DOI: 10.1111/wre.

12310

INSIGHTS

What good is weed diversity?

J STORKEY & P NEVE
Rothamsted Research, Harpenden, Hertfordshire, UK

Received 8 February 2018

Revised version accepted 21 March 2018
Subject Editor: Matt Liebman, IA, USA

resistant species and that the diversity of the weed seed-

Summary
bank will be indicative of the overall sustainability of
Should the declining diversity of weed communities in the cropping system. Common to these hypotheses is
conventionally managed arable fields be regarded as a the idea that the intensification of agriculture has been
problem? The answer to this question has tended to accompanied by a homogenisation of cropping systems
divide researchers into those whose primary focus is on and landscapes, accounting for both declines in weed
conserving farmland biodiversity and those whose goals diversity and the reduced resilience of cropping systems
are dictated by weed control and maximising yield. (including the build-up of herbicide resistance). As such,
Here, we argue that, regardless of how weeds are per- weed communities represent a useful indicator of the
ceived, there are common ecological principles that success of rediversifying systems at multiple scales,
should underpin any approach to managing weed com- which will be a central component of making agriculture
munities, and, based on these principles, increasing in- and weed control more sustainable.
field weed diversity could be advantageous agronomi-
cally as well as environmentally. We hypothesise that a Keywords: niche differentiation, herbicide resistance,
more diverse weed community will be less competitive, sustainable intensification, Broadbalk experiment,
less prone to dominance by highly adapted, herbicide- species richness.

STORKEY J & NEVE P (2018). What good is weed diversity? Weed Research. https://doi.org/10.1111/wre.12310.

control and the concept of conserving weeds within

Introduction
cropped fields is, at best, incomprehensible and, at
The number of weed species that are typically found in worst, an insult to the efforts of weed scientists over
conventionally managed crop fields is now a fraction the past half-century to reduce the serious yield losses
of the levels recorded in the 1950-1970s, owing to inflicted by weeds (Oerke, 2006). This divergence in
increased fertiliser and herbicide use, simpler rotations peoples’ perception of weeds represents a philosophical
and loss of field boundaries and semi-natural features fault line running through the weed science community
in the landscape (Andreasen et al., 1996; Fried et al., and often reflects differences in scientific background
2009; Meyer et al., 2013). For some, this loss of in- (ecology vs. agronomy). Here, we argue that, rather
field biodiversity is a concern, reflecting an erosion of than perpetuating this dichotomy, we should recognise
the natural capital and ecosystem services on which that the objectives of maintaining arable biodiversity
sustainable production is founded. However, for and preventing cropping systems becoming dominated
others, it is seen as a measure of successful weed by a few highly competitive, herbicide-resistant weeds

Correspondence: J Storkey, Sustainable Agriculture Sciences, Rothamsted Research, West Common, Harpenden, Hertfordshire AL5 2JQ, UK.
Tel: (+44) 1582 938550; E-mail: [email protected]
© 2018 The Authors. Weed Research published by John Wiley & Sons Ltd on behalf of European Weed Research Society
This is an open access article under the terms of the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited.
2 J Storkey & P Neve

both rely on a common set of ecological and manage- postulated as the direct proximate cause of variation in
ment principles that should unite our research. species richness (Cardinale et al., 2009). Where
Two important points of clarification are needed to increased weed diversity is an emergent property of spa-
frame our argument. Firstly, the focus here is on weed tio-temporal heterogeneity in resource supply, therefore,
species richness and evenness, not density or total bio- we would expect decreased crop competition. There is
mass. There may be a benefit to increasing the total now some support for this hypothesis in the literature
amount of plant resources for higher trophic groups (Cierjacks et al., 2016), in contrast to experiments that
on farmland (including seeds for farmland birds and use artificially assembled weed communities which can
pollen and nectar for pollinators and biocontrol be confounded by the ‘sampling effect’ (Pollnac et al.,
agents), but the argument that weed biomass should be 2009). Here, we provide further support for the RPDH
managed in crops specifically to provide these ‘ecosys- using data from the Broadbalk long-term experiment
tem services’ is not our primary focus. Secondly, our at Rothamsted (Fig. 1). The Broadbalk experiment
emphasis is on the common weed flora rather than includes herbicide-free plots with contrasting fertiliser
those rare and threatened arable specialists that have treatments and naturally assembled weed communities
suffered the steepest declines in regional populations that we have compared with equivalent weed-free plots
over recent decades and require specific conservation to calculate yield loss. A classic ‘hump-backed’ rela-
measures (Storkey et al., 2012). Our argument for tionship of weed species richness is observed along the
increasing weed diversity is primarily agronomic and fertility gradient with the highest species richness
two pronged. Firstly, we contend that a more diverse observed on intermediate fertility plots where soil
weed community will be less competitive in any given resources are most evenly balanced (Moss et al., 2004).
crop and, secondly, that weed diversity is indicative of When the data were grouped on the basis of weed spe-
the wider sustainability of the whole cropping system. cies richness, a strong negative relationship with crop
yield loss was observed.
The concept of habitat heterogeneity can be extended
A more diverse weed community is less
beyond fine-scale spatial patchiness of soil resources to
competitive
capture a range of other management interventions that
Ecological niche theory argues that phenotypic differ- act as filters on the local weed species pool (Booth &
entiation between species results in contrasting ability Swanton, 2002); examples include crop sowing date and
to capture the resources required for growth and that intensity of cultivation. In functional terms, a greater
co-existence is supported by spatial and temporal diversity of crop types, nutrient inputs and cultivation
heterogeneity in resource availability (Silvertown et al., practices will lead to a greater breadth of weed ‘re-
1999; Chesson, 2000). While selective herbicides have sponse’ traits, consequently reducing the dominance of
been a major driver of recent declines in weed diver- competitive ‘effect traits’ that impact crop yield (Navas,
sity, the homogenisation of habitats through the use of 2012). Robust evidence supporting this hypothesis of
inorganic fertilisers and simplified crop rotations has weed competition based on functional traits is currently
also selected for fewer dominant species with similar lacking in the literature. In providing this evidence,
resource requirements to the crop (Borgy et al., 2012). long-term cropping experiments that result in naturally
Although persistent seedbanks will continue to buffer assembled gradients of weed functional diversity are
the negative impacts of management, it is likely that preferable to short-term studies on artificially selected
the remaining diversity in conventionally managed weed mixtures. Additional, trait-based, analysis of exist-
crops is now driven to a much larger extent than in ing experiments such as Broadbalk and the establish-
the past by repeated recolonisation from field edges or ment of new long-term cropping system experiments
neighbouring ruderal habitats. This may explain the that focus on weed diversity and crop yield loss should,
large numbers of ‘chance’ records observed in modern therefore, be a key research aim in the future.
weed surveys (Baessler & Klotz, 2006) and why diver-
sity is higher in complex landscapes with smaller fields
A more diverse weed community is an
(Gabriel et al., 2005; Gaba et al., 2010).
indicator of agronomic and environmental
Smith et al. (2010) explored the hypothesis that one
sustainability
consequence of a reducing niche breadth for weeds is
that interspecific competition with the crop intensifies The simplification of cropping systems and increased
as the diversity of available resource pools decreases – inputs of agro-chemicals have led to the dominance of a
the so-called Resource Pool Diversity Hypothesis few competitive, highly adapted and widely distributed
(RPDH). In unmanaged systems, the supply and weed species. This is exemplified by Alopecurus myosur-
imbalance of limiting resources have also been oides Huds. in north-west Europe (Delye et al., 2010),

© 2018 The Authors. Weed Research published by John Wiley & Sons Ltd on behalf of European Weed Research Society
 Weed diversity 3

80

70

60

Yield loss (%)

50

40

30

20

10

0
0 5 10 15 20 25
Number of weed species recorded

Fig. 1 Relationship between weed species richness and crop yield loss on the Broadbalk winter wheat experiment (begun in 1843). Weed
species richness is assessed on herbicide-free plots annually, and weed diversity varies in response to contrasting fertiliser treatments.
Winter wheat yield loss from weed competition can be calculated as a percentage of the equivalent plots with the same fertiliser treat-
ments but where weeds are controlled with herbicides. Data are presented for 19 years collected between 1991 and 2014 (the plots were
fallowed in some years during this period) and sorted by weed species richness. For each level of weed species richness, the average yield
loss is presented with error bars indicated the standard error of the mean (r2 = 0.59, P < 0.001). Plots with no nitrogen but with added
phosphorus and potassium are excluded from the analysis as the abundance of leguminous weeds leads to facilitation and greater yield
in the weedy plots compared to the weed-free plots.

Amaranthus palmeri S. Watson in southern and central environmental impacts. This is a multifaceted and
United States (Ward et al., 2013) and Lolium rigidum complex challenge that will require trading off multiple
Gaudin in Australia (Owen et al., 2014). As populations criteria using a range of metrics (Garnett et al., 2013).
of these species increase, growers become more reliant As well as being a direct threat to agronomic sustain-
on fewer herbicides, selection pressure increases and ability, we hypothesise that recent declines in weed spe-
dominant weeds quickly adapt to new management cies richness are correlated with the wider loss of
including becoming resistant to herbicides (Neve et al., cropping system resilience and that weed communities
2014). Many of the world’s cropping systems are now may represent a useful proxy for agronomic and envi-
afflicted by herbicide resistance, such that as weed diver- ronmental sustainability at the field, farm and land-
sity has declined, in many cases weed biomass has not; scape scale. There is potential, therefore, to develop an
this challenges the ongoing sustainability of the whole indicator of sustainable intensification built on the taxo-
system. Herbicide resistance is just one of multiple fac- nomic or functional diversity of the weed seedbank
tors currently threatening the agronomic and environ- that represents a legacy of previous management
mental sustainability of modern agriculture. Other across the whole cropping system. Our hypothesis is
pressures on the system include declining soil health, that fields with low seedbank diversity have a less sus-
pollution of water courses, greenhouse gas emissions tainable cropping system, both agronomically and
and declining functional biodiversity (including pollina- environmentally, than a field with a more diverse weed
tor populations). The loss of agro-ecosystem diversity community. Challenging this thesis will require a
has contributed to these problems; the cultivation of a research effort to establish relationships between weed
narrow range of functionally similar crops on large con- diversity and other metrics of SI, including soil health
tiguous areas is predicated on highly mechanised sys- and functional biodiversity and the development of
tems with high inputs of inorganic fertiliser and protocols that identify the appropriate measure of
pesticides, leading to open, leaky systems with minimal weed diversity. Because the emerged weed community
organic inputs to the soil. in any given year is determined by the management
In response to these stresses, there is a call for the specific to the crop being grown, as well as stochastic
‘sustainable intensification’ (SI) of agriculture that processes, sporadic assessments of the above-ground
increases food production without further adverse flora are not appropriate for this purpose. Rather,

© 2018 The Authors. Weed Research published by John Wiley & Sons Ltd on behalf of European Weed Research Society
4 J Storkey & P Neve

continuous assessments over the whole cropping sys- ‘availability’ from resource ‘imbalance’ on productivity-
tem or sampling of the seedbank are required to cap- diversity relationships. Ecology Letters 12, 475–487.
ture the response of the weed flora to the range of CHESSON P (2000) Mechanisms of maintenance of species
diversity. Annual Review of Ecology and Systematics 31,
management practices applied across the whole crop-
343–366.
ping system. CIERJACKS A, POMMERANZ M, SCHULZ K & ALMEIDA-CORTEZ
In conclusion, as weed biologists working at a single J (2016) Is crop yield related to weed species diversity and
institution but whose research focusses on environmen- biomass in coconut and banana fields of northeastern
tal and production endpoints respectively, we are con- Brazil? Agriculture Ecosystems & Environment 220, 175–
vinced that the loss of weed diversity and the 183.

DELYE C, MICHEL S, BERARD A et al. (2010) Geographical
escalation of resistance to herbicides are mediated by
variation in resistance to acetyl-coenzyme A carboxylase-
an identical underlying cause: the simplification of
inhibiting herbicides across the range of the arable weed
agroecosystems and their associated weed management
Alopecurus myosuroides (black-grass). New Phytologist 186,
strategies. Given this, we propose that the goals of 1005–1017.
designing weed management systems that maximise FERNANDEZ-QUINTANILLA C, QUADRANTI M, KUDSK P &
production and maintain ecosystem functioning are BARBERI P (2008) Which future for weed science? Weed
entirely compatible and mutually reinforcing. We Research 48, 297–301.
would, therefore, echo the call made by previous FRIED G, PETIT S, DESSAINT F & REBOUD X (2009) Arable
authors (Fernandez-Quintanilla et al., 2008; Jordan & weed decline in Northern France: crop edges as refugia
for weed conservation? Biological Conservation 142,
Davis, 2015) for weed scientists to integrate their work
238–243.
within the transdisciplinary framework that is required GABA S, CHAUVEL B, DESSAINT F, BRETAGNOLLE V & PETIT S
to meet the challenge of sustainable intensification and (2010) Weed species richness in winter wheat increases with
the transformation of cropping systems. In so doing, landscape heterogeneity. Agriculture Ecosystems &
we would move weed science from being a parochial Environment 138, 318–323.
discipline towards an integral part of a broader GABRIEL D, THIES C & TSCHARNTKE T (2005) Local diversity
research effort focussed on transforming the current, of arable weeds increases with landscape complexity.
Perspectives in Plant Ecology Evolution and Systematics 7,
flawed paradigm of modern intensive agriculture.
85–93.
GARNETT T, APPLEBY MC, BALMFORD A et al. (2013)
Acknowledgements Sustainable Intensification in Agriculture: Premises and
Policies. Science 341(6141), 33–34.
JS is supported by research programme NE/N018125/1 JORDAN NR & DAVIS AS (2015) Middle-way strategies for
LTS-M ASSIST – Achieving Sustainable Agricultural sustainable intensification of agriculture. BioScience 65,
Systems, funded by NERC and BBSRC (BBS/E/C/ 513–519.
000I0140), and would like to thank the DISCO-WEED MEYER S, WESCHE K, KRAUSE B & LEUSCHNER C (2013)
working group supported by CESAB-FRB for stimu- Dramatic losses of specialist arable plants in Central
lating discussions. PN is supported by the Biotechnol- Germany since the 1950s/60s-a cross-regional analysis.
Diversity and Distributions 19, 1175–1187.
ogy and Biological Sciences Research Council through
MOSS SR, STORKEY J, CUSSANS JW, PERRYMAN SAM &
the Smart Crop Protection (SCP) strategic programme HEWITT MV (2004) The Broadbalk long-term experiment
(BBS/OS/CP/000001) and the Black-grass Resistance at Rothamsted: what has it told us about weeds? Weed
Initiative (BB/L001489/1) at Rothamsted Research. Science 52, 864–873.
NAVAS M-L (2012) Trait-based approaches to unravelling
the assembly of weed communities and their impact
References on agro-ecosystem functioning. Weed Research 52,
ANDREASEN C, STRYHN H & STREIBIG JC (1996) Decline of 479–488.
the flora in Danish arable fields. Journal of Applied NEVE P, BUSI R, RENTON M & VILA-AIUB MM (2014)
Ecology 33, 619–626. Expanding the eco-evolutionary context of herbicide
BAESSLER C & KLOTZ S (2006) Effects of changes in resistance research. Pest Management Science 70, 1385–
agricultural land-use on landscape structure and arable 1393.
weed vegetation over the last 50 years. Agriculture OERKE EC (2006) Crop losses to pests. Journal of Agricultural
Ecosystems & Environment 115, 43–50. Science 144, 31–43.
BOOTH BD & SWANTON CJ (2002) Assembly theory applied to OWEN MJ, MARTINEZ NJ & POWLES SB (2014) Multiple
weed communities. Weed Science 50, 2–13. herbicide-resistant Lolium rigidum (annual ryegrass) now
BORGY B, GABA S, PETIT S & REBOUD X (2012) Non-random dominates across the Western Australian grain belt. Weed
distribution of weed species abundance in arable fields. Research 54, 314–324.
Weed Research 52, 383–389. POLLNAC FW, MAXWELL BD & MENALLED FD (2009) Weed
CARDINALE BJ, HILLEBRAND H, HARPOLE WS, GROSS K & community characteristics and crop performance: a
PTACNIK R (2009) Separating the influence of resource neighbourhood approach. Weed Research 49, 242–250.

© 2018 The Authors. Weed Research published by John Wiley & Sons Ltd on behalf of European Weed Research Society
 Weed diversity 5

SILVERTOWN J, DODD ME, GOWING DJG & MOUNTFORD JO STORKEY J, MEYER S, STILL KS & LEUSCHNER C (2012) The
(1999) Hydrologically defined niches reveal a basis for impact of agricultural intensification and land-use
species richness in plant communities. Nature 400, 61–63. change on the European arable flora. Proceedings
SMITH RG, MORTENSEN DA & RYAN MR (2010) A new of the Royal Society B-Biological Sciences 279, 1421–1429.
hypothesis for the functional role of diversity in WARD SM, WEBSTER TM & STECKEL LE (2013) Palmer
mediating resource pools and weed-crop competition in Amaranth (Amaranthus palmeri): a review. Weed
agroecosystems. Weed Research 50, 37–48. Technology 27, 12–27.

© 2018 The Authors. Weed Research published by John Wiley & Sons Ltd on behalf of European Weed Research Society