, (D) a five-digit perman
ent
numeric identifier, (E) the years for which the measurement was determined, (F) a brief summary of the quantity, (G) the method of determination, (H) a l
ink to
the source data, and (I) a link to a processed version of the data saved as a .csv file. When possible, a time series of the data is presented.
(K) Every entry in the database also has a statement of the data use protection associated with the relevant data. When possible, this links directly to
the data
protection statement from the original source. In other cases, it points to the formal definition of the license by a disinterested third party.
(L) Finally, each entry lists the username of the administrator who curated the quantity. Their contact information is available on the
anthroponumbers.org
‘‘About’’ page.
ll
OPEN ACCESS
Descriptor
Patterns
3
, 100552, September 9, 2022
3
recent entry in the linked dataset (
Figure 1
I) gives a monthly
average value of 416.43 ppm for December of 2021, this value
does not account for error in the measurement, fluctuations
throughout December, or seasonal oscillations in atmospheric
CO
2
. Therefore, we report a rounded value of 416 ppm. CO
2
mea-
surements are quite accurate, but other measurements and infer-
ences recorded in the Human Impacts Database are less so. We
therefore strive to give an assessment of the uncertainty for all
values. This can be in the form of a confidence interval, as for
the entry for the global mean sea-level rise since 1900 due to ther-
mal expansion, which reports a 90% confidence interval, or
bounds on the value, as for the number of contemporary animal
extinctions since 1500 CE, which reports only a lower bound. In
addition to error assessment, we also aim to provide legible units
for all entries. Although atmospheric CO
2
is commonly reported in
ppm units, we also report this value in other equivalent units,
including the mole and mass fractions of CO
2
and the total mass
of CO
2
in the atmosphere in kg CO
2
(
Figure 1
C). Whenever
possible, entries will report values in multiple units to make quan-
tities accessible to readers coming from diverse backgrounds.
Furthermore, in many cases, the global value is aggregated from
local measurements. We flag entries for which regional data
broadly defined are available in the database GitHub repository.
Following the numerical value is the permanent Human Im-
pacts Database identifier, which we abbreviate as HuID (
Fig-
ure 1
D). The HuID is a randomly generated five-digit integer
that serves as a permanent and static identifier that can be
used for in-line referencing. Rather than identifying a single
value, we consider the HuID a pointer to a particular
entry
,so
that HuID 81043 can be used to reference the atmospheric
CO
2
concentration in 2021 and 1980 (
Figure 1
E). For example,
to reference the present-day atmospheric CO
2
concentration,
one could report the value as ‘‘
z
416 ppm (HuID
81043:2021).’’ In addition, since each entry comes from a single
source, we may have more than one HuID reporting similar quan-
tities. For example, HuIDs 69674 and 72086 report recent mea-
surements of the temperature of the upper ocean.
The‘‘Summary’’ field (
Figure1
F)gives a succinct description of
the quantity and its relationship to ‘‘human impacts’’ broadly
construed, along with other pertinent information. This could
include a more detailed definition of terms used in the quantity,
suchasthe entryfor‘‘seaice extent lossinMarch,’’ whichdefines
the term ‘‘sea ice extent,’’ or useful historical information about
the measurement. In our example of atmospheric CO
2
concen-
tration, the summary explains that the measurement is made at
the Mauna Loa observatory and points out the seasonal oscilla-
tions that are observed. The following ‘‘Method’’ field describes
themethodbywhichthequantitywas measured,inferred, oresti-
mated (
Figure 1
G). This field also provides an assessment of the
uncertainty in the value, which may include a description of how
confidence intervals were computed or a list of critical assump-
tions that were made to estimate missing data.
All fields through ‘‘Method’’ (
Figures 1
A–1G) depend on
manual curation and interpretation by database administrators.
The following two fields, ‘‘Source’’ and ‘‘Dataset’’ (
Figures 1
H
and 1I), provide direct links to the primary source reference
and the relevant data. Both of these fields are direct links (shown
as insets in
Figure 1
). The ‘‘Source’’ field can point to either the
published scientific literature or the resource page of a govern-
mental, industrial, or non-governmental organization data depo-
sition URL. The ‘‘Dataset’’ field links directly to either a CSV
format of the data or to a folder with global and regional values
within the corresponding GitHub repository. As discussed in
Note S1
, the vast majority of these data files have been con-
verted into a ‘‘tidy-data’’ format
38
by database administrators,
which maximizes programmatic readability.
When possible, a graphical time series of the data is also pre-
sented as an interactive plot (
Figure 1
J). These plots enable
users to quickly apprehend time-dependent trends in the data
without downloading or processing the dataset. The data sour-
ces we rely on in building the database are remarkably varied,
coming from governmental, industrial, and primary scientific
sources, each with their own specific data use protection pol-
icies. Each entry (
Figure 1
K) also provides a link to the data
use policy for each individual dataset. While not available for
every entry, the majority of quantities we have curated in the Hu-
man Impacts Database contain measurements over time. The
last field gives the username of the administrator who generated
this entry (
Figure 1
L). Their affiliation and contact information are
available on the database’s ‘‘About’’ page. We invite the reader
to contact the administrators collectively—through our ‘‘Con-
tact’’ page or directly through our personal emails as provided
on the ‘‘About’’ page—with questions, concerns, or suggestions.
While
Figure 1
is a representative example, each quantity in
the Human Impacts Database tells a different story. Easy and
centralized access to different entries allows users to learn about
the magnitude of human impacts and also study the interactions
between different human activities, which, as we discuss in the
next section, are deeply intertwined.
Global magnitudes
In
Figure 2
, we provide an array of quantities that we believe to be
key in developing a ‘‘feeling for the numbers’’ associated with hu-
man impacts on the Earth system. All of the quantities in
Figure 2
aredrawnfromentriesinthedatabaseandgroupedintothesame
categories used in the database: land, water, flora and fauna, at-
mosphere and biogeochemical cycles, and energy (see color
scheme at the top of
Figure 2
). Although the impacts considered
here necessarily constitute an incomplete description of human
interaction with the planet, these numbers encompass many
that are critically important, such as the volume of liquid water re-
sulting from ice melt (
Figure 2
B), the extent of urban and agricul-
tural land use (
Figure 2
H), global power consumption (
Figure 2
N),
and the heat uptake and subsequent warming of the ocean
surface(
Figure2
S).Inmanycases,therawnumbersareastound-
inglylargeandcantherefore bedifficult tofathom.Rather thanre-
porting only bare ‘‘scientific’’ units, we present each quantity
(when possible) in units that are intended to be relatable as
‘‘per capita’’ values to a broad audience who are members of
(or familiar with) typical Western lifestyles. Consider, for example,
the 18 TW global power consumption (
Figure 2
N). For most audi-
ences, it can be difficult to conceptualize what a watt is, let alone
the sheer magnitude of a
terawatt
. However, most prospective
users of this database likely have a familiarity with the warmth
of a 100 W light bulb. With this in mind, we can do a simple con-
version to say that the global average power use per person is
comparable to constantly running
z
23 light bulbs per person,
making the impact a bit more tangible.
ll
OPEN ACCESS
Descriptor
4
Patterns
3
, 100552, September 9, 2022
Figure 2. Human impacts on the planet and their relevant magnitudes
Relative units and the broad organizational categories are shown in the top left. Source information and contextual comments for each subpanel are pr
esented in
Note S2
.
ll
OPEN ACCESS
Descriptor
Patterns
3
, 100552, September 9, 2022
5
(legend on next page)
ll
OPEN ACCESS
Descriptor
6
Patterns
3
, 100552, September 9, 2022
Exploring these numbers reveals a number of intriguing quan-
tities and relationships. For example, agriculture repeatedly
appears as a major contributor to many human impacts, domi-
nating both global land (
Figure 2
H) and global water use (
Fig-
ure 2
L) and accounting for approximately a third of global tree
cover area loss (
Figure 2
O). In addition, an enormous mass of ni-
trogen is synthetically fixed through the Haber-Bosch process,
primarilyto produce fertilizer (
Figure 2
F), which is a major cause
of emissions of N
2
O(
Figure 2
K), which is a potent greenhouse
gas. About 45 billion livestock are raised on agricultural lands
(
Figure 2
E), which, together with rice paddies, produce a majority
of anthropogenic methane emissions (the greenhouse gas CH
4
;
Figure 2
K). On the other hand, urban land area accounts for a
very small fraction of land area use (
z
1%,
Figure 2
H), and the
expansion of cities and suburbs accounts for only
z
1% of
global tree cover area loss (
Figure 2
O). This is not to say, howev-
er, that urban centers are negligible in their global impacts. As ur-
ban and suburban areas currently house more than half of the
global human population (
Figure 2
J), many human impacts are
linked to industries that directly or indirectly support urban pop-
ulations’ demand for food, housing, travel, electronics, and other
goods. For example, the pursuit of urbanization is the dominating
factor in the mass of earth moved on an annual basis (
Figure 2
W).
Collectively, the
z
8 billion humans on Earth (
Figure 2
J)
consumenearly20TWofpower,equivalentto23100Wlightbulbs
perperson(
Figure2
N).Around80%ofthisenergyderivesfromthe
combustion of fossil fuels (
Figure 2
P). This results in a tremendous
mass of CO
2
being emitted annually (
Figure 2
K), of which only
z
50% remains in the atmosphere (HuID 70632). A sizable portion
oftheemissionsareabsorbedbytheoceans(HuID99089),leading
toasteadyincreaseinoceanacidity(
Figure2
G)andposingrisksto
marine ecosystems.
39
Furthermore, increasing average global
temperatures, primarily caused by greenhouse gas emissions,
contribute to sea-level rise not only in the form of added water
frommeltingice(
Figure2
Band2M),butalsoduetothermalexpan-
sion of ocean water (
Figure 2
M), which accounts for
z
30% of
observed sea-level rise.
40
These are just a few ways in which
one can traverse the impacts illustrated in
Figure 2
, revealing the
remarkable extent to which these impacts are interconnected.
We encourage the reader to explore this figure in a similar manner,
blazing their own trail through the values.
Regional distribution
While
Figure 2
presents the magnitude of human impacts at a
global scale, it is important to recognize that these impacts—
both their origins and their repercussions—are variable across
the globe. That is, different societies vary in their preferences for
food (e.g., Americans consume relatively little fish) and modes of
living (e.g., apartments versus houses), have different levels of
economic development (e.g., Canada compared with Malaysia),
rely on different natural resources to build infrastructure (e.g.,
wood versus concrete) and generate power (e.g., nuclear versus
coal), and promote different extractive or polluting industries
(e.g., lithium mining versus palm oil farming). Some of these
regional differences are evident in
Figure 3
, which summarizes
regional breakdowns of several drivers of global human impacts,
e.g., livestock populations and greenhouse gas emissions.
Just as impactful human activities like coal power generation
and swine farming are more common in some regions than
others (
Figure 2
), the impacts of human activities affect some re-
gions more than others.
42
Figure 3
displays a coarse regional
breakdown of the numbers from
Figure 2
for which regional dis-
tributions could be determined from the literature. The region
definitions used in
Figure 3
are similar to the definitions set forth
by the Food and Agricultural Organization (FAO) of the United
Nations, assigning the semi-continental regions of North Amer-
ica, South America, Africa, Europe (including Russia), Asia, and
Oceania. Here, we specify both the total contribution of each re-
gion and the per-capita value, given the population of that region
as of the year(s) in which the quantity was measured.
Much as in the case of
Figure 2
, interesting details emerge
from
Figure 3
. For example, Asia dominates global agricultural
water withdrawal (excluding natural watering via rainfall), using
about 62% of the total, while North America takes the lead in in-
dustrial water withdrawal. Interestingly, on a per-capita basis,
North America withdraws the most water for all uses: agricul-
tural, industrial, and domestic.
North America also emits more CO
2
per capita than any other
region, with Oceania and Europe coming second and third,
respectively. This disparity can be partially understood by
considering the regional distribution of fossil fuel consumption,
the dominant source of CO
2
emissions (
Figure 3
J). While Asia
consumes more than half of the total fossil fuel energy, per-cap-
ita consumption is markedly lower than in North America, Eu-
rope, and Oceania (
Figure 3
J). Interestingly, the story is different
when it comes to methane. Oceania and South America are the
largest emitters of anthropogenic CH
4
, mainly due to a standing
population of cattle that rivals that of humans in those regions
(
Figure 3
D) and produces this potent greenhouse gas through
enteric fermentation.
33
Regional disparities are also apparent
in the means of energy production. While consuming only 4%
of the total power, South America generates about 14% of the
renewable energy. Nuclear power generation, on the other
hand, is dominated by North America and Europe, while Oce-
ania, which has a single research-grade nuclear reactor, gener-
ates nearly zero nuclear energy.
Investigatingthecausesofforestlossbygeographicregionlike-
wise highlights interesting differences. At a global level, all drivers
of forest loss are comparable in magnitude, except for urbaniza-
tion, which accounts for
z
1% of total annual tree cover area
loss (
Figure 2
O). Despite comparable magnitudes, different
drivers of forest loss have different long-term consequences.
30
Forest loss due to wildfires and forestry often result in regrowth,
while commodity-driven harvesting and urbanization tend to be
drivers of long-lasting deforestation.
43
,
44
Central and South Amer-
ica account for about 65% of commodity-driven deforestation
Figure 3. Regional distribution of anthropogenic effects
(A) Several quantities from
Figure 2
were selected, and the relative magnitudes were broken down by subcontinental area.
(B–J) Donut charts in all sections show the relative contributions of each quantity by region. Ball-and-stick plots show the per-capita breakdown of
each quantity
across geographic regions. All data for global and per-capita breakdowns correspond to the latest year for which data were available. The regional br
eakdown for
deforestation uses the regional convention as reported in the source data.
41
ll
OPEN ACCESS
Descriptor
Patterns
3
, 100552, September 9, 2022
7
(legend on next page)
ll
OPEN ACCESS
Descriptor
8
Patterns
3
, 100552, September 9, 2022
(meaning clear-cutting and human-induced fires with no substan-
tialregrowth of treecover), whereas a majorityofforestlossdue to
shifting agriculture occurs in Africa (where regrowth does occur).
Together, wildfires in North America, Russia, China, and South
Asiamakeupnearly90%oflossesduetofire.
41
Whileurbanization
is the smallest driver of tree cover loss globally, it can still have
strong impacts at the regional level, perturbing local ecosystems
and biodiversity.
45
,
46
Time series
When available, the Human Impacts Database includes time-
series data for each quantity. Just as the regional distributions
of impactful human activities help us understand differences
between societies and regions, studying the history of these
activities highlights recent technological and economic devel-
opments that intensify or reduce their impacts. When consid-
ering the history of human impacts on the Earth, it is natural
to start by considering the growth of the human population
over time. As shown in
Figure 4
, the global human population
grew nearly continually over the past 80 years, with the current
population nearing 8 billion. Historically, most of the global
human population lived in rural areas (about 70% as of 1950,
HuID 93995). Recent decades have been marked by a substan-
tial shift in how humans live globally, with around half of the
human population now living in urban or suburban settings
(
z
55%, HuID 93995).
Given the growth of the human population, it is reasonable to
consider that human population may be the most natural scale to
measure human impacts.
47
To assess this possibility, we plotted
per-capita impacts over several decades (
Figure 4
). If impacts
are growing in direct proportion to the human population, per-
capita impacts would be constant over time. Indeed, this is
roughly true for per-capita water withdrawals over the past
40 years (
Figure 4
B). Deviations from proportionality may indi-
cate important changes in human activities. For example, in
recent decades, per-capita chicken populations grew by nearly
2-fold, while per-capita cattle populations shrunk by roughly
25%, reflecting a modest transition away from beef and toward
chicken as a source of animal protein in global diets (HuIDs
40696 and 79776).
One very visible impact accompanying the shift of the human
population to urban environments is the increase in production of
anthropogenic mass: materials such as concrete, steel, lumber,
and plastics used to build roads, buildings, machines, pack-
aging, and other useful human-made items. Since these mate-
rials are degraded very slowly, anthropogenic mass has been
accumulating over time. In addition, the mass of concrete, ag-
gregates like asphalt, and bricks per capita has been increasing
since the 1950s (
Figure 4
D). Concrete, in particular, has
increased from less than 10 tons per person in the 1950s to
almost 30 tons per person in the 2010s. This increase in per-cap-
ita anthropogenic mass means that the increase in production of
these materials is outpacing the growth of the human population.
These material production trends have been enabled, in part,
by a sustained increase in power generation. As evident from
Figure 4
, total power consumption has been increasing roughly
proportionally with the human population. Per-capita consump-
tion has also increased across all generation types, including
fossil fuels, hydropower, nuclear, and renewables. The growth
among nuclear and renewables has been especially dramatic,
and nuclear power now roughly equals hydropower production.
Production of crops, aquaculture, and populations of livestock
are all likewise correlated with growth in the human population
(
Figures 4
C and 4E). The total number of livestock has increased
with the human population, primarily due to increasing chicken
populations as discussed above. The dominant means of
aquatic food production has also shifted over this time: until
roughly 1980, nearly all seafood was captured wild, but since
then aquaculture has grown to account for roughly ½ of aquatic
food production (HuID 61233,
Figure 4
E).
Turning our focus to greenhouse gases, we see that annual
anthropogenic CO
2
emissions have been increasing with the
population (
Figure 4
G). Burning of fossil fuels is the dominant
contributor to anthropogenic emissions and has increased
slightly on a per-capita basis over the past 60 years. In contrast,
as the pace of global deforestation has slowed,
48
,
49
emissions of
CO
2
due to land-use change have decreased per capita. These
two trends roughly neutralize each other, leading to little overall
change in CO
2
emissions per capita since the 1960s. Akin to CO
2
emissions due to land-use change, CH
4
emissions show a sub-
linear trend with human population, partially due to a decline in
ruminant livestock per capita (
Figures 4
C and 4H).
DISCUSSION
Quantitative literacy is necessary for ‘‘understanding’’ in nearly all
branches of science. As our collective knowledge of anthropo-
genic impacts expands, it has become challenging to sift through
the literaturetocollect specific numbers usefulfor bothcalculation
and communication. We have attempted to reduce this barrier to
entryonseveralfronts.Wehavecanvassedthescientificliterature,
governmental, industrial, and international reports to assemble a
broad, quantitative picture of how human activities have affected
the Earth’s atmosphere, oceans, rivers, lands, biota, chemistry,
and geology. In doing so, we have created an online, searchable
database housing an array of quantities and data that describe
different facets of the human-Earth interface. We view this data-
base as an accessory, rather than a replacement, for the myriad
scientific databases that exist and are publicly available on the
internet (some of which are listed on the database website
www.
anthroponumbers.org/catalog/databases
). While these data-
bases are invaluable resources for accessing scientific data, the
Human Impacts Database is built from the ground up with the
intention of being broadly accessible to scientists and the curious
general public alike to help build the collective quantitative literacy
of the Anthropocene. Beyond the database, we have assembled
Figure 4. Temporal dynamics of key human impacts
(A) Several quantities from
Figure 2
were selected, and the magnitudes were plotted as a function of either time (for cumulative quantities such as anthropomass)
or human population.
(B–H) Ball-and-stick plots show the per-capita breakdown as decadal averages to give a more reflective view of cultural and technological shifts than
year-to-year
variation.
ll
OPEN ACCESS
Descriptor
Patterns
3
, 100552, September 9, 2022
9
thesedataintoacomprehensivesnapshot,releasedalongsidethis
writing as a standalone graphical document (
Data S1
), with all un-
derlying data, associated uncertainties, and referencing housed in
the Human Impacts Database. While necessarily incomplete,
these resources provide a broad view of the ways in which human
activities are having an impact on the Earth on multiple fronts.
One insight that emerges from a holistic consideration of these
diverse human activities together is that they are deeply inter-
twined and driven by a small number of pivotal factors: the size
of the human population, the composition of our diets, and our
demand for materials and energy to build and power our increas-
ingly complex and mechanized societies. Understanding the
scale of human agriculture and water and power usage provides
a framework for understanding most of the numerical gallery pre-
sented in
Figure 2
. Perhaps unsurprisingly, we find that feeding
the growing human population is a major driver of a large swath
of human impacts on Earth, dominating global land (
Figure 2
H,
HuID 29582) and water use (
Figure 2
L; HuIDs 84545, 43593,
95345), as well as significantly contributing to tree cover loss
(
Figure 2
O, HuID 24388), earth moving (
Figure 2
W; HuIDs
19415, 41496), and anthropogenic nitrogen fixation (
Figure 2
F;
HuIDs 60580, 61614), to name a few such examples. The Human
Impacts Database provides a resource to explore relationships
between values temporally, globally, and locally, and go beyond
the standalone values often reported in isolation or cast solely
through the lenses of impact, population, affluence, and technol-
ogy (I = PAT) relationships.
It is common in this setting to argue that the bewildering
breadth and scale of human impacts should motivate some spe-
cific remediation at the global or local scale. We, instead, take a
more modest "just the facts" approach. The numbers presented
here show that human activities affect our planet to a large de-
gree in many different and incommensurate ways, but they do
not provide a roadmap for the future. Rather, we contend that
any plans for the future should be made in the light of a compre-
hensive and quantitative understanding of the interconnected
ways in which human activities impact the Earth system globally
(
Figure 2
), locally (
Figure 3
), and temporally (
Figure 4
). Achieving
such an understanding will require the synthesis of a broad liter-
ature across many disciplines. While the quantities we have cho-
sen to explore are certainly not exhaustive, they represent some
of the key axes that frequently drive scientific and public
discourse and shape policy across the globe.
Earth is the only habitable planet we know of, so it is crucial to
understand how we got here and where we are going. That is,
how (and why) have human impacts changed over time? How
are they expected to change in the future? For every aspect of
human entanglement with the Earth system—from water use to
land use, greenhouse gas emissions, mining of precious min-
erals, and so on—there are excellent studies measuring impacts
and predicting their future trajectories. Of particular note are the
data-rich and explanatory reports from the Intergovernmental
Panel on Climate Change
50
,
51
and the efforts toward defining
‘‘planetary boundaries.’’
52
We hope that the Human Impacts
Database and the associated resources with this work provide
a reference to explore the human-induced interdependencies
between many axes of the human-Earth system and will engage
the scientific community, ultimately helping humanity coexist
stably with the only planet we have.
EXPERIMENTAL PROCEDURES
Resource availability
Lead contact
Requests for further information should be directed to and will be fulfilled by
the lead contact, Griffin Chure (
griffinchure@gmail.com
).
Materials availability
No materials were used in the generation of this work, other than the code and
data as described below. We have collated all data shown in
Figures 2–4
,
along with all information in
Note S2
as a printable, ‘‘graphical snapshot’’
(
Data S1
).
Data and code availability
For every dataset included in the database, there is a folder in the GitHub re-
pository
https://github.com/rpgroup-pboc/human_impacts
(DOI: 10.5281/
zenodo.4453276) that includes the source data, the processed data, and the
code to generate the ‘‘tidy’’ data from the source data. Each folder also in-
cludes a README file that includes information about the dataset. In addition,
all of the code used to generate the figures can be found in the GitHub repos-
itory under the ‘‘figures’’ folder. We strongly encourage the scientific commu-
nity to fork this repository, submit pull requests, and open new constructive is-
sues through the GitHub repository interface.
The database and the FAIR principles of data reuse
The primary goal of the Human Impacts Database is to provide a resource for
the rapid discovery quantities related to the human-Earth system while mini-
mizing the grunt work needed to access (and understand) the underlying
data. This means that facilitating data reuse and reproducibility of any analyses
is paramount to the importance of the database. To that end, we abide by the
FAIR Guiding Principles for Scientific Data Management and Stewardship
(
www.go-fair.org/fair-principles/
). These principles are guidelines to maximize
the findability, accessibility, interoperability, and reusability of original scienti-
fic data. The database closely follows these principles, as is briefly out-
lined below:
d
Findability: The underlying data can be easily searched and navigated,
permitting rapid discovery. Individual entries are assigned a unique
integer identifier that serves as a permanent referencing tool and are pro-
vided with rich metadata about the method of determination, original
source,data useprotectionpolicy,and quantitativevalue in diverse units.
d
Accessibility: The original source of the underlying data is always re-
ported hyperlinked when legally permissible. The transformation, colla-
tion, or manipulation of the underlying data that was necessary to add it
to the Human Impacts Database is preserved under a publicly acces-
sible, version-controlled, GitHub repository (
github.com/rpgroup-
pboc/human_impacts
) and is permanently accessible via
https://doi.
org/10.5281/zenodo.4453276
. This protects against permanent loss
of the data even if an entry is deleted from the database.
d
Interoperability: The data are provided in a human readable format
with an emphasis on description of the data and their source. The
vast majority of datasets are transformed programmatically to follow
a ‘‘tidy,’’ long-form format that facilitates computational analysis of
the data. As the values are hand curated and the target audience is
a curious human, we have not developed an API for programmatic
access of the database, and do not have plans to do so in the fore-
seeable future.
d
Reusability: All entries in the database and the corresponding GitHub re-
pository are extensively annotated with rich metadata, preventing the
need for guesswork as to how the data were collected or what the col-
umn names refer to in the original or processed data. Furthermore, all
data held in the database and repository follow the legal guidelines as
presented by their original owner. This licensing is directly linked to in
each entry.
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at
https://doi.org/10.1016/j.
patter.2022.100552
.
ll
OPEN ACCESS
Descriptor
10
Patterns
3
, 100552, September 9, 2022
ACKNOWLEDGMENTS
We are incredibly grateful for the generosity of a wide array of experts for their
advice, suggestions, and criticism of this work. Specifically, we thank Suzy
Beeler, Joseph Berry, Lars Bildsten, Justin Bois, Chris Bowler, Matthew
Burgess, Ken Caldeira, Jo
̈
rn Callies, Sean B. Carroll, Ibrahim Cisse
́
, Joel Co-
hen, Michelle Dan, Bethany Ehlmann, Gidon Eshel, Moi Exposito-Alonso,
Paul Falkowski, Daniel Fisher, Thomas Frederikse, Neil Fromer, Eric Galbraith,
Lea Goentoro, Evan Groover, John Grotzinger, Soichi Hirokawa, Greg Huber,
Christina Hueschen, Bob Jaffe, Elizabeth Kolbert, Thomas Lecuit, Raphael
Magarik, Jeff Marlow, Brad Marston, Jitu Mayor, Elliot Meyerowitz, Lisa Miller,
Dianne Newman, Luke Oltrogge, Nigel Orme, Victoria Orphan, Marco Pasti,
Pietro Perona, Noam Prywes, Stephen Quake, Hamza Raniwala, Manuel
Razo-Mejia, Thomas Rosenbaum, Benjamin Rubin, Alex Rubinsteyn, Shyam
Saladi, Tapio Schneider, Murali Sharma, Alon Shepon, Arthur Smith, Matthieu
Talpe, Wati Taylor, Julie Theriot, Tadashi Tokieda, Cat Triandifillou, Sabah Ul-
Hasan, Tine Valencic, Ned Wingreen, and Emily Zakem. We also thank Yue Qin
for sharing data related to global water consumption. Many of the topics in this
work began during the Applied Physics 150C course taught at Caltech during
the early days of the COVID-19 pandemic. This work was supported by the Re-
snick Sustainability Institute at Caltech and the Schwartz-Reisman Collabora-
tive Science Program at the Weizmann Institute of Science. G.C. acknowl-
edges support by the NSF Postdoctoral Research Fellowships in Biology
Program (grant no. 2010807).
AUTHOR CONTRIBUTIONS
Conceptualization, G.C., R.A.B., R.M., and R.P.; investigation, G.C., R.A.B.,
A.I.F., N.S.S., M.K., I.L.G., and Y.M.B.; data curation, G.C., R.A.B., N.S.S.,
M.K., and I.L.G.; software, G.C.; writing – original draft, G.C., R.A.B., A.I.F.,
N.S.S., I.L.G., R.M., and R.P.; writing – review & editing, G.C., R.A.B., and
R.P.; visualization, G.C. and R.A.B.; project administration, G.C., R.A.B.,
and R.P.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: March 4, 2022
Revised: May 26, 2022
Accepted: June 23, 2022
Published: August 3, 2022
WEB RESOURCES
BioNumbers Database,
https://bionumbers.hms.harvard.edu
Human Impacts Database,
http://anthroponumbers.org
REFERENCES
1. Fischer, W.W., Hemp, J., and Johnson, J.E. (2016). Evolution of oxygenic
photosynthesis. Annu. Rev. Earth Planet Sci.
44
, 647–683.
https://doi.org/
10.1146/annurev-earth-060313-054810
.
2. Hodgskiss, M.S.W., Crockford, P.W., Peng, Y., Wing, B.A., and Horner,
T.J. (2019). A productivity collapse to end Earth’s Great Oxidation. Proc.
Natl. Acad. Sci. USA
116
, 17207–17212.
https://doi.org/10.1073/pnas.
1900325116
.
3. Gumsley, A.P., Chamberlain, K.R., Bleeker, W., So
̈
derlund, U., de Kock,
M.O., Larsson, E.R., and Bekker, A. (2017). Timing and tempo of the great
oxidation event. Proc. Natl. Acad. Sci. USA
114
, 1811–1816.
https://doi.
org/10.1073/pnas.1608824114
.
4. Sessions, A.L., Doughty, D.M., Welander, P.V., Summons, R.E., and
Newman, D.K. (2009). The continuing puzzle of the great oxidation event.
Curr. Biol.
19
, R567–R574.
https://doi.org/10.1016/j.cub.2009.05.054
.
5. Estes, J.A., Terborgh, J., Brashares, J.S., Power, M.E., Berger, J., Bond,
W.J., Carpenter, S.R., Essington, T.E., Holt, R.D., Jackson, J.B.C., et al.
(2011). Trophic downgrading of planet earth. Science
333
, 301–306.
https://doi.org/10.1126/science.1205106
.
6. Springer, A.M., Estes, J.A., van Vliet, G.B., Williams, T.M., Doak, D.F.,
Danner, E.M., Forney, K.A., and Pfister, B. (2003). Sequential megafaunal
collapse in the North Pacific Ocean: an ongoing legacy of industrial
whaling? Proc. Natl. Acad. Sci. USA
100
, 12223–12228.
https://doi.org/
10.1073/pnas.1635156100
.
7. Hale, S.L., and Koprowski, J.L. (2018). Ecosystem-level effects of
keystone species reintroduction: a literature review: effects of keystone
species reintroduction. Restor. Ecol.
26
, 439–445.
https://doi.org/10.
1111/rec.12684
.
8. Holdo, R.M., Sinclair, A.R.E., Dobson, A.P., Metzger, K.L., Bolker, B.M.,
Ritchie, M.E., et al. (2009). A disease-mediated trophic cascade in the
Serengeti and its implications for ecosystem. PLoS Biol.
7
, e1000210.
https://doi.org/10.1371/journal.pbio.1000210
.
9. Ceballos, G., Ehrlich, P.R., Barnosky, A.D., Garcı
́a,
A., Pringle, R.M., and
Palmer, T.M. (2015). Accelerated modern human–induced species losses:
entering the sixth mass extinction. Sci. Adv.
1
, e1400253.
https://doi.org/
10.1126/sciadv.1400253
.
10. Davidson, A.D., Hamilton, M.J., Boyer, A.G., Brown, J.H., and Ceballos, G.
(2009). Multiple ecological pathways to extinction in mammals. Proc.
Natl. Acad. Sci. USA
106
, 10702–10705.
https://doi.org/10.1073/pnas.
0901956106
.
11. Fortin, D., Beyer, H.L., Boyce, M.S., Smith, D.W., Duchesne, T., and Mao,
J.S. (2005). Wolves influence Elk movements: Behavior shapes a trophic
cascade in Yellowstone National Park. Ecology
86
, 1320–1330.
https://
doi.org/10.1890/04-0953
.
12. Tronstad, L.M., Hall, R.O., Koel, T.M., and Gerow, K.G. (2010). Introduced
lake trout produced a four-level trophic cascade in Yellowstone Lake.
Trans. Am. Fish. Soc.
139
, 1536–1550.
https://doi.org/10.1577/t09-151.1
.
13. Palumbi, S.R. (2001). Humans as the world’s greatest evolutionary force.
Science
293
, 1786–1790.
https://doi.org/10.1126/science.293.5536.1786
.
14. Balmaseda, M.A., Trenberth, K.E., and K
alle
́
n, E. (2013). Distinctive
climate signals in reanalysis of global ocean heat content. Geophys.
Res. Lett.
40
, 1754–1759.
https://doi.org/10.1002/grl.50382
.
15. Cheng, L., Trenberth, K.E., Fasullo, J., Boyer, T., Abraham, J., and Zhu, J.
(2017). Improved estimates of ocean heat content from 1960 to 2015. Sci.
Adv.
3
, e1601545.
https://doi.org/10.1126/sciadv.1601545
.
16. Bar-On, Y.M., Phillips, R., and Milo, R. (2018). The biomass distribution on
Earth. Proc. Natl. Acad. Sci. USA
115
, 6506–6511.
https://doi.org/10.
1073/pnas.1711842115
.
17. Elhacham, E., Ben-Uri, L., Grozovski, J., Bar-On, Y.M., and Milo, R. (2020).
Global human-made mass exceeds all living biomass. Nature
588
,
442–444.
https://doi.org/10.1038/s41586-020-3010-5
.
18.
IPCC (2019). The Ocean and Cryosphere in a Changing Climate
.
19. Peng, H., Ke, C., Shen, X., Li, M., and Shao, Z. (2020). Summer albedo var-
iations in the Arctic Sea ice region from 1982 to 2015. Int. J. Climatol.
40
,
3008–3020.
https://doi.org/10.1002/joc.6379
.
20. Marcianesi, F., Aulicino, G., and Wadhams, P. (2021). Arctic sea ice and
snow cover albedo variability and trends during the last three decades.
Polar Sci
28
, 100617.
https://doi.org/10.1016/j.polar.2020.100617
.
21. Mouginot, J., Rignot, E., Bjørk, A.A., van den Broeke, M., Millan, R.,
Morlighem, M., No
el, B., Scheuchl, B., and Wood, M. (2019). Forty-six
years of Greenland Ice Sheet mass balance from 1972 to 2018. Proc.
Natl. Acad. Sci. USA
116
, 9239–9244.
https://doi.org/10.1073/pnas.
1904242116
.
22. MacGregor, J.A., Fahnestock, M.A., Catania, G.A., Aschwanden, A.,
Clow, G.D., Colgan, W.T., Gogineni, S.P., Morlighem, M., Nowicki,
S.M.J., Paden, J.D., et al. (2016). A synthesis of the basal thermal state
of the Greenland Ice Sheet. J. Geophys. Res. Earth Surf.
121
, 1328–
1350.
https://doi.org/10.1002/2015jf003803
.
23. Kwok, R. (2018). Arctic sea ice thickness, volume, and multiyear ice
coverage: losses and coupled variability (1958–2018). Environ. Res. Lett.
13
, 105005.
https://doi.org/10.1088/1748-9326/aae3ec
.
24. Massom, R.A., Scambos, T.A., Bennetts, L.G., Reid, P., Squire, V.A., and
Stammerjohn, S.E. (2018). Antarctic ice shelf disintegration triggered by
ll
OPEN ACCESS
Descriptor
Patterns
3
, 100552, September 9, 2022
11