cascading ecological responses to an in-stream restoration project in a midwestern river

R E S E A R C H A R T I C L E

Cascading Ecological Responses to an In-StreamRestoration Project in a Midwestern RiverKaleb K. Heinrich,1,2,3 Matt R. Whiles,1 and Charlotte Roy4

Abstract

River restoration projects are increasingly common, butassessments of ecological responses and overall success ofthe vast majority of efforts are lacking. Information onpotential positive ecological effects of restoration effortscan be used to justify further projects and refine meth-ods. We examined responses of multiple trophic levels,aquatic insects and riparian birds, to a series of rock weirsinstalled in an Illinois river to stabilize the channel. Wequantified adult insect emergence and performed weeklypoint counts of birds in spring at four weir and four non-weir (control) sites. Emerging insect abundance was higherat control sites, but species richness and diversity werehigher at weir sites. Total insect emergence production didnot differ between site types, but emergence productionof larger-bodied taxa was higher at weir sites. Ordinationsand analysis of similarity indicated differences in insect and

bird assemblages between site types. Birds showed a pos-itive numerical response to large-bodied emerging insects,and total bird abundance was higher at weir sites. Clutchsize and feeding rates of a focal bird species, Prothonotariacitrea (Prothonotary Warbler), did not differ between sites,but the number of hatchlings and fledglings was higherat control sites. Molothrus ater (cowbird) parasitism washigher at weir sites, likely because of increased edge habitatassociated with weir construction activities. Results showpositive ecological impacts of in-stream restoration andprovide justification for further efforts. However, forestdisturbance associated with construction could offset somebenefits to some species, and thus refinements to proce-dures may be necessary.

Key words: bird reproduction, ecological subsidies, insectemergence, Newbury weirs, riparian.

Introduction

River restoration activities have increased rapidly in the UnitedStates over the last 30 years, with expenses totaling over $1billion/year since 1990 (Bernhardt et al. 2005). However, littlepost-restoration assessment and monitoring occurs (Bernhardtet al. 2005). Considering the abundant financial resourcesgoing into river restoration, documenting ecological and/orsocietal outcomes is important to justify further projects(Palmer & Bernhardt 2006). Further, information generatedfrom post-restoration studies can be used to guide and refinefuture efforts.

The Cache River has been a focal point for stream restora-tion activities. A series of rock weirs, modeled after New-bury and Gaboury (1993), was installed in the Cache Riverto stabilize the channel and control entrenchment in 2001and 2003–2004. Walther and Whiles (2008) found that these

1Department of Zoology and Center for Ecology, Southern Illinois University,Carbondale, IL 62901-6501, U.S.A.2Present address: Stream Ecology Center, Department of Biological Sciences, IdahoState University, Pocatello, ID, 83209, U.S.A.3Address correspondence to K. K. Heinrich, email [email protected] Wildlife Populations and Research Group, Minnesota Department ofNatural Resources, Bemidji, MN 56601, U.S.A.

© 2013 Society for Ecological Restorationdoi: 10.1111/rec.12026

structures also benefited in-stream communities because ofincreased stability and habitat heterogeneity. Biomass ofaquatic insects was higher on rock weirs compared to ambientscoured clay substrata, suggesting weirs might enhance insectemergence production, which could benefit riparian predators(Walther & Whiles 2008).

Adult aquatic insects are important links between streamsand adjacent riparian habitats, as they facilitate flow of energyand nutrients from aquatic systems to terrestrial food webs(i.e. subsidies) (Sabo & Power 2002; Baxter et al. 2005).In particular, adult aquatic insects can be important preyfor insectivorous birds (Burdon & Harding 2008). Aquaticinsects can constitute the highest proportion of insectivorousbird diets during the autumn–spring defoliation season, whenterrestrial prey is scarce, and they can constitute an estimatedapproximately 26% of the annual energy budget for entire birdassemblages (Baxter et al. 2005).

As a result of the sometimes tight linkages between birds,stream insects, and riparian habitats, riparian birds are increas-ingly used as indicators of stream health and biotic integrity(Bryce 2006; Larsen et al. 2010). Use of avian communities toassess stream integrity is less labor intensive than using inver-tebrates or fish, and is particularly useful for communicatingresults to the public (Bryce et al. 2002). Birds are a logicalfocal group for study in the Cache River basin, in southern

72 Restoration Ecology Vol. 22, No. 1, pp. 72–80 JANUARY 2014

Responses to In-Stream Restoration

Illinois, because they are exceptionally diverse in this region,including numerous species that are threatened or of conser-vation concern.

The goal of our study was to quantify the potential influ-ences of a common stream restoration practice, constructionof rock weirs, on aquatic insect emergence production andriparian birds. Our specific objectives were to quantify andcompare: (1) adult aquatic insect abundance, emergence pro-duction, and taxonomic composition; (2) abundance, richness,and diversity of insectivorous birds; and (3) clutch size, fooddelivery rates, fledgling rates, and nest success of a focal insec-tivorous bird at weir and non-weir (control) sites.

Methods

Study Site

The Cache River is located in southern Illinois near theconfluence of the Ohio and Mississippi Rivers. It is recognizedas having nationally and internationally important wetlandhabitats, hosting greater than 100 threatened and endangeredspecies (IDNR 1997). The Post Creek cutoff, completedapproximately 1915, divided the river into upper and lowerwatersheds and resulted in severe stream entrenchment andlateral gully formation in the upper Cache, which is threateningriparian wetlands (Demissie & Xia 1991). Our study wasconducted in the upper Cache River, which drains a 632-km2

mosaic of agricultural lands, forests, and wetlands. From the 25weirs (approximately 15 m long) now located along the upperCache River at approximately 200 m intervals, we selected fourweir sites and four control sites (Fig. 1). Control sites werelocated approximately 100 m apart, midway between weirs.Substrata differ significantly between site types. Walther andWhiles (2008) found that mean particle size (longest axis) atweir sites was 175 mm, whereas non-weir sites were dominatedby scoured clay. To increase sample sizes, two additionalweir and control sites were used to examine food deliveryrates, clutch size, fledgling rates, and nest success of a focalinsectivorous bird species, Prothonotaria citrea (ProthonotaryWarbler). Home-range sizes for insectivorous bird specieswere determined from the literature and accounted for in thesite selection process (Poole 2005).

Emergence Production

We sampled aquatic insect emergence opportunistically(approximately every 4.5 ± 1.0 weeks) when discharge andweather allowed from January 2009 to June 2010 (n = 16). Wesampled more intensively during spring (21 March to 20 June;n = 7) to correspond with expected insect emergence peaksand bird nesting. Summer samples (n = 4) were collectedfrom 21 June to 20 September, fall samples (n = 2) from21 September to 20 December, and winter samples (n = 3)from 21 December to 20 March. For each sampling event,we tethered six floating emergence traps (0.10-m2 samplingarea, 250-μm mesh) randomly at all weir and control sitessimultaneously for 3–5 days. We collected insects from traps

Figure 1. Top: picture of a weir site during spring, 2010. Bottom:picture of a non-weir site during winter, 2010.

every 24–48 hours using a BioQuip Hand-Held Vac/Aspirator(BioQuip Products, Rancho Domingues, CA, U.S.A.). Sam-ples were frozen and later identified to family (Merrit et al.2008), measured, and expressed as emergence production[g dry mass (DM) m−2 d−1] using published length-massregressions (Rogers et al. 1977; Sample et al. 1993; Staglianoet al. 1998; Sabo et al. 2002). Ephemeroptera, Plecoptera,and Trichoptera (EPT) are recognized as indicators of healthyaquatic habitats (Barbour et al. 1999) and were our focus forsome analyses.

Water temperature data were collected with data loggersplaced at two weir and two control sites during sampling peri-ods, and data averaged for each site during sampling periods.Canopy cover was estimated at each site during spring, whenleaves were present, with a spherical densitometer. Compassdirection readings were recorded on the stream margin adjacentto each site.

Birds

We conducted weekly point counts from March to June (tocorrespond with spring migration and breeding periods) in2009 and 2010 to estimate bird abundance (birds/ha), speciesrichness, and species diversity at weir and control sites usedfor emergence sampling. A 50-m fixed-radius point count was

JANUARY 2014 Restoration Ecology 73


conducted between 06:00 and 09:00 hours for 5 minutes ateach site, including all bird species detected visually andaurally. Observers stood on the stream bank, adjacent to whereemergence traps were set, and included the stream channel inthe surveyed area. Repeated surveys were used to generatedetection probabilities to assess detection differences betweensite types. All detected species were included in analyses;most were insectivores (approximately 71.7%), and even thosenot generally considered insectivores can shift their diets inresponse to high resource availability (Bird & Smith 1964).Further, young of many bird species require a diet rich inprotein to promote rapid growth (Reinecke 1979).

We examined clutch size, food delivery rates, fledging rates,and nest success during the breeding season in spring 2009 and2010. For these analyses, we focused on the Prothonotary War-bler, because they are abundant insectivorous cavity nesters.Twelve nest boxes, similar to those of Fleming and Petit(1986), were placed along the stream bank in a 4 × 3 arraycentered on each of the weir and control sites that were used tomeasure emergence, including two additional weir and controlsites to accommodate additional nest boxes. Nest boxes wereplaced approximately 30 m apart and attached to greased con-duit to decrease predation (Hoover 2006). Box openings wereapproximately 38 mm to reduce parasitism by Molothrus ater(Brown-headed Cowbirds) (Hoover 2003). We deployed nestboxes in early March, before arrival of spring migrants, andmonitored them every 3–4 days to estimate nest success, theprobability that a nest produced a fledgling (Mayfield 1975).

When cowbird eggs were discovered in boxes, they wereimmediately removed because brood parasitism can increasefood delivery rates and reduce fledgling quality, thus con-founding any effect due to the restoration. Cowbird nestlingswill divert food from host nestlings (Hoover & Reetz 2006).Food delivery rates (number of trips/time) were observed for1 hour the morning of day 7; the day of highest energeticdemand (Poole 2005).

Statistical Analyses

An a priori α of 0.05 was used for all statistical analyses;p values between 0.05 and 0.1 were considered marginallysignificant. Water temperature and % canopy cover were com-pared using one-way analysis of variance (ANOVA). Insectemergence abundance and production and insectivorous birdabundance, species richness, and species diversity were com-pared with two-way repeated-measures ANOVA using SASPROC MIXED. Site was the main effect and each samplingdate was the repeated-measures effect (version 5; SAS Insti-tute, Cary, NC, USA). Insect abundance, emergence produc-tion, and bird abundance data were square-root transformedto reduce heteroscedasticity. Additionally, simple linear cor-relation was used to examine relationships between insectemergence and insectivorous bird density by matching closestpoint count date, within 5 days, to emergence sampling date(n = 7). Insect emergence and bird density values were aver-ages of all weir and control sites on each date. Bird detectionprobabilities were estimated for each point count survey date

using the program PRESENCE 3.1 (Hines 2006). Modelswere compared with and without site type using Akaike infor-mation criterion (AIC) statistics to compare detection prob-abilities. We also used non-metric multidimensional scaling(NMDS, Minchin 1997), complimented with analysis of sim-ilarity (ANOSIM) and similarity percentage (SIMPER), toexamine community structure of insects and birds. ANOSIMwas used to test for differences between site types (McCune& Grace 2002) and SIMPER breakdown was used to identifytaxa that were primarily responsible for observed differencesin bird assemblages between site types. We used PRIMER 6.0(Clarke & Gorley 2006) to perform NMDS, ANOSIM, andSIMPER analyses. Lastly, we used unpaired t-tests to com-pare food delivery rates, clutch size, fledging rates, and nestsuccess of Prothonotary Warblers.

Results

Average discharge and gage height of the upper Cache Riverduring our study were 10.9 m3/s and 4.1 m, respectively,with high variability during the study period. Average watertemperature measured on 12 sampling dates at weir sites(19.1◦C) and control sites (18.9◦C) was similar. Canopycover was approximately 69 and approximately 89% for weirsand controls, respectively, and differed between treatments(F = 6.22, p = 0.04).

Insect Emergence

Ten insect orders representing 60 families were collected.Emergence abundance (individuals/m2) ranged from 2.3 to210.0 at weir sites and 3.7 to 441.3 at control sites, and waslowest in winter and highest in summer at both site types(Fig. 2). Mean emergence abundance (individuals/m2 ± 1 SE)was higher at control sites (Table 1). ANOVA indicated sig-nificant effects of site type, date, and a site type × dateinteraction, with higher abundance at control sites during theAugust to October period, but similar abundance at other times(Fig. 2). Average abundances at weir and control sites inspring 2009 were 86.1 ± 20.7 and 88.1 ± 31.9 individuals/m2,respectively. During spring 2010, emergence abundance aver-aged 151.3 ± 51.7 individuals/m2 at weirs and 238.5 ± 127.2individuals/m2 at control sites. Mean emerging insect taxa rich-ness (d , Margalef Index) and diversity (H ′, Shannon Index)were higher at weirs compared to control sites, althoughboth interacted with sampling date (d site × date, F = 2.11,p = 0.018; H ′ site × date, F = 5.32, p < 0.001), with highervalues at weir sites throughout the year except during winter.Chironomidae made up a greater proportion of total abundanceand emergence production in control versus weir sites (88 vs.70% and 72 vs. 46%, respectively). EPT taxa constituted 17%of total abundance and 37% of emergence production at weirsites compared to 6 and 20%, respectively at control sites.

Emergence production ranged from 1.3 to 94.7 mg DMm−2 d−1 at weir sites and 2.2 to 100.3 mg DM m−2 d−1

at control sites. As with abundance, emergence production

74 Restoration Ecology JANUARY 2014


Figu

re2.

Tota

lin

sect

emer

genc

eab

unda

nce

(a),

emer

genc

epr

oduc

tion

(b),

tota

lE

PT(E

phem

erop

tera

,Pl

ecop

tera

,an

dT

rich

opte

ra)

taxa

emer

genc

eab

unda

nce

(c),

and

emer

genc

epr

oduc

tion

(d)

atw

eir

and

cont

rol

site

sin

2009

and

2010

.A

ster

isks

(*)

indi

cate

sign

ifica

ntdi

ffer

ence

sbe

twee

nw

eir

and

cont

rol

site

s(p

<0.

05).

Plus

sign

s(+

)in

dica

tem

argi

nally

sign

ifica

ntdi

ffer

ence

s(p

<0.

1>

0.05

).E

rror

bars

indi

cate

1SE

.



Table 1. Mean (±SE) insect diversity (Shannon Index, H ′), abundance, emergence production; and mean EPT abundance and emergence production atweir and control sites.

Site H′a Mean Abundancea Mean Emergence Production EPT Abundancea EPT Emergence Productiona

Weir 1 1.55 97.58 (21.68) 52.92 (10.84) 16.31 (4.43) 21.67 (6.19)Weir 2 1.31 84.63 (19.43) 54.41 (12.96) 15.71 (5.01) 17.00 (4.35)Weir 3 1.22 83.70 (18.89) 44.71 (9.12) 14.18 (3.92) 17.31 (5.36)Weir 4 1.32 67.73 (14.11) 31.84 (5.44) 11.66 (2.98) 12.14 (2.70)

Mean weirs 1.35 83.42 (9.24) 45.97 (5.00) 14.47 (2.04) 17.02 (2.39)Control 1 0.76 86.83 (26.19) 42.34 (12.42) 9.76 (4.26) 13.05 (5.92)Control 2 0.74 133.61 (31.87) 49.02 (10.49) 7.95 (3.22) 10.10 (4.52)Control 3 0.55 124.60 (39.38) 40.39 (11.05) 4.31 (1.79) 6.77 (3.28)Control 4 0.56 119.45 (34.48) 38.29 (7.65) 4.69 (1.51) 4.28 (1.51)

Mean controls 0.65 116.13 (16.41) 42.51 (5.17) 6.68 (1.45) 8.55 (2.06)

Abundance, individuals/m2; emergence production, mg dry mass m−2 d−1.aDifference (p < 0.05) between weir and non-weir sites.

was lowest during winter and peaked in spring and summer(Fig. 2). Mean emergence production over the study perioddid not differ at weir and control sites (Table 1), but theinteraction between site and date was marginally significant(Fig. 2). Emergence production during spring 2009 was53.9 ± 16.3 mg DM m−2 d−1 at weirs and 27.5 ± 12.9 mgDM m−2 d−1 at control sites. In spring of 2010, averageemergence production was 74.4 ± 15.6 mg DM m−2 d−1

at weir sites and 68.9 ± 24.4 mg DM m−2 d−1 at controlsites. Average individual mass of emerging insects was 1.4×higher at weir sites (0.66 ± 0.06 mg) compared to control sites(0.47 ± 0.03 mg) (site, F = 9.77, p = 0.004; date, F = 1.67,p = 0.074). Mean EPT emergence abundance was greaterat weir sites compared to control sites (Table 1), with siteand date effects, and an interaction between site and date,whereby temporal patterns at weir and control sites differed.Mean EPT emergence production was also greater at weirsites, with site and date effects, but no significant interaction(Fig. 2).

Bird Abundance, Richness, and Diversity

We detected 113 bird species during the study and overallbird abundance was higher at weir sites (Table 2). Models

of detection probabilities without site type had more supportthan with site type, indicating no difference in detectabilitybetween weirs and control sites. Average detection probabil-ity for all species did not differ between site types (Table 3).In 2009, the site effect was marginal (F = 3.53, p = 0.070)and date effect was significant (F = 3.85, p < 0.001), withno site × date interaction. In 2010, site and date effectswere significant (site, F = 10.44, p = 0.004; date, F = 1.98,p = 0.036) with no interaction. Analyses of only insectiv-orous bird species showed similar patterns in 2009 (site,F = 4.72, p = 0.037; date, F = 12.15, p < 0.001) and 2010(site, F = 7.66, p = 0.012; date, F = 3.35, p < 0.001).

Insect emergence production and insectivorous bird den-sity were positively correlated across spring sample dates(adj-R2 = 0.31, p = 0.009) and a stronger correlation (adj-R2 = 0.72, p = 0.001) was evident between larger-bodied EPTtaxa emergence production and insectivorous bird density(Fig. 3). Bird species richness and diversity did not varybetween site types in 2009, but date was important (d date,F = 5.68, p < 0.001; H ′ date, F = 7.28, p < 0.001). However,species richness and diversity were higher at weir sites thancontrol sites in 2010, with significant site and date effects(d site, F = 4.46, p = 0.048; d date, F = 5.82, p < 0.001; H ′site, F = 3.95, p = 0.063; H ′ date, F = 5.95, p < 0.001).

Table 2. Mean (±SE) avian abundance (birds/ha), species richness (Margalef Index, d ), and species diversity (Shannon Index, H ′) at weir and controlsites in 2009 and 2010.

2009 2010

Site Mean Abundancea d H′ Mean Abundanceb db H′b

Weir 1 18.6 (1.4) 3.44 2.19 20.2 (1.2) 3.95 2.36Weir 2 22.5 (1.3) 4.21 2.44 25.7 (1.3) 4.53 2.56Weir 3 23.3 (1.4) 3.68 2.29 24.9 (1.3) 4.43 2.54Weir 4 23.7 (1.2) 4.28 2.50 26.9 (1.1) 4.43 2.52

Mean weirs 22.0 (0.7) 3.90 2.35 24.4 (0.9) 4.34 2.49Control 1 20.3 (1.0) 4.09 2.41 18.9 (0.8) 3.80 2.34Control 2 16.6 (0.9) 3.45 2.17 18.9 (1.2) 3.90 2.35Control 3 21.1 (1.4) 4.12 2.40 22.8 (1.2) 4.14 2.45Control 4 21.9 (1.5) 3.90 2.31 20.6 (1.4) 3.96 2.37

Mean controls 20.0 (0.6) 3.89 2.32 20.3 (0.6) 3.95 2.38

aMarginal difference (p < 0.10) between weir and non-weir sites.bDifference (p < 0.05) between weir and non-weir sites.



Table 3. Model comparisons for detection probabilities (p) at weir sites and control sites in the Cache River using Program Presence.

Year AICc �AICc AICc Weight Weir Estimates Control Estimates

2009p (constant) 7104.06 0.00 0.91p (site type) 7108.71 4.65 0.09 0.275 ± 0.009 0.287 ± 0.009

2010p (constant) 7232.80 0.00 0.94p (site type) 7238.22 5.42 0.06 0.283 ± 0.009 0.289 ± 0.010

Detection probabilities are not influenced by site type.

Figure 3. Relationship between mean bird densities (insectivores only)and mean insect emergence production (a), and mean insectivorous birddensities and mean EPT (Ephemeroptera, Plecoptera, and Trichoptera)emergence (b) on spring sampling dates in the upper Cache River in2009 and 2010.

Assemblage Structure

Ordinations and ANOSIM performed on insect abundance(R = 1, p = 0.03, stress 0.01) and emergence production(R = 0.94, p = 0.029, stress 0.05) revealed distinct assem-blages at weir and control sites. SIMPER analysis of dissim-ilarity indicated that 38 insect families accounted for approx-imately 90% of observed differences in abundance betweensite types. Chironomidae accounted for approximately 14% ofobserved differences, with higher average abundance at control

sites. Representatives from the families Baetidae, Hydropsy-chidae, Heptageniidae, Empididae, and Simuliidae were moreabundant at weir sites and accounted for approximately 35%of observed differences. These six families made up approx-imately 50% of dissimilarity between sites. Additionally,SIMPER analysis performed on production data indicated 36families made up approximately 90% of observed differences,with Hydropsychidae, Heptageniidae, Baetidae, and Chirono-midae being the largest contributors. These groups, along withthe Leptophlebiidae, Simuliidae, Philopotamidae, Ephemeri-dae, and Perlidae accounted for approximately 50% of theobserved differences.

Ordinations also showed separation of bird assemblagesassociated with site types (Fig. 4). SIMPER analysis of dissim-ilarity indicated 43 species accounted for approximately 90%of the observed differences in 2009. Notable insectivores thatwere more abundant at weir sites in 2009 were the Prothono-tary Warbler, Polioptila caerulea (Blue-gray Gnatcatcher), andBaeolophus bicolor (Tufted Titmouse). In 2010, 53 speciesaccounted for approximately 90% of observed differences.Bird species that contributed to observed differences and weregenerally more abundant at weir sites included the Chaeturapelagica (Chimney Swift), Vireo flavifrons (Yellow-throatedVireo), Dendroica coronata (Yellow-rumped Warbler), andMyiarchus crinitus (Great Crested Flycatcher).

Prothonotary Warbler Reproduction

During the study, 95 nest boxes contained nesting materials.Prothonotary Warbler eggs were laid in 42 nests (24 in 2009and 18 in 2010). We observed provisioning rates for 31fledglings. The number of eggs laid in boxes was similarbetween site types. However, number of hatchlings andfledglings was higher at control sites (Table 4). Conversely, thenumber of cowbird eggs in nest boxes was approximately 1.7×higher at weir sites. Feeding rates (trips/hour) were marginallyhigher at control sites, but when considered per hatchling (tripsh−1 hatchling−1), were similar between sites. Percentage ofhatchlings that fledged at weir sites was approximately 20%lower than control sites (Table 4). Nest survival was 0.52 atweir sites and 0.73 at control sites.

Discussion

We demonstrate that a common in-stream restoration proce-dure (rock weirs) can have far reaching effects by enhancing



Figure 4. NMDS ordinations of bird abundance at weir and control sitesduring 2009 (a) and 2010 (b).

habitat for aquatic insects in the Cache River. This habitatmodification enhanced diversity and emergence production ofsome aquatic insect groups, particularly EPT taxa. Total insectabundance was higher at control sites because of one group,chironomid midges, which thrive in the ambient scoured claystreambed (Walther & Whiles 2008). However, midges arerelatively small and thus total emergence production was sim-ilar between site types because of tradeoffs between numbersand individual body size. Further, the small size of midgesmay make them less favored prey for some birds (Krebs et al.1977).

Weirs in the Cache promote insect diversity and produc-tion of larger-bodied groups through a variety of mechanisms.Mid- to large-sized low-gradient rivers, such as the Cache,often have limited stable substrata; stable substrata such aslarge rocks and woody debris often yield higher invertebrate

biomass than ambient fine sediments, and thus are importantforaging habitats for fishes and other insectivores (Benke et al.1984). Increased habitat heterogeneity associated with largersubstrata also provides a wider range of microhabitats thatpromotes insect species richness (Minshall 1984; Jacobsenet al. 1997). Weirs in the upper Cache River also representartificial riffles that influence water velocity and turbulence.Invertebrates respond to and segregate across velocity gradi-ents (Malas & Wallace 1977), and invertebrate communitiesin pools and riffles of the same system can be very different(Wallace et al. 1995), enhancing reach-scale diversity (i.e. betadiversity).

Our results suggest that an earlier study in the CacheRiver underestimated potential contributions of chironomids toemergence production at control sites. On the basis of benthicsamples, Walther and Whiles (2008) estimated that emergenceproduction would be approximately 2× higher from weirscompared to the streambed. However, we saw no difference intotal insect emergence production between site types becauseof large contributions of chironomids at control sites. Althoughsome chironomids may benefit from the weirs, many speciesof chironomids thrive in homogeneous substrata, where theycan dominate aquatic insect communities (Flinn et al. 2005).

Beyond in-stream responses, construction of weirs mayenhance biological connectivity between the river and for-est. Bird density was positively related to emerging aquaticinsects, particularly larger bodied insect taxa. Furthermore,bird richness and diversity were higher at weir sites in 1 yearof our study. Weirs may function as feeding “hot spots” forinsectivorous birds and possibly other riparian insectivores. Assuch, weirs like those in the Cache River have the potential toenhance stream to riparian energy and nutrient subsidies, andoverall connectivity between stream and riparian food webs(Murakami & Nakano 2002; Iwata et al. 2003; Baxter et al.2005). However, our study was correlative in nature, and thusthe patterns we observed do not establish causation.

We may have underestimated the positive effects of weirs ifthe dispersal of insects emerging from weirs created spillovereffects between treatment and control sites. This spatialconnection could mask treatment effects. Fletcher et al. (2007)found that restoration may affect areas at least 200–300 maway from a restored site because birds that increased inrestorations were also more abundant adjacent to restoredareas compared to areas near control sites. Jackson and Resh(1989) found that adult aquatic insects dispersed 150 m fromthe stream channel and EPT taxa can disperse greater than1 km (Kovats et al. 1996; Briers et al. 2004). Although some

Table 4. Means (±SE) of nest data for Prothonotary Warbler (host) nest boxes from 2009 and 2010.

Host Eggst = 0.55p = 0.29

Cowbird Eggst = 1.67p = 0.05

Hatchlingst = 1.97

p = 0.029

Fledglingst = 3.01

p = 0.003

Feeding Rate/ht = 1.62p = 0.06

Feeds Nestling−1 h−1

t = 0.35p = 0.37

% Hatchlings Fledgedt = 2.05p = 0.02

Weir n = 20 4.0 (0.3) 1.0 (0.2) 3.1 (0.4) 3.0 (0.4) 9.7 (0.7) 2.9 (0.2) 81.9 (8.7)Control n = 22 4.2 (0.4) 0.6 (0.1) 4.0 (0.3) 4.2 (0.2) 11.5 (0.8) 2.8 (0.2) 100.0 (0.0)

Differences were considered significant when p < 0.05 and marginally significant when 0.05 < p < 0.1. The test statistic (t) and number of nests (n) are indicated.



adult aquatic insects are capable of dispersing considerabledistances, overall positive responses to weirs by birds inour study was likely because densities of larger-bodied adultaquatic insects attenuated with distance from weirs.

Variation among point counts likely occurred because ofhabitat structure, ambient noise, wind conditions, and otherfactors (Wolf et al. 1995). Both site types had featuresthat could reduce detectability. Most notable was the noisefrom the water flowing over weirs, and somewhat denserforest associated with control sites. Regardless, detectionprobabilities at both site types were similar, supporting ourhypothesis that birds were responding to enhanced availabilityof larger insect prey at weirs.

On the basis of optimal foraging theory (MacArthur &Pianka 1966), birds should select larger, more profitable preyitems when they are available (Krebs et al. 1977). Thecorrelation between emergence production and bird densitiesfurther supports our hypothesis that birds were responding toinsect prey availability. Similar bird responses to emergingaquatic insect biomass were documented along a prairie stream(Gray 1993); these represent numerical responses, wherebypredators immigrate to an area in response to enhanced preydensities.

Although our results indicate overall positive responsesby insects and birds, and enhanced connections betweenstream and riparian food webs, reproductive success of theProthonotary Warbler was lower at weir sites. Two other cavitynesting bird species we examined also had lower reproductivesuccess at weir sites, although sample sizes were small(Carolina Chickadee n = 3 and Carolina Wren n = 2). Otherbird species, like open cup nesters, may respond differently.Installation of rock weirs involves the removal of shrubs andsome trees, which may reduce natural cavity availability andcanopy cover and increase edge habitat. Cowbirds prefer edgehabitats (Brittingham & Temple 1983), and this is consistentwith the higher parasitism rates and lower nest success weobserved at weir sites.

We did not see evidence of nest destruction by cowbirdsor predators, but both have been documented in the CacheRiver watershed (Hoover 2003). Hoover and Robinson (2007)found that when parasitized hosts eject parasitic eggs, femalecowbirds will destroy entire clutches or broods in retaliation.They also observed cowbirds farming hosts or forcing renest-ing attempts to create more opportunities for nest parasitism.Higher rates of cowbird parasitism may alter foraging behav-ior by increasing demands of parents by cowbird chicks. If theforest disturbance associated with weir construction increasedcowbird parasitism, rates of parasitism will likely wane as theriparian forest adjacent to weirs reestablishes (Brittingham &Temple 1996). Bock and Jones (2004) also noted that in areaswith more anthropogenic disturbance, bird density and repro-ductive success were often negatively related. Future projectsshould work to minimize disturbance to riparian forests, as thiscould offset some ecological benefits of restoration.

Assessments of restoration success include measuringphysicochemical characteristics or biological responses, butstandard criteria for deeming restorations successful do not

exist (Ruiz-Jaen & Aide 2005). Our results, along with futurestudies, can help develop standard approaches for evaluatingrestoration success in an ecological context, and refine, pro-mote, and justify further restoration efforts. Our investigationsshowed both positive (enhanced production of large insects)and negative (decreased hatchling production) effects associ-ated with restoration, allowing for a comprehensive assessmentof overall ecological success and improvement of methods.Our results also underscore the importance of measuring eco-logical responses at multiple spatial scales.

Implications for Practice

• Ecological assessments of stream channel restorationsshould examine multiple in-stream and riparian responsevariables.

• Linkages between in-stream habitat, stream communitydiversity and productivity, and riparian communitiescan be useful for justifying stream channel restorationprojects.

• Future channel restoration projects should minimizedisturbances to riparian vegetation, as this may offsetsome ecological benefits.

Acknowledgments

We thank the members of the SIUC Freshwater EcologyLaboratory, as well as S. Baer, E. Hellgren, L. Harrison, M.Guetersloh, J. Waycuilis, J. Hoover, and D. Lesmeister forvaluable assistance, advice, and support. We thank S. Yates,two anonymous reviewers, and M. Palmer for comments andsuggestions that greatly improved the manuscript. Fundingwas provided by the Illinois Department of Natural Resourcesthrough a USFWS State Wildlife Grant.

LITERATURE CITEDBarbour, M. T., J. Gerritsen, B. D. Snyder, and J. B. Stribling. 1999.

Rapid bioassessment protocols for use in streams and wadeable rivers:periphyton, benthic macroinvertebrates and fish, Second Edition. EPA841-B-99-002. U.S. Environmental Protection Agency, Office of Water,Washington, D.C.

Baxter, C. V., K. D. Fausch, and W. C. Saunders. 2005. Tangled webs:reciprocal flows of invertebrate prey link streams and riparian zones.Freshwater Biology 50:201–220.

Benke, A. C., T. C. Van Arsdall Jr., D. M. Gillespie, and F. K. Parrish. 1984.Invertebrate productivity in a subtropical blackwater river: the importanceof habitat and life history. Ecological Monographs 54:25–63.

Bernhardt, E. S., M. A. Palmer, J. D. Allan, G. Alexander, K. Barnas,S. Brooks, et al. 2005. Synthesizing US river restoration efforts. Science308:636–637.

Bird, R. D., and L. B. Smith. 1964. The food habits of the Red-wingedBlackbird, Agelaius phoeniceus , in Manitoba. Canadian Field Naturalist78:179–186.

Bock, C. E., and Z. F. Jones. 2004. Avian habitat evaluation: should countingbirds count? Frontiers in Ecology and the Environment 2:403–410.

Briers, R. A., J. H. R. Gee, H. M. Cariss, and R. Geoghegan. 2004.Inter-population dispersal by adult stoneflies detected by stable isotopeenrichment. Freshwater Biology 49:425–431.



Brittingham, M. C., and S. A. Temple. 1983. Have cowbirds caused forestsongbirds to decline? Bioscience 33:31–35.

Brittingham, M. C., and S. A. Temple. 1996. Vegetation around parasitizedand non-parasitized nests within deciduous forest. Journal of FieldOrnithology 67:406–413.

Bryce, S. A. 2006. Development of a bird integrity index: measuring avianresponse to disturbance in the Blue Mountains of Oregon, USA.Environmental Management 38:470–486.

Bryce, S. A., R. M. Hughes, and P. R. Kaufmann. 2002. Development of a birdintegrity index: using bird assemblages as indicators of riparian condition.Environmental Management 30:294–310.

Burdon, F. J., and J. S. Harding. 2008. The linkage between riparian predatorsand aquatic insects across a stream-resource spectrum. FreshwaterBiology 53:330–346.

Clarke, K. R., and R. N. Gorley. 2006. PRIMER v6: user manual/tutorial.PRIMER-E, Plymouth, United Kingdom.

Demissie, M., and R. Xia. 1991. Channel stabilizing structures for the upperCache River. Illinois State Water Survey Contract Report 511.

Fleming, W. J., and D. R. Petit. 1986. Modified milk carton nest boxfor studies of prothonotary warblers. Journal of Field Ornithology 57:313–315.

Fletcher, R., A. Cilimburg, and R. Hutto. 2007. Evaluating habitat restoration atO’dell Creek using bird communities: 2006 report. Final report submittedto PPL-Montana.

Flinn, M. B., M. R. Whiles, S. R. Adams, and J. E. Garvey. 2005.Macroinvertebrate and zooplankton responses to water-level managementand moist-soil plant production in upper Mississippi River backwaterwetlands. Archiv fur Hydrobiologie 162:187–210.

Gray, L. J. 1993. Response of insectivorous birds to emerging aquatic insectsin riparian habitats of a tallgrass prairie stream. American MidlandNaturalist 129:288–300.

Hines, J. E. 2006. PRESENCE2-Software to estimate patch occupancyand related parameters. USGS-PWRC (available from http://www.mbr-pwrc.usgs.gov/software/presence.html).

Hoover, J. P. 2003. Multiple effects of brood parasitism reduce the reproductivesuccess of prothonotary warblers, Protonotaria citrea . Animal Behaviour65:923–934.

Hoover, J. P. 2006. Water depth influences nest predation for a wetland-dependent bird in fragmented bottomland forests. Biological Conserva-tion 127:37–45.

Hoover, J. P., and M. J. Reetz. 2006. Brood parasitism increases provisioningrate, and reduces offspring recruitment and adult return rates, in a cowbirdhost. Oecologia 149:165–173.

Hoover, J. P., and S. K. Robinson. 2007. Retaliatory mafia behavior by aparasitic cowbird favors host acceptance of parasitic eggs. Proceedingsof the National Academy of Sciences of the United States of America104:4479–4483.

IDNR (Illinois Department of Natural Resources). 1997. Cache river areaassessment. Vol. 1, Part 1: Hydrology, Air Quality, and Climate. IDNR,Office of Scientific Research and Analysis, Springfield, Illinois.

Iwata, T., S. Nakano, and M. Murakami. 2003. Stream meanders increaseinsectivorous bird abundance in riparian deciduous forests. Ecography26:325–337.

Jackson, J. K., and V. H. Resh. 1989. Distribution and abundance of adultaquatic insects in the forest adjacent to a northern California stream.Environmental Entomology 18:278–283.

Jacobsen, D., R. Schultz, and A. Encalada. 1997. Structure and diversityof stream invertebrate assemblages: the influence of temperature withaltitude and latitude. Freshwater Biology 38:247–261.

Kovats, Z. E., J. J. H. Ciborowski, and L. D. Corkum. 1996. Inland dispersalof adult aquatic insects. Freshwater Biology 36:265–276.

Krebs, J. R., J. T. Erichsen, T. Webber, M. I. Charnov, and E. L. Charnov.1977. Optimal prey selection in the Great Tit (Parus major). AnimalBehaviour 25:30–38.

Larsen, S., A. Sorace, and L. Mancini. 2010. Riparian bird communities asindicators of human impacts along Mediterranean streams. EnvironmentalManagement 45:261–273.

MacArthur, R. H., and E. R. Pianka. 1966. On optimal use of a patchyenvironment. American Naturalist 100:603–609.

Malas, D., and J. B. Wallace. 1977. Strategies for coexistence in threespecies of net-spinning caddisflies (Trichoptera) in second-order southernAppalachian streams. Canadian Journal of Zoology 55:1829–1840.

Mayfield, H. F. 1975. Suggestions for calculating nest success. The WilsonBulletin 87:456–466.

McCune, B., and J. B. Grace. 2002. Analysis of ecological communities. MjMSoftware Design, Gleneden Beach, Oregon.

Merrit, R. W., K. W. Cummins, and M. B. Berg, editors. 2008. An introductionto the aquatic insects of North America. 4th edition. Kendall/Hunt,Dubuque, Iowa.

Minchin, P. R. 1997. An evaluation of the relative robustness of techniques forecological ordination. Vegetatio 69:89–107.

Minshall, G. W. 1984. Aquatic insect-substratum relationships. Pages 358–400in V. H. Resh and D. M. Rosenberg, editors. The ecology of aquaticinsects. Praeger, New York.

Murakami, M., and S. Nakano. 2002. Indirect effects of aquatic insectemergence on a terrestrial insect population through by birds predation.Ecology Letters 5:333–337.

Newbury, R. W., and M. N. Gaboury. 1993. Exploration and rehabilitationof hydraulic habitats in streams using principles of fluvial behavior.Freshwater Biology 29:195–210.

Palmer, M. A., and E. S. Bernhardt. 2006. Hydroecology and river restoration:ripe for research and synthesis. Water Resources Research 42:W03S07.

Poole, A.. 2005. The Birds of North America Online. Cornell Labora-tory of Ornithology, Ithaca, NY (available from http://bna.birds.cornell.edu/BNA/).

Reinecke, K. J. 1979. Feeding ecology and development of juvenile BlackDucks in Maine. Auk 96:737–745.

Rogers, L. E., R. L. Buschbom, and C. R. Watson. 1977. Length-weightrelationships of shrub-steppe invertebrates. Annals of the EntomologicalSociety of America 70:51–53.

Ruiz-Jaen, M. C., and T. M. Aide. 2005. Restoration success: how is it beingmeasured? Restoration Ecology 13:569–577.

Sabo, J. L., and M. E. Power. 2002. Numerical response of lizards toaquatic insects and short-term consequences for terrestrial prey. Ecology83:3023–3036.

Sabo, J. L., J. L. Bastow, and M. E. Power. 2002. Length-mass relationshipsfor adult aquatic and terrestrial invertebrates in a California watershed.Journal of the North American Benthological Society 21:336–343.

Sample, B. E., R. J. Cooper, R. D. Greer, and R. C. Whitmore. 1993. Estimationof insect biomass by length and width. American Midland Naturalist129:234–240.

Stagliano, D. M., A. C. Benke, and D. H. Anderson. 1998. Emergence ofaquatic insects from 2 habitats in a small wetland of the southeasternUSA: temporal patterns of numbers and biomass. Journal of the NorthAmerican Benthological Society 17:37–53.

Wallace, J. B., J. R. Webster, and J. L. Meyer. 1995. Influence of log additionson physical and biotic characteristics of a mountain stream. CanadianJournal of Fisheries and Aquatic Sciences 52:2120–2137.

Walther, D. A., and M. R. Whiles. 2008. Macroinvertebrate responses toconstructed riffles in the Cache River, Illinois, USA. EnvironmentalManagement 41:516–527.

Wolf, A. T., R. W. Howe, and G. J. Davis. 1995. Detectability of forest birdsfrom stationary points in northern Wisconsin. Pages 19–32 in C. J. Ralph,J. R. Sauer and S. Droege, editors. Monitoring bird populations by pointcounts USDA Forest Service General Technical Report PSW-GTR-149,Albany, California.


cascading ecological responses to an in-stream restoration project in a midwestern river

Documents