View the SRWP Facebook Page  View the SRWP Twitter Feed  Signup for the SRWP Email Newsletter

3.2.3 Benthic Macroinvertebrates Community Structure


Goal: B. Protect and enhance native aquatic and terrestrial species, especially sensitive and at-risk species and natural communities

Objective: 2. Protect and enhance native aquatic invertebrate communities

WAF Attribute: Biotic Condition

What is it?


Freshwater benthic macroinvertebrates (BMI) are small animals without backbones that live on and under submerged rocks, logs, sediment, debris and aquatic plants during some period in their life. BMI include the immature forms of aquatic insects such as mayfly and stonefly nymphs, as well as crustaceans such as crayfish, molluscs such as clams and snails, and aquatic worms.

Many BMI are highly sensitive to changes in their aquatic environment and thus can act as continuous monitors of the condition of the water they live in. Human activities that interfere with or disrupt natural processes in a watershed can have significant impacts on the types and numbers of BMI that live there. We can assess the biological health of a watershed by looking at the types of BMI that either thrive or do not thrive in it. BMI represent an extremely diverse group of aquatic animals, with a wide range of responses to stressors such as organic pollutants, sediments, and toxicants. If only a few types of benthic macroinvertebrates live there, or if the macroinvertebrates present are primarily ones that are insensitive to disturbed systems, there is some kind of problem present.

Why is it Important?

The best way to assess the ability of a watershed to support living things is to look at those living things. Unlike chemical monitoring, for example, which provides information about water quality at the time of measurement, monitoring of living organisms (biomonitoring) can provide information about past and/or episodic pollution and the cumulative effects of a suite of watershed impacts. BMI represent ideal biomonitors for assessing the overall health of watersheds for a number of reasons:

  • They are widespread
  • They are easy to collect and identify
  • They are relatively sedentary and long-lived, so reflect the longer-term effects of activities within their watershed
  • Some species of BMI are highly sensitive to pollution

BMI-related metrics (e.g., taxa richness and diversity, specific taxa pollution sensitivities/tolerances, etc.) have been used by varied US agencies for many years as "bioindicators" of water quality, providing integrated information on toxic chemical concentrations, DO levels, nutrients, and habitat quality. Beyond their usefulness as bioindicators BMI are themselves an important part of aquatic food chains, especially for fish. Many BMI feed on algae and bacteria, which are on the lower end of the food chain. Some shred and eat leaves and other organic matter that enters the water. Because of their abundance and position as "middlemen" in the aquatic food chain, BMI play a critical role in the natural flow of energy and aquatic nutrients in streams, lakes and wetlands.

What is the target or desired condition?

The desired condition is to a have rich and diverse community of BMI across the watershed, reflecting maintenance of natural river/stream processes and clean water that allows persistence of particularly sensitive species. A variety of BMI metrics (e.g., diversity, sensitive taxa, functional feeding groups, rare species, etc.) can all be used to provide some assessment of watershed condition and the status of aquatic invertebrate populations. One group of BMI, the "EPT" taxa (Ephemeroptera - mayfly, Plecoptera - stonefly, Trichoptera - caddisfly) are often used because they decrease in richness in the presence of pollution. We selected two key BMI metrics (Total Taxa Richness and EPT Taxa Richness) that are commonly used for assessments of aquatic macroinvertebrate communities and that could also be generated easily using readily available agency monitoring data, and supplemented using information from volunteer groups and university programs that have undertaken BMI sampling in the watershed.

What can influence or stress condition?

Some BMI taxa require very good water quality, whereas others tolerate a wide range of environmental conditions. Although BMI can move about to some extent and even drift downstream, they generally cannot move quickly to avoid adverse conditions. Deteriorating water and/or habitat quality and pollutants can be expected to kill or at least stress less tolerant BMI taxa and encourage other more tolerant taxa to proliferate. Once BMI are lost from a waterway, they may take years or decades to recover, both because the system is recovering and because they would have to be recruited from elsewhere.

What did we find out/How are we doing?

In general, most Feather River subwatersheds would seem to currently be in fair condition based on the key BMI metrics that were evaluated across 2004-2008 (Table 1 and Figure 1). Not all subwatersheds had data in 2008, but many had at least some data available between these five years. Average subwatershed condition scores were based on Total Taxa Richness (scoring based on comparison with the most taxa-rich site in the Feather River database) ranged from 30 for the Lower Yuba to 58 for the Middle Yuba (the Lower Yuba score was however based on only 2 samples).

Average scores for EPT Taxa Richness (based on defined target thresholds) ranged from 0 to 100 across the subwatersheds (Table 1). While most subwatersheds had fairly highs scores for the EPT metric Deer Creek, East Branch North Fork Feather and Lower Yuba all had scores of zero. However, Deer Creek values were based only on Level 2 sampling (taxonomic identification at the level of "Family") which would lower its score, and Lower Yuba was based on only two samples. The lower score for East Branch North Fork Feather suggests that further work should be directed there to better determine if this EPT result suggests some current impairment, particularly as our supporting trend analysis indicated a significant decline in Total Taxa Richness for this subwatershed since the mid-1990's, a trend that was also displayed by the Middle Fork Feather. Deer Creek conversely demonstrated a significant positive trend in Total Taxa Richness between 2000 and 2008; this trend is however based on Level 2 sampling only (see How sure are we about our findings section).

Table 1: Report Card scores for Benthic Macroinvertebrate communities for subwatersheds

Goal Measurable Objective Subwatershed Total Taxa Richness Score EPT Taxa Richness Score Trend
B. Protect and enhance native aquatic and terrestrial species, especially sensitive and at-risk species and natural communities 2) Protect and enhance native aquatic invertebrate communities NFF 48 93 ?
EBNFF 45 0
MFF 48 47 down_arrow_angled.png
LF n/a n/a ?
NY 51 97 ?
MY 58 100 ?
SY 52 91 ?
DC 36 0 up_arrow_angled.png
LY 30 0 ?
UB 44 37 ?
LB n/a n/a ?

Figure 1: Subwatershed distribution of Total BMI Taxa richness scores — number of different types of BMI present. Not all sampling locations are represented on the map due to missing spatial coordinates for 2 sites in the Upper Bear and 1 site in the North Yuba.


Temporal and spatial resolution

Seasonal sampling of benthic macroinvertebrates has occurred at varied sites (Table 2) in the Feather River Watershed as far back as 1995, but early sampling was restricted to the North Fork Feather (which represents the longest time series for the basin). More intensive and widespread sampling has occurred subsequent to 1999, with some level of BMI data available from different monitoring programs for at least nine Feather River subwatersheds (Figure 1).

Table 2: Number of samples collected each year for BMI in Feather River subwatersheds since 1995. Note that some of these are replicate samples within a site, and are not necessarily independent sites. Sites represent a mix of Level 2 and Level 3 sampling efforts.

Subwatershed 1995 1997 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
Deer Creek 0 0 0 7 20 22 12 23 24 24 14 25
East Branch NF Feather 0 0 70 2 70 0 38 1 0 25 21 0
Lower Yuba 0 0 0 0 0 0 0 0 0 1 1 0
MF Feather 0 0 21 6 26 0 16 7 2 6 2 0
Middle Yuba 0 0 0 0 0 0 0 3 3 2 4 2
NF Feather 3 3 20 0 20 0 12 2 0 21 6 0
North Yuba 0 0 0 0 0 0 0 0 1 7 1 0
South Yuba 0 0 0 0 0 0 0 18 6 6 7 4
Upper Bear 0 0 0 0 0 0 0 0 0 0 5 3
Unknown 0 0 0 0 0 0 0 3 2 18 21 0
Total 3 3 111 15 136 22 78 57 38 110 82 34

How sure are we about our findings?

The data used for this indicator come from a mix of different sampling programs across the Feather River Watershed, some that are agency-based and some that are volunteer-based, with associated "measurement error" uncertainties related to consistency of sampling protocols employed, processing thoroughness and correct identification of BMI taxa (i.e., Level 3 analyses vs. Level 2 analyses — Richards and Rogers 2006). We have attempted to control for this somewhat by using BMI metrics that may be easier to evaluate consistently and do not necessarily require a high level of invertebrate identification expertise (although the value of this information is enhanced with higher levels of taxonomic resolution). In general, however, evaluating the status and trends of BMI can be very challenging and requires consistently sampled and analyzed data over time as macroinvertebrate populations are naturally highly variable both spatially and temporally (seasonally and annually) (USEPA 2006). Our use of a mix of Level 3 and Level 2 sampled sites within our scoring increased both the spatial and temporal extent of our comparative analyses but also increased the uncertainty around the interpretation of pooled results. We do know that this uncertainty has a directional bias however, as subwatershed evaluations which incorporate Level 2 samples will be biased low (that is they will be conservative estimates, more likely to judge water quality to be poorer).

An additional element of uncertainty for our comparisons relates to our interpretation of the overall sampling frame. For our indicators we have opportunistically mined and pooled BMI data that have been collected by a number of different agencies/groups with possibly different objectives, target populations and sampling frames (e.g., they could have focused on different stream orders, types of streams or times of year). Without specific knowledge of different sample design elements, we have made assumptions about the probability of selecting each site and the appropriate weighting of the observation. As a default for our analyses we have treated each of the BMI sites as though they were a simple random sample from the overall population of streams. This may well be incorrect.

Finally, although BMI community composition is an informative indicator based on the presence of certain types of BMI, the absence of specific types of BMI at specific locations is difficult to interpret. Short-term perturbations may remove sensitive, or most, types of BMI from a location and even after the location has recovered physically and chemically, it may be a long-time before those types return.

Table 3: Summary of BMI condition and trend based on Total Taxa and EPT Taxa by subwatershed.

      Report Card Score    
Subwatershed Average Taxa Richness Average EPT Taxa Richness Total Taxa EPT Taxa Sampling Effort Trend
Deer Creek 22.0 10.7 36 0 C up_arrow_angled.png
EB NF Feather 27.2 11.9 45 0 B down_arrow_angled.png
Lower Yuba 18.0 11.5 30 0 A ?
MF Feather 29.4 15.3 48 47 B down_arrow_angled.png
Middle Yuba 35.1 20.9 58 100 A ?
NF Feather 29.5 18.5 48 93 B ?
North Yuba 31.2 18.8 51 97 B ?
South Yuba 31.7 18.4 52 91 B ?
Upper Bear 25.6 14.5 44 37 B ?

As BMI sampling has been inconsistent across the years the subwatershed values for these metrics and derived "report card" scores are the average of the annual estimates from the last 5 years of available data (2004-2008). Sampling effort indicates whether information to derive the score was based on: (A) Level 3 sampling effort only; (B) a mix of Level 3 and Level 2 sampling; or (C) Level 2 sampling effort only. Full statistical summaries for the BMI metrics ( means, 95% C.I., SE, minimum and maximum values for the annual estimates between 2004-2008, as well as for 1999-2004, are presented in the Analysis section - Tables 4 and 5).

Technical Information

Data Sources

Data for our BMI community metrics were compiled from past sampling that has been undertaken within varied Feather River subwatersheds by the USFS, DWR, FRCRM, Friends of Deer Creek, UC Davis Center for Watershed Sciences, Wolf Creek Community Alliance, and SYRCL. Data from specific sites were assigned subwatershed identifiers, permitting aggregation of information to that scale.


BMI data are available from a variety of groups that have undertaken aquatic sampling in streams throughout the Feather River Watershed. Some of this sampling has been led by government agencies that have undertaken rigorous Level 3 analysis of BMI samples (Harrington 2003, Richards and Rogers 2006), with identification to the taxonomic level of "Genus/species". Other sampling in the Feather River Watershed has been undertaken by citizens groups and university projects which have generally undertaken only Level 2 analyses, with taxonomic identification at the level of "Family". Given differences in how groups have analyzed BMI data historically we sought to capture some simple yet informative BMI metrics that could provide information about the status of BMI communities across the watershed using information from both Level 3 and Level 2 sampling. Data from multiple sources were therefore pooled as possible and summarized for our scoring using two standard metrics of BMI community condition:

1) Total Taxa Richness is the total number of macroinvertebrate taxa (Family/Genera), insect and non-insects at a sampling site. Total Taxa Richness provides an index of the general health of the BMI community and is expected to be higher in subwatersheds with better habitat diversity, suitability, and water quality (Plafkin et al., 1989). Absent a defined California standard for desired BMI total taxa richness in aquatic systems, or alternatively readily available information from a pristine (reference) watershed for comparison, we instead used the highest Total Taxa Richness value obtained at any of the historical sampling sites as a "good" target and gave this the highest score (100); a Total Taxa Richness value of zero was given the poorest score (0). We then used an interpolated straight line function between this range of values to score the average Total Taxa Richness found across the subwatersheds.

To take full advantage of the range of BMI data available from all sampling undertaken across the Feather River Watershed we combined both Level 3 and Level 2 sampling efforts to generate subwatershed averages for Total Taxa Richness and compared this to our defined target (which was based on the best Level 3 sampled site) for the Feather River Watershed. We recognize, however, that this is a biased low interpretation of the Level 2 sites (i.e., more likely to conclude that water quality is poorer in relation to the defined target criteria). Observation of 12 families using the Level 2 sampling implies a minimum of 12 genera, but possibly more. However, without Level 3 sampling we cannot be sure how many more. Hence we use the most conservative estimate and assume only the relative quality of the sampling information available to inform taxa status. For each Feather River subwatershed we indicated if the level of effort at BMI sites across the years was: A) all Level 3 sampling, B) a mix of Level 3 and Level 2 sampling, or C) all Level 2 sampling.

2) EPT Taxa Richness is the total number of EPT taxa (Family and Genera) found within the insect orders Ephemeroptera (mayflies), Plecoptera (stoneflies) and Trichoptera (caddisflies). These are insect orders considered particularly sensitive to pollution and habitat disturbance so that the presence and abundance of EPT taxa provides an indication of overall water quality. Sites at which EPT taxa are more prevalent are considered to have cleaner water and provide better habitat conditions. EPT Richness is one of the most commonly used biometrics used to describe macroinvertebrate community structure and to assess possible stream degradation (Resh and Jackson 1993). Although EPT Taxa Richness would be expected to vary regionally, Harrington et al. (1999) suggest a standard (based on Level 3 sampling) that could be used for California streams, where EPT Taxa Richness > 19 indicates good water quality, 12-19 indicates fair water quality, and < 12 indicates poor water quality. We have adopted this standard as a target for desired BMI condition where subwatersheds with an average EPT Taxa Richness of < 12 were scored as 0, those with > 19 were scored as 100 and those with EPT values between 12 and 19 were scored as an extrapolated straight line function between 12 and 19. To take full advantage of the range of BMI data available from sampling undertaken across the Feather River Watershed (including Level 2 sampling efforts) we used the defined EPT threshold criteria consistently across the Feather River Watershed but combined all Level 2 and Level 3 sampling efforts to generate subwatershed averages. Recognizing that the data relating to EPT status would be poorer at Level 2 sites relative to Level 3 sites we used the same categorizations of historical sampling effort in each subwatershed as described for the Total Taxa Richness metric.

For each of our BMI metrics data from all individual sampling sites were aggregated to produce an average annual estimate for each subwatershed in which sampling had occurred. Where there was more than one record per site per year, these were first averaged. For Total Family Richness estimates and confidence intervals were compared to the highest value reported at any site within the historical dataset (61 observed taxa, Sept. 20, Canyon Creek, South Yuba), for EPT Taxa Richness estimates and confidence intervals were compared to a defined threshold of water quality condition. Estimates and associated confidence intervals were transformed to a 0-100 scale. The 95% confidence interval for the estimate is presented, along with the minimum, maximum, and number of observations (n, Table 3). Depending on how sites were selected, it may be better in the future to first average the results by stream and then average the streams within a subwatershed, but at this point too little replicate information is available per stream to make this a worthwhile approach to consider.

Summaries of statistics for BMI metrics and derived scores are presented for Total Taxa Richness and ETP Taxa Richness in Table 4 and Table 5 respectively. While our assessments determined that there was large site-to-site variability in BMI metrics, the annual means for a subwatershed did not vary as much.

Table 4. Average Total Taxa Richness values and derived scores in each of four subwatersheds in two time
periods of comparison (1999-2003 vs. 2004-2008)

Subwatershed Mean (Total Taxa) 95% C.I. N (years) Minimum value Maximum value Score (mean) Score (95% C.I.)
Deer Creek (1999-2003) 19.6 2.4 4 17.7 21.3 32 3.9
Deer Creek (2004-2008) 22.0 1.5 5 20.1 23.0 36 2.4
EB NF Feather (1999-2003) 32.0 0.7 4 31.4 32.5 53 1.2
EB NF Feather (2004-2008) 27.2 10.4 3 24.1 32.0 45 17.1
Lower Yuba (1999-2003) NA NA NA NA NA NA NA
Lower Yuba (2004-2008) 18.0 50.8 2 14.0 22.0 30 29.5
MF Feather (1999-2003) 36.4 8.3 4 28.6 39.4 60 13.6
MF Feather (2004-2008) 29.4 20.6 4 17.2 45.0 48 33.7
Middle Yuba (1999-2003) NA NA NA NA NA NA NA
Middle Yuba (2004-2008) 35.2 1.8 5 32.7 36.3 58 2.9
NF Feather (1999-2003) 38.3 6.4 3 35.3 40.3 63 10.6
NF Feather (2004-2008) 31.2 22.8 3 19.5 37.5 48 37.3
North Yuba (1999-2003) NA NA NA NA NA NA NA
North Yuba (2004-2008) 31.2 38.4 3 20.7 49.0 51 51.2
South Yuba (1999-2003) NA NA NA NA NA NA NA
South Yuba (2004-2008) 31.7 4.6 5 28.0 37.3 52 7.5
Upper Bear (1999-2003) NA NA NA NA NA NA NA
Upper Bear (2004-2008) 25.6 76.0 2 19.7 31.6 42 42.0

We used these five year blocks to represent reasonable time intervals for BMI reporting purposes. Subwatersheds with NA indicated did not have data collected within the particular time block. Formal trend analysis was undertaken for Total Taxa Richness for a subset of the subwatersheds that had greater than five years of data across the years (not necessarily consecutive, see Trends Analysis section). “95% C.I.” refers to 95% confidence intervals.

Table 5. EPT Taxa Richness values and derived scores in each of four subwatersheds in two time periods of comparison (1999-2003 vs. 2004-2008)

Subwatershed Mean (EPT Taxa) 95% C.I. N (years) Minimum value Maximum value Score (mean) Score (Lower 95% C.I.) Score (Upper 95% C.I.)
Deer Creek (1999-2003) 10.2 1.0 4 9.6 11 0 0 0
Deer Creek (2004-2008) 10.7 0.6 5 10.3 11.5 0 0 0
EB NF Feather (1999-2003) 15.6 5.7 4 11.0 19.5 51 0 100
EB NF Feather (2004-2008) 11.9 4.2 3 10.0 13.2 0 0 59
Lower Yuba (1999-2003) NA NA NA NA NA NA NA NA
Lower Yuba (2004-2008) 11.5 31.8 2 9.0 14.0 0 0 100
MF Feather (1999-2003) 21.3 4.3 4 17.3 23.0 100 72 100
MF Feather (2004-2008) 15.3 15.3 4 4.5 25.0 47 0 100
Middle Yuba (1999-2003) NA NA NA NA NA NA NA NA
Middle Yuba (2004-2008) 20.9 3.7 5 17.0 24.0 100 74 100
NF Feather (1999-2003) 21.8 6.5 3 18.8 23.5 100 47 100
NF Feather (2004-2008) 18.5 19.3 3 11.0 26.5 93 0 100
North Yuba (1999-2003) NA NA NA NA NA NA NA NA
North Yuba (2004-2008) 18.8 16.1 3 13.4 26.0 97 0 100
South Yuba (1999-2003) NA NA NA NA NA NA NA NA
South Yuba (2004-2008) 18.4 3.4 5 16.8 23.3 91 42 100
Upper Bear (1999-2003) NA NA NA NA NA NA NA NA
Upper Bear (2004-2008) 15.0 41.3 4 12.3 18.0 43 0 100

Trend Analysis

We attempted to fit a trend line to any dataset with at least five years of data (these did not have to be consecutive). This was done using simple linear regression with “year” as the independent variable and “BMI metric” as the dependent variable. We then tested the hypothesis that the slope of the line was equal to zero (i.e., no trend). Six of the subwatersheds had sufficient data to fit a trend line for the BMI metrics. Three of these subwatersheds displayed significant trends for Total Taxa Richness (Table 5), one of these (Deer Creek) had a positive (upward) trend (Figure 2C), two of these (East Branch North Fork Feather, and North Fork Feather) displayed negative (downward) trends (Figure 2A). For the other subwatersheds there was no evidence of a significant directional (i.e., non-zero) trend. None of the six subwatersheds displayed any significant trends for EPT Taxa Richness (Table 6).

Table 6. Linear regression estimates for Total Taxa Richness in the six Feather River subwatersheds with at least five years of BMI data. Significant regressions are indicated by a bold asterisk.

  Deer  Creek EB North Fork Feather Middle Fork Feather Middle Yuba North Fork  Feather South Yuba
Estimate 0.536 -0.969 -1.408 0.658 -1.462 1.445
SE 0.134 0.293 1.264 0.362 0.506 1.050
T value 3.989 -3.305 -1.114 1.821 -2.888 1.376
P value 0.005 * 0.021* 0.308 0.166 0.028* 0.262
Trend Positive Negative No indication of trend No indication of trend Negative No indication of trend

Table 7. Linear regression estimates for EPT Taxa Richness in the six Feather River subwatersheds with at least five years of BMI data. No significant trends were evident.

  Deer  Creek EB North Fork Feather Middle Fork Feather Middle Yuba North Fork  Feather South Yuba
Estimate 0.077 -0.492 -1.406 -0.133 -0.551 0.75
SE 0.076 0.443 0.85 1.091 0.407 0.916
T value 1.01 -1.109 -1.654 -0.122 -1.354 0.819
P value 0.346 0.318 0.149 0.91 0.225 0.473
Trend No indication of trend No indication of trend No indication of trend No indication of trend No indication of trend No indication of trend

Figure 2. Annual observations and significant linear trend lines for  (A) East Branch North Fork Feather and (B) North Fork Feather (negative trends) and Deer Creek (positive trend)





Harrington, J. M. 2003. California stream bioassessment procedures. California Department of Fish and Game, Water Pollution Control Laboratory, Rancho Cordova, California.

Harrington, J.M., P. Ode, A. Montvalo, D. Post, C. Sheehy, and M. Dawson. 1999. An Index of Biological Integrity for first to third order Russian River tributary streams. California Department of Fish and Game, Office of Spill Prevention and Response Water Pollution Control Laboratory, Rancho Cordova, Ca.

Plafkin, J. L. , M.T. Barbour, K.D. Porter, S.K. Gross, and R.M. Hughes. 1989. Rapid bioassessment protocols for use in streams and rivers: Benthic macroinvertebrates and fish. USEPA/444/4-89/001. Office of Water Regulations and Standards, U.S. Environmental Protection Agency , Washington, D.C.

Resh, V. H. and J. K. Jackson. 1993. Rapid assessment approaches in benthic macroinvertebrate biomonitoring studies. Pages 195-233 in: D. M. Rosenberg and V. H. Resh (eds.), Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York.

Richard, A. B., and D.C. Rogers. 2006. List of freshwater macroinvertebrate taxa from California and adjacent states including standard taxonomic effort levels. Southwest Association of Freshwater invertebrate Taxonomists (SAFIT).

U.S. Environmental Protection Agency (USEPA). 2006. Benthic macroinvertebrates in wadeable streams. Washington, DC.