ArticlePDF AvailableLiterature Review

A review of the relationship between human population density and biodiversity


Abstract and Figures

To explore the impacts of increasing human numbers on nature, many studies have examined relationships between human population density (HPD) and biodiversity change. The implicit assumption in many of these studies is that as population density increases so does the threat to biodiversity. The implications of this assumption are compounded by recent research showing that species richness for many taxonomic groups is often highest in areas with high HPD. If increasing HPD is a threat to conservation, this threat may be magnified owing to the spatial congruence between people and species richness. Here, I review the relationships between HPD and measures of biodiversity status focussing in particular on evidence for the spatial congruence between people and species richness and the threat that increasing HPD may pose to biodiversity conservation. The review is split into two major sections: (i) a quantitative assessment of 85 studies covering 401 analyses, including meta-analyses on discrete relationships; and (ii) a discussion of the implications of the quantitative analyses and major issues raised in the literature.
Content may be subject to copyright.
Biol. Rev. (2007), 82, pp. 607–645. 607
A review of the relationships between human
population density and biodiversity
Gary W. Luck
Institute for Land, Water and Society, Charles Sturt University, PO Box 789, Albury, NSW, 2640 Australia
(Received 8August 2006; revised 13 August 2007; accepted 20 August 2007)
To explore the impacts of increasing human numbers on nature, many studies have examined relationships between
human population density (HPD) and biodiversity change. The implicit assumption in manyof these studies is that as
population density increases so doesthe threat to biodiversity. The implications of this assumption are compounded by
recent research showing that species richness for many taxonomic groups is often highest in areas with high HPD.
If increasing HPD is a threat to conservation, this threat may be magnified owing to the spatial congruence between
people and species richness. Here, I review the relationships between HPD and measures of biodiversity status
focussing in particular onevidence for the spatial congruence between people and species richness and the threat that
increasing HPD may pose to biodiversity conservation. The review is split into two major sections: (i) a quantitative
assessment of 85 studies covering 401 analyses, including meta-analyses on discrete relationships; and (ii) a discussion
of the implications of the quantitative analyses and major issues raised in the literature.
Our understanding of the relationships between HPD and biodiversity is skewed by geographic and taxonomic
biases in the literature. Most research has been conducted in the Northern Hemisphere and focussed primarily on
birds and mammals, largely ignoring relationships with other taxonomic groups. A total of 127 analyses compared
HPD with the species richness of particular taxonomic groups. A meta-analysis of these results found a significant
positive population correlation indicating that, on average, species-rich regions and human settlements co-occur.
However, there was substantial unexplained heterogeneity in these data. Some of this heterogeneity was explained by
the size of the sampling unit used by researchers – as this increased so did the strength of the correlation between HPD
and species richness. The most convincing result for a taxonomic group was a significant positive population
correlation between HPD and bird species richness. Significant positive population correlations were also found for
HPD versus the richness of threatened and geographically restricted species. Hence, there is reasonably good
evidence for spatial congruence between people and species-rich regions. The reasons for this congruence are only
just beginning to be explored, but key mutual drivers appear to include available energy and elevation.
The evidence forincreasing HPD as a threatto conservation was weak, owing primarily to the extreme heterogeneity
in the approaches used to address this issue. There was some suggestion of a positive relationship between HPD and
species extinction, but this result should be interpreted with caution owing to the wide diversity of approaches used to
measure extinction. Identifying strong links between human development and species extinction is hampered in part by
the difficultyof recording extinction events. The most convincing indication of the negative impact of increasing HPD
was a significant negative population correlation between density and the size of protected areas.
The magnitude and implications of spatial congruence between people and biodiversity are now being
explored using the principles of complementarity and irreplaceability. Human development as a threat to
conservation is usually assessed within a complex, interdisciplinary modelling framework, although population
size is still considered a key factor. Future population growth and expansion of human settlements will present
increasing challenges for conserving species-rich regions and maximising the benefits humans gain from nature.
Key words: human population density, biodiversity, species richness, meta-analysis, protected areas, extinctions.
Address for correspondence: Tel: ]61 2 6051 9945. Fax: ]61 2 6051 9897. E-mail:
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
I. Introduction ...................................................................................................................................... 608
(1) Aims and structure of the review ............................................................................................... 609
II. Quantitative assessment methods ..................................................................................................... 609
(1) Scope and data collection .......................................................................................................... 609
(2) Meta-analysis – general methods ............................................................................................... 610
(3) Meta-analysis – specific methods ............................................................................................... 611
III. Quantitative assessment results ......................................................................................................... 613
(1) General trends ............................................................................................................................ 613
(a) Location ................................................................................................................................. 613
(b) Spatial extent and grain size ................................................................................................ 613
(c) Taxonomic groups and measures of change ........................................................................ 613
(d) Spatial autocorrelation .......................................................................................................... 613
(e) Biased sampling effort ........................................................................................................... 613
(f) Non-linearity ......................................................................................................................... 613
(2) Spatial congruence between people and biodiversity ................................................................ 615
(a) Species richness ..................................................................................................................... 615
(b) Threatened species ................................................................................................................ 617
(c) Restricted species ................................................................................................................... 618
(3) HPD as a threat to biodiversity conservation ............................................................................ 618
(a) Extinctions ............................................................................................................................. 618
(b) Introduced species ................................................................................................................. 618
(c) Protected areas/important conservation areas ..................................................................... 619
(d) Individual species .................................................................................................................. 619
IV. Implications of results and discussion of major themes .................................................................. 620
(1) Geographic, taxonomic and sampling biases ............................................................................ 620
(2) Publication bias ........................................................................................................................... 620
(3) Spatial congruence between people and biodiversity ................................................................ 620
(4) HPD as a threat to biodiversity conservation ............................................................................ 622
(5) Planning for people and nature ................................................................................................. 623
(6) Implications of future human demographics ............................................................................. 624
V. Conclusions ....................................................................................................................................... 624
VI. Acknowledgements ............................................................................................................................ 625
VII. References ......................................................................................................................................... 626
In his essay on the principle of population, Malthus (1798)
discussed the potential for human population growth to
exceed the capacity of the resources required to sustain it.
This thesis has been explored by various authors with
particular emphasis on the adverse consequences that
unbridled growth may have on the environment (Ehrlich,
1971; Meadows et al., 1972; Cohen, 1996; Wilson, 1999;
McKee, 2003). Despite these explorations, clear and
predictable links between human population dynamics
and environmental change remain elusive largely because
of the complexity of the human enterprise and its many and
varied impacts on nature.
Humans have been implicated in the extinction of species
across millennia. Some authors contend that human
colonisation of previously unoccupied continent and island
landmasses resulted in the loss of various taxa, particularly
large fauna species (Martin & Steadman, 1999; Lyons,
Smith & Brown, 2004; Burney & Flannery, 2005; Miller
et al., 2005). Although the role that humans played in faunal
collapse and the interactions that occurred with factors such
as climate are in debate (Brook & Bowman, 2002; Burney &
Flannery, 2005), the notion that even low-density popula-
tions with undeveloped technology can still have a sub-
stantial impact on other species is a pervasive one.
Few would argue against the proposition that contem-
porary human populations also have major impacts on their
local and global environment (Foley et al., 2005). Our
impact can be measured by the interacting factors of
population size, consumption, and the methods used for
resource acquisition (Ehrlich & Holdren, 1971). Conserva-
tion policy and ecosystem resilience also influence the
impact humans have on nature. Yet human population
dynamics is often considered a primary driver of environ-
mental change and linking the two is a useful first step in
understanding the complexity of human impacts on natural
systems (Meyer & Turner, 1992). To this end, a number of
studies have related human population density (HPD) to
biodiversity status or change (e.g. species richness, forest
cover and extinction probability). Spatial or temporal
variation in HPD is generally a more useful measure of
impact than changes in total population size or population
growth rates because the impact of the former is reliant on
the area the population occupies (although if sample area is
Gary W. Luck608
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
constant then values equate to HPD) while the impact of
the latter is reliant on total population size.
The inclusion of HPD in many studies is often based on
the assumption that an increase in density represents
a threat to conservation. The implications of this assump-
tion are compounded by recent research showing that,
across various continents and spatial extents, HPD is often
strongly positively correlated with species richness for
a number of taxonomic groups (Hunter & Yonzon, 1993;
Balmford et al., 2001; Arau
´jo, 2003; Chown et al., 2003;
Real et al., 2003; Gaston & Evans, 2004; Luck et al., 2004;
´zquez & Gaston, 2006). That is, areas that contain the
most people also contain the most species. If increasing
HPD is a threat to species conservation, this threat may be
exacerbated by the spatial congruence between people and
biodiversity. This has substantial implications; therefore, it is
crucial and timely to review the evidence in the literature
for the degree of spatial congruence between HPD and
species richness and the threat increasing density may pose
to species conservation.
(1) Aims and structure of the review
This review has two primary aims: (i) to conduct
a quantitative assessment of studies that correlate HPD
with a measure of biodiversity change or status (i.e. species
richness, the richness of threatened or geographically
restricted species, extinction, the richness of introduced
species, protected area coverage, and individual species) and
include statistical syntheses of key relationships using meta-
analysis; and (ii) to explore the implications of the
quantitative assessment in an expanded narrative that also
discusses some of the major themes characteristic of the
literature on this topic. In the quantitative assessment, I
primarily address spatial relationships between HPD and
particular measures of biodiversity status (see Section II).
Examinations of changes across space are much more
common than assessments of changes over time. The
relationships between human population dynamics and
forest cover or deforestation are not formally reviewed since
they have been dealt with extensively elsewhere (e.g. Geist &
Lambin, 2002; Carr, 2004). However, representative studies
on HPD and forest cover are included in Appendix 1 to
represent accurately the breadth of research on the
reviewed topic.
The quantitative assessment results are organised into
three main sections: general trends (Section III.1); spatial
congruence between people and biodiversity (Section III.2);
and HPD as a threat to biodiversity conservation (Section
III.3). Analyses included in Section III.2 focus primarily on
spatial relationships between HPD and species richness,
while those in Section III.3 provide much stronger evidence
of HPD as a threat to conservation. Meta-analyses are
conducted on discrete relationships between HPD and
particular measures of biodiversity status taking an
exploratory approach. This is appropriate considering the
diversity of methods used to assess HPD relationships across
studies. The aims of the meta-analyses are to determine: the
population effect size for a given relationship (i.e. the
measure of central tendency); if this effect size is
significantly different from zero; the level of heterogeneity
occurring across analyses; possible explanations for this
heterogeneity; and if heterogeneity is reduced within
particular subsets of the data. I also assess the likelihood
of publication bias. In Section IV, I summarise the
implications of the quantitative assessment for our under-
standing of HPD–biodiversity associations and provide an
expanded discussion on: spatial congruence between people
and biodiversity; HPD as a threat to biodiversity conserva-
tion; planning for people and nature; and implications of
future human demographics.
(1) Scope and data collection
The scope of the review was confined to studies that
correlated HPD (or a closely related measure; e.g.
settlement density) with variables that can be meaningfully
related to the status of biodiversity, or where data were
provided that allowed me to conduct correlations. Studies
examining relationships with other measures of population
or human development (e.g. population size, growth rates or
urbanisation) were excluded unless I was able to convert
these measures to HPD.
I searched the following databases using variations of the
key phrase ‘human population density’: Biological
Abstracts, ISI Web of Knowledge, Ovid, Proquest and
ScienceDirect. I also searched the reference lists of papers
identified through these databases to locate additional
studies. Studies that did not measure spatial or temporal
change in biodiversity status or contain data appropriate for
quantitative analysis were discarded. The extent of the
review was limited to papers published in English between
1985 and 2005 (including Va
´zquez & Gaston, 2006, which
was published online in 2005).
Some authors suggest vetting the quality of papers
included in literature reviews (e.g. Gates, 2002), although
others caution against this practice owing to its subjectivity
and the fact that no study is methodologically perfect
(Hunter & Schmidt, 2004). Subjective decisions regarding
the quality of studies can introduce bias into the review process
logical flaws in papers that report findings contrary to their
own world view (Mahoney, 1977; Englund, Sarnelle &
Cooper, 1999). For studies included in this review (Appendix
1), 96% were published in peer-reviewed journals, which I
consider to be a reasonable assessment of the quality of
the work.
I located 85 studies covering 401 analyses (studies often
included multiple analyses of various contexts or taxonomic
groups) that explicitly linked HPD or a surrogate (or
provided data that could be used to calculate HPD) with
a measure(s) of biodiversity change or status. ‘Analysis’ was
used as the sampling unit and the data collected (variables)
for each analysis are described in Table 1 (and presented in
Appendix 1). All analyses were observational (i.e. none
Human population density and biodiversity 609
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
included experimental manipulation). Approximately 50%
of studies provided data on the variation of HPD across the
study area and this was most commonly presented as
a range. Therefore, it is virtually impossible to make
meaningful interpretations of how incremental changes in
HPD may affect biodiversity across different contexts
without the raw data from these studies (which was beyond
the scope of this review). Hence, I excluded these values of
HPD from the data presented.
(2) Meta-analysis – general methods
Meta-analysis is widely used in particular disciplines (e.g.
psychology and medicine), with increasing acceptance
among ecologists (Fernandez-Duque & Valeggia, 1994).
Nevertheless, its use in ecology is controversial considering
the broad diversity of approaches, systems and taxa that
characterise the body of ecological literature (Gurevitch,
Curtis & Jones, 2001). Yet, a meta-analytical approach to
specific questions yields a much more cogent understanding
of general trends than a narrative review or the statistically
inappropriate procedure of ‘vote-counting’ whereby signif-
icant and non-significant results are simply tallied across
studies (Gurevitch et al., 2001; Gates, 2002; Hunter &
Schmidt, 2004).
The Pearson product-moment correlation coefficient was
used as the effect size for each analysis because most studies
reported on the correlation between HPD and change in
biodiversity (Appendix 2 lists all effect and sample sizes; the
impact of spatial autocorrelation is not considered in these
effect sizes). The Pearson coefficient is commonly used as an
effect size in ecological meta-analyses (e.g. Bender, Contreras
& Fahrig, 1998; Gates, 2002; Lampila, Mo
¨nen &
Desrochers,2005; Leimu & Koricheva, 2006) and is a useful
measure because it encapsulates the strength and direction
of a relationship and its squared value is the amount of
variance explained by a given predictor variable (Møller &
Jennions, 2002; Lampila et al., 2005). Spearman and
Table 1. An explanation of the variables collected from each analysis (presented in Appendix 1)
Variable Explanation
Location The location of the analysis to the level of country and above (e.g. continent and globe).
Extent The extent of the area over which data were collected [e.g. a global study (the location)
may be confined to developing countries (the extent)]. Grouped into seven levels: global;
sub-global (including Northern and Southern Hemisphere and subsets of disjunct countries);
continental (including Australia); sub-continental (including sub-Saharan Africa and Latin America);
national (countries, island archipelagos or regions with areas >90,000 km
); regional
(regions and countries with areas O90,000 km
); and local (e.g. cities).
Grain size
The size of the sampling unit used to collect the data [e.g. a study confined to the contiguous
United States (the extent) may collect data for each state (the grain size)]. Grouped into
20 levels and ranked from smallest (1; 100 m
plots) to largest (20; countries) based on
mean actual or estimated size of the sampling unit.
Sample size (N) The number of sampling units. This can vary across analyses within a study because of filtering
so I provide an upper bound in Appendix 1, although actual sample size is used in the
meta-analyses. Occasionally, I had to estimate the sample size when the author(s)
did not provide it.
Response variable The variables that were examined for correlations with HPD. Grouped into two largely mutually
exclusive categories: (i) 12 broad ‘taxonomic’ categories: amphibians, biodiversity hotspots,
birds, fish, forest, herptiles, invertebrates, multi-taxa (Ptwo taxonomic groups combined),
plants, protected areas, reptiles and wetlands (including watersheds); and (ii) 10 type categories:
broad community (encompassing most members of a particular taxonomic group), common species,
endemics, extinctions, functional groups (e.g. carnivores or herbivores), introduced species
(including those native to the country but introduced to the area), restricted species (with a
restricted geographic distribution), single species, taxonomic subgroups (a subgroup of a broader
taxonomic group; e.g. snakes or primates), and threatened species.
Measure of change Grouped into 10 categories: abundance (e.g. abundance of individual species), deforestation
(e.g. per cent area deforested or rate of forest loss), density (density of individuals),
extinction (e.g. rate or risk), forest cover, proportion (proportion of species in particular groups),
protected area (e.g. size of protected reserve), richness (species richness or richness of
certain groups), status (species status), and other.
Direction of relationship (P/N) An indication of whether the relationship between HPD and the component of biodiversity
was positive (P) or negative (N).
Magnitude of relationship Measured using the following values if provided: bivariate correlation (BC), partial correlation (PC),
bivariate regression (BR), partial regression (PR) and multivariate regression (MR).
Grain size categories from smallest to largest: 1 (100 m
); 2 (400 m
); 3 (small settlement); 4 (6.25 km
); 5 (lake); 6 (25 km
); 7 (100 km
8 (250 km
); 9 (quarter degree, small county, province, township); 10 (city); 11 (conservation reserve); 12 (county, district, half degree); 13
(one degree); 14 (biodiversity hotspot); 15 (latitudinal band); 16 (United States ecoregion); 17 (endemic bird area); 18 (island); 19 (U.S.
states, Mexican states); and 20 (country, geographic range, geopolitical unit, wilderness area).
Gary W. Luck610
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
Kendall rank-order correlation coefficients were trans-
formed to Pearson coefficients using the equations docu-
mented in Rupinski & Dunlap (1996). Other effect sizes
(e.g. For tvalues) were transformed to Pearson coefficients
where appropriate using MetaWin 2.0 (Rosenberg, Adams &
Gurevitch, 2000). For studies reporting only results of
bivariate regressions, the square root of the coefficient of
determination (r
) was taken. Correlation coefficients
were subject to the Fisher Z-transformation prior to
MetaWin 2.0 (Rosenberg et al., 2000) was used to
calculate the population effect size (weighted mean Z-
transformed coefficient: r), 95% confidence internals (CI )
using boot-strap re-sampling procedures (4999 iterations),
and total heterogeneity across analyses (Q
). Meta-analyses
were based on random-effects models, which are suitable
for ecological data (Gurevitch & Hedges, 1999; Gurevitch
et al., 2001). I also used MetaWin to explore the partitioning
of heterogeneity using continuous and categorical meta-
analytical models (see below).
Publication bias can occur when non-significant results
have a higher rejection rate from scientific journals or by
authors deciding not to submit such results. This can lead to
a positive bias in reported effect sizes (i.e. larger effect sizes
are those most likely to be published), which results in an
artefactually magnified population effect size (Begg &
Berlin, 1988; Gurevitch et al., 2001). Bias can also occur
when unexpectedly large effect sizes are found especially
when sample size is small (also leading to non-publication;
Gurevitch et al., 2001). A common graphical method for
assessing the likelihood of publication bias is the funnel plot
(Light & Pillemer, 1984). Wang & Bushman (1998) highlight
a number of problems with funnel plots including
interpreting patterns in the data and the fact that the
method provides no assessment of the assumptions of
normality. The authors suggest using normal quantile plots
to circumvent these problems. In a normal quantile plot,
the quantiles of the effect size distribution are plotted
against the quantiles of a standard normal distribution.
Data should fall along a straight line within 95% confidence
bounds and with no obvious gaps if there is no publication
bias and assumptions of normality are met (see Wang &
Bushman, 1998 for details).
I assessed publication bias using normal quantile plots,
Rosenthal’s ‘fail-safe’ number (Rosenthal, 1979) and a rank
correlation test for meta-analytical data (Begg, 1994; Begg &
Mazumdar, 1994). The fail-safe method calculates the
number of unpublished studies with a mean effect size of
zero that are required to reduce the significance level of
a population effect size from a set of published studies to
a¼0.05 (i.e. just significant; see Hunter & Schmidt, 2004
for details). A rule of thumb is that if the fail-safe number is
greater than 5N]10 (where Nis the number of studies in
the meta-analysis) it is highly unlikely that such a number of
unpublished studies exist (Rosenthal, 1979). The rank
correlation test examines the relationship between the
standardised effect size and the sample size across studies. A
significant correlation may indicate publication bias, but the
test has lower power when N<25 so these results should be
interpreted with caution.
(3) Meta-analysis – specific methods
In the first instance, all studies and analyses addressing
a particular topic that included a correlation coefficient or
from which a coefficient could be calculated were included
in the meta-analysis irrespective of non-independence
issues. Studies that reported no relationship between HPD
and the response variable (i.e. they did not provide
a correlation coefficient or a measure suitable for trans-
formation, but stated qualitatively that there was no
relationship) were also included and given the nominal
coefficient value of zero (this is a conservative approach to
protect against the exclusion of non-significant results and
in most meta-analyses constituted <10% of all analyses).
These are referred to in the results as ‘full meta-analyses’.
In many cases a second meta-analysis was conducted that
removed non-independent analyses. These are referred to
as ‘filtered meta-analyses’ and can be considered an
assessment of the sensitivity of the results in relation to
non-independence (Gates, 2002). Non-independent analy-
ses for a given study were defined as those that were based
on a subgroup of a broader taxonomic group already
included in the study or where sampling grain or extent
were varied, but no other measures changed. For example,
if a study reported a correlation between HPD and bird
species richness and a correlation between HPD and the
richness of avian frugivores, the latter analysis was excluded
from the filtered meta-analysis. If a study examined
correlations across various grain sizes (reporting on the
results for each size), only one sampling grain size was used
in the filtered meta-analysis (chosen at random). Analyses
were considered independent if a study examined different
taxonomic groups despite all other factors (e.g. location and
grain size) being the same.
In certain cases where there was significant heterogeneity
across all analyses, continuous and categorical meta-
analytical models were used to explore the partitioning of
variance among potential influencing factors (see Rosenberg
et al., 2000 for an explanation of continuous and categorical
models). These are explained in full in the results. If
sample sizes permitted, meta-analyses were also conducted
on pre-determined subsets of data within a particular topic.
For example, under threatened species I conducted a meta-
analysis on the relationship between HPD and threatened
bird species. This was to determine the level of heterogeneity
occurring among analyses with a much narrower (more
refined) classification.
Meta-analyses were conducted on the relationship
between HPD and: species richness (excluding the rich-
ness of restricted, threatened or introduced species); the
richness/proportion of threatened species; the richness/
proportion of restricted species; extinctions (rate and
probability); the richness/proportion of introduced species;
and area of protected land. Table2 details the analyses
included and the approach taken in each meta-analysis.
Information was also collected on the relationships between
HPD and the species richness of particular taxonomic (e.g.
reptiles) and functional (e.g. carnivores) groups, although
owing to data limitations meta-analyses were not always
Human population density and biodiversity 611
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
Table 2. Results of all meta-analyses (source studies and analyses are listed in the footnotes and Appendix 2). Full meta-analyses
include all analyses regardless of independence and filtered meta-analyses exclude non-independent analyses. Meta-analyses on
particular taxonomic groups are based on subsets of data from the filtered meta-analyses. Sample size indicates the number of
analyses in each meta-analysis with the number of studies in brackets. Total Nis the sum of sample sizes for all analyses included in
the meta-analysis, ris the population effect size, Q
is the total heterogeneity among analyses and CI is the confidence intervals of
the population effect size. Results are also shown for the rank correlation test and fail-safe N. The fail-safe Ncan be compared to the
cut-off N, which is the number of analyses*5 ]10 (see Section II.2). ***P<0.001, **P<0.01, *P<0.05.
(see footnotes) Meta-analysis
size Total NrQ
correlation Fail safe NCut-off N
Species richness
A Full 127(25) 95580 0.407 328.3*** 0.327–0.487 0.155 11367 645
B Filtered 60(25) 35945 0.566 105.8*** 0.465–0.666 0.095 4828 310
C Birds 15(12) 10132 0.586 20.8 0.381–0.779 0.068 241 85
D Mammals 16(9) 7782 0.606 30.0* 0.441–0.780 [0.156 522 90
E Plants 9(8) 1099 0.324 12.4 0.076–0.628 [0.586 19 55
Threatened species
F Full 44(18) 26505 0.347 78.8*** 0.233–0.466 [0.073 876 230
G Filtered 39(18) 20582 0.316 55.0* 0.199–0.433 [0.038 492 205
H Birds 12(10) 4278 0.415 19.1 0.206–0.632 0.011 89 70
I Mammals 10(8) 2108 0.062 7.0 [0.143–0.305 0.024 45 60
Restricted species
J Full 20(5) 14209 0.276 24.7 0.207–0.351 [0.296 476 110
K Filtered 14(5) 9042 0.273 20.5 0.184–0.365 [0.384 234 80
L Full 19(9) 1169 0.304 33.4* 0.143–0.509 [0.426 126 105
M Filtered 10(9) 763 0.370 14.1 0.163–0.628 [0.565 40 60
Introduced species
N Full 12(8) 623 0.612 8.2 0.372–0.867 [0.303 62 70
O Plants 8(8) 500 0.689 4.3 0.390–0.996 [0.238 25 50
Protected areas
P Full 33(9) 2030 [0.526 40.6 [0.661–[0.403 0.048 794 175
Q Filtered 12(9) 1936 [0.537 15.9 [0.754–[0.375 [0.074 191 70
Studies (and analyses) included in each meta-analysis. Cross reference with Appendices 1 and 2.
A. 3(6–10, 13–18), 4(22–26, 28, 30–39), 14(65–70), 15(71,73,75,78), 16(79–80, 83–84, 87–88), 18(92, 94), 19(95), 23(113), 24(116–131), 25(132–133),
26(135–136, 139–142), 28(144–148), 29(149), 42(183), 46(202), 50(211–212, 215–216, 219–221), 54(236–241, 252–257), 61(292), 64(299–308), 70(331–332,
334), 72(341–343), 73(347–348), 80(365–366), 81(374), 82(375, 377, 379–383).
B. 3(6–9, 13), 4(22–26), 14(65–70), 15(71, 73, 75, 78), 16(83), 18(92), 19(95), 23(113), 24(116), 25(132–133), 26(135), 28(144), 29(149), 42(183), 46(202),
50(211–212, 215–216), 54(236–241, 252–257), 61(292), 64(299–303), 70(334), 72(343), 73(348), 80(365), 81(374), 82(375). Another meta–analysis was
conducted with study 24(116) excluded since this used a substantially different measure of HPD. This resulted in a slightly larger population effect size
(0.561) and slightly less total heterogeneity (105.7).
C. 3(7), 4(24), 15(75,78), 16(83), 23(113), 26(135), 28(144), 29(149), 50(211, 215), 54(238, 254), 64(301), 81(374).
D. 3(9), 4(25), 14(65–70), 19(95), 46(202), 50(212, 216), 54(240, 256), 64(300), 82(375).
E. 15(71), 18(92), 25(132–133), 42(183), 61(292), 64(299), 70(334), 72(343). 3(6) was identified as an outlier because of a relatively large sample size (N¼
2434) and is not included in the results presented. Inclusion does not change the qualitative conclusions of the meta–analysis.
F. 3(19–21), 4(29), 16(82, 86, 90), 22(108–109, 112), 26(138), 29(154), 45(197–198), 46(200, 203), 50(217–218), 52(224), 54(247–251, 263–267), 55(268),
56(270, 272, 274, 276), 60(289–291), 65(310–313), 75(353), 80(367), 82(376). Analyses 82 and 138 were identified as outliers apparently because they had
both large sample sizes and large effect sizes, however they are included in the results presented.
G. 3(21), 4(29), 16(86), 22(108–109, 112), 26(138), 29(154), 45(197–198), 46(203), 50(217–218), 52(224), 54(247–251, 263–267), 55(268), 56(270, 272, 274,
276), 60 (289–291), 65(310–313), 75(353), 80(367), 82(376).
H. 16(86), 26(138), 29(154), 45(197), 50(217), 54(247, 264), 56(270, 274), 60(290), 65(310), 75(353).
I. 45(198), 46(200), 50(218), 54(250, 266), 56(272, 276), 60 (291), 65(312), 82(376).
J. 3(11–12), 4(27), 16(81, 85, 89), 54(242–246, 258–262), 75(353, 355–357). Analyses 27 and 81 were identified as outliers apparently because they had both
large sample sizes and large effect sizes, however they are included in the results presented. Results with the outliers excluded are presented in the text.
K. 3(12), 4(27), 16(89), 54(242–246, 258–262), 75(353).
L. 10(55–58), 13(64), 23(115), 39(179–180), 44(191–196), 69(329–330), 73(349), 75(354), 79(364).
M. 10(55), 13(64), 23(115), 39(179–180), 44(191), 69(329), 73(349), 75(354), 79(364).
N. 15(72, 74, 76–77), 18(93), 25(134), 57(277–278), 61(293), 70(333), 71(340), 76(358), 27(143) and 77(360) were identified as outliers because of relatively
large sample sizes (N¼2262 and 3000, respectively) and are not included in the results presented. Inclusion does not change the qualitative conclusions of
the meta-analysis.
O. 15(72), 18(93), 25(134), 57(277), 61(293), 70(333), 71(340), 76(358). Proportion studies included.
P. 10(59), 39(177), 50(210, 213–214), 59(286), 64(309), 69(328), 78(363), 81(369, 372), 82(384). This includes the data from Harcourt et al. (2001) for 22
African countries (see Appendix 2).
Q. 10(59), 39(177), 50(210, 213–214), 59(286), 64(309), 69(328), 78(363), 81(369, 372), 82(384). This includes the data from Harcourt et al. (2001) for the
entire African continent (see Appendix 2).
Gary W. Luck612
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
(1) General trends
Analyses (N¼396; excluding Honduras, and Azores,
Macaronesian and Mediterranean islands) were mostly
conducted across the globe (25%; including those spanning
either the Northern or Southern Hemisphere) or confined
to North America, Africa or Europe (Fig. 1A). Few were
conducted in South or Latin America, Asia or Australia
(2%–7%). At the country level (N¼218), analyses were
dominated by work conducted in the United States (29%).
Our understanding of HPD – biodiversity relationships is
therefore biased by global studies or those conducted in the
Northern Hemisphere.
(b)Spatial extent and grain size
Analyses were mostly conducted at the national level (46%
of 401 analyses) or at larger extents. Few were conducted at
the local level (1%; Fig. 1B). Grain size was skewed towards
very large grains (e.g. countries and geographic ranges: 20%
of 389 analyses) or medium-size units such as counties,
districts or areas ranging from 2,500 km
to 10,000 km
(26%; Fig. 1C). Clearly, more work needs to be done at local
levels using small grain sizes. The results of such analyses
are likely to show less spatial congruence between HPD and
biodiversity (e.g. Turner, Nakamura & Dinetti, 2004). I
assess the relationships between grain size and spatial
congruence below.
(c)Taxonomic groups and measures of change
Not surprisingly, the well-studied groups of mammals and
birds dominated the broad taxonomic classifications with
26% and 24% of 401 analyses, respectively, including these
groups (Fig. 2A). Other common groups studied were plants
(11%) and forests (10%). There were 335 analyses that
could be grouped under a type classification. Of these, 24%
encompassed a broad community, 20% examined threatened
species, 15% covered taxonomic subgroups, 14% single
species, and 12% functional groups (Fig. 2B). Interestingly,
only 5% of analyses included introduced species and it
appears more work needs to be done on the relationship
between HPD and species introductions and establishment
(see below). The measures of change were dominated by
species richness (44%), followed by the proportion of species
(d)Spatial autocorrelation
Spatial autocorrelation occurs when sites close together in
space are more similar than those further apart and this
spatial structure may be an explanation for observed
correlations between dependent and independent variables.
If the objective is to identify the underlying causes of the
spatial pattern then it is crucial that spatial autocorrelation
be considered. The importance of spatial autocorrelation
has only recently been acknowledged and this is reflected in
the fact that only 16% of the 85 studies in this review
addressed the issue either in the presentation of their data
(e.g. not including Pvalues for coefficients), sampling
strategy or statistical analysis, and all of these were
published after 2000 (Balmford et al., 2001; Laurance
et al., 2002; Chown et al., 2003; Hope et al., 2003; Fairbanks,
2004; Gaston & Evans, 2004; Gilbert et al., 2004; Kok,
2004; Luck et al., 2004; Ceballos et al., 2005; Evans &
Gaston, 2005; Evans, Warren & Gaston, 2005; Lavergne
et al., 2005; Va
´zquez & Gaston, 2006; also see Selmi &
Boulinier, 2001 who examined the data of Chown, Gremmen
& Gaston 1998 using spatial and non-spatial models).
Some of these studies examined the relationship between
HPD and species richness comparing the results from non-
spatial and spatial models (Chown et al., 2003; Gaston &
Evans, 2004; Evans & Gaston, 2005; Evans et al., 2005;
´zquez & Gaston, 2006). In most cases the relationship
between HPD and species richness was weakened when
spatial autocorrelation was considered, although it still re-
mained significant for particular comparisons (see Gaston &
Evans, 2004; Evans & Gaston, 2005; Evans et al., 2005;
´zquez & Gaston, 2006). Such analyses are crucial if the
aim is to infer that HPD influences species richness (e.g.
increasing the number of introduced species) or when trying
to identify key drivers of human and species distribution
patterns (e.g. net primary productivity).
(e)Biased sampling effort
An issue that, arguably, is even more problematic than
spatial autocorrelation is the impact of biased sampling
effort on the interpretation of relationships. This is very
rarely addressed in the literature (Balmford et al., 2001;
Luck et al., 2004). Researchers sourcing species distribution
data from atlases or similar broad-scale surveys must be
cogent of the fact that sampling effort is often highest near
densely settled areas. Indeed, Luck et al. (2004) reported
a very strong correlation between bird species richness
(sourced from atlas data) and sampling effort (r
although after controlling for sampling intensity with partial
correlations they still found a relatively strong correlation
between HPD and bird species richness. Sampling effort
can be dealt with post hoc using modelling or statistical
procedures, but a better approach, if possible, is to address
the issue during data collection.
Appendix 1 classifies relationships between HPD and
biodiversity status as either positive, negative or no re-
lationship. This simple classification does not recognise that
some relationships may be non-linear. Only a handful of
studies dealt explicitly with non-linearity in their data
ˇek, 1998; Balmford et al., 2001; Laurance et al., 2002;
Cardillo et al., 2004; Silva & Smith, 2004; Ceballos et al.,
2005; Evans & Gaston, 2005; Evans et al., 2005; Lavergne
et al., 2005; Va
´zquez & Gaston, 2006). These studies
showed, for example, non-linear relationships between
deforestation and HPD with a plateau in deforestation
Human population density and biodiversity 613
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
above a given density (possibly because much of the forest
has already been cleared; Laurance et al., 2002), and non-
linear, quadratic or hump-shaped relationships between
density and species richness or net primary productivity (e.g.
Balmford et al., 2001; Ceballos et al., 2005; Lavergne et al.,
Some studies compared the fit of linear and quadratic
models to the relationship between HPD and species
richness of various groups. These studies typically showed
that quadratic models provide a better fit to the data (Evans &
Gaston, 2005; Evans et al., 2005; Va
´zquez & Gaston, 2006).
This means that species richness peaks at mid-range
Percentage of analyses
North America
South/Central America
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2
Grain size
Fig. 1. The percentage of all analyses conducted for each: (A) location classified by globe or continent and ranked from most to
least analyses (N¼396); (B) extent ranked from largest to smallest (N¼401); and (C) grain size ranked from smallest to largest (N¼
Gary W. Luck614
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
values of HPD. Intuitively this makes sense, as highly
urbanised areas with extreme population densities are likely
to support very few species (Turner et al., 2004), although
much more research is required before any general
conclusions can be drawn.
(2) Spatial congruence between people and
(a)Species richness
A total of 127 analyses related the species richness of
various groups with HPD (excluding threatened, restricted
and introduced species). Patterns in the location and extent
of richness analyses were generally congruent with those
found for all analyses (see above). Grain size was heavily
skewed towards mid-range units (counties and 2,500 km
10,000 km
; 45%). Birds (27%) and mammals (22%) were
the most common groups studied and most analyses
included the entire assemblage of species in a particular
taxonomic group (52%, including native species only).
There was no strong evidence of publication bias in these
analyses. The normal quantile plot showed that data had
a reasonable straight-line fit with all points falling within the
95% confidence bounds (Fig. 3A). The rank correlation
between the standardised effect size and sample size was
weak, and the fail-safe number was substantially larger than
the cut-off value (Table2). However, the quantile plot had
a distinct S shape suggesting these data come from more
than one population (Wang & Bushman, 1998).
Heterogeneity in the data is emphasised by the results of
the meta-analysis. Although a significant positive effect size
between HPD and species richness was found (r¼0.407,
CI ¼0.327–0.487), total heterogeneity (Q
) was substantial
(328.3, d.f. ¼126, P<0.001). The filtered meta-analysis
excluding all non-independent analyses (see Section II.3)
yielded a larger population effect size (0.566), but there was
still substantial heterogeneity among the results (Table 2).
I explored three possible, logical causal factors that may
have led to the observed heterogeneity: a country’s level of
development; grain size; and taxonomic classification. I
differentiated between developed and developing countries,
as it is reasonable to expect a tighter relationship between
HPD and the local environment in developing countries
because their populations are more reliant on resources
extracted locally. Indeed, Huston (2001) hypothesised that
Broad community
Threatened species
Single species
Functional groups
Introduced species
Restricted species
Common species
Type categories
Protected areas
Taxonomic categories
Percentage of analyses
Fig. 2. The percentage of analyses conducted for each: (A) taxonomic category ranked from most to least analyses (N¼401); and
(B) type category ranked from most to least analyses (N¼335).
Human population density and biodiversity 615
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
there would be a greater degree of spatial congruence
between HPD and biodiversity in developing compared to
developed regions. The classification of developed and
developing regions was based on United Nations rankings
Grain size is a logical variable to explore since it is likely
that correlations between HPD and species richness are
stronger with larger grain sizes. This is because larger grain
sizes are more likely to contain both densely populated and
species-rich regions. Splitting data by taxonomic group is
also reasonable, as not all groups may exhibit the same
relationship with HPD (Luck et al., 2004).
A categorical model was used on the filtered data to
determine the degree of heterogeneity explained by level of
development (Rosenberg et al., 2000). The population effect
size for developed regions (r¼0.547, CI ¼0.351–0.743,
N¼21) was slightly larger than that for developing regions
(r¼0.538, CI ¼0.381–0.695, N¼19) contrary to
expectations. Also, this variable did not explain a significant
degree of variance in the data (Q
¼0.01, d.f. ¼1, P¼
0.930; Q
¼68, d.f. ¼38, P¼0.002: where Q
is the
amount of variance or heterogeneity explained by between-
group differences whereas Q
is the amount explained by
within-group differences). That is, a region’s level of
Fig. 3. Normal quantile plots for: (A) species richness (N¼127); (B) threatened species (N¼44); (C) restricted species (N¼20); (D)
extinctions (N¼19); (E) introduced species including the two outliers (see Section III.3; N¼14); (F) introduced species with the two
outliers excluded (N¼12); and (G) protected areas (N¼33).
Gary W. Luck616
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
development could not explain the heterogeneity in the
results of HPD versus species richness.
A continuous model with grain size as the potential
explanatory variable explained a significant proportion of
variance in the data (Q
¼7.2, d.f. ¼1, P¼0.007),
although significant heterogeneity still remained unex-
plained (Q
¼105.8, d.f. ¼58, P<0.001; Fig. 4). As
expected, the correlation between HPD and species
richness was stronger as grain size increased, but this
relationship only explained a small proportion of total
variance in the data.
A categorical model splitting the data into discrete
taxonomic groups (amphibians, birds, invertebrates, mam-
mals, multi-taxa, plants and reptiles) did not explain
a significant proportion of variance in the data (Q
6.5, d.f. ¼6, P¼0.370). I explored this issue further by
conducting meta-analyses on pre-defined subsets of the data
split along taxonomic lines. These analyses found no
significant heterogeneity among the effect sizes for HPD
and the richness of birds or plants, although there was still
significant heterogeneity in the mammal data. Population
effect sizes were positive and significant for each group,
although substantially weaker for plants (Table 2). The
analyses for HPD versus plant species richness showed the
strongest indications of publication bias with a large, albeit
non-significant rank correlation ([0.586, P>0.05, N¼9;
note the power of this test is weak when N<25) and
a relatively small fail-safe number.
The influence of grain size on the outcomes of HPD –
biodiversity assessments warrants further attention. Only
five of the reviewed studies altered grain size while keeping
everything else constant. Borralho et al. (1996) found that
the negative relationship between the relative abundance of
the Egyptian mongoose Herpestes ichneumon and HPD was
stronger at a smaller grain size. Chown et al. (2003) found
that the positive relationship between the species richness of
birds overall, and for narrowly and widely distributed
species and threatened species, increased in strength as
grain size increased from 0.25–0.5 to 1°. Kok (2004) also
altered grain size from 56 km
to 2025 km
, but there was
no consistent pattern with results including HPD. Rowland
et al. (2003) reported no relationship between HPD and
wolverine Gulo gulo abundance at the sub-basin grain size,
but found a negative relationship at the much smaller
watershed grain indicating how grain size can have
important implications for management decisions.
(b)Threatened species
These analyses are included under the section on spatial
congruence because most studies simply examined the
correlation between HPD and threatened-species richness.
When positive correlations exist it does not necessarily
imply that increasing HPD is the reason for an increased
number of threatened species in a given region, although
high human density may complicate conservation strategies.
Analyses on threatened species tended to be conducted
using larger grain sizes when compared to all analyses (58%
of 66 analyses ranked in the top five grain sizes compared to
37% for all analyses). Once again birds and mammals
dominated the taxonomic groups, although there was more
of a skew towards birds (38%) than for all analyses
combined (27% of 350 analyses with non-plant and non-
animal categories removed).
A total of 44% of analyses used the proportion or
percentage of threatened species as the measure of change.
This may be preferable to using the number of threatened
species because threatened species richness will likely
increase with the total number of species, confounding the
relationship between threatened species and HPD. Never-
theless, important information may be gained from using
the number of threatened species if total species richness is
controlled for.
The normal quantile plot for threatened species was
weakly suggestive of publication bias with a slight U shape
(Fig. 3B), but this is not supported by the rank correlation
test or fail-safe number (Table 2). Two analyses occurred
outside the 95% confidence bounds. These analyses appear
to be outliers because they have both large sample sizes and
large effect sizes (analyses 82 and 138, Appendix 2).
There was a significant positive relationship between
HPD and the richness/proportion of threatened species in
the full and filtered analyses, but both exhibited signifi-
cant heterogeneity (Table 2). A possible reason for this
Fig. 4. The relationship between grain size and effect size for analyses correlating human population density with species richness
(N¼60; non-independent analyses excluded).
Human population density and biodiversity 617
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
heterogeneity is differing results among analyses that used
the number of threatened species as the response variable
compared to those that used the proportion or percentage
of threatened species. Analyses measuring HPD versus the
number of threatened species had a larger population effect
size (r¼0.473, CI ¼0.255–0.683, N¼13) than those
measuring the proportion of threatened species (r¼0.245,
CI ¼0.115–0.386, N¼26). A categorical model indicated
a significant difference between these groups (Q
d.f. ¼1, P¼0.05) although a large, albeit non-significant
amount of heterogeneity remained unexplained (Q
47.6, d.f. ¼37, P¼0.11).
Data were not suitable to examine using a categorical
model based on taxonomic classification, but I explored
relationships with data subsets of the most commonly
represented taxonomic groups. The heterogeneity in the
results for HPD versus threatened birds approached
significance (P¼0.06), but there was no significant
heterogeneity for threatened mammals, although the
population effect size for mammals was not significantly
different from zero (r¼0.062, CI ¼[0.143–0.305; Table 2).
In general, human settlements tend to overlap with areas
that contain a high number of threatened species, although
this relationship is weakened when total species richness is
accounted for (i.e. using proportion or percentage of
threatened species as the response variable) and for some
taxonomic groups. More work is required to determine the
direct contribution of HPD to species decline (see below).
However, assuming that high HPD is a general threat to
conservation, the protection of many threatened species
may be compromised by their spatial congruence with
human settlements.
(c)Restricted species
Restricted species are those with a limited geographic
distribution. These species are of potential concern because,
if dense human settlements pose a threat to their
conservation and are spatially congruent with their
distribution, they have limited options in avoiding such
threats by exploiting areas of their geographic range where
human impacts are reduced. This issue has received little
attention in the literature with only five studies (encompass-
ing 20 analyses) examining the relationship between HPD
and restricted species (Appendix 1). A high proportion of
analyses were on birds (47%), yet birds may not be the most
appropriate taxonomic group to explore this issue since they
include some of the most mobile species with relatively
broad geographic ranges. The relationship between HPD
and taxa with limited mobility and relatively extreme
geographic restrictions (e.g. small reptiles and certain
invertebrates) warrants further attention.
There was no strong indication of publication bias in
these data, although two analyses occurred outside the 95%
confidence bounds (Fig. 3C). It appears these analyses are
outliers because they have both large sample sizes and large
effect sizes (analyses 27 and 81, Appendix 2). The full and
filtered analyses showed a significant positive relationship
between restricted species richness and HPD with no
significant heterogeneity among the analyses even when
studies measuring the number or proportion of restricted
species were combined (Table2). However, removal of the
outliers in the filtered meta-analysis resulted in significant
heterogeneity among the effect sizes (Q
¼28.4, d.f. ¼12,
(3) HPD as a threat to biodiversity conservation
Documenting species extinctions in a local or global context
is notoriously difficult. Therefore, it is not surprising that
only 11 studies covering 32 analyses attempted to link HPD
with some measure of extinction (i.e. extinction rate, risk or
probability, or the number or proportion of extinct species).
These results are dominated by the study of Woodroffe
(2000), which examined the relationships between HPD
and the probability of extinction for various mammal
species across Africa, Brazil, India and the United States.
This study was criticised on the grounds that it used
historical patterns of carnivore extinctions at a time when
carnivores were persecuted. Linnell, Swenson & Andersen
(2001) found that current large-carnivore populations were
either increasing or showed no response despite increases in
HPD across North America and Europe. They argued that
favourable policy change and effective management have
more influence on species conservation than changes in
A total of 19 analyses were suitable to include in a meta-
analysis, although the following results should be inter-
preted with caution owing to substantial differences in the
approaches used to assess extinction. There was no
indication of publication bias in the normal quantile plot
(Fig.3D), although the rank correlations were relatively high
and the fail-safe numbers were relatively low (Table2). The
full and filtered meta-analyses suggested a significant
positive relationship between HPD and extinction, although
there was significant heterogeneity among the effect sizes
for the full meta-analysis.
The most convincing studies so far that attempt to link
HPD with species extinction are those of Brashares, Arcese &
Sam (2001), Harcourt, Parks & Woodroffe (2001) and Parks &
Harcourt (2002). These studies were conducted at a very
local level comparing the extinction rates of certain taxa
within conservation reserves to reserve size and surrounding
HPD (within a 50–100 km radius). The results of these
studies indicated that surrounding HPD and/or reserve
area are generally good predictors of extinction in reserves.
Furthering this work requires the assessment of a greater
range of predictor variables to tease apart landscape and
human factors, and also accounting for any ecological or
phylogenetic predisposition to extinction (e.g. slow repro-
ductive rate).
(b)Introduced species
Introduced species are included under the section on HPD
as a threat to conservation owing to the well-documented
adverse impacts that introduced species can have on native
ecosystems (Mooney et al., 2005). Only 10 studies
Gary W. Luck618
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
comprising 18 analyses linked HPD with the richness or
proportion of introduced species. Most of these (72%) used
a large grain size (islands, large states and countries), and in
contrast to the results discussed above, 56% of analyses
were on plants. Only three analyses used proportion of
introduced species as the measure of change, yet the
number of introduced species is likely to be related to total
species richness and the latter needs to be accounted for
when assessing the relationship between HPD and exotics
(some researchers did this using partial correlations; e.g.
´a, Villasen
˜or & Vibrans, 2004).
The normal quantile plot for introduced species showed
two analyses as substantial outliers as a result of extremely
large sample sizes (Fig. 3E, F; analyses 143 and 360, Appendix
2). With these outliers removed, the full meta-analysis showed
a very strong positive population effect size between HPD
and introduced species richness with no significant heteroge-
neity (Table 2; a filtered analysis is not included as this was the
same as the full analysis). Including the outliers does not alter
this conclusion since the effect sizes (correlation coefficients)
for the two analyses were 0.62 and 0.70. A meta-analysis of
data on introduced plants was also included to assess changes
in the level of heterogeneity, which was substantially reduced
in this subset (Table2). Because of the nature of these data
and the relatively small fail-safe numbers, conclusions about
the true relationship between HPD and introduced species
should be cautious.
Many introduced species that have successfully estab-
lished populations in new regions benefit from landscape
modification by humans (Mooney et al., 2005). It is therefore
surprising that more work hasn’t been done on the
connection between HPD and exotics, although much has
been done on related factors such as urbanisation
(McKinney, 2002c).
(c)Protected areas/important conservation areas
A total of nine studies comprising 21 analyses compared
HPD and protected areas or important conservation areas
(important bird areas and wilderness areas were included in
this latter category). Most analyses (76%) contrasted HPD
with the percentage or absolute area of land protected.
Two meta-analyses on the relationship between HPD and
area of protected land were conducted. The full meta-
analysis included results from Harcourt et al. (2001) on 22
African countries, while the filtered meta-analysis in-
cluded only the correlation between HPD and protected
area for the entire African continent from the same study
(Appendix 2).
The normal quantile plot (Fig. 3G), rank correlation tests
and fail-safe numbers suggested no publication bias in the
data, although the quantile plot identified a substantial
outlier to the population. This was analysis #328 that had
a sample size of 864 (Appendix 2), which was more than
twice as large as the next nearest sample size. The results
with or without this outlier were almost the same, so I only
present results with the outlier included. These showed
a significant negative population effect size between HPD
and area of protected land with most of the heterogeneity in
the data being explained by sampling error (Table 2). As
HPD increases, the amount of land set aside for con-
servation substantially decreases.
It is not surprising that the size or absolute cover of
protected areas declines with increasing HPD. The majority
of land comprising or close to human settlements is
designated for urban development or agriculture. Moreover,
it is mostly of high productivity (Luck, 2007). Protected
areas are usually characterised by land of low productivity
unsuitable for other human uses (Pressey, 1994, 1995),
hence large protected areas are commonly isolated from
human development. The trend is disconcerting owing to
the strong correlations between HPD and species richness
for many taxonomic groups. High-density and species-rich
regions are often afforded the least protection (in the sense
of land cover). Protected areas close to human settlements
suffer from ‘double jeopardy’ (sensu Harcourt et al., 2001):
they are small, which makes them susceptible to external
impacts, and they are surrounded by high HPD potentially
undermining their capacity to afford adequate protection to
their associated ecosystems. Such a conundrum presents
a substantial challenge to conservation managers wishing to
preserve the integrity of protected areas, and to conserva-
tion planners hoping to maximise the effectiveness and
completeness of the reserve system and to ensure that at
least some large reserves conserve highly productive and
species-rich regions.
(d)Individual species
There were 47 analyses across 13 studies that examined the
relationships between HPD and individual species. The
majority of analyses were on mammals (80%), with only two
analyses on a taxonomic group other than mammals or
birds. Twenty-two analyses compared HPD with the
abundance or density of the focal species, and 77% found
a negative relationship. Although analyses that include
entire species assemblages often report a positive relation-
ship between HPD and species richness, studies of the
abundance of individual species commonly report a negative
correlation. These data are not conducive to meta-analysis
owing to the extreme heterogeneity among the approaches
used in the various studies.
It is possible that researchers bias the results for
individual species by selecting species likely to be sensitive
to human impacts. Indeed, studying the relationship
between humans and common, widespread species is
probably relatively unattractive unless such species are
considered pests. Of the five analyses showing a positive
relationship between HPD and species abundance, four of
these were on pest species. Greater research needs to be
done on the connection between humans and more
disturbance-tolerant species to improve our understanding
of the HPD – biodiversity relationship. Moreover, and
rather surprisingly, little has been done on the link between
HPD and the abundance of pest species.
Only one of the reviewed studies correlated HPD with
the reproductive success of a particular species. Hansen and
Rotella (2002) found that the nesting success of yellow
warblers Dendroica petechia was negatively related to HPD,
while there was no relationship with the success of American
Human population density and biodiversity 619
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
robins Turdus migratorius. Such studies are crucial in
understanding the threat posed by human populations
and this area of research remains wide open.
(1) Geographic, taxonomic and sampling biases
Our understanding of the relationships between HPD and
biodiversity are skewed by biases in the literature. Most
analyses were conducted in the Northern Hemisphere with
work at the country level dominated by studies in the
United States. Few analyses were conducted at local levels
or using small grain sizes. The potential ‘direct’ effects of
change in HPD (e.g. habitat clearance for housing) are
better studied across local to regional extents with suitably
sized sampling units that allow a clear demarcation between
developed and undeveloped areas. Documenting ‘indirect’
effects (e.g. abundance of invasive species) may require
analyses using larger grain sizes across local to global
Vertebrates, particularly mammals and birds, were used
as the main indicators of biodiversity response to changes in
HPD. A wider breadth of taxonomic groups needs to be
studied to improve our understanding of potential threats.
Moreover, little work has been done on the response of
particular functional groups. There were 41 analyses that
included functional groups, but these were dominated by
just three studies accounting for 73% of analyses (Jokima
¨ki &
Suhonen, 1998; Balmford et al., 2001; Drake & Pereira,
2002). It is possible that certain functional traits of
organisms increase their susceptibility to threats posed by
human populations. For example, large mammals may have
suffered disproportionately with increases in HPD (e.g.
Parker & Graham, 1989; Liu et al., 1999b; Woodroffe, 2000;
Walsh et al., 2003; Lamprey & Reid, 2004; Cardillo et al.,
2005; Graham, Beckerman & Thirgood, 2005), but the
generality of this relationship is debatable and other
ecological traits appear to contribute to a species’ suscep-
tibility to decline (Ceballos & Ehrlich, 2002; Cardillo et al.,
2004). Further research on specific functional guilds will
help elucidate the consequences of HPD for ecosystem
functioning and the provision of ecosystem services. Also,
the relationships between HPD and introduced and invasive
species needs to be examined in much greater detail
considering the major threat invasive species pose to
indigenous biodiversity (Pimentel, 2002, Mooney et al.,
Analyses of the abundance of individual species,
dominated by mammals, showed overwhelmingly a decline
in abundance with increases in HPD (although these data
were not suited to a meta-analysis). I hypothesise that these
results are biased by an emphasis on sensitive species likely
to show negative responses to increasing human develop-
ment. From a conservation perspective it is logical to focus
on such species, yet studies on disturbance-tolerant species
will help to broaden our knowledge of the links between
HPD and biodiversity change. Surprisingly, I could find
only one study that made explicit links between HPD and
the reproductive success of species (Hansen & Rotella,
2002). To improve our understanding of the threats posed
by human populations it is crucial to expand the number of
studies on the relationships between HPD and species
fecundity, survival and mortality rates; an area that has been
largely ignored.
Almost all studies reviewed examined spatial variation in
HPD and biodiversity; few analyses incorporated temporal
variation (e.g. Kummer & Sham, 1994; Palo, Lehto &
Uusivuori, 2000; Fairbanks, 2004). Temporal analyses offer
a particularly rigorous examination of the threats humanity
poses to conservation. Tracking changes over time in local
or regional diversity, for example, and linking these with
changes in HPD across the same spatial extent and over the
same time period would contribute valuable insights into
the role human numbers plays in biodiversity change. Such
analyses are hampered by a lack of quality time series data,
particularly for variables such as species richness or
Finally, little work has been done on what is arguably
the most pervasive evidence of the negative impact of
increasing HPD: species extinction and extinction risk.
Studies conducted at local levels provide reasonably strong
evidence for HPD as a driving factor in species extinction
(e.g. Thompson & Jones, 1999; Brashares et al.,2001;
Harcourt et al., 2001; Parks & Harcourt, 2002), yet the
results from studies at global levels are equivocal (e.g.
Woodroffe, 2000; Linnell et al., 2001; Cardillo et al., 2004).
(2) Publication bias
In all of the meta-analyses conducted, evidence for
publication bias was generally weak. This result is likely
influenced by the fact that studies that qualitatively reported
no relationship were included with an effect size of zero.
Identifying the magnitude of publication bias is difficult
with small sample sizes that complicate the interpretation of
normal quantile plots and result in low power for the rank
correlation test. The normal quantile plots identified
outliers to the effect size population distribution or
indicated that data may come from more than one
population. This is underscored by the significant hetero-
geneity present in many of the meta-analyses. This level of
heterogeneity is not unexpected in ecological research and
combining such diverse studies in a meta-analysis is
considered inappropriate by some critics (see Gurevitch
et al., 2001 and references therein). However, it is much
more useful to quantify the degree of heterogeneity and
explore its underlying causes in a transparent and defensible
manner than remain unaware of it, or worse ignore it,
which is often done in narrative reviews.
(3) Spatial congruence between people and
The results of the meta-analyses indicated strong population
correlations between HPD and combined species richness
Gary W. Luck620
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
or the richness of particular taxonomic groups (Table 2).
However, there was substantial heterogeneity in the data for
combined species richness suggesting certain factors medi-
ate this relationship. One of these factors is sampling grain
size. As grain size increased, so did the strength of the
positive correlation coefficients. Combine this with the fact
that many studies used mid to large grain sizes then it is
probably not surprising that positive correlations between
HPD and species richness were recorded. Heterogeneity
among the effect sizes was reduced when examining more
narrowly defined taxonomic subsets of the data. Therefore,
current evidence is supportive of a positive correlation
between HPD and the richness of particular taxonomic
groups (birds, mammals and plants) especially when
sampling units are >2500 km
Given these results, it is appropriate to explore possible
reasons for the spatial congruence between people and
species richness at the regional level. Gaston (2005)
proposed that HPD and species richness both respond
similarly to the underlying driver of energy availability.
Recent work supports this hypothesis (Gaston & Evans,
2004; Evans & Gaston, 2005; Evans et al., 2005; Luck,
2007), although a number of complex, interacting factors
are likely to drive human settlement patterns and species
distributions. Historical and contemporary case studies
demonstrate that human settlements may be influenced by
geological and environmental factors particularly as they
relate to suitability for agriculture (Kirch et al., 2004;
Vitousek et al., 2004), cultural and economic change (Liu
et al., 1999a; Homewood et al., 2001), land cover, elevation
and net primary productivity (Yue et al., 2003), disease
control (Lamprey & Reid, 2004), presence of and access
to natural amenities (Rasker & Hansen, 2000; Hansen
et al., 2002; Schnaiberg et al., 2002; Gustafson et al.,
2005; Radeloff, Hammer & Stewart, 2005) and established
patterns of settlement (e.g. expansion of existing urban
centres – Hammer et al., 2004; Radeloff et al., 2005) to name
a few.
Such an array of complex, interacting factors suggests
that any attempts to understand the relationships between
human population and species distribution patterns, and
how current and future human settlements may impact
biodiversity, face daunting challenges. Yet broader scale
studies offer some hope for identifying key drivers of human
distribution. For example, global studies show that more
people live at lower elevations and within 100 km of
a shoreline than expected by chance (Cohen & Small, 1998;
Small & Nicholls, 2003). I am unaware of any global studies
that examine the influence of energy on human population
distribution, but Luck (2007) found a strong positive cor-
relation between net primary productivity and HPD across
Australia, and Evans & Gaston (2005) reported the same for
temperature-HPD across Britain.
Morris & Kingston (2002) offer a thought-provoking
analysis that posits that the distribution of humans can be
explained by habitat selection theory and density-dependent
feedbacks on fitness, as is the case with other animals. The
distribution of human settlement size follows predictable and
consistent patterns such as the rank-size distribution
(whereby a straight line with a negative slope is generated
when plotting (log)settlement population size against (log)-
rank of that settlement – see Reed, 2002). Such broad and
predictable patterns offer hope for reconciling the spatial
congruence between people and biodiversity and forecasting
the consequences of demographic change for species
conservation. The distribution of species richness also follows
consistent patterns, such as latitudinal (see Rohde, 1999;
Gaston, 2000; Willig, Kaufman & Stevens, 2003; Hillebrand,
2004 for reviews) and elevational gradients (e.g. Lees,
Kremen & Andriamampianina, 1999; Brown, 2001; Sanders,
2002; McCain, 2004). Species richness generally increases
from the poles to the equator and decreases with altitude
(although peaks may occur in mid-range values). Numerous
hypotheses have been generated to explain these patterns,
although the most readily accepted are variation in energy
availability, area, spatial or habitat heterogeneity, evolution-
ary time and geometric constraints (Rahbek & Graves,
2001; Willig et al., 2003).
It is not surprising that ecologists invoke variables such as
energy availability when attempting to explain the spatial
congruence between people and biodiversity since its
influence is well established in the ecological literature, yet
these are tentative and initial explorations. Our under-
standing of the distributional overlap between people and
biodiversity would be greatly enhanced by examining the
role of other potential, mutual drivers of species and human
distribution patterns (e.g. spatial heterogeneity). The
importance of identifying such drivers cannot be overstated:
it is fundamental to understanding the consequences of
changes in HPD for biodiversity conservation.
An alternative explanation to spatial congruence is that
human land transformation near settlements increases
species richness. McKinney (2002a) established that in the
United States, HPD was the best predictor of net gain in
plant species richness whereby the number of new species
introduced to an area was greater than the number lost.
Fairbanks (2004) found that land transformation had
a positive effect on bird species richness in South Africa,
driven mostly by an increase in generalist species. Any
suggestions that humans create environments that are able
to support high species richness must differentiate between
the proportion of native and non-native species, and
compare their results against some benchmark that
represents the number of species that occurred in the area
prior to settlement [e.g. historical distribution records or
species richness from ecologically comparable areas (e.g. the
nearest conservation reserve)].
There are two broad implications of spatial congruence:
(i) the conservation of many species populations may be
threatened by their proximity to human settlements; and (ii)
with appropriate management there are substantial oppor-
tunities for large numbers of people to interact with a wide
array of species. To address the first implication, more work
needs to be done to determine the population status and
dynamics of individual species adjacent to and far from
human population centres. We should also be exploring
conservation planning scenarios that alleviate this spatial
overlap. However, if we accept the importance of the second
implication, planning must not result in complete isolation
of people from nature. The way forward is to develop
Human population density and biodiversity 621
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
strategies that allow for interaction while minimising threat.
I discuss this issue below. Key questions for future research
are presented in Table 3.
(4) HPD as a threat to biodiversity conservation
Positive population effect sizes between HPD and species
extinction and the richness of introduced species suggest
increasing density is a threat to the conservation of native
biodiversity (Table 2). This supports the implicit assump-
tion of human population growth as a threat to nature, yet
the evidence for this is tenuous owing primarily to a lack
of data and substantial variability in the approaches
used by researchers to address this issue. Moreover, the
meta-analyses on extinction and introduced species were
suggestive of publication bias. Much stronger evidence for
the adverse consequences of human population growth on
conservation come from the meta-analysis on protected
areas that showed clearly that as HPD increases the
amount of land designated for conservation decreases.
Combined with the above results this means that the most
species-rich regions are generally afforded the least
Although widely used, variables such as HPD are
imperfect indicators of humanity’s threat to biodiversity,
as they generally ignore patterns in population distribution
or the activities of the populace (Cincotta, Wisnewski &
Engelman, 2000). The distribution problem is especially
acute for studies using large grain sizes like countries where
average population density across the entire region is
matched with biodiversity status, yet the majority of people
may be confined to a few large cities. The impact of
humans in areas of high population density is generally
accepted and may extend many kilometres beyond
settlement boundaries (Myers, 1994; Repetto, 1994), but
impact can vary from minor to substantial (e.g. broad-scale
agriculture) in areas of low HPD depending on pre-
dominant land use. Moreover, certain indicators of
Table 3. Key questions to guide future research on the relationships between human populations and biodiversity
Topic Key questions for future research
Spatial congruence between people
and biodiversity
Are there key, mutual drivers of human and species distribution patterns?
How does variation in spatial extent or grain size alter the spatial congruence
between people and biodiversity? In particular, more work needs to be
conducted at local levels; even though this may cloud broader patterns it is
where many settlement planning and conservation decisions are made.
What are the impacts of adjacency to dense human populations for select species
(e.g. measures of reproductive success or survival)?
What opportunities exist for the conservation of high levels of species richness near
dense human populations, allowing for greater interaction between people and nature
(and the maintenance of local ecosystem services)?
What is the relationship between HPD and other measures of diversity (e.g. genetic or
functional diversity)?
HPD as a threat to conservation What is the form of the relationship between HPD and a given measure of
biodiversity status? Is it linear, exponential, or does it follow a decelerating curve
(e.g. quadratic), as seems to be the case in some situations (e.g. where the rate of
non-native species introductions decreases with increasing HPD: see Rapoport, 1993;
McKinney, 2001a). Do thresholds exist?
How does the threat to biodiversity change at varying levels of HPD? How do socio-economic
variables and conservation policy influence this relationship?
How does population dispersion alter the HPD – environment relationship?
How do ecological traits of species alter their susceptibility to the threats posed by HPD?
Are such traits consistent across taxa? Do particular functional groups appear to suffer
more with increasing HPD?
What are the relationships between HPD and other demographic measures
(e.g. household density) or socio-economic variables proposed to be key influences on
biodiversity status?
Planning for people and nature What are the possibilities for alleviating spatial conflict between people and nature given
greater constraints on site selection (i.e. incorporating more complex measures of threat
and improving the likelihood of species persistence)?
What conservation planning strategies might be needed in the future to address changes in
human demographics and environmental variation (e.g. climate change – see van Rensburg
et al., 2004)?
What are the socio-economic conditions that lead to favourable conservation policies?
What policies are required to ensure the conservation of species in areas of high HPD
and where are they most desperately needed?
What levels of HPD and types of settlement design are complementary to supporting
the greatest number of native species?
Gary W. Luck622
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
biodiversity status (e.g. species abundance) can vary owing
to factors completely unrelated to HPD (e.g. Hoare & Du
Toit, 1999). It seems obvious to look beyond simple
measures of human demographics when assessing the
consequences of the human enterprise for nature, and
many studies have incorporated demographic, social,
economic and ecological factors into more complex models
of the human-environment relationship (e.g. Allen &
Barnes, 1985; Hecht, 1993; Kummer & Sham, 1994; Skole
et al., 1994; Bawa & Dayanandan, 1997; Masera, Ordo
& Dirzo, 1997; Pys
ˇek, 1998; Liu et al., 1999b).
In addition to population dynamics, anthropogenic
factors that impact on biodiversity include level of
development and economic activity (James, 1994; Czech,
2000; Czech, Krausman & Devers, 2000) and policy
(Repetto, 1988, Linnell et al., 2001). These socio-economic
drivers of change need to be considered in light of the traits
and conditions that make species and ecological systems
susceptible to decline (Forester & Machlis, 1996; Cardillo
et al., 2004). Addressing this complexity requires an
interdisciplinary modelling approach (Forester & Machlis,
1996). Such approaches appear to be the way forward for
determining the threat the human enterprise poses to
biodiversity. Comparative modelling may also be useful in
teasing apart the influence of the various, interacting
drivers. For example, Taylor & Irwin (2004) compared the
capability of population-only (including the direct effects of
population and ecological factors) and population-economic
models (including the direct effects of economic and
ecological factors and the indirect effects of population) in
explaining the distribution of exotic plants in the United
States. They found that the population-economic model
explained more variance in the distribution of exotics than
the population-only model.
Developing complex models that incorporate all the
social, economic and demographic drivers of change and
the interactions between them is challenging. It is also
becoming increasingly complicated owing to the growth in
global trade and movement of natural and human-made
resources. For example, deforestation in a given country
may be influenced by numerous factors including: (i) local
human population dynamics, cultural traditions, and social
and financial circumstance; (ii) national economic and
conservation policy; (iii) international commodity markets;
and (iv) local, national and international consumer demand
(which in turn is a factor of population size and affluence;
Geist & Lambin, 2002). Nevertheless, population demog-
raphy is recognised as a key factor in these interactive
relationships (Carr, 2004). HPD is a better demographic
measure to use than total population size or population
growth rates (see Section I). It is easy to interpret, widely
applicable and relatively accurate given the advent and
growth of comprehensive census data. For example,
Cardillo et al. (2004) argued that HPD is a useful indicator
of threat because it is more readily quantified than direct
threats such as habitat loss or hunting, particularly over
broad extents. At local extents, HPD may be a very good
measure of impact if it is related to factors such as hunting
pressure (e.g. Brashares et al., 2001; Walsh et al., 2003;
Altrichter & Boaglio, 2004).
It is reasonable to argue that including too many
variables and their interactions in explanatory or predictive
models is likely to yield a complexity that is virtually
impossible to interpret. The alternative is to continue to
focus on probable key drivers like population (Meyer &
Turner, 1992). If the latter approach is followed, it would be
wise to address the links between population and other key
drivers of change such as consumer demand. For example,
recent studies have invoked density of households (rather
than people) as a key measure of threat to biodiversity (e.g.
Liu et al., 2003; Linderman et al., 2005), yet these studies
generally don’t utilise a comparative modelling approach to
determine if household density is a better measure of threat
than HPD, or address the interactions that occur between
these variables (although see Liu et al., 1999a; Brown &
Laband, 2006).
(5) Planning for people and nature
Although correlations between HPD and species richness
indicate that species-rich areas occur near human settle-
ments, they don’t imply that there is complete distributional
overlap between all of these species and people. That is, it
may be possible to locate sites in areas of low HPD where
collectively almost all species occur. This possibility has
been explored by a number of researchers using the
principles of complementarity and irreplaceability via
conservation planning scenarios (Fjeldsa
˚& Rahbek, 1998;
Balmford et al., 2001; Arau
´jo, Williams & Turner, 2002;
Chown et al., 2003; Luck et al., 2004; Van Rensburg et al.,
2004; Diniz-Filho et al., 2006; O’Dea, Arau
´jo & Whittaker,
2006; Va
´zquez & Gaston, 2006). The results of these studies
indicate that opportunities to reduce the degree of overlap
between people and regional biodiversity vary from
substantial to limited, although some of these analyses are
only initial explorations of the problem. For example, Luck
et al. (2004) only required each species to be represented
once in a minimum set of sites, which in most cases is hardly
sufficient for the conservation of viable populations. More
complex conservation planning scenarios need to be
explored examining a wider range of limitations on site
selection and including multiple sites for species conserva-
tion. It is likely that such studies will show a greater degree
of spatial congruence between HPD and site selection than
these initial explorations.
There is great value in incorporating regional threats in
planning strategies aimed at locating or managing sites for
conservation. Adding sites to the conservation reserve
system not only needs to consider the complementarity of
the species or ecosystem assemblages, but the vulnerability
of the site to factors such as surrounding land use or HPD
(Wilson et al., 2005). Management of established sites needs
to focus on internal dynamics, external threats and the
interactions between the two. It is crucial to determine how
the landscape context of a reserve influences its conserva-
tion value and vulnerability, and also how landscape change
might alter these relationships. From the handful of studies
conducted, it is clear that adjacent land use or HPD may
have a substantial impact on the ecological integrity of
Human population density and biodiversity 623
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
conservation reserves (e.g. Rivard et al., 2000; Brashares
et al., 2001; Harcourt et al., 2001; Hansen & Rotella,
2002; Parks & Harcourt, 2002; Wiersma, Nudds & Rivard,
Although conservation planning and policy change can
help alleviate the threats posed by human populations to
biodiversity, this should be reconciled with the need to
facilitate interactions between people and nature. For the
most part, highly urbanised areas have diminished indig-
enous biodiversity (e.g. McIntyre, 2000; Marzluff, 2001;
McKinney, 2002c; Hope et al., 2003; Eppink, van der Bergh
& Rietveld, 2004; Turner et al., 2004; Williams, McDonnell
& Seager, 2005; Chace & Walsh, 2006). Yet there is some
evidence to suggest substantial opportunities for conserva-
tion in urban areas. For example, Ku
¨hn, Brandl & Klotz
(2004) reported that German city grid cells had significantly
higher native plant species richness than non-city grid cells.
Chace & Walsh (2006) note that although urbanisation
leads to lower bird species richness, urban areas often
support greater bird biomass. This indicates an abundance
of resources for certain species, although these are mostly
generalists or introduced species. Interestingly, some studies
that have examined urban-rural/forest gradients show that
species diversity peaks at mid-levels of development (e.g.
Nuhn & Wright, 1979; Racey & Euler, 1982; Jokima
¨ki &
Suhonen, 1993; Blair, 1999; Germaine & Wakeling, 2001).
Although this is not consistent across all studies (Marzluff,
2001), it demonstrates the possibility of conserving high
species richness around human settlements with sympa-
thetic management practices.
The Millennium Ecosystem Assessment (2005) demon-
strated unequivocally the dire state of many of our
ecological systems and clearly implicated humans as the
main drivers of change. Future growth in the human
population will undoubtedly exacerbate further global
decline in ecosystem health. The stabilisation and gradual
reduction of human numbers is key to avoiding ecosystem
collapse. Yet simply reducing population size and density
in certain areas without an understanding of people-
environment interactions may lead to unexpected results.
For example, Fisher et al. (2003) demonstrated that
a reduction in HPD in the Lake Patzcuaro Basin in Mexico
led to severe land degradation owing to a disruption in the
management of a human-modified environment dependent
on human labour for maintenance. In Puerto Rico, forest
cover has increased despite rises in HPD, as agricultural land
has been replaced by forest owing to changes in the economy
and land management policy (Lugo, 2002). Guyette &
Spetich (2003) found that in Arkansas, fire frequency initially
increased with HPD during the early stages of European
settlement, but as the human population continued to grow
fires became less frequent owing to cultural and land
management changes leading to a policy of fire suppression.
Most landscapes on Earth are now severely modified by
human activity. Simply removing humans from these
landscapes may lead to unwanted and unexpected results.
To avoid this, we must have a comprehensive understand-
ing of human-environment relationships and implement
management strategies that aim to address mitigating
(6) Implications of future human demographics
In the coming decades there are likely to be three major
changes to the human population: (i) global population size
will increase to approximately 9 billion by 2050, with
95% of growth occurring in developing nations; (ii) the
percentage of the population living in urban areas will
increase from 47% in 2000 to 60% in 2030; and (iii) the
population will age substantially, especially in developed
countries with low growth rates (United Nations, 2002,
2005). The other major factor likely to affect the distri-
bution of human populations is climate change, which
could see some low-lying coastal or island communities
displaced and substantial movements of people as they seek
out more temperate climates.
Future population growth and movement to urban
centres means even greater pressure on species-rich and
highly productive areas, and further displacement of people
from nature. The appropriate management of urban areas
to ensure species preservation will become even more
crucial over time. It is difficult to estimate the effects of
changing age structure, but it highlights the need to
understand how socio-economic factors influence biodiver-
sity status.
Only a few studies predict the likely outcomes of people-
biodiversity relationships into the future. For example,
McKee et al. (2004) predicted that future population growth
is likely to result in an average increase of 14% in the
number of threatened species across growing nations. Van
Rensburg et al. (2004) suggested that the likely changes
wrought by population growth and climate change in South
Africa would increase pressure on conservation reserves.
It is crucial to identify regions of rapid human growth and
implement forward-thinking conservation policies that
address predicted changes. Growing urban centres should
incorporate plans to ensure species protection within
settlement boundaries.
Future global population growth, regardless of changes in
consumption, will almost certainly increase the threat posed
by humans to nature conservation. Realistic, humane and
socially acceptable policies aimed at stabilising and
eventually reducing human population size must be
implemented now. The extent of the policies must take
both a global and local perspective to reduce our impact on
the immediate environment and overall global environ-
ment. Human population size will peak eventually, probably
within this century (Lutz, Sanderson & Scherbov, 2001).
The key questions are how much biodiversity can we save as
the population continues to grow, what are the likely
population dynamics post-peak (e.g. relative stability, minor
fluctuations or precipitous decline) and what implications
will these dynamics have for conservation?
(1) The literature on the relationships between HPD and
biodiversity is extremely heterogeneous, although there are
some consistent trends (e.g. HPD and species richness).
Gary W. Luck624
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
Significant positive population correlations were found
between HPD and species richness, the richness of
threatened, geographically restricted and introduced spe-
cies, and extinctions. A significant negative correlation
occurred between HPD and protected area coverage.
Heterogeneity among analyses was reduced when more
discrete subsets of the data were examined.
(2) Despite substantial heterogeneity, some important
conclusions are possible (Fig. 5). At broad scales, there is
a strong positive relationship between HPD and the
richness of many taxonomic groups leading to spatial
congruence between human development and diverse
ecosystems. Such congruence is both a threat to species
conservation and an opportunity to facilitate interaction
between people and nature given appropriate management.
Identifying key, mutual drivers of human and species
distribution patterns is crucial to our understanding of this
phenomenon. One likely candidate is energy availability,
but other factors require investigation.
(3) Human population growth results in increasing land
transformation and the introduction of exotic species. Land
designated for conservation is substantially reduced near
human settlements. Through these processes, and directly,
HPD influences biodiversity status. Species sensitive to
anthropogenic change are lost or decline in abundance,
while disturbance-tolerant and generalist species may
prosper. The impact of human populations is magnified
through socio-economic factors such as per capita consump-
tion, although technological development, behavioural
change and policy implementation may help reduce overall
(4) Ultimately, we must ensure management for the
environment given the ambitions of the human enterprise.
This requires greater emphasis on the key anthropogenic
drivers of environmental change. In the end, it will be the
appropriate management of people, not plants and animals,
which determines the future state of our planet.
Thanks to Karl Evans, Jessica Gurevitch, Nick Klomp,
Mike McKinney, Gayle Smythe and two anonymous
referees for providing valuable and constructive comments
on drafts of the manuscript.
Human population density
Energy availability
Biodiversity status
(e.g. species richness or
extinction probability)
Socio-economic factors
(e.g. gross national
Exotic species
Land transformation
(e.g. reduction in size of
protected areas)
May influence human
settlement patterns
Influences species
HPD correlated with
exotic species
Exotic species influence
biodiversity status. Ecosystem
characteristics (e.g. resilience)
influence exotic establishment
HPD has direct impacts on biodiversity (e.g.
hunting).Biodiversity characteristics (e.g.
natural amenities) may influence human
settlement patterns
Human populations transform land. Certain
land transformations may encourage further
population growth
Land transformation
influences biodiversity
Interactions occur between
human population dynamics
and social and economic
factors Social and economic factors
(e.g. level of development)
alter biodiversity status.
Natural resources often
underpin development
Fig. 5. A schematic representation of the proposed relationships between human population density and biodiversity focussing
particularly on the negative impacts of population growth. The evidence for each of these relationships varies in the literature. The
diagram includes biodiversity feedback loops, but not interconnections between energy availability, exotic species establishment,
land transformation and socio-economic factors.
Human population density and biodiversity 625
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
ALLEN,J.C.&BARNES, D. F. (1985). The causes of deforestation in
developing countries. Annals of the Association of American
Geographers 75, 163–184.
ALTRICHTER,M.&BOAGLIO, G. I. (2004). Distribution and relative
abundance of peccaries in the Argentine Chaco: associations
with human factors. Biological Conservation 116, 217–225.
´JO, M. B. (2003). The coincidence of people and biodiversity
in Europe. Global Ecology and Biogeography 12, 5–12.
´JO, M. B., WILLIAMS,P.H.&TURNER, A. (2002). A sequential
approach to minimise threats within selected conservation areas.
Biodiversity and Conservation 11, 1011–1024.
A., WILLIAMS,P.&RAHBEK, C. (2001). Conservation conflicts
across Africa. Science 291, 2616–2619.
BARNES, R. F. W. (1990). Deforestation trends in tropical Africa.
African Journal of Ecology 28, 161–173.
BAWA,K.S.&DAYANANDAN, S. (1997). Socioeconomic factors and
tropical deforestation. Nature 386, 562–563.
BEGG, C. B. (1994). Publication bias. In The Handbook of Research
Synthesis (eds. H. Cooper and L. V. Hedges), pp. 399–409.
Russell Sage Foundation, New York
BEGG,C.B.&BERLIN, J. A. (1988). Publication bias: a problem in
interpreting medical data. Journal of the Royal Statistical Society,
Series A 151, 419–463.
BEGG,C.B.&MAZUMDAR, M. (1994). Operating characteristics of
a rank correlation test for publication bias. Biometrics 50, 1088–
´LANGER,L.&GRENIER, M. (2002). Agriculture intensification
and forest fragmentation in the St. Lawrence valley, Que
Canada. Landscape Ecology 17, 495–507.
BENDER,D.J.,CONTRERAS,T.A.&FAHRIG, L. (1998). Habitat loss
and population decline: a meta-analysis of the patch size effect.
Ecology 79, 517–533.
BILSBORROW,R.&GOERES, M. (1994). Population, land–use and
the environment in developing countries: what can we learn
from cross–national data? In The Causes of Tropical Deforestation
(eds. K. Brown and D. W. Pearce), pp. 106–133. UCL Press
Limited, London
BLAIR, R. B. (1999). Birds and butterflies along an urban gradient:
surrogate taxa for assessing biodiversity? Ecological Applications 9,
distribution of the Egyptian mongoose Herpestes ichneumon (L.) in
Portugal. Mammal Review 26, 1–8.
BRASHARES, J. S., ARCESE,P.&SAM, M. K. (2001). Human
demography and reserve size predict wildlife extinction in West
Africa. Proceedings of the Royal Society of London Series B 268, 2473–
BROOK,B.W.&BOWMAN, D. M. (2002). Explaining the Pleistocene
megafaunal extinctions: models, chronologies, and assumptions.
Proceedings of the National Academy of Sciences U.S.A.99, 14624–
BROW N, J. H. (2001). Mammals on mountainsides: elevational
patterns of diversity. Global Ecology and Biogeography 10, 101–109.
BROW N,R.M.&LABAND, D. N. (2006). Species imperilment and
spatial patterns of development in the United States. Conservation
Biology 20, 239–244.
BUIJS,J.A.&VA N WIJNEN, J. H. (2001). Survey of feral rock doves
(Columba livia) in Amsterdam, a bird–human association. Urban
Ecosystems 5, 235–241.
BURNEY,D.A.&FLANNERY, T. F. (2005). Fifty millennia of
catastrophic extinctions after human contact. Trends in Ecology
and Evolution 20, 395–401.
BUTLER, B. J., SWENSON,J.J.&ALIG, R. J. (2004). Forest
fragmentation in the Pacific Northwest: quantification and
correlations. Forest Ecology and Management 189, 363–373.
(2005). Multiple causes of high extinction risk in large mammal
species. Science 309, 1239–1241.
J. & MAC E, G. M. (2004). Human population density and
extinction risk in the world’s carnivores. PLoS Biology 2, 909–914.
CARR, D. L. (2004). Proximate population factors and deforesta-
tion in tropical agricultural frontiers. Population and Environment
25, 585–612.
CEBALLOS,G.&EHRLICH, P. R. (2002). Mammal population losses
and the extinction crisis. Science 296, 904–907.
(2005). Global mammal conservation: what must we manage?
Science 309, 603–607.
CHACE,J.F.&WALSH, J. J. (2006). Urban effects on native
avifauna: a review. Landscape and Urban Planning 74, 46–69.
CHOWN, S. L., GREMMEN,N.J.M.&GASTON, K. J. (1998).
Ecological biogeography of southern ocean islands: species–area
relationships, human impacts, and conservation. The American
Naturalist 152, 562–575.
S. L. & VAN JAARSVELD, A. S. (2003). Energy, species richness,
and human population size: conservation implications at
a national scale. Ecological Applications 13, 1233–1241.
population in the biodiversity hotspots. Nature 404, 990–992.
CLEMANTS,S.&MOORE, G. (2003). Patterns of species richness in
eight northeastern United States cities. Urban Habitats 1, 4–16.
COHEN, J. E. (1996). How Many People Can the Earth Support? W. W.
Norton and Company Incorporated, New York.
COHEN,J.E.&SMALL, C. (1998). Hypsographic demography: the
distribution of human population by altitude. Proceedings of the
National Academy of Sciences U.S.A.95, 14009–14014.
COWLISHAW,G.&DUNBAR, R. (2000). Primate Conservation Biology.
University of Chicago Press, Chicago
CROPPER,M.&GRIFFITHS, C. (1994). The interaction of
population growth and environmental quality. The American
Economic Review 84, 250–254.
CZECH, B. (2000). Economic growth as the limiting factor for
wildlife conservation. Wildlife Society Bulletin 28, 4–14.
CZECH, B., KRAUSMAN,P.R.&DEVERS, P. K. (2000). Economic
associations among causes of species endangerment in the
United States. BioScience 50, 593–601.
B., CARVALHO,P.&BASTOS, R. P. (2006). Anuran species
richness, complementarity and conservation conflicts in Brazil-
ian cerrado. Acta Oecologica 29, 9–15.
(1997). Geographic distribution of endangered species in the
United States. Science 275, 550–553.
´GUEZ,J.P.&ROBERTS, W. M. (2001). Synoptic
tinkering: integrating strategies for large–scale conservation.
Ecological Applications 11, 1019–1026.
´ZAR, J. A., GANGOSO, L., FORERO,M.G.&JUSTE, J. (2005).
Presence, richness and extinction of birds of prey in the
Gary W. Luck626
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
Mediterranean and Macaronesian islands. Journal of Biogeography
32, 1701–1713.
DRAKE,M.T.&PEREIRA, D. L. (2002). Development of a fish–
based index of biotic integrity for small inland lakes in central
Minnesota. North American Journal of Fisheries Management 22,
EHRLICH, P. R. (1971). The Population Bomb. Ballantine Books,
New York.
EHRLICH,P.R.&HOLDREN, J. P. (1971). Impact of population
growth. Science 171, 1212–1217.
importance of data-selection criteria: meta-analyses of stream
predation experiments. Ecology 80, 1132–1141.
Modelling biodiversity and land use: urban growth, agriculture
and nature in a wetland area. Ecological Economics 51, 201–216.
˜OR,J.L.&VIBRANS, H. (2004).
The rich generally get richer, but there are exceptions:
correlations between species richness of native plant species
and alien weeds in Mexico. Diversity and Distributions 10,
EVANS,K.L.&GASTON, K. J. (2005). People, energy and avian
species richness. Global Ecology and Biogeography 14, 187–196.
EVANS, K. L., WARREN,P.H.&GASTON, K. J. (2005). Does energy
availability influence classical patterns of spatial variation in
exotic species richness? Global Ecology and Biogeography 14, 57–65.
FAIRBANKS, D. H. K. (2004). Regional land–use impacts affecting
avian richness patterns in southern Africa – insights from
historical avian atlas data. Agriculture, Ecosystems and Environment
101, 269–288.
FERNANDEZ-DUQUE,E.&VALEGGIA, C. (1994). Meta-analysis –
a valuable tool in conservation research. Conservation Biolog y 8,
MONROY,V.H.&BANERJEE, S. K. (2003). A re-examination of
human–induced environmental change within the Lake Pa
cuaro Basin, Michoaca
´n, Mexico. Proceedings of the National
Academy of Sciences U.S.A.100, 4957–4962.
˚,J.&RAHBEK, C. (1998). Continent–wide conservation
priorities and diversification processes. In Conservation in
a Changing World (eds. G. M. Mace, A. Balmford and J. R.
Ginsberg), pp. 139–160. Cambridge University Press, Cam-
RAMANKUTTY,N.&SNYDER, P. K. (2005). Global consequences
of land use. Science 309, 570–574.
FORESTER,D.J.&MACH LIS,G.E.(1996).Modelinghumanfactors
that affect the loss of biodiversity. Conservation Biology 10, 1253–1263.
GASTON, K. J. (2000). Global patterns in biodiversity. Nature 405,
GASTON, K. J. (2005). Biodiversity and extinction: species and
people. Progress in Physical Geography 29, 239–247.
GASTON,K.J.&EVANS, K. L. (2004). Birds and people in Europe.
Proceedings of the Royal Society of London Series B 271, 1649–1655.
GATES, S. (2002). Review of methodology of quantitative reviews
using meta-analysis in ecology. Journal of Animal Ecology 71,
GAVASHELISHVILI, A. (2004). Habitat selection by East Caucasian
tur (Capra cylindricornis). Biological Conservation 120, 391–398.
GEIST,H.J.&LAMBIN, E. F. (2002). Proximate causes and
underlying driving forces of tropical deforestation. BioScience 52,
GERMAINE,S.S.&WAKELING, B. F. (2001). Lizard species
distributors and habitat occupation along an urban gradient
in Tuscon, Arizona, USA. Biological Conservation 97, 229–237.
GIBBS, J. P. (2000). Wetland loss and biodiversity conservation.
Conservation Biology 14, 314–317.
Long–distance dispersal and human population density allow the
prediction of invasive patterns in the horse chestnut leafminer
Cameraria ohridella.Journal of Animal Ecology 73, 459–468.
´GOIRE, J–C., Heitland,
W. , S TRAW, N., TILBURY,C.&AUGUSTIN, S. (2005). Forecasting
Cameraria ohridella invasion dynamics in recently invaded
countries: from validation to prediction. Journal of Applied Ecolog y
42, 805–813.
predator–prey conflicts: ecological correlates, prey losses and
patterns of management. Biological Conservation 122, 159–171.
GREEN,G.M.&SUSSMAN, R. W. (1990). Deforestation history of
the eastern rain forests of Madagascar from satellite images.
Science 248, 212–215.
GUREVITCH,J.&HEDGES, L. V. (1999). Statistical issues in
ecological meta-analyses. Ecology 80, 1142–1149.
GUREVITCH, J., CURTIS,P.S.&JONES, M. H. (2001). Meta-analysis
in ecology. Advances in Ecological Research 32, 199–247.
(2005). The relationship between environmental amenities and
changing human settlement patterns between 1980 and 2000 in
the Midwestern USA. Landscape Ecology 20, 773–789.
GUYETTE,R.P.&SPETICH, M. A. (2003). Fire history of oak–pine
forests in the Lower Boston Mountains, Arkansas, USA. Forest
Ecology and Management 180, 463–474.
VOSS, P. R. (2004). Characterizing dynamic spatial and temporal
residential density patterns from 1940–1990 across the north
central United States. Landscape and Urban Planning 69, 183–199.
HANSEN,A.J.&ROTELLA, J. J. (2002). Biophysical factors, land use,
and species viability in and around nature reserves. Conservation
Biology 16, 1112–1122.
R. L. & KRASKA, M. P. V. (2002). Ecological causes and
consequences of demographic change in the new west. BioScience
52, 151–162.
HARCOURT, A. H. (1996). Is the gorilla a threatened species? How
should we judge? Biological Conservation 75, 165–176.
HARCOURT,A.H.&PARKS, S. A. (2003). Threatened primates
experience high human densities: adding an index of threat to
the IUCN Red List criteria. Biological Conservation 109, 137–149.
density as an influence on species/area relationships: double
jeopardy for small African reserves? Biodiversity and Conservation
10, 1011–1026.
HECHT, S. B. (1993). The logic of livestock and deforestation in
Amazonia. BioScience 43, 687–695.
HILLEBRAND, H. (2004). On the generality of the latitudinal
diversity gradient. The American Naturalist 163, 192–211.
HOARE,R.E.&DUTOIT, J. T. (1999). Coexistence between people
and elephants in African savannas. Conservation Biolog y 13,
Human population density and biodiversity 627
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
SCOTT, G. I. (2004). Linkages between tidal creek ecosystems
and the landscape and demographic attributes of their water-
sheds. Journal of Experimental Marine Biology and Ecology 298,
Long–term changes in Serengeti–Mara wildebeest and land
cover: pastoralism, population, or policies? Proceedings of the
National Academy of Sciences U.S.A 98, 12544–12549.
Socioeconomics drive urban plant diversity. Proceedings of the
National Academy of Sciences U.S.A 100, 8788–8792.
HUNTER,J.E.&SCHMIDT, F. L. (2004). Methods of Meta-analysis:
Correcting Error and Bias in Research Findings: Second Edition. Sage
Publications, California.
HUNTER,M.L.&YONZON, P. (1993). Altitudinal distribution of
birds, mammals, people, forest, and parks in Nepal. Conservation
Biology 2, 420–423.
HUSTON, M. A. (2001). People and biodiversity in Africa. Science
293, 1591.
JAMES, P. C. (1994). On economic growth and ecological decay.
Conservation Biology 8, 1161–1162.
¨KI,J.&SUHONEN, J. (1993). Effects of urbanization on the
breeding bird species richness in Finland: a biogeographical
comparison. Ornis Fennica 70, 71–77.
¨KI,J.&SUHONEN, J. (1998). Distribution and habitat
selection of wintering birds in urban environments. Landscape and
Urban Planning 39, 253–263.
¨KI, J., SUHONEN, J., INKI,K.&JOKINEN, S. (1996).
Biogeographical comparison of winter bird assemblages
in urban environments in Finland. Journal of Biogeography 23,
KEANE, A., BROOKE,, P. J. K. (2005).
Correlates of extinction risk and hunting pressure in gamebirds
(Galliformes). Biological Conservation 126, 216–233.
KERR,J.T.&CURRIE, D. J. (1995). Effects of human activity and
global extinction risk. Conservation Biology 9, 1528–1538.
M., SHERROD, D. R., COIL, J., HOLM,L.&SHARP, W. D. (2004).
Environment, agriculture, and settlement patterns in a marginal
Polynesian landscape. Proceedings of the National Academy of Sciences
U.S.A 101, 9936–9941.
KIRKLAND,G.L.&OSTFELD, R. S. (1999). Factors influencing
variation among states in the number of federally listed mammals
in the United States. Jour nal of Mammalogy 80, 711–719.
KOK, K. (2004). The role of population in understanding
Honduran land use patterns. Journal of Environmental Management
72, 73–89.
(2005). Percentage canopy cover –using Landsat imagery to
delineate habitat for Myanmar’s endangered Eld’s deer (Cervus
eldi). Animal Conservation 8, 289–296.
KUMMER,D.&SHAM, C. H. (1994). The causes of tropical
deforestation: a quantitative analysis and case study from the
Philippines. In The Causes of Tropical Deforestation (eds. K. Brown
and D. W. Pearce), pp. 146–158. UCL Press Limited, London.
¨HN, I., BRANDL,R.&KLOTZ, S. (2004). The flora of German
cities is naturally species rich. Evolutionary Ecology Research 6,
¨NEN,M.&DESROCHERS, A. (2005).
Demographic responses by birds to forest fragmentation.
Conservation Biology 19, 1537–1546.
LAMPREY,R.H.&REID, R. S. (2004). Expansion of human
settlement in Kenya’s Maasai Mara: what future for pastoralism
and wildlife? Journal of Biogeography 31, 997–1032.
LAN,D.&DUNBAR, R. (2000). Bird and mammal conservation
in Gaoligongshan Region and Jingdong County, Yunnan, China:
patterns of species richness and nature reserves. Oryx 34, 275–286.
Predictors of deforestation in the Brazilian Amazon. Journal of
Biogeography 29, 737–748.
Environmental and human factors influencing rare plant local
occurrence, extinction and persistence: a 115-year study in the
Mediterranean region. Journal of Biogeography 32, 799–811.
model for species richness gradients: bounded range overlap of
butterflies and other rainforest endemics in Madagascar.
Biological Journal of the Linnean Society 67, 529–584.
LEIMU,R.&KORICHEVA, J. (2006). A meta-analysis of tradeoffs
between plant tolerance and resistance to herbivores: combining
the evidence from ecological and agricultural studies. Oikos 112,
LIGHT,R.J.&PILLEMER, D. B. (1984). Summing up: The Science of
Reviewing Research. Harvard University Press, Cambridge MA.
J. (2005). Modelling the spatio–temporal dynamics and
interactions of households, landscapes, and giant panda habitat.
Ecological Modelling 183, 47–65.
LINNELL,J.D.C.,SWENSON,J.E.&ANDERSEN, R. (2001). Predators
and people: conservation of large carnivores is possible at high
human densities if management policy is favourable. Animal
Conservation 4, 345–349.
LIU, J., OUYANG, Z., TAN, Y., YANG,J.&ZHANG, H. (1999a).
Changes in human population structure: implications for
biodiversity conservation. Population and Environment 21, 45–58.
H. (1999b). A framework for evaluating the effects of human
factors on wildlife habitat: the case of giant pandas. Conservation
Biology 13, 1360–1370.
LIU, J., DAILY, G. C., EHRLICH,P.R.&LUCK, G. W. (2003). Effects
of household dynamics on resource consumption and bio-
diversity. Nature 421, 530–533.
LUCK, G. W. (2007). The relationships between net primary
productivity, human population density and species conserva-
tion. Journal of Biogeography 34, 201–212.
Alleviating spatial conflict between people and biodiversity.
Proceedings of the National Academy of Sciences U.S.A.101, 182–186.
LUGO, A. E. (2002). Can we manage tropical landscapes? – An
answer from the Caribbean perspective. Landscape Ecology 17,
LUTZ, W., SANDERSON,W.&SCHERBOV, S. (2001). The end of world
population growth. Nature 412, 543–545.
LYONS, S. K., SMITH F. A . & B ROWN, J. H. (2004). Of mice,
mastodons and men: human–mediated extinctions on four
continents. Evolutionary Ecology Research 6, 339–358.
MAHONEY, M. J. (1977). Publication prejudices: an experimental
study of confirmatory bias in the peer review system. Cognitive
Therapy and Research 1, 161–175.
Gary W. Luck628
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
MALTHUS, T. R. (1798). An Essay on the Principle of Population, as it
Affects the Future Improvement of Society with Remarks on the Speculations
of Mr. Godwin, M. Condorcet, and Other Writers. J. Johnson, London.
MARTIN,P.S.&STEADMAN, D. W. (1999). Prehistoric extinctions on
islands and continents. In Extinctions in Near Time: Causes, Contexts,
and Consequences (ed. R. D. E. MacPhee), pp. 17–55. Kluwer
Academic/Plenum, New York.
MARZLUFF, J. M. (2001). Worldwide urbanization and its effects on
birds. In Avian Ecology and Conservation in an Urbanizing World (eds.
J. M. Marzluff, R. Bowman and R. Donnelly), pp. 19–47.
Kluwer, Boston.
˜EZ,M.J.&DIRZO, R. (1997). Carbon
emissions from Mexican forests: current situation and long–
term scenarios. Climatic Change 35, 265–295.
MCCAIN, C. M. (2004). The mid–domain effect applied to
elevational gradients: species richness of small mammals in
Costa Rica. Journal of Biogeography 31, 19–31.
MCINTYRE, N. E. (2000). Ecology of urban arthropods: a review
and call to action. Annals of the Entomological Society of America 93,
MCKEE, J. K. (2003). Sparing Nature: The Conflict Between Human
Population Growth and Earth’s Biodiversity. Rutgers University Press,
New Jersey
Forecasting global biodiversity threats associated with human
population growth. Biological Conservation 115, 161–64.
MCKINNEY, M. L. (2001a). Effects of human population, area, and
time on non–native plant and fish diversity in the United States.
Biological Conservation 100, 243–252.
MCKINNEY, M. L. (2001b). Role of human population size in
raising bird and mammal threat among nations. Animal
Conservation 4, 45–57.
MCKINNEY, M. L. (2002a). Do human activities raise species
richness? Contrasting patterns in United States plants and
fishes. Global Ecology and Biogeography 11, 343–348.
MCKINNEY, M. L. (2002b). Effects of national conservation
spending and amount of protected area on species threat rates.
Conservation Biology 16, 539–543.
MCKINNEY, M. L. (2002c). Urbanization, biodiversity, and
conservation. BioScience 52, 883–890.
MCKINNEY, M. L. (2002d). Why larger nations have dispropor-
tionate threat rates: area increases endemism and human
population size. Biodiversity and Conservation 11, 1317–1325.
MCMASTER, R. T. (2005). Factors influencing vascular plant
diversity on 22 islands off the coast of eastern North America.
Journal of Biogeography 32, 475–492.
W. (1972). T he Limits to Growth. University Books, New York.
MENDOZA,E.&DIRZO, R. (1999). Deforestation in Lacandonia
(southeast Mexico): evidence for the declaration of the
northernmost topical hot–spot. Biodiversity and Conservation 8,
MERCER,E.&ARU NA , P. B. (2000). Assessing the impacts of forests
on human welfare: preliminary results from the mid–Atlantic
integrated assessment. Environmental Monitoring and Assessment 63,
MEYER,W.B.&TURNER II, B. L. (1992). Human population
growth and global land–use/cover change. Annual Review of
Ecology and Systematics 23, 39–61.
Millennium Ecosystem Assessment (2005). Living beyond our
means: Natural Assets and Human Wellbeing.
CLARKE,S.J.&JOHNSON, B. J. (2005). Ecosystem collapse in
Pleistocene Australia and a human role in megafaunal
extinction. Science 309, 287–290.
KORMOS, C. (2003). Wilderness and biodiversity conservation.
Proceedings of the National Academy of Sciences U.S.A.100, 10309–
MØLLER,A.P.&JENNIONS, M. D. (2002). How much variance can
be explained by ecologists and evolutionary biologists? Oecologia
132, 492–500.
SCHEI,P.J.&WAAGE, J. K. (eds.) (2005). Invasive Alien Species: A
New Synthesis. Island Press, Washington D.C.
MORRIS,D.W.&KINGSTON, S. R. (2002). Predicting future threats
to biodiversity from habitat selection by humans. Evolutionary
Ecology Research 4, 787–810.
MYERS, N. (1994). Population and biodiversity. In Population – The
Complex Reality (ed. F. Graham–Smith), pp. 117–136. Royal
Society, London.
H. I. (1994). The conflict between wildlife and local people
living adjacent to protected areas in Tanzania: human density as
a predictor. Conservation Biology 8, 249–255.
NUHN,T.P.&WRIGHT, C. G. (1979). An ecological survey of ants
in a landscape suburban habitat. The American Midland Naturalist
102, 353–362.
Crouching tigers, hidden prey: Sumatran tiger and prey
populations in a tropical forest landscape. Animal Conservation
6, 131–139.
´JO,M.B.&WHITTAKER, R. J. (2006). How well do
important bird areas represent species and minimize conserva-
tion conflict in the tropical Andes? Diversity and Distributions 12,
PALO, M., LEHTO,E.&UUSIVUORI, J. (2000). Modeling causes of
deforestation with 477 subnational units. In World Forests from
Deforestation to Transition? Vol. II (eds. M. Palo and H. Vanhanen),
pp. 101–124. Kluwer Academic Publishers, Boston.
PARKER,I.S.C.&GRAHAM, A. D. (1989). Men, elephants and
competition. Symposium of the Zoological Society of London 61, 241–252.
PARKS,S.A.&HARCOURT, A. H. (2002). Reserve size, local human
density, and mammalian extinctions in U.S. protected areas.
Conservation Biology 16, 800–808.
Pimentel, D. (ed.) (2002). Biological Invasions: Economic and
Environmental Costs of Alien Plant, Animal, and Microbe Species.
CRC Press, Florida.
PRESSEY, R. L. (1994). Ad hoc reservations: forward or backward
steps in developing representative reserve systems? Conservation
Biology 8, 662–668.
PRESSEY, R. L. (1995). Conservation reserves in NSW: crown jewels
or leftovers? Search 26, 47–51.
ˇEK, P. (1998). Alien and native species in central European
urban floras: a quantitative comparison. Journal of Biogeography
25, 155–163.
ˇERA, T. (2002). Patterns of invasion in
temperate nature reserves. Biological Conservation 104, 13–24.
´, M., KROPA
ˇ, Z., TICHY
J. (2005). Alien plants in temperate weed communities:
prehistoric and recent invaders occupy different habitats. Ecology
86, 772–785.
Human population density and biodiversity 629
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
RACEY,G.D.&EULER, D. L. (1982). Small mammal and habitat
response to shoreline cottage development in central Ontario,
Canada. Canadian Journal of Zoology 60, 865–880.
RADELOFF, V. C., HAMMER,R.B.&STEWART, S. I. (2005). Rural and
suburban sprawl in the U.S. Midwest from 1940 to 2000 and its
relation to forest fragmentation. Conservation Biology 19, 793–805.
RAHBEK,C.&GRAVES, G. R. (2001). Multiscale assessment of
patterns of avian species richness. Proceedings of the National
Academy of Sciences U.S.A.98, 4534–4539.
RAPOPORT, E. H. (1993). The process of plant colonization in small
settlements and large cities. In Humans as Components of Ecosystems:
The Ecology of Subtle Human Effects and Populated Areas (eds. M. J.
McDonnell and S. T. A. Pickett, S.T.A.), pp. 190–207. Springer-
Verlag, New York.
RASKER,R.&HANSEN, A. (2000). Natural amenities and
population growth in the Greater Yellowstone Region. Human
Ecology Review 7, 30–40.
(2003). Relative importance of environment, human activity and
spatial situation in determining the distribution of terrestrial
mammal diversity in Argentina. Journal of Biogeography 30,
REED, W. J. (2002). On the rank–size distribution for human
settlements. Journal of Regional Science 42, 1–17.
REPETTO, R. (1988). The Forest for the Trees? Gover nment Policies and the Misuse
of Forest Resources. World Resources Institute, Washington, D.C.
REPETTO, R. (1994). The ‘‘Second India’’ Revisited: Population, Poverty,
and Environmental Stress Over Two Decades. World Resources
Institute, Washington, D.C.
D. J. (2000). Changing species richness and composition in
Canadian national parks. Conservation Biology 14, 1099–1109.
ROHDE, K. (1999). Latitudinal gradients in species diversity and
Rapoport’s rule revisited: a review of recent work and what can
parasites teach us about the causes of the gradients? Ecography
22, 593–613.
Statistical Software for Meta-Analysis Version 2. Sinauer Associates
Inc., Massachusetts.
ROSENTHAL, R. (1979). The ‘‘file drawer problem’’ and tolerance
for null results. Psychological Bulletin 86, 638–641.
COPELAND,J.P.&EDELMANN, F. B. (2003). Evaluation of
landscape models for wolverines in the interior northwest,
United States of America. Journal of Mammalogy 84, 92–105.
RUPINSKI,M.T.&DUNLAP, W. P. (1996). Approximating Pearson
product-moment correlations from Kendall’s tau and Spear-
man’s rho. Educational and Psychological Measurement 56, 419–429.
SANDERS, N. J. (2002). Elevational gradients in ant species richness:
area, geometry, and Rapoport’s rule. Ecography 25, 25–32.
level of threat to restricted–range bird species can be predicted
from mapped data on land use and human population. Biological
Conservation 123, 317–326.
Explaining human settlement patterns in a recreational lake
district: Vilas County, Wisconsin, USA. Environmental Management
30, 24–34.
SELMI,S.&BOULINIER, T. (2001). Ecological biogeography of
Southern Ocean Islands: the importance of considering spatial
issues. The American Naturalist 158, 426–437.
SILVA,L.&SMITH, C. W. (2004). A characterization of the non–
indigenous flora of the Azores Archipelago. Biological Invasions 6,
(1994). Physical and human dimensions of deforestation in
Amazonia. BioScience 44, 314–322.
SMALL,C.&NICHOLLS, R. J. (2003). A global analysis of human
settlement in coastal zones. Journal of Coastal Research 19, 584–
PETERJOHN, B. (2005). Plant species invasions along the
latitudinal gradient in the United States. Ecology 86, 2298–2309.
Conserving Africa’s rain forests: problems in protected areas
and possible solutions. Biological Conservation 123, 45–54.
TAYLOR,B.W.&IRWI N, R. E. (2004). Linking economic activities
to the distribution of exotic plants. Proceedings of the National
Academy of Sciences U.S.A.101, 17725–17730.
THOMPSON,K.&JONES, A. (1999). Human population density and
prediction of local plant extinction in Britain. Conservation Biolog y
13, 185–189.
(2005). Priority areas for the conservation of coastal marine
vertebrates in Chile. Biological Conservation 126, 420–428.
TURNER, W. R., NAKAMURA,T.&DINETTI, M. (2004). Global
urbanization and the separation of humans from nature.
BioScience 54, 585–590.
United Nations (2002). World Urbanization Prospects: The 2001
Revision. Department of Economic and Social Affairs, Population
Division, New York.
United Nations (2005). World Population Prospects: The 2004 Revision.
Department of Economic and Social Affairs, Population
Division, New York.
GASTON,K.J.&CHOWN, S. L. (2004). Conservation during
times of change: correlations between birds, climate and people
in South Africa. South African Journal of Science 100, 266–272.
´ZQUEZ, L–B. & Gaston, K. J. (2006). People and mammals in
Mexico: conservation conflicts at a national scale. Biodiversity and
Conservation 15, 2397–2414.
CHADWICK, O. A. (2004). Soils, agriculture, and society in
precontact Hawai’i. Science 304, 1665–1669.
VOGELMANN, J. E. (1995). Assessment of forest fragmentation
in southern New England using remote sensing and
geographic information systems technology. Conservation Biology
9, 439–449.
WHITE,L.J.T.&WILKIE, D. S. (2003). Catastrophic ape decline
in western equatorial Africa. Nature 422, 611–614.
WANG,M.C.&BUSHMAN, B. J. (1998). Using the normal quantile
plot to explore meta-analytic data sets. Psychological Methods 3,
WIERSMA, Y. F., NUDDS,T.D.&RIVARD, D. H. (2004). Models to
distinguish effects of landscape patterns and human population
pressures associated with species loss in Canadian national
parks. Landscape Ecology 19, 773–786.
Gary W. Luck630
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
Factors influencing the loss of an endangered ecosystem in
an urbanising landscape: a case study of native grasslands
from Melbourne, Australia. Landscape and Urban Planning 71,
Latitudinal gradients of biodiversity: pattern, process, scale, and
synthesis. Annual Review of Ecology, Evolution and Systematics,34,
WILSON, E. O. (1999). The Diversity of Life. W. W. Norton and
Company Incorporated, New York.
POSSINGHAM,H.&WESTON, C. (2005). Measuring and
incorporating vulnerability into conservation planning. Environ-
mental Management 35, 527–543.
WOODROFFE, R. (2000). Predators and people: using human
densities to interpret declines of large arnivores. Animal
Conservation 3, 165–173.
Yue, T. X., Wang, Y. A., Chen, S. P., Liu, J. Y., Qiu, D. S., Deng, X.
Z., Liu, M. L. & Tian, Y. Z. (2003). Numerical simulation of
population distribution in China. Population and Environment 25,
Human population density and biodiversity 631
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
Appendix 1. Data for all studies linking human population density (HPD) with a measure of biodiversity status. See Table 1 for an explanation of each variable. Extent and
grain size were estimated if the author(s) provided no specific information. Abbreviations used in the column headings: S (source #); A (analysis #, cross reference with Table
2 and Appendix 2); N(sample size: upper bound only); BC (bivariate correlation: Pearson, Spearman rank or Kendall); PC (partial correlation); BR (bivariate regression); PR
(partial regression); MR (multivariate regression; the variance explained by a multivariate model that included HPD irrespective of its individual contribution). Abbreviations
used in the table: P, positive; N, negative; NR, no relationship; spp., species; threat., threatened; conc., concentrated; end, endangered; SH, Southern Hemisphere; NH,
Northern Hemisphere; NP, National Park. All correlation and regression coefficients were rounded to two decimal places. Within a study, a blank space means the same value
as that above it (except for columns BC – MR where a blank space ¼no value provided). Lack of information is indicated by ‘?’. See footnotes for further details, list of
citations and scientific names for all species
S A Location Extent Grain size NResponse variable Measure of change P/N BC PC BR PR MR
1 1 Global Developing countries Country
39 Forests Deforestation
P 0.06 0.35
2 Country
25 P 0.17 0.50
2 3 Argentina
400 200 km 79 km
153 Collared peccary Relative abundance
4White-lipped peccary N
5 Chacoan peccary N
3 6 Europe Europe 2500 km
2434 Plants Species richness P 0.59
7 Birds P 0.19
8 Herptiles P 0.56
9 Mammals P 0.47
10 European endemics P 0.44
11 Narrow European endemics P 0.10
12 Restricted-range spp. P 0.17
13 All spp. P 0.55
14 Vertebrates P 0.38
15 Snakes P 0.46
16 Amphibians P 0.61
17 Mammalian carnivores P 0.05
18 Non-threat.; conc. in Europe P 0.31
19 Threat.; conc. in Europe P 0.30
20 Threat.; not conc. in Europe P 0.15
21 Globally threat. spp. P 0.21
4 22 Africa Sub-saharan Africa 1°1957 All spp. Species richness P 0.54
23 Amphibians P 0.35
24 Birds P 0.59
25 Mammals P 0.43
26 Snakes P 0.43
27 Narrow-range endemics P 0.39
28 Widely distributed spp. P 0.54
29 Threat. spp. P 0.36
30 Avian frugivores P 0.48
31 Nectarivores P 0.45
32 Aerial mammalian frugivores P 0.49
33 Large herbivores P 0.29
34 Waterbirds P 0.61
35 Aerial avian insectivores P 0.57
36 Woodpeckers P 0.45
37 Large carnivores P 0.50
38 Terrestrial mammalian insectivores P 0.38
39 Aerial mammalian insectivores P 0.50
5 40 Africa Tropical Africa Country 23 Closed broad-leaved forests Area deforested/annum
Gary W. Luck632
Biological Reviews 82 (2007) 607–645 Ó2007 The Author Journal compilation Ó2007 Cambridge Philosophical Society
41 % area deforested/annum P 0.46 0.21
6 42 Global Tropical regions Country 68 Tropical forests Annual deforestation rate P 0.68 0.42
43 Africa 32 P 0.51 0.33
44 Latin America ? P 0.48
45 Asia ? P 0.96
7 46 Canada Southern Quebec Municipality 59 Forests Woodlot density P 0.53
47 Woodland density P 0.49
48 Surface area of woodlots N 0.45
49 % area forested N 0.60 0.40
50 Discontinuity index N 0.91 0.77
8 51 Global Developing countries Country 68 Forests % cover N
52 Rate of deforestation NR
9 53 Portugal Portugal District 16 Egyptian mongoose Relative abundance N 0.26
54 County 76
N 0.35 0.29 0.82
10 55 Ghana
Ghana Reserve 6 Large mammals Extinction rate P 0.93 0.87 0.98
56 Carnivores P 0.81
57 Ungulates P 0.89
58 Primates P 0.23
59 Conservation reserves Reserve area N 0.76
11 60 Netherlands Amsterdam Town council 14 Rock dove Abundance P 0.91
12 61 USA Pacific northwest Census unit
1446 Forests Fragmentation index
P 0.18 0.80
62 Western Oregon 605 P 0.90
63 Western Washington