impact of simulated drought on ecosystem biomass production: an experimental test in stream...

Impact of simulated drought on ecosystem biomassproduction: an experimental test in stream mesocosmsM A R K E . L E D G E R *, F R A N C O I S K . E D WA R D S *w , L E E E . B R O W N *z,A L E X A N D E R M . M I L N E R *§ and G U Y W O O D WA R D }*School of Geography, Earth and Environmental Sciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK,

wCentre for Ecology and Hydrology, Maclean Building, Benson Lane, Crowmarsh Gifford, Wallingford OX10 8BB, UK, zSchool of

Geography, University of Leeds, Woodhouse Lane, Leeds LS2 9JT, UK, §Institute of Arctic Biology, University of Alaska, Fairbanks,

Alaska 99775, USA, }School of Biological and Chemical Sciences, Queen Mary, University of London, London E1 4NS, UK

Abstract

Climate models predict widespread shifts in precipitation patterns and increases in the frequency of extreme events

such as droughts, but consequences for key processes in affected ecosystems remains poorly understood. A 2-year

manipulative experiment used a series of stream mesocosms to test the effect of recurrent drought disturbance on the

composition and secondary production of macroinvertebrate consumer assemblages and functional groups. On

average, secondary production in drought-disturbed communities (mean 4.5 g m�2 yr�1) was less than half of that that

in controls (mean 10.4 g m�2 yr�1). The effects of the drought differed among functional feeding groups, with

substantial declines for detritivore shredders (by 69%) and engulfing predators (by 94%). Contrasting responses

were evident among taxa within most functional feeding groups, ranging from extirpation to irruptions in the case of

several small midge larvae, but production of most species was suppressed. Taxon-specific responses were related to

body mass and voltinism. The ratio of production to biomass (community P/B) increased under drought, reflecting a

shift in production from large long-lived taxa to smaller taxa with faster life cycles. This research provides some of the

first experimental evidence of the profound effects that droughts can have on both the structure and functioning of

aquatic ecosystems.

Keywords: disturbance, ecosystem functioning, macroinvertebrates, processes, secondary production, streams

Received 22 September 2010; revised version received 13 January 2011 and accepted 20 February 2011

Introduction

Climate models predict widespread shifts in regional

precipitation patterns (Beniston et al., 2007; IPCC, 2007)

that are likely to change the frequency of extreme

events, with potentially profound effects on ecosystems

(Gurvich et al., 2002). In many regions, climate change is

expected to cause untimely or unusually severe

droughts (Kundzewicz et al., 2008; Poff & Zimmerman,

2010) that will alter the hydrology and physical dis-

turbance regimes of many freshwater environments

(Milly et al., 2005; Woodward et al., 2010a, b). In rivers

and streams, climate-induced drought may be exacer-

bated by water extraction to satisfy growing domestic

and agricultural demand (Schindler & Donahue, 2006;

Chessman, 2009), with potentially far reaching conse-

quences for the structure and functioning of riverine

communities (Daufresne & Boet, 2007; Vorosmarty et al.,

2010).

Droughts typically reduce hydrological connectivity

and habitat availability in streams and increase the

deposition of fine sediments (Wood & Armitage, 1999;

Dewson et al., 2007a). Declining flows also typically

increase water temperature (Matthews, 1998), reduce

the availability of dissolved oxygen (Everard, 1996), and

alter the cycling of key nutrients (Dahm et al., 2003). The

spatial and temporal scale of dewatering and substra-

tum drying can vary from regular short disturbances of

habitat patches to infrequent but prolonged reach-scale

events (Stanley & Fisher, 1997). During severe droughts,

surface flow may cease in some patches, exposing the

bed to drying (Stanley & Fisher, 1997), although deeper

hyporheic sediments together with other patches on the

bed surface, may remain wet and act as potential

refugia for some organisms (James et al., 2008). The loss

of suitable habitat renders many freshwater organisms

particularly vulnerable to drought episodes (Beniston

et al., 2007; Beche et al., 2009). However, ecological

studies into the effects of drought have emerged only

gradually over the past 20 years, perhaps in part

reflecting the considerable logistical challenges inherent

in researching these events. Since droughts occurCorrespondence: Dr Mark Ledger, tel. 1 44 121 414 5540, fax 1 44

121 414 5528, e-mail: [email protected]

Global Change Biology (2011) 17, 2288–2297, doi: 10.1111/j.1365-2486.2011.02420.x

2288 r 2011 Blackwell Publishing Ltd

mailto:[email protected]

unpredictably in many systems, research inevitably has

tended to be phenomenological and opportunistic (e.g.

Ledger & Hildrew, 2001), often beset by confounding

gradients or lacking adequate reference or preimpact

data (Boulton, 2003; James et al., 2008), and controlled

manipulative experiments are rare (but see e.g. Dewson

et al., 2007b; Ledger et al., 2008). Nevertheless, several

reviews have emphasized marked effects of droughts

on structural attributes of aquatic communities (e.g.

Boulton, 2003; Lake, 2003; Dewson et al., 2007a), typi-

cally with reductions in species richness (e.g. Wood &

Petts, 1999) but contrasting effects on population abun-

dances (see Dewson et al., 2007b). In the face of drought,

invertebrates usually show low resistance whereas

resilience is more varied, with responses often appar-

ently species-specific (Lake, 2003; Clarke et al., 2010),

being related to their body mass and life-history traits

(Chadwick & Huryn, 2005, 2007). However, far less is

known about the consequences of shifts in community

structure for key ecosystem processes (Bertrand et al.,

2009). Recent research has examined the effect of low

flows on leaf-litter decomposition rates and invertebrate

drift (Dewson et al., 2007b), but how biodiversity loss

might alter ecosystem productivity in drought-affected

aquatic systems remains largely unexplored (but see

Chadwick & Huryn, 2007).

Climate models predict increased occurrences of

drought in many UK lowland rivers (Whitehead et al.,

2006) and chalk streams, the focus for our research, are

particularly susceptible because they are already sub-

jected to widespread water extraction. In these ground-

water-dominated systems, low flows can persist for

prolonged periods; against this background flow fluc-

tuations can cause repeated dewatering of some stream

bed patches, particularly those at the margins or on

riffle crests, while those in deeper water remain rela-

tively undisturbed and serve as local sources of recolo-

nists (Ledger et al., 2008). Drought in chalk streams can

alter the taxonomic composition and abundance of

benthic invertebrate communities adapted to life in

these rich, normally perennial waters (e.g. Ladle & Bass,

1981; Wood & Armitage, 2004; Wright et al., 2004), but

consequences of these events for ecosystem processes,

most notably secondary production, remain largely

unknown.

Secondary production, defined as the rate of total

population biomass of consumers accrued per unit time

and area, is a central pathway of energy flux in ecosys-

tems (Benke, 1979, 1984, 1993) and an important baro-

meter of environmental change and stress on functioning

in aquatic systems (Lugthart & Wallace, 1992; Burrell &

Ledger, 2003; Benke & Huryn, 2010). Estimates of

production incorporate individual growth rates, de-

velopmental time and standing stock biomass,

which together reflect the overall performance of po-

pulations and their responses to environmental stress

(Huryn & Wallace, 2000; Woodcock & Huryn, 2007).

Stream invertebrates are often classified into functional

feeding groups, based on their mode of resource acqui-

sition (Cummins, 1973; Ledger et al., 2002), and effects

of drought on the productivity of these populations

could influence the rates of organic matter processing

in these communities. In this study, the effect of drought

stress on benthic macroinvertebrate secondary produc-

tion was tested in a 2-year field experiment, via direct

manipulation of flow regimes. A series of replicate

stream mesocosms fed by a chalk stream were used to

simulate drought episodes (see Ledger et al., 2008, 2009).

Mesocosm-scale experiments provide the means to iso-

late key factors from confounding gradients inherent in

field survey approaches and to make direct compari-

sons between replicated communities under different

flow regimes. Previous research has shown that the

artificial channels used in this field trial maintain many

of the key physical and chemical conditions character-

istic of natural stream reaches (Harris et al., 2007), and

support realistic assemblages typical of natural chalk

streams (Ledger et al., 2008, 2009; Brown et al., 2011). The

experiment was used to test two hypotheses based on

the expected community and ecosystem level impacts

of disturbances, consistent with theoretical predictions

based on the theory of r- and K-selection (e.g. Pianka,

1970) and body-mass allometries (e.g. Peters, 1983;

Brown et al., 2004). Hypothesis one proposed that severe

drought disturbance would reduce overall invertebrate

secondary production per unit area through temporary

loss of biomass in disturbed habitat patches (cf. Pimm,

1991) and the corresponding null hypothesis was that

equal secondary production would be maintained in

disturbed and control patches. Hypothesis two pro-

posed that drought would drive shifts in community

composition and alter the distribution of production

within functional groups, with small more r-selected

species with fast life histories replacing larger, longer

lived more K-selected taxa (cf. Pianka, 1970). The sec-

ond null hypothesis was that the composition and

distribution of production within functional groups

would remain unchanged in the face of the droughts.

Materials and methods

Study site

The research was conducted over 24 months (March 2000–

February 2002) in stream mesocosms located at the Freshwater

Biological Association River laboratory, East Stoke, Dorset, UK

(5014004800N, 211100600W) (Harris, 2006; Harris et al., 2007). Four

blocks of three mesocosms were each sited immediately

D R O U G H T I M PA C T S O N S T R E A M S 2289

r 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 2288–2297

adjacent to a chalk stream and received water and suspended

particles (including algae, detritus, and invertebrates) through

a 110 mm diameter feeder pipe (6 m length) (see Appendix S1

in the Supporting Information). Each block of mesocosms

consisted of three stainless-steel-lined linear channels (width

0.33 m, length 12 m, depth 0.30 m). Two of the mesocosms in

each block were used in this study, with data from the third

mesocosm reported elsewhere (see Ledger et al., 2008). Water

flow through the mesocosms was controlled by valves at the

closed upper end of each channel. Water drained freely from

mesocosms under gravity, via an open outlet positioned 10 cm

above a downstream channel to prevent any potential cross-

contamination among the mesocosms. Channels were filled

with a 20 cm layer of stony substrate of the same substratum

particle size distribution (85% of particle volume 11–25 mm)

and geological parent material (chert) to that of the source

stream (Ledger et al., 2008), providing both benthic and inter-

stitial substrata in which suitably adapted species may find

refuge during drought (Ledger et al., 2009). However, the

mesocosms did not have extensive hyporheic zones, consistent

with many natural chalk streams (Harris, 2006; Trimmer et al.,

2010). Physicochemical conditions were highly congruent

among mesocosms (Harris et al., 2007) and closely paralleled

those of the source stream (Ledger et al., 2009). During the

main study period, water temperature (mean 12.2 1C) varied

seasonally, with summer maxima (18.7 1C in June 2000) and

winter minima (6.0 1C in December 2001). Inflowing water was

nutrient rich (mean PO4: 0.16 mg L�1; NO3: 5.62 mg L�1 from

March 2000–February 2002) with high pH (mean 8.1) and

conductivity (mean 460 mS cm�1) (Harris et al., 2007). Outside

simulated drought periods, discharge in the mesocosms was

stable (mean 0.005 m3 s�1, range 0.002–0.012 m3 s�1), with

mean water velocity and depth over the gravel of 0.20 m s�1

and 81 mm, respectively, and water residence times were short

(mean 66 s).

Experimental design and application

Unfiltered stream water was diverted into all mesocosms,

initiating the natural development of benthic assemblages

over 2 months. Thereafter, a drying disturbance was applied

at approximately monthly (mean 33-day) intervals (high fre-

quency disturbance treatment of Ledger et al., 2008) by closing

inlet valves and allowing water to drain gradually from the

mesocosms, exposing the benthos to a short (6-day) period of

flow cessation (i.e. 21 perturbations in total). During the

simulated drying periods, surface flows ceased and drying

of exposed substrata occurred in patches, whereas the inter-

stices beneath the bed surface remained wet, and small pools

persisted at intervals along the length of the dewatered

channels (Ledger et al., 2008). Surfaces of exposed substrata

dried at natural ambient rates such that the stress experienced

by organisms stranded in the mesocosms was consistent with

those in adjacent drying stream reaches (Harris, 2006). In the

control mesocosms, flows were continuous for the duration of

the experimental period (March 2000–February 2002). A

blocked experimental design was used such that each of the

four blocks of mesocosms contained one drought-disturbed

channel and one control channel (4 blocks� 2 treatments 5 8

channels in total) (Zar, 1999).

Sampling and processing

Benthic macroinvertebrates were sampled monthly from each

mesocosm between March 2000 and February 2002, immedi-

ately (1 h) before disturbances were applied. On each occasion,

three Surber samples (0.0225 m2, 300mm mesh aperture) were

taken from each replicate mesocosm to limit the extent of

destructive sampling (Harris, 2006). Macroinvertebrates were

sorted from debris, identified to the lowest practicable taxo-

nomic unit (usually species), and counted. Data from each of

the three samples were pooled to provide a single estimate of

abundance (m�2) for each mesocosm on each sampling occa-

sion (i.e. channels, not sample-units, were replicates). For

secondary production estimation, macroinvertebrate body

lengths (all individuals sampled, n 5 63 092) were measured

to the nearest 0.1 mm using an ocular graticule and dissecting

microscope.

Secondary production

Individual biomass (mg dry mass) was calculated for all

macroinvertebrate specimens using published length-mass

regressions (see Edwards et al., 2009). Secondary production

of all invertebrates was calculated using the size-frequency

method (Hynes & Coleman, 1968), excepting rare taxa (o1%

total numbers). Individuals collected over 12 sample dates

(monthly over a year) were grouped into discrete size cate-

gories so as to maximize the number of size classes while

ensuring that abundance decreased from one size category to

the next. The resulting size-class frequency distribution, the

hypothetical ‘average cohort’ (Hamilton, 1969), thus repre-

sented survivorship over the year, assuming organisms spent

an equal length of time in each size cohort (Benke, 1979). Mean

annual density (abundance m�2) was calculated for each size

class of each taxon. The number of individuals lost to mortality

between each successive size class was calculated for each

taxon and then expressed as a loss in biomass, with the sum of

these between size classes representing the production (P) of

the average cohort. Production was multiplied by the number

of size classes used for each taxon, as the method assumes the

number of average cohorts is equal to the number of size

classes (Benke, 1984). Because the method is based on devel-

opment over 1 year, the final estimate of production was

corrected using the cohort production interval (CPI) in days

(Benke, 1979), which represents larval development time and

therefore production was multiplied by 365/CPI. Values of

CPI were determined by examining the cohort structure over

the experiment. The mean biomass over the entire year (B) was

then calculated for each taxon, and used to calculate the taxon-

specific P/B ratio (Benke, 1984). For rare taxa, production was

estimated by multiplying mean annual biomass by an annual

P/B value of the most closely related taxon. Production was

estimated for each replicate control and treatment channel and

for the first year and the second year of the experiment

separately.

2290 M . E . L E D G E R et al.


Data analysis

Repeated-measures analysis of variance (RM-ANOVA) was used

to test the main effect of drought treatment, mesocosm block

(between-subject factors), year (within-subject factor), and

their interaction, on total annual secondary production and

that of each functional feeding group (ln-transformed). Both

treatment and mesocosm block were fixed effects in the

analysis. The RM-ANOVAs revealed consistent effects of drought

across the 2 years of the experiment (i.e. nonsignificant inter-

action between treatment and year) in all cases and therefore

data are presented graphically as mean annual secondary

production (i.e. the mean of year 1 and year 2). Taxa were

assigned to functional feeding groups, which are classifica-

tions of macroinvertebrates based on their role in the proces-

sing of organic matter, with reference to Moog (1995) as

follows: engulfing predators, piercing predators, collector-

gatherers, filterers, grazer-scrapers and shredders.

For each taxon, the mean of the percentage difference in

annual production between drought-affected channels and

controls was determined for each experimental block. One-

sample t-tests were then used to test whether taxon-specific

changes in secondary production differed significantly from

zero. Sequential Bonferroni corrections were applied to pre-

serve an alpha of 0.05 (Rice, 1989). Responses to drought were

also examined in relation to the potential number of life cycles

per year, based on voltinism classifications from Tachet et al.

(2000). A nonparametric Wilcoxon test was used to test for

significant differences in the responses of short-lived (41 cycle

year) and longer-lived taxa (� 1 cycle year) and a Kruskal–

Wallis test was used to ascertain the effect of body size (mg dry

mass) on macroinvertebrate responses to drought.

Results

Total secondary production varied significantly

between the treatments (RM-ANOVA, F1,3 5 17.58,

P 5 0.025) but there was no effect of mesocosm block

(F3,3 5 2.95, P 5 0.199) or year (F1,3 5 0.02, P 5 0.893),

nor any interaction between treatment and year

(F1,3 5 6.08, P 5 0.09; Appendix S2). Secondary pro-

duction in undisturbed mesocosms was 10 395 �2170 mg m�2 yr�1 and standing biomass was 3204 �365 mg m�2 with a community P/B ratio of 3.2 (Appen-

dix S3). Ninety-one percent of production was derived

from 15 taxa (each accounting for 41% total produc-

tion), with the remainder (59 taxa) combined contribut-

ing only a further 9% (Fig. 1; Appendix S3). In contrast,

in drought-affected mesocosms secondary production

(4509 � 790 mg AFDM m�2 yr�1) was significantly lower

(by 52 � 9%) than in controls, whereas standing bio-

mass was more strongly reduced (by 65 � 5% to

1113 � 151 mg m�2), and the community P/B was great-

er (4.1) (Appendix S3).

Primary consumers in the undisturbed community con-

sisted of grazer-scrapers (3417 � 965 mg AFDM m�2 yr�1),

collector-gatherers (3094 � 590 mg AFDM m�2 yr�1), filter

feeders (1195 � 217 mg AFDM m�2 yr�1) and shredders

(1471 � 475 mg AFDM m�2 yr�1, Fig. 2, Appendix S3).

Predators consisted of piercing (950 � 469 mg AFDM

m�2 yr�1) and engulfing (268 � 59 mg AFDM m�2 yr�1)

feeders (Fig. 2, Appendix S3). The effect of the drought

treatment differed among functional feeding groups

(Fig. 2, Table 1), with statistically significant (RM-ANOVA,

Po0.05) reductions in production for engulfing predators

Fig. 1 Mean ( � 1 SE) annual secondary production for macro-

invertebrates in drought treatments and controls. For each treat-

ment, taxa were ranked from left to right in order of decreasing

production.

Fig. 2 Mean (1 1 SE) annual secondary production for macro-

invertebrates across functional feeding groups in drought treat-

ments and controls. Asterisks denote statistically significant

differences between treatments (RM-ANOVA, *Po0.05, **Po0.005).



(�75% of controls), shredders (�69%), filterers (�60%),

and collector-gatherers (�55%), but not for grazers or

piercers (Fig. 2; Table 1).

Directional responses to disturbance differed mark-

edly among the fauna (Fig. 3; Appendix S3), with 49%

of taxa showing significantly lower production in dis-

turbed treatments than in controls, 9% showing higher

production (one sample t-tests, Po0.05; Fig. 3), and the

remainder showing no significant change (P40.05; Fig.

3). The percentage of taxa with reduced secondary

productivity differed among functional feeding groups:

shredders (100% of taxa reduced), engulfers (80%),

collectors (45%), piercers (45%), grazers (40%) and

filterers (25%). Only collector and grazer functional

groups contained taxa with increased productivity re-

lative to controls (Fig. 3).

Drought responses within functional groups

Production of the dominant shredders Gammarus pulex

(L.) (1291 � 463 mg AFDM m�2 yr�1) and Sericostoma

personatum (Spence in Kirby & Spence) (108 � 35 mg

AFDM m�2 yr�1) was strongly reduced by drought (by

64% and 99%, respectively, Fig. 4). Large engulfing

predators were also suppressed (Erpobdella octoculata

(L.), Polycentropus flavomaculatus (Curtis), Sialis lutaria

(L.), whereas the opposite was true for the much

smaller Tanypodinae larvae (Fig. 4). Contrasting re-

sponses were evident within the piercers, with beetle

larvae (Laccobius sp., Orectochilus villosus O.F. Muller)

and Turbellaria spp. more strongly affected than fly

larvae (Bezzia sp., Empididae and Tabanus sp.). Simi-

larly, among collector-gatherers, production by the

dominant snails (Potamopyrgus antipodarum (J.E. Gray)

and mayflies (Ephemera danica Muller, Serratella ignita

(Poda) was strongly reduced (Fig. 4), whereas produc-

tion by other collectors was weakly affected (Asellus

aquaticus (L.), Tubificidae) or increased in the case of the

much smaller Chironominae (Fig. 4). Grazer-scraper

production was dominated by the snail Radix balthica

(L.) (2903 � 969 mg AFDM m�2 yr�1), which contribu-

ted 28% of total secondary production and 85% of the

production of the grazer-scraper functional group.

Orthoclad chironomid larvae and the snail Valvata sp.

together contributed a further 3% of total production.

Of the grazers, R. balthica was strongly reduced by

Table 1 Summary of repeated measures analysis of variance (RM-ANOVA) testing the main effects of drought treatment, mesocosm

block (between-subject factors) and year (within-subject factors), and the interaction between treatment and year, on secondary

production (mg m�2 yr�1) for macroinvertebrates in six functional feeding groups

Functional group Source of variation Degrees of freedom F P

Collectors Drought 1,3 12.35 0.013

Block 3,3 0.15 0.922

Year 1,3 1.54 0.303

Treatment� year 3,3 3.78 0.147

Filterers Drought 1,3 14.69 0.031

Block 3,3 1.42 0.391

Year 1,3 14.50 0.032


Grazers Drought 1,3 4.26 0.131

Block 3,3 2.78 0.212

Year 1,3 6.20 0.089


Engulfers Drought 1,3 16.36 0.027

Block 3,3 0.488 0.715

Year 1,3 0.01 0.932


Shredders Drought 1,3 38.07 0.009

Block 3,3 22.68 0.015

Year 1,3 26.47 0.014


Piercers Drought 1,3 1.74 0.278

Block 3,3 1.78 0.325

Year 1,3 1.38 0.325


Statistically significant P values are in bold (see Appendix S2 for full RM-ANOVA tables).

2292 M . E . L E D G E R et al.


drought (by 50%) and both Valvata sp. and Ancylus

fluviatilis O.F. Muller were eliminated from the meso-

cosms, whereas the small orthoclads increased mark-

edly (by 52%, Fig. 4). The dominant filterers were

reduced (Pisidium sp.) or not significantly affected

(Hydropsyche siltalai Dohler) by the treatment (Fig. 4).

On average, the response of semivoltine (o1 year life

cycle per year, n 5 2 taxa) and univoltine (1 cycle per

year, n 5 32 taxa) taxa was significantly different from

that of multivoltine taxa (41 cycle per year, n 5 18 taxa;

Wilcoxon test, Po0.05) (Fig. 5a). Taxa with large body

mass were more susceptible to drought than smaller

taxa (Kruskal–Wallis H 5 11.49, P 5 0.009) (Fig. 5b).

Discussion

Coupled climate-hydrological models predict an in-

creased frequency of extreme hydrologic events, includ-

ing severe droughts, over the next century (Milly et al.,

2005; Beniston et al., 2007). Unpredictable droughts are

thought to be particularly deleterious for biota adapted

to life in running waters (Bonada et al., 2007; Beche et al.,

2009), but the impacts of these events on key stream

ecosystem processes are still poorly understood (Boul-

ton, 2003). The results of the present study, using

experimental mesocosms, indicate that periods of

drought can alter biomass production profoundly: 2

years of simulated flow intermittency halved overall

macroinvertebrate secondary production. Drought

effects also differed among functional feeding groups,

with strongest reductions for shredders and engulfing

predators. These changes may have wider effects on

organic matter processing and food web structure,

respectively. The standing stock biomass and produc-

tion of some taxa increased in response to drought,

raising the possibility that compensatory dynamics

might mitigate its effects, but these were often far out-

weighed by substantial reductions in productivity for

most populations.

The effect of drought on secondary production is a

function of the body size and life-history traits of

component species (Chadwick & Huryn, 2007). In-

creases in the community P/B ratios of disturbed

channels were observed that reflected shifts in commu-

nity composition, with production by small short-lived

taxa (41 cycle per year), notably chironomids and other

Diptera, replacing larger taxa with longer life cycles

(� 1 cycle per year). Resistance to drought is likely to

decline with increasing body size because large indivi-

duals should have less access to physical refugia in wet

benthic sediments (Lancaster & Hildrew, 1993). By

contrast, small short-lived (multivoltine) taxa can

quickly recruit individuals into space liberated by dis-

turbance, and have a lower probability of exposure than

larger univoltine or semivoltine macroinvertebrates

(e.g. Ledger & Hildrew, 2001). However, drought stress

strongly constrained the populations of the majority of

taxa, thereby limiting the potential for compensation

within and among functional groups.

Shredder secondary production was particularly

strongly suppressed (by 69%) by drought, reflecting

reduced biomass for all members of the group. The

most abundant shredders in the undisturbed meso-

cosms, G. pulex and S. personatum, are key processors

of detrital resources in many streams (Jonsson & Mal-

mqvist, 2000; Woodward et al., 2008). Their marked

vulnerability to drought, together with other members

of the group, may well have increased accumulation of

detritus and, hence, the availability of this key basal

resource in the postdrought environment (Chadwick &

Huryn, 2005). The immediate effect of drought on the

functioning of detritus-detritivore pathways could

therefore have wider implications for the trophic

Fig. 3 Distribution of drought effects on secondary production

for taxa in six functional feeding groups. Taxa were classified

according to their statistically significant positive ( 1 ) negative

(�) or lack of (0) response to droughts, as revealed by one-

sample t-tests.



economy and dynamic stability of riverine commu-

nities, because secondary production in many river

systems is based largely on terrestrial detritus (Ledger

& Winterbourn, 2000), and detrital pathways can dam-

pen the potentially destabilizing effects of autochtho-

nous based pathways in food webs (Rooney et al., 2006).

In terms of its influence on autochthonous algal-

based pathways in the food web, drought reduced

biomass and production of many grazers, including

the dominant snail R. balthica. Many of these taxa were

‘temporarily attached’ (sensu Tachet et al., 2000), living

in close association with the algae-coated upper sur-

faces of stones, and can become stranded where the

surface flows periodically recede (Harris, 2006). Radix is

a large and potent herbivore, capable of strong top-

down control of algal production, and also inhibits

colonization by herbivorous chironomids (Ledger

et al., 2006). In drought-disturbed channels, observed

declines in the biomass of Radix may to some degree

explain increased abundance of filamentous diatoms

(see Ledger et al., 2008), and hence the irruption of

small chironomid grazers in the postdrought assem-

blages. Although the increased production of chirono-

mids under drought treatments partially compensated

for the declines observed in other herbivores, losses of

key grazers suggests that severe droughts are likely to

lessen the overall intensity of herbivory in benthic

habitats. Droughts with no loss of surface flow may

have very different effects on algae–herbivore interac-

tions, however. For example, Power et al. (2008) showed

that the absence of floods in drought years favoured

large herbivorous caddisflies, which became more

abundant suppressing algal abundance on stones.

Effects of the drought on predator production dif-

fered markedly between functional groups, with engul-

fers being severely reduced and piercers less affected.

These contrasting responses may be related to body size

and other traits that determine refugium use. The large

engulfing predators which dominated production in

undisturbed channels, such as E. octoculata, S. fuliginosa

Fig. 4 The mean (� 1 SE) effect of drought disturbances, expressed as the percentage difference from control secondary production, for

core taxa (41% production in any functional group) in six functional feeding groups (a–f). (a) shredders, (b) engulfers, (c) piercers, (d)

filterers, (e) grazers and (f) collectors. Statistically significant (Po0.05) responses, as evidenced by one-sample t-tests, are denoted by

closed bars.

2294 M . E . L E D G E R et al.


and P. flavomaculatus, were strongly reduced as a con-

sequence of limited access to wet interstices in drought-

affected channels, whereas the much smaller (by one

to two orders of magnitude) predatory chironomids

increased production under drought conditions, pre-

sumably exploiting the surfeit of small prey (e.g., non-

predatory chironomids), and the absence of larger

predators, in refugia (Harris, 2006). The piercers were

taxonomically distinct from engulfers and their relative

resilience to drought may be explained by their rela-

tively small size, narrow body form and refuge-seeking

habits (Tachet et al., 2000; Harris, 2006). The loss of

larger predators can have destabilizing effects on the

food webs of aquatic communities, potentially releasing

prey populations and prompting trophic cascades

(Power et al., 2008) and a range of other indirect effects

(Montoya et al., 2009). Although a significant decline in

overall biomass and production of all predators was not

observed, the considerable declines seen in large-bod-

ied engulfing predators is likely to lessen pressure on

focal prey species and may to some degree account for

the increased biomass and production of taxa, notably

chironomids, observed under drought conditions.

Collector-gatherers were well represented in the

benthos and there was a striking variety of responses

to the drought within the group. Chironomids typically

benefited but losses of the larger and most productive

species, P. antipodarum and E. danica, which frequently

became stranded on the substratum surface in dewa-

tered patches (Harris, 2006), more than outweighed

these increases. These shifts in community structure,

from relatively large long-lived species to small multi-

voltine taxa with a rapid capacity for increase following

droughts, are broadly consistent with the predictions of

ecological theory (e.g. Pianka, 1970).

Several recent studies have emphasized the need for

more rigorous experimental approaches to detect the

mechanisms and processes of drought in running waters

(e.g. Dewson et al., 2007a, b; James et al., 2008). In this

study, flows in a series of outdoor mesocosms adjacent to

a chalk stream were manipulated to simulate drying

episodes, in order to gain a higher degree of replicability

and control than is possible with field surveys (see Harris

et al., 2007). Experiments conducted at relatively small

spatial scales under controlled conditions can neverthe-

less lack the realism of field studies (Ledger et al., 2009).

Drought was simulated in an array of relatively large

artificial stream channels and conditions were consistent

with those in natural streams in several key respects.

First, macroinvertebrate assemblages in the mesocosms

were characteristic of those of the parent river, forming

complex networks of interacting species (Ledger et al.,

2009; Brown et al., 2011). Second, key physical and

chemical conditions in the mesocosms, including those

of the drought environment, were analogous to the

source stream (Harris, 2006; Harris et al., 2007).

In order to determine the effects of habitat drying on

annual secondary production within an experimental

setting it was necessary to run the experiments, includ-

ing treatment applications, across both cool and warm

seasons, and in this regard our manipulations were

consistent with the occurrence of supraseasonal drought

and/or excessive water extraction. Samples collected

immediately following each disturbance event showed

that ecological resistance to drying episodes was related

to disturbance intensity, with highest mortality of macro-

invertebrates in warm periods when high ambient tem-

peratures rapidly dried exposed sediments, and lowest

mortality during cool periods when water pooled at the

substratum surface (Harris, 2006). The necessity to con-

duct the experiment at relatively small spatial scales may

Fig. 5 Mean (� 1 SE) effect of drought on secondary produc-

tion for macroinvertebrate taxa in relation to (a) the potential

number of life-cycles per year and (b) mean individual body

mass.



render our results conservative, however, because in the

drought simulations, previously disturbed patches (the

mesocosms) were readily recolonized from undisturbed

habitats in the nearby source stream. In this regard, the

experiment reflects the effects of droughts that cause

recurrent dewatering at relatively small spatial scales:

for example, habitat patches within river reaches.

Nevertheless, these dynamics serve as a precursor to

more extreme events that cause extensive drying across

the wider riverscape. Ultimately, the effects of drought

on ecosystem functioning are likely to depend upon the

many facets of drought regimes, including their dura-

tion, frequency and intensity (Poff & Zimmerman,

2010). In our view, many more systematic experimental

studies are now needed to tease out the relationships

between flow regime and processes in order to fully

understand both the general and contingent effects of

drought on aquatic systems.

Conclusion

This study provides evidence that drought conditions can

lead to strong and seemingly predictable reductions in

the secondary production of stream macroinvertebrates,

with consequences for the quantity and distribution of

energy flow from basal resources to higher consumers.

The suppression of secondary production by drought

could have marked effects that ramify through the food

web, for example, by constraining vertebrate predator

populations (e.g. fish, birds), as well as influencing the

rate of other key ecosystem processes, such as primary

production, nutrient cycling, herbivory and detrital de-

composition rates. The challenge now is to use a range of

additional large-scale and long-term experiments to ex-

plore how community structure, ecosystem functioning

and their interaction, are affected not only by drought,

but also by other key components of global climate

change including alterations to thermal regimes and

nutrient availability (Woodward et al., 2010a).

Acknowledgements

The authors thank the staff at the Centre for Ecology andHydrology and the Freshwater Biological Association (FBA)River Laboratory, Dorset, UK, for supporting this research. Theresearch was funded by a FBA/NERC postdoctoral fellowshipawarded to MEL and NERC grant NER/B/S/2002/215. Mr BGodfrey, Dr R Harris and Dr B Ledger provided assistance in thefield. Three anonymous referees provided valuable comments onan earlier draft of the manuscript.

References

Beche LA, Connors PG, Resh VH, Merenlender AM (2009) Resilience of fishes and

invertebrates to prolonged drought in two California streams. Ecography, 32, 778–

788.

Beniston M, Stephenson DB, Christensen OB et al. (2007) Future extreme events in

European climate: an exploration of regional climate model projections. Climate

Change, 81, 71–95.

Benke AC (1979) Modification of the Hynes method for estimating secondary

production with particular significance for multivoltine populations. Limnology

and Oceanography, 24, 168–171.

Benke AC (1984) Secondary production of aquatic insects. In: The Ecology of Aquatic

Insects (eds Resh VH, Rosenberg DM), pp. 289–322. Praeger Scientific, New York.

Benke AC (1993) Concepts and patterns of invertebrate production in running waters.

Verhandlungen der Internationalen Vereingung fur Theoretische und Angewandte Limno-

logie, 25, 15–38.

Benke AC, Huryn AD (2010) Benthic invertebrate production-facilitating answers to

ecological riddles in freshwater ecosystems. Journal of the North American Bentholo-

gical Society, 29, 264–285.

Bertrand KN, Gido KB, Dodds WK, Murdock JN, Whiles MR (2009) Disturbance

frequency and functional identity mediate ecosystem processes in prairie streams.

Oikos, 118, 917–933.

Bonada N, Doledec S, Statzner B (2007) Taxonomic and biological trait differences of

stream macroinvertebrate communities between mediterranean and temperate regions:

implications for future climatic scenarios. Global Change Biology, 13, 1658–1671.

Boulton AJ (2003) Parallels and contrasts in the effects of drought on stream macro-

invertebrate assemblages. Freshwater Biology, 48, 1173–1185.

Brown JH, Gilooly JF, Allen AP, Savage VM, West GB (2004) Toward a metabolic theory

of ecology. Ecology Letters, 85, 1771–1789.

Brown LE, Edwards F, Milner AM, Woodward G, Ledger ME (2011) Food web

complexity and allometric scaling relationships in stream mesocosms: implications

for experimentation. Journal of Animal Ecology, doi: 10.1111/j.1365-2656.2011.

01814.x.

Burrell GP, Ledger ME (2003) Growth of a stream-dwelling caddisfly (Olinga feredayi:

Conosucidae) on surface and hyporheic food resources. Journal of the North

American Benthological Society, 22, 92–104.

Chadwick MA, Huryn AD (2005) Response of stream macroinvertebrate production to

atmospheric nitrogen deposition and channel drying. Limnology and Oceanography,

50, 228–236.

Chadwick MA, Huryn AD (2007) Role of habitat in determining macroinvertebrate

production in an intermittent-stream system. Freshwater Biology, 52, 240–251.

Chessman BC (2009) Climatic changes and 13-year trends in stream macroinvertebrate

assemblages in New South Wales, Australia. Global Change Biology, 15, 2791–2802.

Clarke A, MacNally R, Bond N, Lake PS (2010) Flow permanence affects aquatic

macroinvertebrate diversity and community structure in three headwater streams in

a forested catchment. Canadian Journal of Fisheries and Aquatic Sciences, 67, 1649–1657.

Cummins KW (1973) Trophic relations of aquatic insects. Annual Review of Entomology,

18, 183–206.

Dahm CN, Baker MA, Moore DI, Thibault JR (2003) Coupled biogeochemical and

hydrological responses of streams and rivers to drought. Freshwater Biology, 48,

1219–1232.

Daufresne M, Boet P (2007) Climate change impacts on structure and diversity of fish

communities in rivers. Global Change Biology, 13, 2467–2478.

Dewson ZS, James ABW, Death RG (2007a) A review of the consequences of decreased

flow for instream habitat and macroinvertebrates. Journal of the North American

Benthological Society, 26, 401–415.

Dewson ZS, James ABW, Death RG (2007b) Stream ecosystem functioning under

reduced flow conditions. Ecological Applications, 17, 1797–1808.

Edwards FK, Lauridsen RB, Armand L, Vincent HM, Jones JI (2009) The relationship

between length, mass and preservation time for three species of freshwater leeches

(Hirudinea). Fundamental and Applied Limnology/Archiv fur Hydrobiologie, 173,

321–327.

Everard M (1996) The importance of periodic droughts for maintaining diversity in the

freshwater environment. Freshwater Forum, 7, 33–50.

Gurvich DE, Diaz S, Falczuk V, Perez-Harguindeguy N, Cabido M, Christopher

Thorpe P (2002) Foliar resistance to simulated extreme temperature events in

contrasting plant functional and chorological types. Global Change Biology, 8,

1139–1145.

Hamilton AL (1969) On estimating annual production. Limnology and Oceanography, 14,

771–782.

Harris RML (2006) The effect of experimental drought disturbance on macroinvertebrate

assemblages in stream mesocosms. PhD Thesis, University of Birmingham, UK.

Harris RML, Milner AM, Armitage PD, Ledger ME (2007) Replicability of physico-

chemistry and macroinvertebrate assemblages in stream mesocosms: implications

for experimental research. Freshwater Biology, 52, 2434–2443.

2296 M . E . L E D G E R et al.


10.1111/j.1365-2656.2011.01814.x

10.1111/j.1365-2656.2011.01814.x

Huryn AD, Wallace JB (2000) Life history and production of stream insects. Annual

Review of Entomology, 45, 83–110.

Hynes HBN, Coleman MJ (1968) A simple method of assessing the annual production

of stream benthos. Limnology and Oceanography, 13, 569–573.

IPCC (2007) Climate Change 2007: Impacts, Adaptation and Vulnerability.. In: Con-

tribution of Working Group II to the Fourth Assessment. Report of the Intergovernmental

Panel on Climate Change, (eds Parry ML, Canziani OF, Palutikof JP, vanderLinden PJ,

Hanson CE), Cambridge University Press, Cambridge, UK, 976 pp.

James ABW, Dewson ZS, Death RG (2008) Do stream macroinvertebrates use instream

refugia in response to severe short-term flow reduction in New Zealand streams?

Freshwater Biology, 53, 1316–1334.

Jonsson M, Malmqvist B (2000) Ecosystem process rate increases with animal species

richness: evidence from leaf-eating, aquatic insects. Oikos, 89, 519–523.

Kundzewicz ZW, Mata LJ, Arnell N et al. (2008) The implications of projected climate

change for freshwater resources and their management. Hydrological Sciences

Journal, 53, 3–10.

Ladle M, Bass JAB (1981) The ecology of a small chalk stream and its responses to

drying during drought conditions. Archiv fur Hydrobiologie, 90, 448–466.

Lake PS (2003) Ecological effects of perturbation by drought in flowing water. Fresh-

water Biology, 48, 1161–1172.

Lancaster J, Hildrew AG (1993) Flow refugia and the microdistribution of lotic macro-

invertebrates. Journal of the North American Benthological Society, 12, 385–393.

Ledger ME, Crowe AML, Woodward G, Winterbourn MJ (2002) Is the mobility of

stream insects related to their diet? Archiv fur Hydrobiologie, 154, 41–59.

Ledger ME, Harris RML, Armitage PD, Milner AM (2008) Disturbance frequency

influences patch dynamics in stream benthic algal communities. Oecologia, 155,

809–819.

Ledger ME, Harris RML, Armitage PD, Milner AM (2009) Realism of model ecosys-

tems: an evaluation of physicochemistry and macroinvertebrate assemblages in

artificial streams. Hydrobiologia, 617, 91–99.

Ledger ME, Harris RML, Milner AM, Armitage PD (2006) Disturbance, biological legacies

and community development in stream mesocosms. Oecologia, 148, 682–691.

Ledger ME, Hildrew AG (2001) Recolonization by the benthos of an acid stream

following a drought. Archiv fur Hydrobiologie, 152, 1–17.

Ledger ME, Winterbourn MJ (2000) Growth of New Zealand stream insect larvae in

relation to food type. Archiv fur Hydrobiologie, 149, 353–364.

Lugthart GJ, Wallace JB (1992) Effects of disturbance on benthic functional structure

and production in mountain streams. Journal of the North American Benthological

Society, 11, 138–164.

Matthews WJ (1998) Patterns in Freshwater Fish Ecology. Chapman & Hall, New York.

Milly PCD, Dunne KA, Vecchia AV (2005) Global pattern of trends in streamflow and

water availability in a changing climate. Nature, 438, 347–350.

Montoya JM, Woodward G, Emmerson MC, Sole RV (2009) Press perturbations and

indirect effects in real food webs. Ecology, 90, 2426–2433.

Moog O (1995) Fauna Aquatica Austriaca, Version 1995. Wasserwirtschaftskataster,

Bundesministerium fur Land- und Forstwirtschaft, Wien.

Peters RH (1983) The Ecological Implications of Body Size. Cambridge University Press,

Cambridge, UK.

Pianka ER (1970) On r and K selection. American Naturalist, 104, 592–597.

Pimm SL (1991) The Balance of Nature? Ecological Issues in the Conservation of Species and

Communities. University of Chicago Press, Chicago.

Poff NL, Zimmerman JKH (2010) Ecological responses to altered flow regimes: a

literature review to inform the science and management of environmental flows.

Freshwater Biology, 55, 194–205.

Power ME, Parker MS, Dietrich WE (2008) Seasonal reassembly of a river food web:

floods, droughts, and impacts of fish. Ecological Monographs, 78, 263–282.

Rice WR (1989) Analyzing tables of statistical tests. Evolution, 43, 223–225.

Rooney N, McKann K, Gellner G, Moore JC (2006) Structural asymmetry and the

stability of diverse food webs. Nature, 442, 265–269.

Schindler DW, Donahue WF (2006) An impending water crisis in Canada’s western

prairie provinces. Proceedings of the National Academy of Science USA, 103, 7210–7216.

Stanley EH, Fisher SG (1997) Ecosystem expansion and contraction in streams.

Bioscience, 47, 427–443.

Tachet H, Bournaud M, Richoux P, Usseglio-Polatera P (2000) Invertebres d’eau Douce:

Systematique, Biologie, Ecologie. CNRS Editions, Paris.

Trimmer M, Maanoja S, Hildrew AG, Pretty JL, Grey J (2010) Potential carbon fixation

via methane oxidation in well-oxygenated riverbed gravels. Limnology and Oceano-

graphy, 55, 560–568.

Vorosmarty CJ, McIntyre PB, Gessner MO et al. (2010) Global threats to human water

security and river biodiversity. Nature, 467, 555–561.

Whitehead PG, Wilby RL, Butterfield D, Wade AJ (2006) Impacts of climate change on

in-stream nitrogen in a lowland chalk stream: an appraisal of adaptation strategies.

Science of the Total Environment, 365, 260–273.

Wood PJ, Armitage PD (1999) Sediment deposition in a small lowland stream –

management implications. Regulated Rivers: Research and Management, 15, 199–210.

Wood PJ, Armitage PD (2004) The response of the macroinvertebrate community to

low-flow variability and supra-seasonal drought within a groundwater dominated

stream. Archiv fur Hydrobiologie, 161, 1–20.

Wood PJ, Petts GE (1999) The influence of drought on chalk stream macroinverte-

brates. Hydrological Processes, 13, 387–399.

Woodcock TS, Huryn AD (2007) The response of macroinvertebrate production to a

pollution gradient in a headwater stream. Freshwater Biology, 52, 177–196.

Woodward G, Benstead JP, Beveridge OS et al. (2010a) Ecological networks in a

changing climate. Advances in Ecological Research, 42, 71–138.

Woodward G, Papantoniou G, Edwards FE, Lauridsen R (2008) Trophic trickles and

cascades in a complex food web: impacts of a keystone predator on stream

community structure and ecosystem processes. Oikos, 117, 683–692.

Woodward G, Perkins DM, Brown LE (2010b) Climate change in freshwater ecosys-

tems: impacts across multiple levels of organisation. Philosophical Transactions of the

Royal Society B, 365, 2093–2106.

Wright JF, Clarke RT, Gunn RJM, Kneebone NT, Davey-Bowker J (2004) Impact of

major changes in flow regime on the macroinvertebrate assemblages of four chalk

stream sites, 1997–2001. River Research and Applications, 20, 775–794.

Zar JH (1999) Biostatistical Analysis, 4th edn. Prentice Hall, London.

Supporting Information

Additional Supporting Information may be found in the

online version of this article:

Appendix S1. Schematic representation (a) and photo-

graphic image (b) of the stream mesocosm facility at the

Freshwater Biological Association River Laboratory, Dorset,

U.K. Four blocks of three stream mesocosms (each channel

12 m length� 0.3 m width) were fed water through pipes

(6 m length) from the parent stream. Water flow (direction

indicated by arrows) in to each mesocosm was controlled by

a valve. Each block contained a control channel and a dis-

turbed channel, with the third channel in each block used in

allied research not reported here.

Appendix S2. Tables for repeated measures ANOVA testing

the main effects of drought treatment, experimental block

(between-subject factors) and year (within-subject factor), and

their interactions, on total macroinvertebrate secondary pro-

duction (a) and that of six functional feeding groups (b–g).

Appendix S3. Summary of benthic macroinvertebrate bio-

mass and secondary production in undisturbed controls and

monthly disturbed stream mesocosms. B mean annual bio-

mass (mg AFDM m-2), P mean annual secondary production

(mg AFDM m�2 year�1) with standard error in parentheses. B

– Bivalvia, C – Coleoptera, D – Diptera, E – Ephemeroptera,

G – Gastropoda, H – Hirudinea, He – Hemiptera, M –

Megaloptera, O – Oligochaeta, T – Trichoptera, Cr – Crusta-

cea, Od – Odonata, P – Plecoptera. Taxa within functional

feeding groups (totals in bold) are ranked alphabetically.

Please note: Wiley-Blackwell are not responsible for the con-

tent or functionality of any supporting materials supplied by

the authors. Any queries (other than missing material) should

be directed to the corresponding author for the article.



impact of simulated drought on ecosystem biomass production: an experimental test in stream...

Documents