Science of the Total Environment 646 (2019) 770–781

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Irrigation canals are newly created streams of semi-arid agricultural regions

Erick A. Carlson a,b,⁎, David J. Cooper a,b, David M. Merritt a,c, Boris C. Kondratieff a,b,d, Reagan M. Waskom e
a
Graduate Degree Program in Ecology, Colorado State University, Fort Collins, CO 80523, United States of America
b
Department of Forest and Rangeland Stewardship, Colorado State University, Fort Collins, CO 80523, United States of America
c
USFS National Watershed, Fish and Wildlife Staff, Fort Collins, CO 80526, United States of America
d
Bioagricultural Sciences and Pest Management, Colorado State University, Fort Collins, CO 80523, United States of America
e
Colorado Water Institute, Colorado State University, Fort Collins, CO 50523, United States of America

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Irrigation canals are more prevalent in

some semi-arid agricultural landscapes.
• Riparian vegetation of irrigation canals
showed similarities to local streams.
• Aquatic macroinvertebrates of irrigation
canals were similar to local streams.
• Land-use affected similarities between
canals and streams.

a r t i c l e i n f o a b s t r a c t

Article history: The natural hydrologic processes that create and maintain the diversity of aquatic and riparian habitats along the
Received 30 March 2018 World's streams and rivers have been profoundly altered by humans. Diversion of surface water to support produc-
Received in revised form 16 July 2018 tion agriculture in arid and semi-arid regions has degraded ecosystems but also created potential habitat along and
Accepted 17 July 2018
in canals specifically designed to transport water. The prevalence of canals and the immense amount of water used
Available online 29 July 2018
for agriculture have created these new artificial stream systems. This study demonstrates the potential for irrigation
Editor: D. Barcelo canals to support riparian and aquatic communities similar to natural streams in urban/residential and agricultural
landscapes. We examined the hydrological and ecological characteristics of streams and irrigation canals in urban
Keywords: and agricultural landscapes in northeastern Colorado, typical of regions dominated by irrigation-supported agricul-
Riparian vegetation ture. Flow patterns in canals depended on their size and had a range of patterns with potential ecological conse-
Aquatic macroinvertebrates quences such as rapidly rising and falling water stage, intermittent dry periods, and delayed peak and base flows
Irrigation canal compared to natural streams. Despite these hydrologic differences, the taxonomic and functional composition of
Community ecology riparian plant and aquatic macroinvertebrate communities indicated that ecological similarities exist between
streams and canals, but are dependent, in part, on their landscape setting with stronger similarities in agricultural
areas. We also tested the influence of characterizing taxa by functional groups using physiology, ecology and life
history traits to explore attributes of habitats including woody canopy structure and water quality. We used a
Habitat Quality Index (HQI) that combined physical and biological measures into a single index. Streams scored
higher on average within agriculture and urban/residential settings compared to canals; however, one third of
urban canals scored above the average of agricultural streams. This multidisciplinary study shows that irrigation
canals can be valuable riparian and aquatic habitat, especially in regions with severely degraded streams.
© 2018 Published by Elsevier B.V.

⁎ Corresponding author at: Department of Biology and Environmental Sciences, Marietta College, Marietta, OH 45750, United States of America.
E-mail address: eac005@marietta.edu (E.A. Carlson).

https://doi.org/10.1016/j.scitotenv.2018.07.246
0048-9697/© 2018 Published by Elsevier B.V.
 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
1. Introduction
Streamflow reductions for hydropower generation and municipal
and agricultural use affect rivers of all sizes across the Earth
(Graf, 1999; Kingsford, 2000). Production agriculture in arid and
semi-arid regions requires water to be diverted from streams or
pumped from the ground to irrigate crops (Khan et al., 2006). The ef-
fects of dewatering streams on aquatic and riparian biota have been
relatively well studied (Bunn and Arthington, 2002; Malmqvist and
Rundle, 2002; Poff and Zimmerman, 2010), however in many regions
canals specifically designed to transport water to agricultural land sup-
port little known aquatic and riparian ecosystems (Patten, 1998). Some
researchers and managers have suggested that some canals may sup-
port species-rich plant and aquatic invertebrate communities like
those of natural streams (Chester and Robson, 2013), but the character-
istics of these communities is not well known.
Riparian ecosystems are highly productive and biologically diverse
relative to surrounding uplands in most regions (Naiman et al., 1993).
Floodplains and riparian zones function to store, transform and cycle
nutrients, organic matter, sediment and water (Jacobs et al., 2007;
Tank et al., 2010). Hydrologic and geomorphic processes shape the
physical landscape, and the disturbance regime controls the coloniza-
tion and persistence of organisms (Shafroth et al., 2002; Katz et al.,
2009) and laterally connect terrestrial and aquatic ecosystems at local
scales and longitudinally connect river networks at landscape scales
(Harvey and Gooseff, 2015).
Streams and riparian areas throughout the world have been directly
modified by floodplain development and channelization and indirectly
through watershed modification, stream flow alteration, and land use
changes (Patten, 1998; Nilsson and Svedmark, 2002; Bunn and
Arthington, 2002). Human activities have altered the physical structure
of rivers and their floodplains through dam and levee construction,
canal diversion structures, channel straightening and bank stabilization.
Impoundments have changed the historic sediment regime and re-
duced flood intensity, frequency and duration, which have altered geo-
morphic processes and the structure, composition and functioning of
riparian plant communities (Merritt and Cooper, 2000). For instance,
Populus spp. dominated riparian forests rely on natural flood distur-
bance for health and regeneration with reduced growth, dieback and
mortality under altered flow regimes or water table depths (Cooper
et al., 1999; Williams and Cooper, 2005). Decades of diversion from nat-
ural rivers has decreased aquatic insect and fish populations (Miller
et al., 2007; Kingsford, 2000), homogenized streamflow, habitats and bi-
otic communities (Bunn and Arthington, 2002; Leigh and Sheldon,
2008), and in extreme cases created rivers that resemble artificial chan-
nels with engineered bed and banks (Urban and Rhodes, 2003; Gregory,
2006). The natural hydrologic processes that create and maintain a di-
versity of aquatic and riparian habitats along the World's streams and
rivers have been profoundly reduced by human actions to guard against
floods, develop floodplains for agriculture and divert water for irrigation
(Malmqvist and Rundle, 2002; Miller et al., 2007; Poff and Zimmerman,
2010).
Human activities completely shape the physical structure and flow
characteristics of canals that influence their riparian and aquatic ecosys-
tems. The metrics commonly used to characterize stream geomorphic
processes including overbank flooding, sinuosity, and channel migra-
tion, are rarely applicable for highly engineered water conveyance ca-
nals that are designed to minimize turbulent flow and maximize
conveyance. In addition, the lack of hydrologic variability and spatial
heterogeneity limits fluvial landform creation (Swamee, 1995). Canal
management, including removal of sediment and woody debris, weed
suppression and controlled burns, may simplify and homogenize ripar-
ian and aquatic ecosystems (Andersen and Nelson, 1997; Meany et al.,
2003; Casas et al., 2011).
Canals are largely decoupled from the surrounding landscape by
constructed berms that limit water, sediment and organic matter
771
exchange (Fig. 1). However, because streams are the source of water
for canals, they create connectivity allowing riparian plant and aquatic
macroinvertebrate propagule introductions (Ernegger et al., 1998;
Koetsier and McCauley, 2015). Canals have long been recognized as sus-
ceptible to the invasion of non-native species through hydrochory
(Egginton and Robbins, 1920), as well as facilitating the introduction
of non-native, weedy species to natural streams thorough agricultural
return flows.
Aquatic macroinvertebrates are a key component of aquatic, riparian
and some upland food webs (Polis et al., 1997; Nakano et al., 1999; Leigh
et al., 2013) and are highly sensitive to flow variation (Nakano and
Murakami, 2001; Miller et al., 2007; Muehlbauer et al., 2014). Their pop-
ulations are influenced by channel physical characteristics, food type
and availability, and hydrologic patterns at relatively small scales
(Williams and Feltmate, 1992). Aquatic macroinvertebrates often use
seasonal and hydrologic cues for their life history development
(Butler, 1984). Flow variability in canals can include rapid drawdowns
and multiple high and low flow events per year creating unnatural
timing that can affect aquatic macroinvertebrate growth and reproduc-
tion (Kennedy et al., 2016). Drainage ditches in agricultural areas are
populated by biota with life history adaptations to highly variable flow
regimes including short life cycles, desiccation resistant life stages and
winged adults (Whatley et al., 2015). Seasonal and random disturbance
events combined with the physical setting can influence colonization
and competition and the formation of communities (Petraitis et al.,
1989; Hobbs and Huenneke, 1992). The presence of aquatic plants in ar-
tificial temporary channels and wetlands further can support regional
diversity local food webs in agricultural watersheds (Bolpagni and
Piotti, 2016).
In natural watersheds, streams are organized from small headwa-
ter channels to larger main stem rivers, while canal systems are the
opposite, with the largest canals fed from rivers, and branching in
dendritic patterns downgradient to reach the complex network of
smaller canals that deliver water to fields. Canals add extensive com-
plexity and channel length to stream networks and significantly
change watershed scale riparian habitat from its pre-settlement con-
dition. Although canals have been critical to the foundation of human
societies in many regions of the world, research on their biological
characteristics has been limited to a few studies in Chile, Spain,
France, and the United States (Habit et al., 1998; Fernald et al., 2007;
Lopez-Pomares et al., 2015; Aspe and Jacque, 2015). The prevalence of
canals can be easily overlooked, as their vegetation may resemble natu-
ral streams (Fig. 2).
Colorado has a 150-year history of irrigation that has facilitated agri-
cultural production and provides support for human population growth
(Evans and Evans, 1991). Stream water is diverted into canals and ap-
plied directly to crops or stored in off-channel reservoirs to extend the
agricultural irrigation season and supply year-round water to cities.
Water diversions from the Cache la Poudre River (a tributary to the
South Platte River in the Missouri River basin) have occurred since
1864 (Strange et al., 1999) with 85% of the mean annual flow being re-
moved from the river for agricultural use. Thousands of kilometers of ca-
nals have been constructed to deliver water and most remain in use. The
ongoing conversion of agricultural land to residential and industrial
uses continues to alter the timing and amount of water flowing in canals
and the management of vegetation along stream banks. Maintenance
along canals varies from no activities to burning, herbicide spraying
and mechanical removal of sediment and vegetation. The range of
canal sizes and flows, along with varied maintenance, supports a wide
variety of riparian vegetation and channel characteristics through the
labyrinth of irrigation canals.
The contribution of canal riparian vegetation and aquatic inverte-
brates to local and regional biodiversity is poorly documented and
understood (Paul and Meyer, 2001; Herzon and Helenius, 2008;
Vermonden et al., 2009; Verdonschot et al., 2011). In this paper, we ad-
dress the following questions: (1) do canals support riparian vegetation
772 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781

Fig. 1. Cross section of typical stream and canal with two geomorphic surfaces (bank and top/floodplain).

and aquatic macroinvertebrate communities similar to natural streams 2. Methods

of similar size, (2) are functional groups of plants and aquatic macroin-
vertebrates similar between canals and streams, (3) is there agreement
between species and functional group data, and (4) do riparian and
aquatic indices support species and functional group patterns between
canals and streams? Canals could support riparian and aquatic ecosys-
tems comparable to natural streams as plants and animals have colo-
nized this new water distribution system.
2.1. Study area and data collection
The study was conducted in north-central Colorado (Fig. 3) a semi-
arid region with extensive irrigated agriculture. Precipitation averages
250 mm during the summer and 135 mm during winter (www.
usclimatedata.com). Total annual precipitation is insufficient to support

Fig. 2. Example streams and canals in agricultural and residential landscapes: A) agricultural stream, Willow Creek, Weld County, Colorado B) agricultural canal, Fort Collins, Larimer
County, Colorado. C) residential/urban canal, Fort Collins, Larimer County, Colorado. D) agricultural stream, Lone Tree Creek, Weld County, Colorado.
 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
desirable crops such as corn, beans, and vegetables. Water for irrigation
is diverted from rivers fed by melting snow in the Rocky Mountains and
applied to crops from April through September.
The study area includes irrigation ditch companies in the South
Platte River Basin Water District 3 that divert water from the Cache la
Poudre River primarily through earthen canals. The canal network in-
cludes 2014 km of canals, providing irrigation water to 54,529 ha of
crops within the study area (CDSS). Canals ranged in size from b1 to
18 m wide and up to 3 m deep. Vegetation management along canals
varies from frequent and intense where crops were being grown, to
minimal in urban and residential areas. We separated sites as agricul-
tural (Ag.) or residential (Res.) using the dominant land use within
100 m of each canal segment. The western region (Larimer County) in-
cluded more residential area while the eastern region (Weld County)
was primarily agricultural. Flow of the Cache la Poudre River is altered
by numerous low-head dams that divert water into irrigation canals,
and by the addition of water from trans-basin diversions (Evans and
Evans, 1991). The floodplain has also been constrained by gravel min-
ing, roads and agriculture (Strange et al., 1999; Wohl, 2001).
2.2. Site selection and layout
To quantify vegetation cover and structure along canals, we strati-
fied the vegetation into five categories: heavy tree cover, light tree
cover, shrub cover, herbaceous cover, and concrete-lined, using sub-
meter resolution aerial and satellite imagery in ArcGIS v.10. Heavy
tree was defined as N50% cover of woody plants over 3 m tall visually es-
timated. Light tree had 25–50% tree cover. Shrub sites had woody vege-
tation b3 m in height covering N50% of the bank vegetation. All
Fig. 3. Study area in north-central Colorado.
773
remaining sites were classified as herbaceous unless the bank and chan-
nel bottom were concrete-lined.
Each canal site was selected from a randomly generated set of points
distributed proportionally across the canal network using the total
length of canals in each canopy cover class to determine the frequency
of sampling. Stream sites were selected to minimize hydrologic modifi-
cations within the reach and maximize undeveloped floodplain width
to represent the least impacted natural stream reaches and are consid-
ered reference sites. Vegetation was sampled at 47 canals and seven
streams during the summer of 2013. Each site had three transect lengths
oriented perpendicular to the channel evenly spaced five bank full chan-
nel widths apart. Aquatic macroinvertebrates were sampled every three
weeks from May through August in 2015 for a maximum of six samples
per site for 20 sites representing all substrate types, channel sizes, and
flow regimes.
2.3. Environmental and hydrologic variables
Channel and floodplain physical characteristics, land use, network,
and hydrologic variables were selected and modified from existing ri-
parian and stream assessment methods (Gonzalez del Tanago and
Garcia de Jalon, 2011) (details in Appendix A). Stream flow data were
obtained from two USGS gages on the Cache la Poudre River for the pe-
riod 1999–2015 (Fort Collins #06752260, and Greeley #06752500).
Streamflow records for Spring Creek were only available for
2013–2015 from the City of Fort Collins flood warning system. Owl
Creek and Willow Creek with additional stream sites lack stream flow
records. Flow in irrigation canals was recorded at the point of diversion
from the Cache la Poudre River. Streams and canals were separated by
774 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
bank full width into large (N4 m) and small (b4 m) for some hydrologic
comparisons. Flow depth was measured and compared to bankfull
depth as a ratio (observed/bankfull).
2.4. Riparian vegetation
Canopy cover of each species was visually estimated for each plot.
Five plots were sampled on the top/floodplain and bank surfaces.
These were oriented perpendicular to a transect that spanned both
sides of the canal or stream. This sampling scheme of 20 plots per tran-
sect was repeated 3 times at each site for a total of 60 plots per site.
Cover was estimated using modified Braun-Blanquet cover classes
(trace, b1%, 1–2%, 2–5%, 5–10%, 10–25%, 25–50%, 50–75%, 75–95%,
N95%) (Bonham, 2013) and averaged for each site to capture site level
vegetation characteristics. Plant species nomenclature follows Weber
and Wittman (2012). Height of species was placed into classes (b1 m,
1–2 m, 2–5 m, 5–10 m, N10 m) (Merritt and Bateman, 2012). Bare
sediment was analyzed using the same cover classes as plant species.
Diversity metrics including total richness, native species richness,
Shannon-Wiener (log e) and Simpson's index (1-λ) were calculated
using all species. Species present in fewer than 5% of sites were removed
(McCune and Grace, 2002) reducing the number of plant species used in
nMDS and cluster analysis from 251 to 126.
We developed a habitat quality index (HQI) that adds quantitative
metrics for site level hydrologic, geomorphic and vegetation character-
istics (detailed in the Appendix A) to the riparian quality index (RQI) of
Gonzalez del Tanago and Garcia de Jalon (2011). Functional redundancy
(FR) (Bruno et al., 2016) was used to identify diversity in functional
groups and is used as a component of an ecosystem's resilience to veg-
etation disturbance. FR was calculated as species richness divided by
functional group richness using the full plant species list.
Plant species were classified into 22 functional groups (Table B.2.). A
priori groups were used because little is known about plant species re-
sponses to fluvial disturbance including inundation, burial and physical
damage for many of the observed species. Groups were created with
three ecological characteristics: origin (native or introduced), growth
form (graminoid, forb, shrub, tree (USDA Plants), and National Wetland
Inventory wetland indicator status (Lichvar, 2012). Cacti and vines were
a small proportion of total species and included in the forb category.
Three wetland indicator groups were used; wetland (obligate and facul-
tative wetland), facultative (facultative, facultative upland) and upland.
Similar attributes have been suggested for use in riparian plant guild
classification (Merritt et al., 2010). Species were replaced by functional
groups and cover summed to create a guild dataset. The guilds were
treated as pseudo species for statistical analysis.
2.5. Aquatic macroinvertebrates
Aquatic macroinvertebrates were collected at each study site every
three weeks from May–August 2015 using a standard D-frame kick
(Merritt et al., 2008) net with a 500-μm capture net. One kicknet sample
was collected from each site on each sampling date by disturbing the
top 5 cm of sediment and submerged vegetation for three minutes. A
composite kicknet of micro-habitats including riffles, pools, banks, sub-
merged and aquatic vegetation were collected as one sample. A sweep-
ing motion under the water using the D-frame net was used to simulate
flow in stagnant pools. Presence of submerged vegetation and aquatic
macrophytes was recorded.
All aquatic macroinvertebrates were removed from samples and
preserved in 80% ethanol. Volume based subsampling was used in five
samples few cases where more than several hundred individuals oc-
curred in a sample (Hickley, 1975). Individuals in most groups were
identified to genus, worms and leaches to family. A total of 97 samples
were collected, five of which contained zero individuals. The number
of samples collected at each site varied between three and seven be-
cause some canals were dry during collection dates and others were
flowing too swiftly for safe access. To account for this difference, abun-
dance counts of taxa were averaged for each site.
We calculated the traditional EPT index (Ephemeroptera [mayflies],
Plecoptera [stoneflies], and Trichoptera [caddisflies] for taxa richness
(EPTr) as well as the proportion of EPT taxa using abundance (EPTa)
(Kerans and Karr, 1994; Wallace et al., 1996; Rosenberg et al., 2008).
The EPTr index was calculated for each site as the number of EPT taxa di-
vided by the total number of taxa present, the EPTa index replaced rich-
ness with abundance in the proportional calculation. Physiological and
ecological traits of genera and families were used to create the func-
tional aquatic dataset (Poff et al., 2006). Counts of individuals from
taxa within a functional group were added together to produce an
abundance for the functional group.
2.6. Statistical analyses
Community analyses for vegetation and aquatic macroinvertebrates
were performed using Primer v.7 software (Clarke et al., 2014). Non-
metric Multi-Dimensional Scaling (nMDS) was used to analyze the
overall structure of communities, and principal components analysis
for physical characterization of riparian and aquatic habitats. For all sta-
tistical tests an alpha b0.05 was used to indicate a significant result.
Diversity metrics, stream biotic and habitat indices were tested for dif-
ferences between all groups using ANOVA and between channels con-
trolling for land use, for example canal vs. stream, using Student's t-test.
2.6.1. Environmental variables
Each environmental variable was normalized by subtracting the
mean and dividing by the standard deviation (Clarke et al., 2014).
Euclidean distance was used to calculate a similarity matrix for environ-
mental variables. Environmental variables were related to both species
and functional group datasets for vegetation and aquatic macroinverte-
brates using distance-based linear modeling with the AICc selection
criteria and by Spearman Rank correlation of environmental and biotic
similarity matrices (Anderson et al., 2008).
2.6.2. Ecological variables
Abundance data for plants and macroinvertebrates were square-
root transformed to down weight the few abundant taxa and included
the predictive value of infrequent species, and a Bray-Curtis similarity
matrix was created (Clarke et al., 2014). An nMDS was calculated to vi-
sualize the data and multivariate dispersion was tested using the
PermDISP routine in Primer with distances measured to the centroid
of the group. Permutational multivariate analysis of variance
(PerMANOVA) was used to test the effects of land use, channel type,
and their interaction on vegetation and aquatic macroinvertebrate com-
position. Due to the unbalanced sampling design, Type III sum of
squares and unrestricted permutation of the raw data were used in
the calculation with 999 permutations. Comparisons that resulted in
fewer than 100 permutations were reported with Monte Carlo
corrected p-values (Anderson et al., 2008). These statistical analyses
were performed on the species/taxonomic and functional datasets for
vegetation and macroinvertebrates.
3. Results
Forty-seven canal sites and seven stream sites were included in this
study with a range of physical attributes (see Appendix B). Land use dif-
fered markedly between two regions in the study area. The western re-
gion is 33.1% agricultural and 29% residential/urban. The eastern region
is 70% agricultural and only 1.6% residential/urban. The length of open
irrigation canals in the more densely populated western region was
370 km compared to 250 km of streams, for a canal/stream length
ratio of 1.5 even as portions of canals are directed into pipes under-
ground in urban areas. Open channels were more common in eastern
 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
agricultural areas with 1644 km of canals to 487 km of streams for a
canal/stream length ratio of 3.4.
3.1. Hydrology
Flow varied by two orders of magnitude across channel types. Re-
cords from stream gages indicate that precipitation driven peak flows
were most prevalent on small streams while snowmelt runoff was
most prevalent in the large stream (Table 1). Small channels had two
patterns: 1) a short early summer peak and low late summer flows,
and 2) a consistent flow level through the entire summer (Fig. 4). Ob-
served flow depths indicated that canals in residential areas were
flowing close to bankfull conditions (higher Ratio) in June and decreas-
ing over the summer, while agricultural canals had the opposite trend,
with the most flow occurring in August as crop irrigation demand peaks.
Flow variability and short-term changes assessed as daily coefficient
of variation (CV) (Table 1) indicated that most streams had greater var-
iability than canals. The exception was Dry Creek whose watershed has
several storm water retention structures. The date of peak flow was
22 days later for the large canal and 51 days earlier for small canals
than similar sized streams. The large canal and large stream had a sim-
ilar number of low flow events and days, while small canals had more
than small streams. The average number of days with flow from April
1–September 30 (183 days) was 174 (SE = 3.1) for the large canal,
47.8 (6.1) for small canals and 183 (0) for streams, all of which were
perennial.
3.2. Environmental characteristics
Canals varied from 1 to 22 m wide and 0.4 to 2.5 m deep at bank full
flow. Streams had a similar range, 1.5 to 30 m wide and 0.6 to 2.5 m
deep at bank full flow. Mean water temperature in canals in agricultural
areas was 5–6 °C warmer than residential streams and canals, and agri-
cultural streams were also 5–6 °C warmer than residential streams.
Temp in canals was more positively correlated with month (r = 0.47,
p b 0.001) than distance to the point of diversion (r = 0.11, p = 0.28)
and was weakly negatively correlated to percent woody canopy
(r = −0.15, p = 0.14).
3.3. Community composition: plants
A total of 3247 plots were analyzed for vegetation composition and
cover at 54 sites (see Appendix B for complete list). Forbs comprised
50% of species present, graminoids 32%, shrubs and trees 14%, and
vines and cacti 4%. Native species provided 51% of the total richness
but contributed only 36% of total plant cover. Two exotic species,
Bromopsis inermis (Leysser) and Phalaris arundinacea (L.), and the native
Carex emoryi (Dewey) dominated canal and stream riparian vegetation
in both the urban and agricultural regions, accounting for 49% of total
cover. Plant diversity metrics and vegetation indices were not signifi-
cantly different between canals and streams (Table 2). Composition, in-
cluding cover had mixed results with significant differences between
canals in both land uses and between streams and canals in the
Table 1
Hydrologic metrics used to characterize flow regimes of canals and streams calculated from April 1 to Sept 30 including coefficient of variation (CV), date of peak flow (in Julian days),
number of high flow events (flows exceeding 90%), number of low flow events (flows b10%), number of days below 10%, and number of days with zero flow. Detailed definitions are
in the Appendix A. 1 located in Larimer County, Colorado; 2 located in Weld County, Colorado.
Cache la Poudre River Cache la Poudre River
(USGS gage 6752260)1 (USGS gage 6752500)2
CV 1.77 1.6
Date of peak 150 155
# High flow events 1.6 1.3
# Low flow events 4.3 3.3
# Low flow days 18 18
Zero flow days 0 0
775
agricultural landscape, however, streams were not significantly differ-
ent between land uses and in the residential areas, streams and canals
were on border of significance (t = 1.344, p = 0.05). Stream and
canal vegetation did not separate into distinct groups in non-metric
multidimensional scaling (Fig. 5). Canopy type influenced some mixing
of groups, for example concrete lined canals, a recently reconstructed
canal and a naturally intermittent stream were grouped in the lower
left of the ordination space with significant bare sediment in Panel A.
The same intermittent stream had a high proportion of upland grasses
drawing the point to the upper left of Panel B. A trend of sites with
woody species (upper half) to those dominated by grasses and forbs
(lower half) creates additional mixing between groups. The percent
cover of woody vegetation (trees and shrubs) was highest for residen-
tial sites while grasses were more common in the agricultural areas
(Fig. 6). The stream sites on the far right have the most diverse assembly
of functional groups including grasses, forbs, shrubs and trees not
matched by any canal sites.
Vegetation composition of canals and streams were different across
both land-use categories (F = 2.018, p = 0.017, df = 53) with compa-
rable variance (F b 0.01, p = 0.991, df = 53). Land-use influenced
species composition with a significant difference (F = 4.248, p =
0.001, df = 53) between riparian vegetation in agricultural and residen-
tial areas. Pairwise PerMANOVA using the two by two matrix of channel
type and land-use indicated that 3 of 4 comparisons had significantly
different species composition (Table 3a) with streams similar in urban
and agricultural areas.
3.4. Community composition: aquatic macroinvertebrates
Aquatic macroinvertebrates were relatively diverse across the study
area with 147 typical regional taxa identified (Table B.3). Thirty-one
taxa were found only in streams including seven EPT taxa that suggest
good water quality such as Seratella micheneri (Traver), and Pteronarcella
badia (Hagen). Thirty-one taxa were collected only from canals and in-
cluded mayfly in the genera Drunella, Ameletus and Rhithrogena. The
most diverse orders found were Diptera with 64 taxa represented, Cole-
optera with 20, and Ephemeroptera with 18. Mean site level richness
was highest for residential streams (37 ± 11.3), followed by agricultural
streams (28 ± 4.2), residential canals (20.1 ± 11.3), and agricultural ca-
nals (18.9 ± 3.8). Many taxa were infrequently collected, and 25% were
collected only once. The variance was similar between sites measured as
multivariate dispersion, (F = 1.59 p = 0.833, df1 = 3, df2 = 16). There
was overlap between site groups in Fig. 7 for both functional and species
datasets with PerMANOVA pairwise tests confirming similarity between
comparisons in Table 3b.
3.5. Functional community composition: plants
A permutational distance-based test for homogeneity of multivari-
ate dispersion (PERMDisp) indicated that functional community
composition of riparian vegetation was similar between channel types
(F = 0.09, p = 0.832) and land-use categories (F = 0.007, p = 0.944).
The number of functional plant groups observed at canal and stream
Spring Larimer and Weld New Mercer Larimer #2
Creek1 Canal1 Canal1 Canal1
0.89 1 0.81 0.79
186 174 135 128
5 3 1.6 2.3
2 3.7 3 3.3
17.6 30 13.3 34
0 12.3 2.6 10.3
776 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781

Fig. 4. Mean daily streamflow (cms) averaged for three irrigation seasons (April-Sept) 2013–2015 in one small stream (Spring Creek), and three small canals. All canal data are from gages
at the point of diversion from the Cache la Poudre River, Larimer County, Colorado.

sites averaged 9.8 and was statistically similar between channel
types (t = 1.68, p = 0.361). The number of functional plant groups
at each site did not change after removing rare species. Introduced
upland forb was the most common functional group and occurred
at 98% of sites, and introduced wetland grasses comprised 24% of
all vegetation cover, the most of any functional plant group. Forbs
were a relatively small proportion of total cover (18.5%), although
they accounted for the highest diversity of any group with 142 spe-
cies, or 60% of all species in the study area.
3.7. Comparison of taxonomic and function data sets
Comparisons of species and functional group composition for ripar-
ian vegetation similarity showed agreement when holding channel type
constant (Table 3a). The two datasets disagreed when holding land-use
constant and comparing channel types. Aquatic macroinvertebrate tax-
onomic and functional group data agreed for all comparisons (Table 3b).
The similarities in vegetation and macroinvertebrates (Table 3c) indi-
cated more agreement between the two habitats when using functional
groups.

3.6. Functional community composition: aquatic macroinvertebrates

Sixty-eight aquatic macroinvertebrate taxa (Table B.4), accounting
for 71% of collected individuals, were classified by functional traits
into six groups by kR means cluster analysis (R = 0.94), results in
Table 4. The number of functional macroinvertebrates groups was sig-
nificantly lower (t = −2.95, p = 0.006) for canals with an average of
4.1 of a possible six groups compared to streams with 5.25 observed.
Group A was commonly absent from canals collected at only 25% of
sites. All sites had some members of Group D, but this was the only func-
tional group represent at all canals and stream sites. The abundance of
each functional group indicated a similarly between canals and streams
(F = 1.54, p = 0.2, df = 19). Multivariate dispersion was similar across
sites (F = 2.5776, p = 0.225, df1 = 3, df2 = 16).
3.8. Environmental conditions and landscape and biotic indices
The average Habitat Quality Index (HQI) score for canals was 2.9
while for streams it was 4.4, a significant difference (t = −2.93, p =
0.01, df = 7). Channels in the agricultural landscape had a lower aver-
age score of 2.7 compared to residential channels, 3.5 on average, also
significantly different (t = −2.22, p = 0.02, df = 42). Residential canals
and agricultural streams were similar (t = 1.00, p = 0.19, df = 4). Other
environmental variables including water temperature and estimated
flow were poor predictors of species and functional group composition
for riparian vegetation.
The EPTr index was similar between site groups but generally higher
in residential areas (Table 2). The proportional abundance of EPT taxa,

Table 2
Summary riparian vegetation values for diversity metrics and indices of wetland prevalence, conservative species cover, native richness and percent cover. Aquatic macroinvertebrate taxa
richness (at the level of genus), mean total abundance, mean EPT richness (EPTr) and mean proportion of abundance of EPT (EPTa) taxa groups. ANOVA results of comparison of four land-
scape X channel type groups (F-statistic and p-value) reported for each metric.

ANOVA F-statistic Riparian vegetation Aquatic macroinvertebrates

(p-value)
Species Native Shannon Simpson's Wetland Cover weighted Native Taxa Mean EPTa EPTr
richness richness diversity (1-λ) PI mean C cover % richness abundance

0.38 (0.77) 0.94 (0.43) 0.34 (0.80) 0.49 (0.69) 0.29 1.12 (0.35) 0.85 (0.47) 3.92 4.85 0.62 0.30
(0.83) (0.03) (0.01) (0.61) (0.83)

CANALS 22.2 10.9 1.6 0.7 2.87 1.61 37.3 19.4 78.8 19.7 19
Residential 23.5 12.8 1.61 0.71 2.90 1.63 37.7 20.1 46.5 22.5 22.4
Agricultural 21.3 9.6 1.59 0.69 2.84 1.59 37 18.9 104.1 17.5 16.4
STREAMS 23.8 14 1.44 0.62 3.21 1.11 26.1 32.5 170 39.7 19.7
Residential 26.3 14.5 1.56 0.61 3.05 1.88 39.5 37 181.8 39.4 22.8
Agricultural 22 13.3 1.35 0.63 3.33 0.53 16 28 158.4 40.1 16.7
 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
Fig. 5. nMDS of vegetation averaged by site (n = 54). Panel A – species level data with
vectors for selected species with a Pearson r N 0.5, 2-D stress was 0.22. Panel B – functional
group data with vectors for functional groups with Pearson r N 0.5, (IUG = introduced up-
land grass, IUF = introduced upland forb, IWG = introduced wetland grass, NMW = na-
tive mesic woody, IMG = introduced mesic grass), 2-D stress was 0.21.
however, was on average more than double for streams (Table 2). Chan-
nel size was a major contributor to EPT for streams where the large
Cache la Poudre River sites had EPTr and EPTa of approximately 32 and
80, while smaller streams were 13 and 3. The mayfly Tricorythodes
explicatus (Eaton) is relatively tolerant of poor water quality and higher
water temperatures (Relyea et al., 2000) and comprised 45% of the EPT
abundance in agricultural streams but was not common in agricultural
canals. Residential canals and streams shared many EPT taxa including
the mayflies Baetis, Ephemerella, and Heptagenia. The HQI was highly
correlated with the EPTa (r = 0.69, t = 4.05, p b 0.001) and moderately
with the EPTr indices (r = 0.47, t = 2.28, p = 0.04). Pearson correlation
for distance along canal network from the point of diversion from the
river (Dist) to EPTr and EPTa was significantly negative while bankfull
width was positive.
Stream vegetation was not hydrophytic according to the wetland
prevalence index as scores were N3.05, while canal vegetation on aver-
age did have predominantly wetland plants with scores b2.90 (Table 2).
Cover weighted mean C was between 1.59 and 1.63 for canals
surrounded by different land use while streams had a much larger
range between 0.53 for agricultural streams and 1.88 for residential
streams (Table 2).
4. Discussion
Irrigation canals are artificial waterways that are simple in physical
form with strict flow schedules determined by the timing of crop
777
irrigation and other calls for water by downstream users. The length
and prominence of these man-made aquatic and riparian ecosystems
in arid and semi-arid regions of the world highlights their importance.
We found that canals supported similar riparian plant and aquatic mac-
roinvertebrate species and functional groups as natural streams in our
study region. Riparian vegetation was influenced by surrounding land-
uses while aquatic macroinvertebrate communities were similar be-
tween and within land-use categories. Agricultural canals had similar
macroinvertebrate taxa, functional group composition, and plant func-
tional group composition, but different plant species composition
when compared to agricultural streams. Residential channels had simi-
lar macroinvertebrate communities yet distinct plant species and func-
tional groups. Habitat quality index (HQI) scores were lower for canals
than natural streams, yet residential canals had only slightly lower
scores than agricultural streams suggesting that irrigation canals could
be important ecosystem elements (albeit functioning at a lower level
than natural aquatic and riparian ecosystems) in human-dominated
landscapes.
Previous studies of the ecological attributes of constructed water-
ways have focused on drainage ditches in agricultural regions (Herzon
and Helenius, 2008; Leslie and Lamp, 2017) and transportation canals
(Harvolk et al., 2014). These artificial waterways have been shown to
support diverse biotic communities (Chester and Robson, 2013). Fish
utilizing irrigation canals were larger, had access to more and diverse
prey items while maintaining synchronous reproduction like fish in riv-
ers (Habit et al., 2005). In many regions artificial channels and wetlands
are temporary habitats but maintain biodiversity and ecological func-
tioning in agricultural landscapes (Bolpagni and Piotti, 2016).
Processes associated with natural streams in the study area such as
overbank flooding, vertical and lateral erosion and sediment deposition
rarely, or do not occur at all under normal canal construction, operation,
and maintenance (Depeweg and Mendez, 2002). The design and con-
struction of irrigation canals limits groundwater supported base flows
and surface water inputs characteristic of a stream's interaction with
its watershed, although inadvertent water leakage from canals can sup-
port extensive wetlands (Sueltenfuss et al., 2013).
Daily streamflow had distinct inter- and intra-annual variability
with peak flows driven by snowmelt and rain events. Peak flow in
small canals occurred earlier than in small streams and the opposite
was true for larger streams. A striking difference was that during the
183-day April–September study period canals averaged only 147 days,
80%, with measurable flows, while all study streams had flow on all
183 days.
4.1. Riparian vegetation and aquatic macroinvertebrate communities
In constructed channels, such as canals, stream flow patterns and
variance can be analogous to natural streams, yet mechanical channel
maintenance and vegetation management can limit the formation of
aquatic habitats and in-channel food resources. The hydrologic and geo-
morphic differences between streams and canals measured in this study
produced no differences in species richness, Shannon diversity, or
Simpson's index for riparian vegetation between natural and con-
structed channels. Aquatic macroinvertebrate communities in agricul-
tural and residential landscapes were also similar to streams. Habit
et al. (1998) found similar species and biomass in irrigation canals and
streams in Chile, as did Koetsier and McCauley (2015) in Idaho, U.S.A.
Other man-made channels have been reported to support diverse
aquatic and riparian communities along agricultural drainage ditches
in temperate regions of the northern hemisphere (Herzon and
Helenius, 2008; Verdonschot et al., 2011).
4.2. Functional groups of riparian plants and aquatic macroinvertebrates
Plant functional groups or guilds have been used to characterize en-
vironmental filters in riparian ecosystems, and wildlife habitat (Merritt
778 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781

Fig. 6. Box plots of average percent cover of major plant growth forms for each channel type. Endpoints indicate maximum and minimum values, the box indicates 25th and 75th
percentiles with the median as the center line.

et al., 2010; Merritt and Bateman, 2012; Bruno et al., 2016). Aquatic between canals and streams. Decades of flow alteration in streams
macroinvertebrates can be grouped using physiological traits, life his- and intermittent flow patterns in canals could have reduced the re-
tory, and physical habitat requirements and are informative about the gional pool of taxa through environmental and trait based filters leading
type and condition of aquatic habitats (Poff et al., 2006). For plants to a higher proportion of generalist taxa that are tolerant of or able to
and aquatic invertebrate biota, functional groups indicated similarities avoid difficult conditions through adult colonization, rapid maturity,
and desiccation resistance (Mackay, 1992; Miller and Golladay, 1996;
Bogan and Boersma, 2012; Whatley et al., 2015).
Table 3 Functional plant attributes differed between residential and agricul-
Selected pairwise comparisons riparian vegetation (A), aquatic macroinvertebrates
tural landscapes. Fewer sites were characterized by woody plants along
(B) composition of landuse (Ag. – agricultural, Res. – residential) and channel types using
PERMANOVA. The agreement of significant differences found between streams and canals agricultural streams and canals compared to those in residential areas.
using the vegetation and aquatic macroinvertebrates datasets using a threshold p-value of This is likely due to vegetation management, particularly vegetation
0.05 (C). t-statistic (t-stat) and p-values (p-val) reported. 1 indicates Monte Carlo clearing, burning and herbicide application as well as a potentially
corrected p-value. more variable and erratic stream flows along canals and limited protec-
A) Species dataset Functional dataset tion of streams in agricultural landscapes. Vertical structure and species
T-stat p-val T-stat p-val
diversity and wetland plant cover were also significantly lower along
1
agricultural streams and canals. This has implications for wildlife that
Ag. stream vs. res. stream 0.948 0.449 1.118 0.2571
require structurally diverse riparian zones to provide cover and food
Ag. canal vs. Res. canal 2.052 0.001 1.768 0.008
Ag. canal vs. Ag. stream 1.083 0.027 1.008 0.392 (Meany et al., 2003; Lopez-Pomares et al., 2015).
Res. canal vs. Res. stream 1.344 0.05 1.752 0.013

B) Taxonomic dataset Functional dataset 4.3. Comparison of taxonomic and function data sets
T-stat p-val T-stat p-val
Comparisons of species composition among sites are commonly
Ag. stream vs. Res. stream 0.78 0.6571 0.68 0.7121 used to identify differences in biological communities (Petchey and
Ag. canal vs. Res. canal 1.35 0.025 1.38 0.029
Gaston, 2002). Grouping species into physiologically and ecologically
Ag. canal vs. Ag. stream 0.86 0.6441 0.70 0.8571
Res. canal vs. Res. stream 1.40 0.0981 1.41 0.1131 similar groups can add information about physical structure, food re-
sources and adaptations of communities to disturbance that are valu-
C) Species or taxonomic dataset Functional dataset able in understanding current or potential ecological functioning
Ag. stream vs. Res. stream Agree Agree (Merritt and Bateman, 2012; Petchey and Gaston, 2002).
Ag. canal vs. Res. canal Agree Agree We found statistical similarity between riparian plant species and
Ag. canal vs. Ag. stream Disagree Agree functional group composition for some comparisons of channel type/
Res. canal vs. Res. stream Disagree Disagree
land-use categories. A disagreement between the two datasets would
 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
Fig. 7. nMDS of macroinvertebrate samples (n = 97). Panel A – taxa level data with vectors
of selected taxa showing correlation to points, 2-D stress was 0.24. Panel B – functional
group data, 2-D stress was 0.15.
indicate that a group of species provide similar ecological value or re-
spond to similar habitat characteristics, thus several species could be
present to provide the ecological functions. The similarity of canals
and streams in residential areas was indicated by an inconclusive statis-
tical result. Aquatic macroinvertebrate taxonomic and functional group
composition were similar across all streams and canals.
4.4. Environmental conditions and landscape and biotic indices
The HQI developed for this study was well correlated to EPTr and
EPTa indices as the physical and chemical characteristics of aquatic eco-
systems such as water temperature, food resources, and in-channel
woody debris are partly dependent on inputs from adjacent riparian
Table 4
Description of functional macroinvertebrate groups determine through cluster analysis of physiological, ecological and life history traits from Poff et al. (2006).
Functional group Description of key characters
A Desiccation resistant predators/collector gatherers
B Large predators with long lived, winged adults
C Collector gatherers of erosional habitats with synchronized life histories
D Small, tolerant collector gatherers of depositional habitats
E Aquatic adults, strong swimmers, predators/herbivores
F Free moving collector gatherers with synchronized life histories
779
ecosystems (Pozo et al., 1997; Moore et al., 2005). These can be even
more important in agricultural landscapes where riparian buffers are
being increasingly utilized to improve water quality and aquatic habitat
(Wooster and DeBano, 2006). This could improve the aquatic food web
and linkages with the surrounding riparian and terrestrial ecosystems
(Nakano et al., 1999; Nakano and Murakami, 2001; Ballanger and
Lake, 2005). Not all riparian areas have woody canopies, and prairie
streams in the region were likely dominated in some areas by herba-
ceous vegetation in the pre-settlement period due to frequent fires
that hindered survival of woody species. Herbaceous canals may act as
modern analogs (Dodds et al., 2004; Vandermyde and Whiles, 2015).
Distance along a canal from the point of stream diversion was hy-
pothesized to negatively impact macroinvertebrate communities, mak-
ing them dissimilar to regional streams. Small lateral canals at terminal
positions in the irrigation network had fewer taxa and were dominated
by chironomids and Simulium, likely in response to their adaptations to
poor water quality and short life history suitable for channels with long
dry periods, short flow duration, and turbid water. However, distance
was poorly correlated with overall community composition.
5. Conclusions
A holistic view of canals used to transport water for agricultural and
urban use, as integrated elements of the landscape is necessary to recog-
nize their contribution to and support of local and regional ecosystems
and biota. The prevalence of canals in semi-arid landscapes around the
world makes them important aquatic habitat to analyze and understand
in the context of river conservation and integrated river basin manage-
ment. The volume of water diverted into canals may exceed that re-
maining in natural streams, and in many regions, canals may be the
dominant lotic ecosystems. The landscape setting influences the quality
of streams and canals and their potential similarity. We found that ca-
nals supported similar riparian vegetation and aquatic macroinverte-
brate communities to streams in agricultural landscapes. These results
support the concept that canal networks have created habitat typical
of streams. This does not indicate that all irrigation canals create habi-
tats comparable to those of natural streams, because adjacent land-use
affected the similarities. Furthermore, the similarities between streams
and canals could be attributed to the degraded conditions of streams
and not to the development of diverse and high functioning riparian
and aquatic ecosystems.
The physical and biotic attributes of these artificial ecosystems sup-
port many rare species (Meany et al., 2003), habitat connectivity, tro-
phic subsidies and regional biodiversity (Fernald et al., 2007; Lopez-
Pomares et al., 2015). The importance of irrigation canals as ecosystems
is further supported by the prevalence of artificial waterways in agricul-
tural and residential landscapes, the degradation of natural streams,
point and non-point source pollution, water extraction, and channeliza-
tion. Irrigation canals have been critical to the development and support
of agriculture in arid and semi-arid regions throughout history in all re-
gions of the world. These byproducts of agricultural systems have cre-
ated riparian and aquatic habitats in sufficient amounts and with
similar biological communities as natural steams and should be consid-
ered integral components of agro-ecosystems.
Representative taxa
Isoperla, Optioservus, Dubiraphia, Atherix pachypus
Agabus, Claassenia sabulosa, Aeshna
Brachycentrus, Hydropsyche, Simulium
Chironomus, Cricotopus, Orthocladius, Tanytarsus
Trichocorixa, Belostoma flumineum, Rhagovelia distincta
Baetis, Siphlonurus occidentalis, Tricorythodes explicatus, Heptagenia
780 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
Acknowledgements
The authors would like to thank the National Science Foundation
IGERT, I-WATER: Integrated Water, Atmosphere, Ecosystem Education
and Research Program at Colorado State University [DGE-0966346]
and Colorado Water Institute (2014CO296B) for financial support and
Colorado State University for equipment. Permission to access irrigation
canals through private land was necessary at every site and deeply ap-
preciated. Volunteers aided in collecting and processing plant and mac-
roinvertebrate samples, a very tedious task. This manuscript benefited
from theoretical and technical editing by a talented writing group at
Colorado State University.
Appendices
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.scitotenv.2018.07.246.
References
Andersen, D.C., Nelson, S.M., 1997. Vegetation characteristics and butterfly use of unlined
and PVC-lined reaches of an irrigation delivery canal, Government Highline Canal,
Colorado, U.S.A. J. Arid Environ. 35, 747–764.
Anderson, M.J., Gorley, R.N., Clarke, K.R., 2008. PERMANOVA+ for PRIMER: Guide to Soft-
ware and Statistical Methods. PRIMER-E Plymouth, U.K.
Aspe, C., Jacque, M., 2015. Agricultural irrigation canals in southern France and new urban
territorial uses. Agric. Agric. Sci. Procedia 4, 29–39.
Ballanger, A., Lake, P.S., 2005. Energy and nutrient fluxes from rivers and streams into ter-
restrial food webs. Mar. Freshw. Res. 57, 15–28.
Bogan, M.T., Boersma, K.S., 2012. Aerial dispersal of aquatic invertebrates along and away
from arid-land streams. Freshwat. Sci. 31, 1131–1144.
Bolpagni, R., Piotti, A., 2016. The importance of being natural in a human-altered
riverscape: role of wetland type in supporting habitat heterogeneity and vegetation
functional diversity. Aquat. Conserv. Mar. Freshwat. Ecosyst. 26, 1168–1183.
Bonham, C.D., 2013. Measurements for Terrestrial Vegetation. John Wiley & Sons, West
Sussex.
Bruno, D., Gutierrez-Canovas, C., Velasco, J., Sanchez-Fernandez, D., 2016. Functional re-
dundancy as a tool for bioassessment: a test using riparian vegetation. Sci. Total En-
viron. 566, 1268–1276.
Bunn, S.E., Arthington, A.H., 2002. Basic principles and ecological consequences of altered
flow regimes for aquatic biodiversity. Environ. Manag. 30, 492–507.
Butler, Malcolm G., 1984. Life histories of aquatic insects. In: Resh, Vincent H., Rosenberg,
David M. (Eds.), The Ecology of Aquatic Insects. Praeger Publishers, New York,
pp. 24–55.
Casas, J.J., Sánchez-Oliver, J.S., Sanz, A., Furné, M., Trenzado, C., Juan, M., Paracuellos, M.,
Suárez, M.D., Fuentes, F., Gallego, I., Gil, C., Ramos-Miras, J.J., 2011. The paradox of
the conservation of an endangered fish species in a Mediterranean region under ag-
ricultural intensification. Biol. Conserv. 144, 253–262.
Chester, E.T., Robson, B.J., 2013. Anthropogenic refuges for freshwater biodiversity: their
ecological characteristics and management. Biol. Conserv. 166, 64–75.
Clarke, K.R., Gorley, R.N., Somerfield, P.J., Warwick, R.M., 2014. Change in Marine Commu-
nities: An Approach to Statistical Analysis and Interpretation. 3rd Ed. Bretonside
Copy, Plymouth, UK.
Cooper, D.J., Merritt, D.M., Andersen, D.A., Chimner, R., 1999. Factors controlling Fremont
cottonwood seedling establishment on the upper Green River, Colorado and Utah.
Regul. Rivers Res. Manag. 15, 419–440.
Depeweg, H., Mendez, V.N., 2002. Sediment transport applications in irrigation canals.
Irrig. Drain. 51, 167–179.
Dodds, W.K., Gido, K., Whiles, M.R., Fritz, K.M., Matthews, W.J., 2004. Life on the edge: the
ecology of Great Plains prairie streams. Bioscience 54, 205–216.
Egginton, G.E., Robbins, W.W., 1920. Irrigation water as a factor in the dissemination of
weed seeds. The Agricultural Experiment Station of the Colorado Agricultural College
Bulletin. 253, pp. 3–25.
Ernegger, T., Grubinger, H., Vitek, E., Csekits, C., Eitzinger, J., Gaviria, S., Kotek, D., Krisaf, H.,
Nachtnebel, H.P., Pritz, B., Sabbas, T., Sschmutz, S., Schreiner, P., Stephan, U., Unfer, G.,
Wychera, U., Neudorfer, W., 1998. A natural stream created by human engineering:
investigations on the succession of the Marchfeld Canal in Austria. Regul. Rivers
Res. Manag. 14, 119–139.
Evans, H.E., Evans, M.A., 1991. Cache la Poudre: The Natural History of a Rocky Mountain
River. University Press of Colorado, Niwot, Colorado.
Fernald, A.G., Baker, T.T., Guldan, S.J., 2007. Hydrologic, riparian, and agroecosystem func-
tions of traditional acequia irrigation systems. J. Sustain. Agric. 30, 147–171.
Gonzalez del Tanago, M., Garcia de Jalon, D., 2011. Riparian Quality Index (RQI): a meth-
odology for characterizing and assessing the environmental conditions of riparian
zones. Limnetica 30, 235–254.
Graf, W.L., 1999. Dam nation: a geographic census of large American dams and their hy-
drologic impacts. Water Resour. Res. 35, 1305–1311.
Gregory, K.J., 2006. The human role in changing river channels. Geomorphology 79,
172–191.
Habit, E., Bertran, C., Arevalo, S., Victoriano, P., 1998. Benthonic fauna of the Itata river and
irrigation canals (Chile). Irrig. Sci. 18, 91–99.
Habit, E., Victoriano, P., Campos, H., 2005. Trophic ecology and reproductive aspects of
Trichomycterus areolatus (Pisces, Trichomycteridae) in irrigation canal environments.
Rev. Biol. Trop. 53, 195–210.
Harvey, J., Gooseff, M., 2015. River corridor science: hydrologic exchange and ecological
consequences from bedforms to basins. Water Resour. Res. 51, 6893–6922.
Harvolk, S., Symmank, L., Sundermeier, A., Otte, A., Donath, T.W., 2014. Can artificial wa-
terways provide a refuge for floodplain biodiversity? A case study from north west-
ern Germany. Ecol. Eng. 73, 31–44.
Herzon, I., Helenius, J., 2008. Agricultural drainage ditches, their biological importance
and functioning. Biol. Conserv. 141, 1171–1183.
Hickley, P., 1975. An apparatus for subdividing benthos samples. Oikos 26, 92–96.
Hobbs, R.J., Huenneke, L.F., 1992. Disturbance, diversity, and invasion: implications for
conservation. Conserv. Biol. 6, 324–337.
Jacobs, S.M., Bechtold, J.S., Biggs, H.C., Grimm, N.B., Lorentz, S., McClain, M.E., Naiman, R.J.,
Perakis, S.S., Pinay, G., Scholes, M.C., 2007. Nutrient vectors and riparian processing: a
review with special reference to African semiarid savanna ecosystems. Ecosystems
10, 1231–1249.
Katz, G.L., Stromberg, J.C., Denslow, M.W., 2009. Streamside herbaceous vegetation re-
sponse to hydrologic restoration on the San Pedro River, Arizona. Ecohydrology 2,
213–225.
Kennedy, T.A., et al., 2016. Flow management for hydropower extirpates aquatic insects,
undermining river food webs. Bioscience 66, 561–575.
Kerans, B.L., Karr, J.R., 1994. A benthic index of biotic integrity (B-IBI) for rivers of the Ten-
nessee Valley. Ecol. Appl. 4, 768–785.
Khan, S., Tariq, R., Yuanlai, C., Blackwell, J., 2006. Can irrigation be sustainable? Agric.
Water Manag. 80, 87–99.
Kingsford, R.T., 2000. Ecological impacts of dams, water diversions and river management
on floodplain wetlands in Australia. Austral. Ecol. 25, 109–127.
Koetsier, P., McCauley, L.M.M., 2015. An irrigation canals as a lotic mesocosm: examining
the relationship between macroinvertebrate benthos and drift. West. North Am. Nat.
75, 259–270.
Leigh, C., Sheldon, F., 2008. Hydrological changes and ecological impacts associated with
water resource development in large floodplain rivers in the Australian tropics.
River Res. Appl. 24, 1251–1270.
Leigh, C., Reis, T.M., Sheldon, F., 2013. High potential subsidy of dry-season aquatic fauna
to consumers in riparian zones of wet-dry tropical rivers. Inland Waters 3, 411–420.
Leslie, A.W., Lamp, W.O., 2017. Taxonomic and functional group composition of
macroinvertebrate assemblages in agricultural drainage ditches. Hydrobiologia 787,
99–110.
Lichvar, R.W., 2012. The National Wetland Plant List. Cold Regions Research and Engineer-
ing Laboratory, Hanover NH.
Lopez-Pomares, A., Lopez-Iborra, G.M., Martin-Cantarino, C., 2015. Irrigation canals in a
semi-arid agricultural landscape surrounded by wetlands: their role as a habitat for
birds during the breeding season. J. Arid Environ. 118, 28–36.
Mackay, R.J., 1992. Colonization by lotic macroinvertebrates: a review of processes and
patterns. Can. J. Fish. Aquat. Sci. 49, 617–628.
Malmqvist, B., Rundle, S., 2002. Threats to running water ecosystems of the world. Envi-
ron. Conserv. 29, 134–153.
McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities. MjM Software,
Gleneden Beach, Oregon, USA www.pcord.com.
Meany, C.A., Ruggles, A.K., Ludbow, B.C., Clippinger, N.W., 2003. Abundance, survival and
hibernation of Preble's meadow jumping mice (Zapus hudsonius preblei) in Boulder
County, Colorado. Southwest. Nat. 48, 610–623.
Merritt, D.M., Bateman, H.L., 2012. Linking streamflow and groundwater to avian habitat
in a desert riparian system. Ecol. Appl. 22, 1973–1988.
Merritt, D.M., Cooper, D.J., 2000. Riparian vegetation and channel change in response to
river regulation: a comparative study of regulated and unregulated streams in the
Green River Basin, USA. Regul. Rivers Res. Manag. 16, 543–564.
Merritt, R.W., Cummins, K.W., Resh, V.H., Batzer, P., 2008. Sampling aquatic insects: col-
lection devices, statistical considerations, and rearing procedures. In: Merritt, R.W.,
Cummins, K.W., Berg, M.B. (Eds.), An Introduction to the Aquatic Insects of North
America, 4th edition Kendall Hunt, Dubuque Iowa, pp. 15–37.
Merritt, D.M., Scott, M.L., Poff, N.L., Auble, G.T., Lytle, D.A., 2010. Theory, methods and tools
for determining environmental flows for riparian vegetation: riparian vegetation-
flow response guilds. Freshw. Biol. 55, 206–225.
Miller, A.M., Golladay, S.W., 1996. Effects of spates and drying on macroinvertebrate as-
semblages of an intermittent and perennial prairie stream. J. N. Am. Benthol. Soc.
15, 670–689.
Miller, S.W., Wooster, D., Li, J., 2007. Resistance and resilience of macroinvertebrates to ir-
rigation water withdrawals. Freshw. Biol. 52, 2494–2510.
Moore, R.D., Spittlehouse, D.L., Story, A., 2005. Riparian microclimate and stream temper-
ature response to forest harvesting: a review. J. Am. Water Resour. Assoc. 41,
813–834.
Muehlbauer, J.D., Collins, S.F., Doyle, M.W., Tockner, K., 2014. How wide is a stream? Spa-
tial extent of the potential “stream signature” in terrestrial food webs using meta-
analysis. Ecology 95, 44–55.
Naiman, R.J., DeCamps, H., Pollack, M., 1993. The role of riparian corridors in maintaining
regional biodiversity. Ecol. Appl. 3, 209–212.
Nakano, S., Murakami, M., 2001. Reciprocal subsidies: dynamic interdependence between
terrestrial and aquatic food webs. Proc. Natl. Acad. Sci. 98, 166–170.
Nakano, S., Miyasaka, H., Kuhara, N., 1999. Terrestrial–aquatic linkages: riparian arthro-
pod inputs alter trophic cascades in a stream food web. Ecology 80, 2435–2441.
Nilsson, C., Svedmark, M., 2002. Basic principles and ecological consequences of changing
water regimes: riparian plant communities. Environ. Manag. 30, 468–480.
 E.A. Carlson et al. / Science of the Total Environment 646 (2019) 770–781
Patten, D.J., 1998. Riparian ecosystems of semi-arid North America: diversity and human
impacts. Wetlands 18, 498–512.
Paul, M.J., Meyer, J.L., 2001. Streams in the urban landscape. Annu. Rev. Ecol. Syst. 32,
333–365.
Petchey, O.L., Gaston, K.J., 2002. Functional diversity (FD), species richness and commu-
nity composition. Ecol. Lett. 5, 402–411.
Petraitis, P.S., Latham, R.E., Niesenbaum, R.A., 1989. The maintenance of species diversity
by disturbance. Q. Rev. Biol. 64, 393–418.
Poff, N.L., Zimmerman, J.K.H., 2010. Ecological responses to altered flow regimes: a liter-
ature review to inform the science and management of environmental flows. Freshw.
Biol. 55, 194–205.
Poff, N.L., Olden, J.D., Vieira, N.K.M., Finn, D.S., Simmons, M.P., Kondratieff, B.C., 2006.
Functional trait niches of North American lotic insects: trait-based ecological applica-
tions in light of phylogenetic relationships. J. N. Am. Benthol. Soc. 25, 730–755.
Polis, G.A., Anderson, W.B., Holt, R.D., 1997. Toward an integration of landscape and food
web ecology: the dynamics of spatially subsidized food webs. Ann. Rev. Ecol. 28,
289–316.
Pozo, J., Gonzalez, E., Diez, J.R., Molinero, J., Elosegui, A., 1997. Inputs of particulate organic
matter to streams with different riparian vegetation. J. N. Am. Benthol. Soc. 16,
602–611.
Relyea, C.D., Minshall, G.W., Danehy, R.J., 2000. Stream insects as bioindicators of fine sed-
iment. Proc. Water Environ. Fed. (6), 663–686.
Rosenberg, D.M., King, R.S., Resh, V.H., 2008. Use of aquatic insects in biomonitoring.
Chapter 7. In: Merritt, R.W., Cummins, K.W., Berg, M.B. (Eds.), An Introduction to
Aquatic Insects of North America, 4th Edition 2008. Kendal Hunt (pp).
Shafroth, P.B., Stromberg, J.C., Patten, D.T., 2002. Riparian vegetation response to altered
disturbance and stress regimes. Ecol. Appl. 12, 107–123.
Strange, E.M., Fausch, K.D., Covich, A.P., 1999. Sustaining ecosystem services in human-
dominated watersheds: biohydrology and ecosystem processes in the South Platte
River Basin. Environ. Manag. 24, 39–54.
Sueltenfuss, J.P., Cooper, D.J., Knight, R.L., Waskom, R.M., 2013. The creation and mainte-
nance of wetland ecosystems from irrigation canal and reservoir seepage in a semi-
arid landscape. Wetlands 33, 799–810.
781
Swamee, P.K., 1995. Optimal irrigation canal sections. Irrig. Drain. Eng. 121, 467–469.
Tank, J.L., Rosi-Marshall, E.J., Griffiths, N.A., Entrekin, S.A., Stephen, M.L., 2010. A review of
allochthonous organic matter dynamics and metabolism in streams. J. N. Am. Benthol.
Soc. 29, 118–146.
Urban, M.A., Rhodes, B.L., 2003. Catastrophic human-induced change in stream-channel
planform and geometry in an agricultural watershed, Illinois, USA. Ann. Assoc. Am.
Geogr. 93, 783–796.
Vandermyde, J.M., Whiles, M.R., 2015. Effects of experimental forest removal on macroin-
vertebrate production and functional structure in tallgrass prairie streams. Freshwat.
Sci. 34, 519–534.
Verdonschot, R.C.M., Keizer-Vlek, H., Verdonschot, P.F.M., 2011. Biodiversity value of agri-
cultural drainage ditches: a comparative analysis of the aquatic invertebrate fauna of
ditches and small lakes. Aquat. Conserv. Mar. Freshwat. Ecosyst. 21, 715–727.
Vermonden, K., Leuven, R.S.E.W., van der Velde, G., van Katwijk, M.M., Roelofs, J.G.M.,
Hendrisk, A.J., 2009. Urban drainage systems: an undervalued habitat for aquatic
macroinvertebrates. Biol. Conserv. 142, 1105–1115.
Wallace, J.B., Grubaugh, J.W., Whiles, M.R., 1996. Biotic indices and stream ecosystem pro-
cesses: results from an experimental study. Ecol. Appl. 6, 140–151.
Weber, W.A., Wittmann, R.C., 2012. Colorado Flora: Eastern Slope. 4thd Edition. Univer-
sity Press of Colorado, Boulder.
Whatley, M.H., Vonk, J.A., van der Geest, H.G., Admiraal, W., 2015. Temporal abiotic vari-
ability structures invertebrate communities in agricultural drainage ditches.
Limnologica 52, 20–29.
Williams, C.A., Cooper, D.C., 2005. Mechanisms of riparian cottonwood decline along reg-
ulated rivers. Ecosystems 8, 382–395.
Williams, D., Feltmate, B., 1992. Aquatic Insects. CAB International Wallingford, U.K.
Wohl, E.E., 2001. Virtual Rivers: Lessons from the Mountain Rivers of the Colorado Front
Range. Yale University Press, New Haven.
Wooster, D.E., DeBano, S.J., 2006. Effect of woody riparian patches in croplands on stream
macroinvertebrates. Arch. Hydrobiol. 165, 241–268.