Judith A. Gerlach Okay, Coordinator of the Difficult Run Project

Virginia Department of Forestry

The Difficult Run watershed is the largest watershed in Fairfax County, Virginia. It has a 58 square mile drainage. Difficult Run mainstem has ten tributaries (Fig. 1) which flow through a diversity of landscapes comprised of farms, more suburban areas and highly developed urban centers. The stream side forests in the Difficult Run watershed have become fragmented and the quality of the riparian forests have been compromised. As development has taken place vegetated landscapes have been replaced with impervious parking lots, walkways, roads and roofs. It is the purpose of this study to investigate the relationship between riparian forest buffer width, land use density, stream morphology and macroinvertebrate diversity.

Three streams from the Difficult Run watershed were selected for this study. The selection was based on the density of development within the stream drainage basins. The width of riparian forest buffer was also taken into consideration. Riparian forest buffers are an integral component of stream ecosystems and perform valuable functions that influence water quality within streams (USDA Forest Service 1991). Streamside tree canopy provides shade which moderates stream water temperatures (Boulton and Lake 1990). This is a very important attribute in urban/ suburban landscapes that have a high percentage of the landscape covered with impervious surfaces which produce heat that is transferred to streams during storm events (Schueler 1994). It has also been demonstrated that riparian forest buffers are the most effective filters for nutrient management and the removal of sediment from surface runoff (USDA Forest Service 1991). Litter and woody debris produced by streamside trees provide a food base for macroinvertebrate species which shred the leaves after they are somewhat decomposed by bacteria (Cummins et al. 1973).

Macroinvertebrate communities have been relied on as indicators of water quality, they are on the lower end of the trophic level within streams providing a base for the instream food chain (Firehouck and Doherty 1995). Species are categorized as pollution tolerant or pollution sensitive (intolerant) dependent upon their ability to survive in water that provides less than ideal habitat. Schueler (1994) reports research that correlates macroinvertebrate diversity and imperviousness, macroinvertebrates were replaced with more pollution tolerant species when imperviousness increased. The premise of this study is that the diversity of macroinvertebrate communities will decrease as imperviousness increases and riparian forest buffers decrease.

The stretch of Little Difficult Run analyzed is surrounded by a 250 acre park with mixed hardwood and pine forest. Beyond the park, the drainage basin is zoned at a maximum density of 1-5 acre lots. Angelico Branch another Difficult Run tributary is located in Kemper Park, an eight acre park surrounded by residential development zoned at a maximum density of 0.25 to 0.5 acre lots. A tributary of Colvin Run referred to as Greenmont Court was the third stream selected, it flows between two townhome communities. The communities have four - eight dwellings per acre, the drainage basin of this tributary is zoned residential interspersed with commercial. All of the stream stretches studied are headwater areas. The morphology and drainage basins are presented in Table 1.

An analysis of stream morphology was performed to determine the similarities and differences of the three streams. The method of stream analysis used employed techniques developed by Dave Rosgen (1994). Characteristics considered are the stream width, bed to bank ratio, the bed substrate and the valley slope. These measurements show the capacity and relative condition of the streams. All of the parameters are measured using a rod and level with the exception of the pebble count. An actual random count of 100 pebbles, rocks, or boulders is done and each particle is measured (Harrelson et al. 1994). The particles are placed in categories and the percent of each representative category is calculated. This analysis shows the classification of the bed substrate which can range from <2 millimeters (sand) to 2048 -4096 millimeters (very large boulder). Knowing the classification of the bed substrate is important because of the need for stable macroinvertebrate community habitat (Death 10). Erosion of bed substrate and sediment deposition create new geomorphic surfaces and have a disturbance effect on a stream ecosystem (Gregory 1991).

Macroinvertebrate Sampling

Macroinvertebrate samples were taken three times at each of the selected streams. The timing of the sampling was set up to be seasonal; winter, summer and late autumn. Hydrology, temperature and available food in each of these seasons differs (Gregory 1991). Deciduous leaf litter can remain instream for up to six months as well as macroinvertebrate maturity can take from one to five years (Wetzel 1983). Replicate sampling over a one year period is recommended by the Izaak Walton League of America (IWLA) Save Our Streams program and this was the sampling protocol used.

The technique involved makes use of a kick seine set just below a riffle section of stream. The seine is held by one person and another person moves up stream from the riffle turning rocks, stones and pebbles with their hands and rolling the substrate with their feet. The sampling can be performed by one individual using a small D-frame net and performing both the task of turning substrate and holding the net in place below the riffle.

The net is removed from the stream and the organisms captured are counted and identified. Organisms are classified as sensitive, somewhat-sensitive or tolerant according to the IWLA rating system, which relies on response to pollution in the stream. The Environmental Protection Agency’s recommended Rapid Bioassessment Protocol (RBP) evaluation was used as an extension of the IWLA rating. The RBP is a more intricate approach to water quality monitoring and requires species identification of macroinvertebrates to the biological Family level (Hilsenhoff 1988). A reference stream with a good water quality rating and located in the same physiographic province as the study sites is used to compare and calculate the rating for the study streams. A rating of <21% is a severely impaired rating, 29-72% is a moderately impaired rating and > 79% is a non-impaired rating.

Sorenson’s coefficient of similarity was applied to analyze the results of the macroinvertebrate sampling. This analytical procedure takes into consideration the total number of species collected in each sampling unit and the number of species the sampled sites have in common. The output is in the form of a percent and displayed as clustering on a dendrogram. The highest value is 1.00 or 100% for those sample units that are identical in the species they have in common. Sorenson’s coefficient of similarity analysis shows the macroinvertebrate similarities the three study streams have in common with the reference stream and with each other.

Stream classification is helpful in determining the vulnerability of a stream to increased stream bank erosion and widening of the channel or entrenchment of the bed. The value that stream morphology analysis has in this study is the ability to predict the stability of macroinvertebrate habitat in each stream analyzed. The findings of the analysis performed are displayed in Table 1. The stream beds ranged from sandy - coarse gravel for Angelico Branch, large cobble - small boulder at the Greenmont Court site. Little Difficult Run bed structure was a coarse gravel - small boulder. Of the three streams, Greenmont Court has the most stable stream bed, it has actually downcut to the point that bedrock is visible in many parts of the channel. The stream with the least stable bed structure is Angelico Branch which has a sandy to coarse gravel bed.

The width/depth ratio is another morphologic feature important to macroinvertebrate species. Little Difficult Run stream has a low width/depth ratio, the interpretation of this is that the banks are not as steep revealing a certain amount of stability rather than undercutting and scouring. The stream at Greenmont Court has the highest width/depth ratio which is interpreted to mean that the banks are steep and depending on the soil structure can be unstable. Another stream element considered in this study is the temperature of the stream water versus the ambient air temperature for each of the stream sample sites. The air temperature was 12.7° C at both Angelico Branch and Greenmont Court sites, however, the instream water temperatures were 6.4° C and 7. 7° C respectively. At the Difficult Run site the ambient air temperature was 13.1° C and instream water temperature was 5.5° C. These temperatures are consistent with air and water temperatures reported by water quality monitors in the local Soil and Water Conservation District monitoring program (Penney 1997). Dissolved oxygen levels are dependent on temperature and the rate of decay for leaf litter is also temperature dependent. It should be noted that historical chemical monitoring data produced by the Fairfax County Health Department (1989-1994) reveal some of the lowest nitrogen and phosphorous levels are present in the Difficult Run watershed as well as consistently high dissolved oxygen levels. Stream monitors sampling different sites in the Difficult Run watershed reported fair to moderate water quality (Penney 1997).

Samples taken from the three streams over a one year period of time contained representatives from 13 Families of macroinvertebrates. The sample from the reference stream contained 14 Families. Of the three study streams, the maximum number of macroinvertebrate Family representatives at any one stream was 9 at each of the streams, but at different times. There are many variables that influence these numbers and they will be addressed in the discussion section. The list of Families and their presence in the respective streams can be found in Table 2.

The cluster analysis using Sorenson’s coefficient of similarity shows both Greenmont Court and Angelico Branch to be most closely related to the reference stream in terms of macroinvertebrate community structure. However, the percent similarity to the reference stream for both Angelico Branch 2 (AB2) and Greenmont Court 1(CRT) is only 0.522 or 52%. Cluster analysis results are presented in Table 3. This can be related to the Rapid Bioassessment Protocol evaluations. The ratings as mentioned earlier are: < 21%, severely impaired; 29-72%, moderately impaired and > 79%, non-impaired. Little Difficult Run received the lowest rating of severely impaired (21%) in January, 1997. The highest rating, moderately impaired (43%) was also for Little Difficult Run in August, 1997, and Angelico Branch received a 43% rating in December 1997. Greenmont Court remained consistent for all three sampling periods with a moderately impaired rating of 36%.

The streams that have the highest coefficient of similarity (0.80) are Angelico Branch (AB1) and Little Difficult Run (LDR1). Angelico Branch (AB1) and Greenmont Court (CRT) have a coefficient of similarity of 0.70. This demonstrates which of the stream samples at each stream had similar macroinvertebrate species. The dendrogram (Fig. 2) displays the cluster relationships between the various stream samples.

The relationship between land use and water quality is not a concept to be easily explained or understood. There are numerable variables that influence water quality. An attempt is made in this study to isolate land use density and riparian forest buffers as influential elements that affect water quality. Macroinvertebrate community structure is the tool used to evaluate the influence of land use density and riparian forest buffers on water quality. It all seems very simple, but quickly becomes complicated because stream ecosystems are just that, systems with many intricate elements.

The first element considered is the influence of stream morphology. Streams that are unstable and are undergoing rapid change carry heavy sediment loads and have dynamic movement of the banks and bed. The volume and velocity of the water carried by the stream will determine just how much structural change will take place. In urban situations hydrology is very flashy and results in large volumes and swift velocities, such is the case for the three streams in this study. Schueler (1994) reports that macroinvertebrate diversity is inversely proportional to the amount of imperviousness in a watershed.

Of the three streams studied, Greenmont Court has the largest drainage area of imperviousness, however, the macroinvertebrate samples collected proved to be the most closely related to the samples of the good quality reference stream. Greenmont Court also has the poorest quality riparian buffer. This is not the relationship expected at the onset of this study. One element very unique to Greenmont Court, not found in the other streams, is the fact that the bed structure and general morphology of the stream indicate that it has degraded to a point of stability. The large bedrock on the stream bottom and the cobble are quite stable and possibly create good stable habitat for macroinvertebrates to colonize. Under the stress of high volume and velocity flows, there is protection under the large rocks. A study by Death and Winterbourn (1995) credit an increase in macroinvertebrate diversity to more stable stream structure. Although analysis of Little Difficult Run stream morphology reveals that the banks are quite stable in the stretch analyzed, the bed material, coarse gravel - small boulder can be moved by the sheer stress of large volume and fast velocity flows. When the bed material rolls around, macroinvertebrates are buried, washed out and generally lost (Schueler 1994). The same would be true for the morphology of Angelico Branch, but there is an additive condition in this stream. In addition to an unstable bed, the banks are extremely scoured and the stream carries a considerable amount of sediment. The sediment deposits not only bury macroinvertebrates, but the movement of the sand and gravel, scrape the epidermis of the organisms making them vulnerable to disease. Another point to consider is that many macroinvertebrate species have gills that get clogged by sediment, making oxygen absorption and survival difficult. Death (1995) reports a decrease in filter feeding macroinvertabrates at disturbed, unstable stream sites.

Little Difficult Run stream had the greatest number of macroinvertebrates in the January, 1997 sample, 202, representing 9 Families. The number is higher than the sample taken from the reference stream which had 117, representing 14 families. The problem with the quality of the sample collected from Little Difficult Run is the high number of tolerant species, particularly black fly larvae. The stream was rated as severely impaired at this time. Greenmont Court had a total number of 21 macroinvertebrates a considerably lower number than Little Difficult Run, but the proportion of pollution sensitive organisms was high (7/8 Families) and a moderate impairment was recorded. Rainfall for the month of January equaled 5.48 centimeters (cm.) The August sample numbers diminished for Little Difficult Run, 68 total organisms, representing 6 Families. The total organisms for Greenmont Court differed by one organism, 20 for August versus 21 for January. Rainfall for the month of August was 13.51 cm., much higher than for the month of January. The rainfall recorded for August would produce flashy hydrology in the urban streams that were studied. It has been observed that 5.17 cm. of rain in a 24 hour period results in overbank flooding in the Difficult Run watershed. The point to be made is that stream bed stability is a likely explanation for the variance in numbers of organisms in Difficult Run from January versus August, whereas, samples from Greenmont Court with a predominantly stable bed structure only varied by one organism.

All of the samples collected in December, 1997 were lower than at other times. Rainfall for the month of December was 8.17cm. Although actual totals were lower, the quality of the samples all produced moderately impaired ratings. Family representation was higher (9 Families) for Little Difficult Run and lower for Greenmont Court (only 3 Families were represented). Angelico Branch followed the same trends as the other streams with fluctuation in total numbers of organisms and Family representation with the lowest numbers being recorded for December and the highest numbers recorded in January.

The riparian forest buffer widths along the study stretch of Little Difficult are ideal, both sides are >30 meters. Angelico Branch has 9.0 meters on one side and 12.10 meters on the other, Greenmont Court does not meet the minimum desired buffer width of 10.6 meters on either side of the stream. Although there was less fluctuation in the number of macroinvertebrate species and the quality of the samples taken at Greenmont Court were rated equal or higher than Little Difficult Run it must be noted that the sheer numbers of organisms were much higher for Little Difficult Run for all three samples. The two positive factors that can account for this are the riparian forest buffer and land use density.

The shade provided by tree canopy influences water temperature, particularly in August when ambient air temperatures are higher. The plentiful leaf litter and woody debris provide food, cover and nesting materials for many macroinvertebrate species. Caddisflies in particular rely on the woody debris to form casings to protect eggs (Wetzel 1983). Although all of the stream buffers observed can provide these materials, Little Difficult Run has the greatest potential to do so, because of its extensive riparian buffers.

There are several points of interest presented by this study. A major concern is the reliance on one parameter or tool for evaluation of water quality in urban streams. The interdependence of riparian forest buffers, land use and stream morphology is a complicated issue, but certainly influential on the structure of macroinvertebrate communities. If macroinvertebrate monitoring is the tool of choice for stream water quality monitoring, stream morphology, land use and riparian forest, buffer quality should be included in the study. It is not good enough to simply collect and interpret macroinvertebrate samples. The reasons for their presence and absence transcends the whole ecosystem which includes the drainage basin, sources of food, water temperature, shelter and nesting substrate as well as habitat stability.

The importance of urban riparian forest buffers was discussed in the context of typical stream ecosystems. In this study a stream with a healthy riparian forest buffer produced sufficient litter to support numbers of macroinvertebrates, but the numbers of sensitive organisms were lacking, evidence of poor water quality. The stream with a minimal riparian forest buffer produced fewer numbers of macroinvertebrates, but those present were of a sensitive rating, indicating a moderate water quality. Since historical data and current water quality monitoring data records (chemical and physical) determine water quality to be moderate to fair in Difficult Run watershed, it is not evident that the water quality is so poor that sensitive organisms cannot exist. It is suggested that stream morphology in urban landscapes is an integral component of macroinvertebrate community sustainability. It was demonstrated that a stream with stable bed structure will support a more sensitive suite of macroinvertebrates, the organisms do not have to contend with unstable bed structure and abrasive sediment. More tolerant organisms can withstand these less than ideal conditions.

A shortcoming of this study is the length of time invested. It is difficult to draw firm conclusions with one year’s data set. Under the circumstances, trends cannot be set and statistical significance is difficult to establish. Although the stream morphology and riparian forest buffers change slowly, other factors such as seasonal hydrology fluctuates regularly, climate affects food production, and reproductive efforts can be disrupted. A longer study period that includes litter accumulation data and more frequent sampling would add more balance to the study.

The Virginia Coastal Resources Program has granted funds for this study as a part of the Difficult Run Riparian Project. Assistance with surveying and sampling was received from the following individuals: Sam Austin, Larry Dunn, Maureen Kellogg and Barbara and Megan White. Their help and dedication under less than ideal circumstances and weather conditions is appreciated.