Effects of agricultural land use on benthic macroinvertebrate communities in Southern Illinois headwater streams

Nathan J. Boyer-Rechlin*†Gregory L. Bruland, Ph.D.*Michael A. Rechlin, Ph.D.*

*Biology and Natural Resources DepartmentPrincipia College

1 Maybeck Pl.Elsah, IL 62028

†2519 Mountaineer DriveFranklin, WV 26807

nathan.boyerrechlin@gmail.com

Land Use Effects on Illinois Benthic Macroinvertebrates

Keywords: biotic index, headwater streams, Illinois, land use, macroinvertebrates

1

SUMMARY:

1) Headwater streams are important in buffering the downstream effects of agricultural sedimentation and nutrient loading. In the Midwestern United States, agricultural development has degraded the quality of freshwater systems from headwaters to the Mississippi Delta. In order to assess the effects of varying degrees of agricultural impacts on Illinois headwaters, benthic macroinvertebrates were sampled from 15 perennial headwater streams in Jersey County, Illinois.

2) Land-use/land-cover (LULC) distributions were calculated for each watershed within three spatial scales: watershed, 150m radial buffer, and 50m radial buffer. Sampled watersheds ranged from 5-80% agricultural cover (cultivated crops & pasture/hay lands).

3) Family diversity and richness were low across the region. A total of 20 taxa were identified, among which only four were Ephemeropteran, Plecopteran, or Trichopteran (EPT) taxa. Hydropsychidae (Trichoptera) was the most dominant family and had a strong positive correlation with % agriculture. Additionally, a PCA was used to analyze community structure trends among the macroinvertebrate functional feeding groups.

4) While community bioassessment metrics, including biotic indices and various diversity metrics are commonly used methods of stream quality monitoring and analysis, these metrics did not correlate well with LULC in this study and they may not be effective metrics for agricultural streams with low benthic macroinvertebrate diversity in this region.

5) Overall, none of the streams sampled were indicative of highly degraded communities and most streams had similar macroinvertebrate composition in spite of variability in LULC. Even within these streams, however, % agriculture had a significant impact on community composition and results suggest that % Hydropsychidae may be an effective means of assessing small-scale impacts of agriculture in the headwater systems of Southern Illinois.

2

Introduction:

Agricultural development negatively impacts freshwater systems in a variety

of ways and often causes the following: (1) increases in sedimentation; (2) greater

flow variation, which can disrupt biological communities and processes; (3) loss of

high quality habitat; (4) contamination from pesticides and (5) increases in nitrogen

(N) and phosphorous (P) concentrations (Allan, 2004, Herringshaw et al., 2011).

Nutrient transport through headwater systems is complex due to the high rates of

uptake and transformation in these streams (MacDonald & Coe, 2006, Ohio EPA,

2003). However, it has been shown that in highly impacted agricultural watersheds,

headwater streams serve as the primary transport mechanism for excess nutrients

and sediments (Borah & Bera, 2003, Ohio EPA, 2003).

These nutrient and sediment stressors decrease biotic diversity and alter

trophic interactions, which in turn can have cascading impacts on functional group

distributions, organic matter transport, and overall stream function throughout the

entire continuum of lotic systems (Cummins, 1973; Vannote et al., 1980; Allan, 2004,

Burcher, Valett & Benfield., 2007). Studies suggest that agricultural development has

a net negative impact on invertebrate diversity and overall community structure

when agriculture is prominent, i.e. greater than 30-40% of watershed land use/land

cover (LULC) (Genito, Gburek & Sharpley, 2002; Stone et al., 2005; Kratzer et al.,

2006; Niyogi et al., 2007). The extent of agricultural impacts and stress is variable,

however, and can be largely influenced by the spatial distributions and types of

agriculture in the watershed, with cultivated croplands having much higher impacts

3

than pasture and hay fields (Allan, 2004; Stone et al., 2005; Kratzer et al., 2006; King

et al., 2005)

Across Illinois, agricultural expansion has severely impacted the landscape.

Almost three quarters of pre-settlement forests and 90% of the state’s prairies have

been converted to agriculture over the past 200 years (Iverson, 1988). It has been

estimated that 15% of the N and 10% of the P inputs leading to the hypoxic zones in

the Gulf of Mexico come from Illinois croplands (Gentry et al., 2000; Alexander et al.,

2008). Jersey County, in west-central Illinois, is unique in that it is home to highly

forested, highly cultivated, and mixed LULC drainages all within close proximity to

each other and without substantial urban impact. Because of its many protected

natural areas and rugged topography, characterized by steep ravines and riverside

bluffs, the southwestern half of Jersey County, has nearly twice as much contiguous

forest (proportionately) as rest of the state (Joselyn & Brown, 1996; Krohe, 1997). In

contrast, the northeastern half of the county resembles the typical Illinois landscape,

dominated by row-crop agriculture. (Fry et al., 2011). The county’s LULC continuum

from forested to agriculturally dominated watersheds makes this region

representative of the varying extents of agricultural impacts found across Illinois.

One common method used to monitor stream quality response to stressors

such as agriculture is bioassessment, including the use of various macroinvertebrate

community metrics (Hilsenhoff, 1987; Lenat, 1993; Barbour et al., 1999; Bode et al.,

2002; Rawer-Jost et al., 2000; Allan, 2004,). Bioassessment provides insights into the

health of riverine systems while avoiding the complexity, costliness, and dramatic

daily and seasonal variance that accompanies traditional water quality testing

4

(Barbour et al., 1999; Bode et al., 2002,). Hilsenhoff’s macroinvertebrate biotic index,

which was developed to show stream quality response to organic and nutrient

pollution, and its various regional adaptations are some of the most commonly used

bioassessment metrics. These biotic indices, as well as other diversity metrics, have

seen widespread application over the past 30-40 years and have been used

extensively by the Illinois Natural History Survey to monitor stream condition across

Illinois (Hilsenhoff, 1982; Hilsenhoff, 1987; Lenat, 1993; Chirhart, 2003; Pond et al.,

2003; Miller, 2004). Additionally, simplified family and order level biotic indices have

become common in volunteer stream monitoring programs across the United States

(Miller, 2004; Marchal, 2005). Because macroinvertebrate communities reflect long-

term disturbances and stress trends, community metrics and biotic indices reflect

the effects of anthropogenic stressors on the function of stream systems as a whole,

including influences on food web dynamics, biodiversity, and organic matter

transport (Cummins, 1973; Vannote et al., 1980).

The purpose of this study was to investigate the relationships among

agricultural land use and benthic macroinvertebrate communities in Jersey County

headwater streams. Secondarily, we aimed to evaluate the use of various

macroinvertebrate community indices and metrics as means of assessing regional

stream quality. It was hypothesized that benthic community structure would be a

function of % agricultural land use. This was evaluated at three spatial scales:

watershed, 150m radial stream buffer, and 50m radial stream buffer. Additionally it

was hypothesized that diversity, evenness, and richness would be negatively

correlated with % agricultural land use and that Hilsenhoff’s (1988) family biotic

5

index (FBI) scores would be positively correlated with agriculture for all spatial

scales. Because of the mosaic of forested and agriculturally developed watersheds in

the county, understanding the relationships between agricultural distributions and

benthic community structure, and the applicability of commonly used bioassessment

metrics in regional headwater streams would be valuable for stream quality

monitoring programs and trend assessment throughout Illinois and the greater

Mississippi River Basin.

Methods

Study area

Jersey County covers 979 km2 in west central Illinois, bordering the

Mississippi and Illinois rivers, and is split between three Illinois natural divisions

(Schwegman et al., 1973). Sample sites were all located in southern and western

Jersey County, within or directly adjacent to the Middle Mississippi Borderland

Division (MMBD). This region is characterized by highly dissected karst topography

with deciduous forests dominating the hills and ravines and cultivated croplands in

the floodplains and ridge top plateaus (Schwegman et al., 1973; Panno, Weibel & Li,

1997). The majority of the county’s springs and high gradient streams fall within this

region (Krohe, 1997; Wetzel & Webb, 2007; Fry et al., 2011).

Site Selection

Benthic macroinvertebrates were sampled from 15 perennial headwater

streams (watershed area <10 km2). For the purposes of this study, streams were

classified as perennial if the sampled reach maintained surface flow throughout the

6

semi-drought conditions in the fall of 2013. All streams had notable (>5m) riparian

vegetation on at least one stream bank. Substrates were primarily cobble and gravel

intermixed with silt and, at a few sites, bedrock.

Streams were selected by isolating possible watersheds in ArcMap (Version

10.2, ESRI, Redlands, CA) using a digital elevation model, an Illinois streams layer,

and the NLCD 2006 layer (Fry et al., 2011). Personal knowledge of regional

watersheds along with suggestions from local experts and site visits to potential

streams were also crucial in picking sample sites. Drainages were selected on a

gradient from low to high agricultural land use (Table 1) (Fry et al., 2011). ArcMap

was used to delineate watersheds using the Hydrology tool in Spatial Analyst, and to

calculate LULC distributions within the watersheds and the 50m and 150m radial

buffers for each stream.

Reach metrics

Riffle velocity was measured at a representative riffle in each reach. Discharge

was calculated at channelized areas with uniform flow, often transition zones

between riffle and pool habitats, by measuring the wetted width, average depth, and

velocity. Because of the extremely low flow, velocity was calculated by averaging the

time it took to float a stick one to two meters over three repetitions. Width was

measured at representative riffle, pool, and channel cross sections. Additionally, a

visual habitat assessment was performed at each reach following the rapid

bioassessment protocols outlined by the U.S. EPA (Barbour et al., 1999).

Temperature, pH, conductivity, total solids, and dissolved oxygen were measured

using the YSI 556 water quality probe (YSI Incorporated, Yellow Springs, OH).

7

Measurements were taken in riffle areas and recorded after allowing the probe to

assimilate to stream conditions for 10 minutes.

Macroinvertebrate sampling

Sampling was completed during base-flow conditions between 15 September

and 16 October 2014. A Surber sampler was used to collect macroinvertebrates from

riffle habitats (stream flow rates >0.02 m s-1) within a 60-meter reach at each stream.

Between two and five samples were collected at each site by disturbing, kicking,

rubbing, and upturning the substrate for one minute. Large sticks and leaves were

removed and all macroinvertebrates and remaining detritus were transferred into

mason jars and mixed with 95% ethanol. The net was rinsed and picked for

remaining macroinvertebrates (>3mm) for roughly five minutes.

Upon return to the lab, jars were drained and filled with 70% ethanol. Each

sample was subsampled by spreading the invertebrates and remaining detritus

across two gridded pans and picking all invertebrates from randomly selected grids

until >150 invertebrates had been picked. Individuals were identified to family and

stored in 70% ethanol or isopropyl alcohol.

Macroinvertebrate metrics

Shannon Diversity (H’), family richness, evenness, and functional feeding

group (FFG) distributions were calculated and used to assess community structure.

For FFG analysis, families were grouped into five groups: shredders, scrapers,

collector-filterers, collector-gatherers, and predators (Cummins, 1973; Voshell, 2002;

Cummins, Merritt & Andrade, 2005). Hilsenhoff’s FBI was also calculated to estimate

the effects of organic and nutrient pollution (Hilsenhoff, 1982; Hilsenhoff, 1988).

8

Statistical analysis

Pearson’s and Spearman correlations were run among the explanatory

variables, LULC distributions, macroinvertebrate abundances, and community

metrics using the SPSS statistical software package (Version 19, IBM, New York, NY).

Correlations were considered significant at p<0.1. A Principal Components Analysis

(PCA) was also run using PCOrd (Version 6, MjM Software, Gleneden Beach, OR) to

investigate multivariate patterns in the family-level and the FFG data as well as to

identify the strongest explanatory variables (McCune & Grace, 2002).

Results

Land use

Agricultural land use (including both cropland and pasture hay) of the

sampled watersheds ranged from 4.7% to 80.2% with cultivated crops composing

roughly two thirds of total agriculture (Table 1). Forest cover ranged from 8.6% to

92.3% and urban development was negligible. All streams had forested riparian

buffers with no stream having less 26% forest within its 50m buffer (Table 1). Forest

cover became more dominant at smaller spatial scales with the 50m buffers having

the highest proportion of forest cover (Table 1). Forest cover was also inversely

correlated with agriculture for the watershed (r2=0.99), 150m buffer (r2=0.99), and

50m buffer (r2=0.72) scales. The LULC distribution for Jersey County, including

sampled watersheds, is shown in Figure 1.

Reach metrics

9

Average discharge was 0.019 m3s-1. The wetted width at riffle sections ranged

from 0.5m to 2.7m and channel width ranged from 3.5m to 13.3m. Habitat

Assessment (HA) scores ranged from 110 to 167, out of 200. Out of all the sub-

categories, substrate embeddedness, channel flow status, and vegetative protection

had the most variation. HA scores had a significant Pearson’s correlation with forest

cover at all spatial scales, with the strongest correlation occurring for the 50m

buffers (r=0.72, p=0.002).

Water quality characterization

Dissolved oxygen (DO) ranged from 6.14 mg L-1 to 10.17 mg/L and had a

negative Pearson’s correlation with agricultural land use at the watershed (r=-0.45,

p=0.09), 150m buffer (r=-0.52, p=0.05), and 50m buffer (r=-0.45, p-0.01) scales. All

sites were mildly acidic, averaging a pH of 5.90. Site 5 had an abnormally low pH of

4.66. Conductivity ranged from 0.514 to 0.728 MS s-1. Stream temperature varied

substantially among sites, ranging from 11.9 to 19.27oC.

Community structure

The distribution of FFGs at all sites is shown in Figure 2. The community

structure varied from 100% shredders at site 2 to collector-filterer dominance at site

13 (Figure 2). According to the PCA, axis one accounted for 55% of the variance and

axis two accounted for 19.2% of the variance (Figure 3). Axis one covered a gradient

from shredder dominance on the right, to scraper and collector-filterer dominance

on the left. Axis 2 showed a gradient from collector-gatherers on the bottom to

predators on the top. The sites fell into three main groupings in the ordination space

with shredder dominated sites on the far right, mixed FFGs in the center and

10

collector-filterer dominance on the left. Additionally, % agriculture and % forest

cover, at the watershed scale, also had strong loadings on axis 1 (Figure 3).

Overall, 18 different families, one sub-phylum, and one sub-class of benthic

invertebrates were identified. Hydropsychidae (Trichoptera), Gammaridae

(Amphipoda), Asellidae (Isopoda), Baetidae (Ephemeroptera) and Chironomidae

(Diptera) were the most common, accounting for 94% of all invertebrates.

Hydropsychidae and Gammaridae were co-dominant accounting for 39% and 28% of

invertebrates, respectively. Sub-phylum Turbellaria, sub-class Oligochaeta, Elmidae

(Coleoptera), Physidae (Gastropoda), Heptageniidae (Ephemeroptera), Veliidae

(Hemiptera), and Calopterygidae (Odonata) were all found in at least three streams.

Six families were only found at one site and in most cases were very rare in their

communities. Perlodidae (Plecoptera), which was only found at site 14, was the only

exception and made up 3.5% of invertebrates collected in that stream.

FFG Relationships

Pearson’s correlation coefficients and p-values are shown for the five FFGs

and LULC variables at all spatial scales in Table 2. Both shredders and collector-

filterers had significant correlations with forest and agriculture at a 95% confidence

level for all spatial scales except for forest in the 50m buffers (Table 2). Shredders

had a negative correlation with agriculture, which was strongest for the 50m buffers.

Collector-filterers had a positive correlation with agriculture, which was also

strongest within the 50m buffers (Table 2).

Scrapers were significantly correlated with forest and agriculture within the

watershed scale, at a 90% confidence level. Scrapers also had a significant

11

correlation at a 95% confidence level with agriculture in the 50m and 150m buffers

and with forest in the 150m buffers (Table 2). Collector-gathers were significantly

correlated with urban land-cover in the 50m buffers at a 90% confidence level

(r=0.48). Predators were not correlated with LULC or any explanatory variables.

Family-level relationships

Pearson’s correlation coefficients and p-values are shown for the five

dominant families and LULC variables at all spatial scales in Table 3. Gammaridae

were positively correlated with forest cover and negatively correlated with

agricultural land use at most spatial scales. For both agricultural land use and forest

cover, the strongest correlations occurred when LULC was calculated at the

watershed scale. Hydropsychidae were negatively correlated with forest cover and

positively correlated with agricultural land use at all spatial scales, excepting forest

cover within the 50 m buffers (Table 3). Chironomidae had significant Spearman

correlations with watershed agriculture (r=0.57, p=0.03) and forest cover (r=-0.60,

0.02). The strongest family-LULC correlation was between Hydropsychidae and

agriculture in the 50m buffers (Figure 4).

Community metrics

Hilsenhoff’s FBI scores ranged from 4.0 to 6.0 (Table 4). FBI was not

correlated with any the LULC data or explanatory variables. Shannon-Weiner family

diversity (H’) ranged from 0.07 to 1.57 (Table 4) and had a positive Pearson’s

correlation with channel width (r=0.54, p=0.04). Additionally, H’ had positive

Spearman correlation with watershed size (r=0.46, p=0.08) and riffle width (r=0.45,

p=0.01). H’ was not correlated with LULC or any other of the explanatory variables.

12

Family richness ranged from 2 to 12 with an average richness of 7. Family evenness

ranged from 0.10 to 0.72 (Table 4). Evenness also had a significant Pearson’s

correlation with channel width (r=0.47, p=0.08).

Discussion

Land use/land cover

Even though agricultural cover was as high as 80% in some sampled

watersheds, none of these watersheds had over 60% cultivated croplands, which are

responsible for the majority of agricultural based nutrient loading and sedimentation

(Allan, 2004; King et al., 2005; Kratzer et al., 2006). Although there are some almost

completely cultivated parts of Jersey County, we were unable to sample watersheds

in these areas because we found no perennial headwater streams in these regions.

This is explained in part by the spatial extent of the county’s karst geology, where

most of its springs are located. The most cultivated watersheds were found in the

flatter, glacial till plain of northeastern Jersey County. While better adapted for

agricultural development, this area lacks the karst geology characteristic of the hillier

and more forested MMBD in the southern and western portions of the county

(Schwegman et al., 1973; Panno et al., 1997).

Because of its karst geology, subsequent groundwater systems and numerous

springs, the majority of perennial headwater streams, including all streams sampled

in this study, were located in the MMBD (Schwegman et al., 1973; Webb et al., 1995;

Wetzel & Webb, 2007). However, because of the dissected topography in this region,

agricultural development is limited to the ridge tops and floodplains (Figure 1),

13

limiting the extent and impact of agriculture across the region. Due to this tendency

for agricultural development to occur in the uplands, away from steep ravines,

agriculture is less prominent, proportionately, close to the streambeds than it is

across whole watersheds. Among our sites, the ratio of agriculture to forest cover

was the lowest in the 50m buffers (0.32) and the highest across the watersheds

(0.84). The comparatively high proportion of forest cover within the 50m and 150m

buffers may have helped reduce the impacts of upland agriculture (Weigel, 2003;

Allan, 2004; Vondracek et al., 2005; Song et al., 2009).

The limited range of agricultural impacts among sampled watersheds,

evidenced by the lack of completely cultivated watersheds and the tendency for

agriculture to occur away from immediate stream buffers, is a potential limitation of

this study. Resampling both perennial and non-perennial streams with a multi-

habitat sampling approach would facilitate a more comprehensive survey of the

county’s streams, allowing for sampling in the most heavily cultivated regions.

Community structure

The dominance of shredders and collectors observed in this study matches

the predictions of the river continuum concept for heterotrophic headwater streams

(Vannote et al., 1980). However, in contrast with Sangunett (2005) and Stone et al.

(2005) who found collector-gatherers to be the dominant FFG in agriculturally

impaired watersheds and small streams, we found mostly collector-filterers. One or

two families dominated both the shredder and collector-filterer FFGs with

Gammaridae and Asellidae being the only shredders and Hydropsychidae accounting

for all but 10 individual collector-filterers. This indicated that Gammaridae and

14

Hydropsychidae are key taxa in fulfilling the ecological rolls of streams in the region.

Additionally the dominance of single, tolerant taxa in these groups suggested that

across all streams, persistent stress and degradation might have impacted the

viability of less tolerant taxa.

Many of the taxa found in this study, notably Hydropsychidae, Baetidae,

Chironomidae, Elmidae, Simuliidae, and sub-class Oligochaeta have been found to be

common across the state (Heatherly et al., 2007). However the high proportion of

Amphipoda and Isopoda observed has not been consistently documented (Sangunett,

2005; Stone et al., 2005; Heatherly et al., 2007). High proportions of these taxa are

often associated with spring communities and their dominance in our streams is

likely related to the numerous hard water springs found in the MMBD (Webb et al.,

1995; Wetzel and Webb, 2007). Additionally, a study of forested streams in northern

Illinois also reported a high proportion of amphipods and isopods, suggesting that

these taxa may play a pivotal role in linking the terrestrial and aquatic systems in

wooded Illinois streams (Vinikour & Anderson, 1984).

The dominance of Hydropsychidae found in this study indicates nutrient

stressed conditions (Voshell, 2002). Cheumatopsyche (Trichoptera: Hydropsychidae),

a common hydropsychid found across Illinois, has been shown to become hyper-

dominant in response to organic pollution and nutrient enrichment. Other genera

display similar responses (Pond et al., 2003; Heatherly et al., 2007). The general lack

of Ephemertopteran, Plecopteran, and Trichopteran (EPT) taxa, other than

Hydropsychidae and Baetidae which are generally more pollution tolerant, also

suggested degraded stream conditions and stressed benthic communities (Lenat &

15

Penrose, 1996; Voshell, 2002). EPT taxa tend to be more intolerant of pollution and

habitat disturbance and EPT richness has been shown to decrease with increased

stress (Barbour et al., 1999; Griffith et al., 2001; Chirhart, 2003; DeWalt, 2003;

Sangunett, 2005). EPT richness varies across Illinois regionally and with stream size,

but a comprehensive survey showed most Illinois streams to have higher EPT

richness than those sampled in this study (DeWalt, 2003; Miller, 2004).

Still, in comparison with other statewide benthic macroinvertebrate surveys,

the streams in this study only appear to be moderately degraded. A similar study

conducted on Chicago area urban headwater streams found lower family richness

than we did (Gorman, 1987). Both that study and a study on agricultural streams in

southwestern Illinois reported significantly different community structures,

dominated by oligochaetes and highly tolerant non-insectan families along with a

few tolerant insectan families, namely Chironomidae and Ceratopogonidae (Gorman,

1987; Stone et al., 2005). These midge and non-insect dominated communities are

common in sediment- and nutrient-stressed agricultural streams (Voshell, 2002;

Stone et al., 2005; Heatherly et al., 2007; Herringshaw et al., 2011). Though most

streams in our study were less diverse than those in some other similar regions of

the state, the communities in the streams in this study are more characteristic of

low-medium impact systems than they are of highly degraded streams (Vinikour &

Anderson, 1984; Sangunett, 2005; Stone et al., 2005)

Despite the similarities in community structure among most sites, the

presence of Heptageniidae at four sites and Perlodidae (Plecoptera) at site 14

suggests that area streams are capable of supporting greater EPT diversity and

16

higher quality communities. Located within an almost completely forested

watershed, site 14 lacked the obvious environmental stressors present across much

of the region and was expected to exhibit higher quality benthic communities.

Plectopterans are some of the most intolerant EPT taxa and the absence of this order

from all other sites suggests that even minimal agricultural disturbance significantly

impacted the viability of Plecopteran populations in regional headwater systems

(Chirhart, 2003).

However, the usefulness of site 14 as a reference site is suspect due to high

number Asellidae, of which some species are often found in urban and other

organically polluted streams (Williams, 1976; Heatherly et al., 2007). Asellids were

only common at one other site and the dynamics influencing the dual presence of

large numbers of isopods and a notable plecopteran population are not well

understood. It is possible that upstream septic systems or hard-water spring

dynamics may help explain the community dynamics in this stream. In spite of this

issue, the presence of Perlodidae at this site still shows that Plecoptera are not

completely extirpated from the region and may serve as a good indicator of higher

quality streams across the county.

Another factor influencing the comparison of streams within Jersey County is

the possibility of similar reference quality streams having different unique

community structures. Sites 2, 10, and 15, were home to an overwhelmingly non-

insect community dominated by Gammaridae and, at site 10, Asellidae. Together,

these two families accounted for over 95% of invertebrates and only one individual

insect, a Baetidae, was found across all three sites. The PCA, which analyzed

17

community structure based on FFG distributions, grouped these three sites, along

with site 3, together in a stand-alone cluster (Figure 3). These were classified as the

shredder-dominated communities. This classification also supports the observation

of a few unique, non-insectan communities because the only shredders found in this

study were Asellidae and Gammaridae.

A study on benthic spring communities in Illinois found that Amphipod

dominated non-insectan communities were related to the presence hard-water

springs (Webb et al., 1995; Wetzel & Webb, 2007). We hypothesize that southwest

Jersey County’s karst geology, and the corresponding hard-water springs might help

explain the differences between some of the benthic communities in our streams.

The possible differences in community structure among nearby or even adjacent

regional streams may prove problematic when using community metrics and biotic

indices to compare water quality and these factors must be considered when

applying these methods.

FFG and family relationships

A decline in shredders is an expected community response to disturbance

(Rawer-Jost et al., 2000). The negative correlation between shredders and

agricultural land use observed in this study supported the hypothesis that upstream

agriculture influences community structure and suggests that the loss of shredders

in stream communities may indicate habitat degradation. However, the family level

correlations with LULC were stronger than those with the FFGs. In streams with

similar functional quality, the distribution of individual families may serve as a better

indicator of fine-scale LULC impacts.

18

The strong correlation between Gammaridae and forest cover in the 150m

buffers and watersheds explained most of the variance in shredder populations and

their correlation with LULC. Due to the possible complications with Gammaridae-

dominated hard-water spring communities discussed earlier, which in this study

often occurred in highly forested watersheds, this correlation with forest cover could

be skewed. Most importantly, our results show that in this region % Hydropsychidae

may be one of the most useful indicators of agricultural disturbance, especially

among low and medium impact streams with similar upstream environmental

conditions.

Community metrics

Along with community composition analysis, Hilsenhoff’s index and

community diversity metrics, including Shannon-Weiner diversity, species richness,

and EPT richness, have all been successfully used to differentiate between high

quality and stressed streams in agricultural regions (Griffith et al., 2001; Genito et al.,

2002; Stone et al., 2005; Song et al., 2009). Although Stone et al. (2005) found a

modified Hilsenhoff’s index useful in differentiating among agricultural streams in

Southern Illinois, our results support the findings of DeWalt (2003) and Sangunett

(2005) who found that Hilsenhoff’s index (and derivative indices) scores varied little

across the state and may not be effective tools for bioassessment in Illinois. Although

some of these metrics may not be consistently effective individually, the lack of

correlation among the LULC variables and any of the bioassessment metrics suggests

that stream quality varied little among sampled sites and was not significantly

impacted by increases in % agriculture, up to 80%.

19

Although these metrics are often successful in depicting water quality

variance among highly dissimilar streams or in regions with a large range of

anthropogenic impacts, they are less successful when applied to areas with a smaller

range of environmental conditions (Kratzer et al., 2006). The higher proportion of

forest-cover within the riparian zones and limited range of agricultural LULC

distributions (<60% cultivated crops) likely helped minimize the effects of stream

quality differences among sites. Additionally, the low diversity found among even

the most pristine watersheds might have limited the ability of diversity metrics to

model the impacts of agricultural pollution in the region. The changes in community

composition in relation to LULC demographics, evident among the family and FFG

distributions, suggest that agricultural impacts had at least a moderate effect on

stream condition and macroinvertebrate communities. However, because of the

inherent low-diversity found across the county’s headwater streams, diversity and

richness metrics may not model these changes as well as community composition

data. Similarly, the lack of diversity likely decreases the resolution of Hilsenhoff’s FBI

and its ability to model water quality fluctuations among moderately degraded

streams.

It is important to note that our use of family-level identification may have had

an impact on the resolution and effectiveness of these metrics. Family level

taxonomic identification can significantly reduce the resolution of diversity and

richness metrics (Song et al., 2009). Although a North Carolina study comparing

order level and genus level macroinvertebrate biotic indices showed no significant

differences between the two, few comparative studies have been conducted in other

20

regions and Hilsenhoff (1988) documented that his FBI tends to produce values

closer to four due to prevalence of certain families (i.e. Hydropsychidae, Gammaridae

and Baetidae), giving lower scores to heavily polluted streams and higher scores to

pristine streams (Marchal, 2005). Nonetheless, it is important to clearly understand

the effectiveness of metrics at this level of taxonomic resolution because of the

prevalence of family and order level macroinvertebrate indices in volunteer stream

monitoring, and initial assessments.

In conclusion, most streams in the MMBD portions of Jersey County were

shown to have similar benthic invertebrate communities despite differing LULC

distributions. In spite of these similarities, changes in agricultural land use were

shown to have an impact on these streams with an increase in % Hydropsychidae

being the best indicator of community response to agriculture. H’, richness,

evenness, and Hilsenhoff’s FBI showed no correlation with the LULC data suggesting

that these bioassessment metrics may not be effective in the region’s low-diversity,

medium impact agricultural streams.

21

Acknowledgements

We thank M. Rhaesa (aquatic ecology consultant) and A. Ringhausen (Great

Rivers Land Trust) for their support of this research. Additionally, we are grateful to

Principia College students L. Pyne, C. Chichester, and K. Krishniswami for helping

with the fieldwork. Finally, we appreciate the support of the entire Principia College

Biology and Natural Resources Department.

22

Literature Cited

Alexander R.B., Smith R.A., Schwarz G.E., Boyer E.W., Nolan J.V. & Brakebill J.W. (2008) Differences in phosphorus and nitrogen delivery to the gulf of Mexico from the Mississippi river basin. Environmental Science and Technology, 42, 822-830.

Allan J.D. (2004) Landscapes and riverscapes: the influence of land use on stream ecosystems. Annual Review of Ecology, Evolution, and Systematics, 35, 257-284.

Barbour M.T., Gerritsen J., Snyder B.D. & Stribling J.B. (1999) Rapid bioassessment for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates and fish, 2nd ed. EPA 841-B-99-002. Office of Water, US Environmental Protection Agency, Washington, DC.

Bode R.W., Novak M.A., Abele L.E., Heitzman D.L. & Smith A.J. (2002) Biological stream monitoring in New York State. Stream Biomonitoring Unit, Bureau of Water Assessment and Management, Division of Water, NYS Department of Environmental Conservation, Albany, NY.

Borah D.K. & Bera M. (2003) Watershed-scale hydrologic and nonpoint-source pollution models: review of mathematical bases. Transactions of the American Society of Agricultural Engineers, 46, 1553-1566.

Burcher C.L., Valett H.M. & Benfield E.F. (2007) The Land-Cover Cascade: Relationships Coupling Land and Water. Ecology, 88, 228-242.

Chirhart J. (2003) Development of a macroinvertebrate index of biological integrity (MIBI) for rivers and streams of the St. Croix River basin in Minnesota. Minnesota Pollution Control Agency, Biological Monitoring Program, St. Paul, Minnesota.

Cummins K.W. (1973) Trophic relations of aquatic insects. The Annual Review of Entomology, 18, 183–206.

Cummins K,W., Merritt R.W. & Andrade P.C.N. (2005) The use of invertebrate functional groups to characterize ecosystem attributes in selected streams and rivers in south Brazil. Studies on Neotropical Fauna and Environment, 40, 69–89.

DeWalt R.E. (2003) Aquatic Insects Report: Biological and Habitat Condition of Illinois Streams. In: Critical Trends Assessment Program 2002 Report. Illinois Department of Natural Resources, Office of Realty and Environmental Planning, Division of Review and Coordination, Springfield, Illinois.

23

Fry J., Xian G., Jin S., Dewitz J., Homer C., Yang L. et al. (2011) Completion of the 2006 National Land Cover Database for the Conterminous United States. Photogrammetric Engineering and Remote Sensing, 77, 858-864.

Genito D., Gburek W.J. & Sharpley A.N. (2002) Response of Stream Macroinvertebrates to Agricultural Land Cover in a Small Watershed. Journal of Freshwater Ecology, 17, 109–119.

Gentry L.E., David M.B., Smith-Starks K.M. & Kovacic D.A. (2000) Nitrogen fertilizer and herbicide transport from tile drained fields. Journal of Environmental Quality, 29, 232-240.

Gorman O.T. (1987) A survey of the fishes and macroinvertebrates of some small streams in Cook, Lake, and Dupage counties, Illinois. Transactions of the Illinois Academy of Science, 80, 253–272.

Griffith M.B., Kaufmann P.R., Herlihy A.T. & Hill B.A. (2001) Analysis of macroinvertebrate assemblages in relation to environmental gradients in rocky mountain streams. Ecological Applications, 11, 489-505.

Heatherly T. 2nd, Whiles M.R., Royer T.V. & David M.B. (2007) Relationships between water quality, habitat quality, and macroinvertebrate assemblages in Illinois streams. Journal of Environmental Quality, 36, 1653–1660.

Herringshaw C.J., Stewart T.W., Thompson J.R. & Anderson P.F. (2011) Land use, stream habitat and benthic invertebrate assemblages in a highly altered Iowa watershed. American Midland Naturalist, 165, 274-293.

Hilsenhoff W.L. (1982) Using a biotic index to evaluate water quality in streams.

Technical Bulletin no. 132, Department of Natural Resources, Madison, Wisconsin.

Hilsenhoff W.L. (1987) An improved biotic index of organic stream pollution. The Great Lakes Entomologist, 20, 31–39.

Hilsenhoff W.L. (1988) Rapid field assessment of organic pollution with a family-level biotic index. Journal of the North American Benthological Society, 7, 65–68.

Iverson L.R. (1988) Land-use changes in Illinois, USA: the influence of landscape attributes on current and historic land use. Landscape Ecology, 2, 45-61.

King R.S., Baker M.E., Whigham D.F., Weller D.E., Jordan T.E., Kazyak P.F. et al. (2005) Spatial considerations for linking watershed land cover to ecological indicators in streams. Ecological Applications, 15, 137-153.

24

Kratzer E.B., Jackson J.K., Arscott D.B., Aufdenkampe A.K., Dow, C.L., Kaplan L.A. et al. (2006) Macroinvertebrate distribution in relation to land use and water chemistry in New York City drinking-water-supply watersheds. Journal of the North American Benthological Society, 25, 954-976.

Krohe J. Jr. (1997) The Illinois big rivers: an inventory of the region’s resources. Critical Trends Assessment Program, Illinois Department of Natural Resources, Springfield, Illinois.

Lenat D.R. (1993) A biotic index for the southeastern United States: derivation and list of tolerance values, with criteria for assigning water-quality ratings. Journal of the North American Benthological Society, 12, 279–290.

Lenat D.R. & Penrose D.L. (1996) History of the EPT taxa richness metric. Bulletin North American Benthological Society, 13, 305-307.

MacDonald L.H. & Coe D. (2007) Influence of headwater streams on downstream reaches in forested areas. Forest Science, 53, 148-168.

Marchal J. (2005) An evaluation of the accuracy of order-level biotic indices for southern Appalachian streams. Bios, 76, 61-67.

McCune B. & Grace J.B. (2002) Principal Components Analysis. In: Analysis of Ecological Communities, pp. 114-121. MjM Software Design, Gleneden Beach, Oregon.

Miller, K.M. (2004) Illinois Streams: Conditions and Trends. In Critical Trends Assessment Program 2003-04 Report. Illinois Department of Natural Resources, Springfield, Illinois.

Niyogi D.K., Koren M., Arbuckle C.J. & Townsend C.R. (2007) Stream communities along a catchment land-use gradient: subsidy-stress responses to pastoral development. Environmental Management, 39, 213-225.

Ohio EPA. (2003) The importance and benefits of primary headwater streams. Division of Surface Water, Ohio Environmental Protection Agency, Columbus, Ohio.http://www.epa.ohio.gov/portals/35/wqs/headwaters/HWH_import_jan2003.pdf

Panno S.V., Weibel C.P. & Li W. (1997) Karst Regions of Illinois. Illinois State Geographical Survey, Champaign, Illinois.

Pond G.J., Call S.M., Brumley, J.F. & Compton M.C. (2003) The Kentucky macroinvertebrate bioassessment index: derivation of regional narrative ratings for assessing wadeable and headwater streams. Division of Water, Water Quality

25

Branch, Kentucky Department for Environmental Protection, Frankfort, Kentucky.

Rawer-Jost C.J., Bohmer J., Blank J. & Rahmann H. (2000) Macroinvertebrate functional feeding group methods in ecological assessment. Hydrobiologia, 422/423, 225-232.

Schwegman J.E., Hutchison M., Shepherd W.M., Fell G.B., Paulson G., & White J. (1973) Comprehensive plan for the Illinois nature preserves system part 2: the natural divisions of Illinois. Illinois Nature Preserves Commission, Rockford, Illinois.

Stone M.L., Whiles M.R., Webber J.A., Williard K.W. & Reeve J.D. (2005) Macroinvertebrate communities in agriculturally impacted southern Illinois streams: patterns with riparian vegetation, water quality, and in-stream habitat quality. Journal of Environmental Quality, 34, 907–917.

Suloway L., Joselyn M. & Brown P.W. (1996) Inventory of Resource Rich Areas in Illinois: An Evaluation of Ecological Resources. Illinois Department of Natural Resources, Springfield, Illinois.

Sangunett B.M. (2005) Reference conditions for streams in the grand prairie natural division of Illinois. MS Thesis. University of Illinois at Urbana-Champaign.

Song M., Leprieur F., Thomas, A., Lek-Ang, S., Chon, T. & Lek, S. (2009) Impact of agricultural land use on aquatic insect assemblages in the Garonne river catchment (SW France). Aquatic Ecology, 43, 999–1009.

Vannote R.L., Minshall G.W., Cummins K.W., Sedell J.R. & Cushing C.E. (1980) The River Continuum Concept. Canadian Journal of Fish and Aquatic Science, 37, 130–37.

Vinikour W.S. & Anderson R.V. (1984) Macroinvertebrates and fishes of Trout Park nature preserve, Elgin, Illinois. Transactions of Illinois Academy of Sciences, 77, 95–102.

Voshell J.R. Jr. (2002) A Guide to Common Freshwater Invertebrates of North America. The McDonald & Woodward Publishing Company, Granville, Ohio.

Vondracek B., Blann K.L., Cox C.B., Nerbonne J.F., Mumford K.G., Nerbonne, B.A. et al. (2005) Land use, spatial scale, and stream systems: lessons from an agricultural region. Environmental Management, 36, 775-791.

Webb D.W., Wetzel M.J., Reed P.C., Phillippe L.R. & Harris, M.A. (1995) Aquatic Biodiversity in Illinois Springs. Journal of the Kansas Entomological Society, 68, 93–107.

26

Wetzel M.J. & Webb D.W. (2007) Springs of Illinois homepage. World Wide Web URL: http://www.inhs.uiuc.edu:80/~mjwetzel/SPOIL.hp.html. 26 August 2014.

Weigel B.M. (2003) Development of stream macroinvertebrate models that predict watershed and local stressors in Wisconsin. Journal of the North American Benthological Society, 22, 123-142.

Williams W.D. (1976) Freshwater Isopods (Asellidae) of North America. Biota of Freshwater Systems Identification Manuel 7, U.S. Environmental Protection Agency, Washington, D.C.

27

TABLE 1: Land use distributions in the watershed, 150m buffer, and 50m buffer spatial scales for each sample site.

Site % Agriculture

% Ag 150m

%Ag 50 m

%Forest %Forest 150m

%Forest 50m

%Urban %Urban150m

% Urban 50m

1 56.7 48.2 32.6 36.7 47.4 63.2 6.2 3.7 2.72 4.7 8.4 12.5 92.3 89.4 84.6 3.0 2.2 2.93 11.3 4.5 2.2 87.1 95.5 97.8 0.7 0.0 0.04 33.9 26.3 16.7 59.7 66.7 73.7 6.3 6.8 9.25 72.5 54.6 35.6 23.0 41.9 60.7 4.4 3.6 3.76 50.6 36.3 21.4 46.0 60.6 73.3 3.4 3.1 5.37 32.9 14.2 4.0 63.6 82.2 88.9 3.5 3.6 7.18 45.4 39.1 22.5 49.8 55.7 68.5 4.6 5.0 8.49 25.7 2.6 2.8 72.6 90.9 97.2 1.6 0.0 0

10 48.6 25.7 11.0 46.9 69.8 38.5 4.5 4.5 3.411 72.4 63.5 37.0 24.6 34.1 61.6 3.0 2.3 1.412 71.4 75.8 63.7 15.6 19.7 32.6 12.6 4.0 2.613 80.2 75.6 64.4 8.6 14.0 26.2 10.1 8.3 5.314 17.5 8.3 2.8 78.2 90.4 96.7 4.2 1.3 0.515 27.5 19.5 6.5 70.1 78.6 93.5 2.5 1.9 0.0

MEAN 43.4 33.5 22.4 51.7 62.5 70.5 4.7 3.4 3.5

TABLE 2: Pearson’s correlations and significance among the 5 functional feeding groups and land use distributions with n=15 (**significant at p<0.05; *significant at p<0.1).

% Agriculture (watershed)

%Agriculture (150m)

%Agriculture (50m)

%Forest (watershed)

%Forest (150m)

%Forest (50m)

Shredders -0.558 (0.031)** -0.538 (0.038)** -0.568 (0.027)** 0.572 (0.026)** 0.575 (0.025)** 0.327 (0.234)

Collector-Filterers

0.574 (0.025)** 0.577 (0.024)** 0.627 (0.012)** -0.593 (0.020)** -0.613 (0.015)** -0.398 (0.141)

Collector-Gatherers

0.236 (0.397) 0.125 (0.657) 0.033 (0.907) -0.222 (0.427) -0.140 (0.619) 0.093 (0.742)

Scrapers 0.487 (0.066)* 0.570 (0.027)** 0.518 (0.048)** -0.496 (0.060)* -0.558 (0.031)** -0.363 (0.184)

Predator -0.026 (0.927) -0.099 (0.726) -0.078 (0.782) 0.031 (0.913) 0.073 (0.795) 0.217 (0.438)

TABLE 3: Pearson’s Correlations and p-values for the 5 dominant families and land use distributions with n=15 (**=significant at p<0.05; *=significant at p<0.1)

%Agriculture (watershed)

%Agriculture (150m)

%Agriculture (50m)

%Forest (watershed)

%Forest (150m)

%Forest (50m)

Gammaridae -.581 (.023)** -.477 (.072)* -0.446 (0.095)* .586 (.022)** .501 (.057)* 0.344 (0.210)

Asellidae -.005 (.985) -.172 (.541`) -0.293 (0.289) .026 (.928) .199 (.476) -0.006 (0.984)

Hydropsychidae .579 (.019)** .603 (.017)** 0.653 (0.008)** -.617 (.014)** -.637 (.011) ** -0.419 (0.120)

Baetidae .048 (.864) -.071 (.802) -0.118 (0.675) -.036 (.899) .048 (.865) 0.181 (0.519)

Chironomidae .203 (.469) .140 (.618) 0.106 (0.707) -.208 (.456) -.148 (.599) -0.042 (0.881)

28

TABLE 4: Hilsenhoff’s family-level biotic index, Shannon-Weiner diversity (H’), richness, and evenness for all sites.

29

Site FBI H’ Richness

Evenness

1 4.4 1.53 9 0.672 4.1 0.07 2 0.103 4.9 1.10 6 0.564 4.6 1.37 8 0.595 5.4 1.28 6 0.726 4.9 1.32 7 0.687 5.6 1.37 6 0.708 4.7 1.57 9 0.719 4.1 1.09 8 0.50

10 6.0 0.73 3 0.6611 4.4 1.46 12 0.5912 4.7 1.02 6 0.5713 4.0 0.83 9 0.3814 5.6 1.48 12 0.5915 4.1 0.42 5 0.26

MEAN 4.8 1.11 7 0.55