Biosolids as a Roadside Soil Amendment

Vegetation is a critical component of the roadside environment. It supplies services that include stabilizing soil, filtering runoff, trapping debris before it enters drains and water supplies, and providing visually appealing scenery for roadway users. These services are threatened when soil is unable to provide an environment suitable to the needs of perennial vegetation, leading to annual, often invasive, species that outcompete and thrive in dry, low nutrient and disturbed habitats dominating the environment. As a means of improving the availability of the macronutrient nitrogen, a limiting nutrient for many plants, we amended roadside soil in Saunderstown, Rhode Island with seven different products: five stabilized biosolids (products whose organic material consisted only of sewage sludge) and two composts (products whose organic material consisted either partially or completely of yard waste). The biosolids were applied at rates of 48 kg N/ha, 144 kg N/ha and 288 kg N/ha (nitrogen available in the first year) while the composts were applied at rates of 15%, 30% and 45% v:v of the first 15 cm of the soil. Control plots received no fertilization beyond the 95.2 kg/ha of 19-19-19 fertilizer included in the hydroseeding mixture. All plots were hydroseeded in September 2012 with a seed mixture of Kentucky bluegrass (Poa pratensis), perennial ryegrass (Lolium perenne) and red fescue (Festuca rubra). Quality of the vegetation within the plots amended with biosolids one year after planting indicated that soil would provide sufficient nitrogen without posing a significant leaching risk. However, results two years after planting indicated that the composts were a better source of long-term nitrogen and soil organic matter compared to biosolids. The vegetation on biosolids amended plots demonstrated a vulnerability to droughtinduced senescence due to low soil moisture associated with insufficient soil organic matter. By the end of the second year, no biosolids-amended plots had turfgrass quality significantly differing from the control plots, while all compost-amended plots had significantly better quality than the control. A combination of composts and biosolids products applied in a manner that increases soil organic matter levels to the already recommended 5% while providing 144 kg N/ha of first-year available nitrogen should provide sufficient short-term and long-term nitrogen to support persistent growth of planted perennial species. Two surveys of the study area, conducted in June and September of 2014, measured how the relative coverage of planted species versus weed species and perennial species versus annual species varied based on amendment and, rate of amendment application, and distance from the road. With the exception of the WRB (anaerobically digested biosolids) amended plots monitored in September 2014 (33% relative planted coverage), all other amendments and rates assessed in June 2014 (39% 61% relative planted coverage) and September 2014 (35% 45% relative planted coverage) had significantly greater relative coverage of planted species than the control (June 2014 – 28%, September 2014 24% relative planted coverage). As for the relative coverage of perennial species, no individual rate of application in either June 2014 (72% 79% relative perennial coverage) or September 2014 (55% 60% relative perennial coverage) produced a relative perennial coverage significantly different than that of the control (June 2014 – 68%, September 2014 – 50% relative perennial coverage). In June 2014, individual biosolids amendments had significantly greater relative coverage of perennial species (RMI [wood-ash stabilized biosolids] – 83%, WW [aerobically composted biosolids] – 82% relative perennial coverage) than the control (68%). No other amendments in either June 2014 (79% 67% relative perennial coverage) or September 2014 (46% 64% relative perennial coverage) had relative coverage percentages significantly different from the control (June 2014 – 68%, September 2014 – 50% relative perennial coverage). Distance from the road had a significant impact on the relative coverage of both planted and perennial species. The area within 1 meter of the road had significantly less planted coverage than the areas of the plots further from the road in all plots except the BBCC (biosolid cocompost) and control. Relative perennial coverage within the first meter from the road was significantly less than the rest of the area within the plots for both composts and biosolids. Digitaria sp. dominated that first meter, with the effect being especially pronounced on the side of the road where traffic was closer to the vegetation. In order to promote the establishment and persistence of perennial species within Rhode Island highway roadsides, the use of composts that have both sufficient first-year plant-available nitrogen (144 kg n/ha) and high levels of organic matter (sufficient to raise soil levels to a minimum of 5% organic matter) are recommended. Regions of the roadside within one meter of the paved surface require the development and implementation of specific management techniques beyond amending the soil with organic materials.

environment. As a means of improving the availability of the macronutrient nitrogen, a limiting nutrient for many plants, we amended roadside soil in Saunderstown, Rhode Island with seven different products: five stabilized biosolids (products whose organic material consisted only of sewage sludge) and two composts (products whose organic material consisted either partially or completely of yard waste). The biosolids were applied at rates of 48 kg N/ha, 144 kg N/ha and 288 kg N/ha (nitrogen available in the first year) while the composts were applied at rates of 15%, 30% and 45% v:v of the first 15 cm of the soil. Control plots received no fertilization beyond the 95.2 kg/ha of 19-19-19 fertilizer included in the hydroseeding mixture. All plots were hydroseeded in September 2012 with a seed mixture of Kentucky bluegrass (Poa pratensis), perennial ryegrass (Lolium perenne) and red fescue (Festuca rubra). Quality of the vegetation within the plots amended with biosolids one year after planting indicated that soil would provide sufficient nitrogen without posing a significant leaching risk.
However, results two years after planting indicated that the composts were a better source of long-term nitrogen and soil organic matter compared to biosolids. The vegetation on biosolids amended plots demonstrated a vulnerability to drought-induced senescence due to low soil moisture associated with insufficient soil organic matter. By the end of the second year, no biosolids-amended plots had turfgrass quality significantly differing from the control plots, while all compost-amended plots had significantly better quality than the control. A combination of composts and biosolids products applied in a manner that increases soil organic matter levels to the already recommended 5% while providing 144 kg N/ha of first-year available nitrogen should provide sufficient short-term and long-term nitrogen to support persistent growth of planted perennial species. Distance from the road had a significant impact on the relative coverage of both planted and perennial species. The area within 1 meter of the road had significantly less planted coverage than the areas of the plots further from the road in all plots except the BBCC (biosolid cocompost) and control. Relative perennial coverage within the first meter from the road was significantly less than the rest of the area within the plots for both composts and biosolids. Digitaria sp. dominated that first meter, with the effect being especially pronounced on the side of the road where traffic was closer to the vegetation.
In order to promote the establishment and persistence of perennial species within Rhode Island highway roadsides, the use of composts that have both sufficient first-year plant-available nitrogen (144 kg n/ha) and high levels of organic matter (sufficient to raise soil levels to a minimum of 5% organic matter) are recommended.
Regions of the roadside within one meter of the paved surface require the development and implementation of specific management techniques beyond amending the soil with organic materials.

INTRODUCTION
It is easy to overlook the roadside due to its very ubiquity. With 1,102 miles of state roads and highways in Rhode Island alone, it is an environment that is easy to take for granted despite the variety of services it provides. According to the AASHTO

Guidelines for Vegetation Management (American Association of State Highway and
Transportation Officials Subcommittee on Maintenance 2011), the services provided by properly maintained roadside environments include: preserving visibility and sightlines for drivers, reducing highway maintenance costs, maintaining the integrity of the shoulder area, preventing guardrails from failing, preserving stream banks, preventing drains from clogging, preserving wetlands, improving runoff water quality, and keeping the roadside aesthetically pleasing. All of these services are threatened by degradation of the vegetation and soil adjacent to highway travel surfaces.
The vegetation community best suited to provide these critical services for the northeastern region of the United States is composed of native and regionally adapted perennial grass species (Booze- Daniels et al. 2000;RISCC 2014). These species have the benefits of tolerating mowing, rooting deeply, stabilizing soil, and staying green and attractive from spring through fall (Dunn et al. 2002). Mowing is necessary to prevent the succession to woody plants that is typical to the region (Parr and Way 1988). Mowing also preserves sightlines for drivers and decreases the attractiveness of the area as habitat for large mammals. The root systems of perennial grasses consist of shorter, thicker and more lignified roots with lower nitrogen concentrations than those of annual grass species. The lower nitrogen and higher lignin content of the perennial vegetation means they decompose more slowly, leaving the soil less vulnerable to erosion (Roumet et al. 2006). Currently, the dominant species on Rhode Island roadsides are the annual grasses large and smooth crabgrass (Digitaria sanguinalis and Digitaria ischaemum). These warm-season annual grasses are alive only during the summer months, leaving roadsides bare and unattractive in the cooler months of the growing season (Brown and Sawyer 2012) and vulnerable to erosion.
The vegetation most commonly seeded along Rhode Island highways are perennial cool-season turf species chosen for their relatively low-price and wide availability. Despite being generally well adapted to the soils and climates of the Northeast, cool-season perennial turf grasses struggle with the unique challenges of the roadside environment (Dunifon et al. 2011). These include soils lacking in nutrients and organic matter, excessive drainage, irregular and potentially harmful mowing, traffic damage, competition from annual, invasive and opportunistic species, and exposure to road runoff that may contain harmful and herbicidal compounds (Haan et al. 2012).
As "disturbance specialists", annual species are especially well suited for these high-stress areas. They are able to colonize low-nutrient and exposed ground, and propagate by setting massive amounts of seed rather than spreading vegetatively, allowing them to withstand the pressures of the environment better than perennial species that have been developed for more highly maintained sports fields and lawns (USDA 1935). The relative suitability of perennial and annual species is likely to be affected by microhabitats within the roadside environment (Karim and Mallik 2008) and potentially by the distance from the travel surface (Spencer et al. 1988). As slope position and distance from vehicle traffic change, so do moisture patterns, organic matter accumulation, bulk density and opportunities for disturbance, impacting the ability of species to establish and persist.
Annual vegetation -particularly those species adapted to disturbed habitats -are better adapted to grow in the dry and coarse soil of roadsides. Annuals have the advantage of being able to seek out and acquire limited soil nutrients without committing significant resources to durable above-ground biomass or thick, persistent roots. Roadside soils, with their low cation exchange capacity and limited ability to retain water, place perennial species at a distinct disadvantage in the competition for nutrients. A number of explanations have been put forth for the difficulty that perennial and native species have persisting on roadsides. These include stress from the bulk density and particle size of soils (Haan et al. 2012), road salt (Friell et al. 2014) , the speed at which annual species establish (Booze- Daniels et al. 2000), the timing of the emergence of annual vegetation (Stevens and Fehmi 2011), and disturbance from travel and construction (Hansen and Clevenger 2005) and other human activities. An additional possibility is that a lack of plant available nutrients, especially mineralize nitrogen, limits the ability of perennial cool-season species to persist and out-compete annual warm-season vegetation (Wakefield et al. 1974, Wakefield et al. 1981Wakefield and Sawyer 1982;Boen and Haraldsen 2011).
If nutrients are the limiting factor, then the addition of nutrients in organic forms, particularly nitrogen, through the incorporation of organic amendments, should increase the coverage, quality and persistence of perennial species as compared to roadside soils lacking these amendments.
It is this theory that drives the hypotheses tested by this study: Hypothesis 1: The addition of stabilized biosolids at a rate of 144 kg N/ha to roadside soil will significantly increase vegetation quality as compared to an unamended control for at least two years after seeding because of increased plantavailable nitrogen in the soil.
Hypothesis 2: The addition of composts to roadside soil at a rate of 30% v:v will significantly increase vegetation quality scores as compared to an unamended control for at least two years after seeding because of increased plant-available nitrogen in the soil.
Hypothesis 3: The addition of stabilized biosolids at a rate of 144 kg N/ha to roadside soil prior to hydroseeding will significantly increase the coverage of perennial species in both June and September in the second year after seeding because of increased plant-available nitrogen in the soil.
Hypothesis 4: The addition of composts to roadside soil at a rate of 30% v:v prior to hydroseeding will significantly increase the coverage of perennial species in both June and September in the second year after seeding because of increased plantavailable nitrogen in the soil.

Conditions for the establishment of perennial grasses on roadsides
The importance of microhabitats for the competitiveness of native species is described by Karim and Mallik (2008). In studying random transects along 14 km of the Trans Canada Highway in the Terra Nova National Park in Newfoundland, Canada, they described four distinct microhabits present between the edge of the road and the edge of the forest. These microhabitats --the shoulder, the side slope, the ditch, and the back slope --sustain significantly different plant communities. They attribute these differences in plant communities to gradients in soil moisture content, bulk density, organic matter depth and pH.
From the mid-1960s until the mid-1980s a number of trials testing seed mixes, mowing heights, the effects of mulches and the effects of fertilizers on the establishment and persistence of roadside vegetation were conducted through the agricultural experiment station at URI in conjunction with the RIDOT. These tests resulted in reports and recommendations that address the unique requirements of vegetation along Rhode Island highways. Wakefield, et al. (1974) provide a number of recommendations using data gathered from at least 8 years of tests. Regarding seed mixtures, they found fine-leaved fescues (Festuca rubra, Festuca tenuifolia) and Colonial bentgrass (Agrostis capillaris) to be the best adapted to establishing in roadside areas, while Kentucky bluegrass (Poa pratensis) did not establish well in droughty soils and Perennial ryegrass (Lolium perenne) had the benefit of rapid germination but was short-lived and tended to reduce the establishment of other species. All species, however, showed improved performance when seeded in quality topsoil over a graveled base. This is a critical finding given the lack of quality topsoil along engineered roadsides. The study recommends a "Park Mix" of species for mowed areas, such as the seed mix included in this study, and a "Slope Mix" of species for unmowed areas, which are outside of the scope of this study. The recommended "Park Mix" is composed of 75% red fescue (Festuca rubra), 15% Kentucky bluegrass (Poa pratensis), 5% Colonial bentgrass (Agrostis tenuis) and 5% perennial ryegrass (Lolium perenne). Additionally, the report recommends that a chemical fertilizer be applied at rates of 112 kg N/ha, 179 kg P/ha and 156 kg N/ha at the time of seeding.  Fertilizer specifications and rates -10-10-10 fertilizer applied at 850 lbs per acre. Fifty percent of the nitrogen in the fertilizer must be slow release.
 Seeding dates -Spring seeding from April 1 to May 31 and Fall seeding from August 15 to October 15.
 Seedbed preparation practicesa raked seedbed with sticks, litter, wire, weeds, cable, cobbles and stones larger than 1 inch removed. Compacted seedbeds are to be scarified to a depth of 5 inches. Soil should be limed at a rate of 1 ton per acre unless otherwise specified.
 Seeding rates and methods -150 lbs per acre seeded either mechanically or hydroseeded  Post-seeding mulching and watering practices -All new seeding is to be covered with a cellulose mulch. The cellulose much can be part of the hydroseeding mixture. Watering is to occur within 72 hours of planting in a manner that ensure that water reaches the root zone without eroding the soil.
 Mowing practices -Vegetation should be mowed twice a year to a uniform height of three inches.

Use of composts and biosolids to improve establishment and persistence of perennial grasses
The usage of composts as fertilizers and amendments for residential, commercial and athletic turf applications is well documented (Henderson et al. 2012;Henry et al. 2002;Landschoot 1996). The goal is often to establish mowing tolerant turf stands that are green for a long part of the year, can withstand foot traffic and are free of nonplanted species. In these applications, inorganic fertilizers are often preferred for their ease of storage and application and the precision with which they can be applied.
Sometimes these fertilizers are "slow release", though they are not meant to persist in the soil for more than a single season. They require regular, well-timed application and are not ideal for minimal management scenarios.
Under minimally managed conditions, biosolids are an attractive option as soil amendments because they are low-cost, plentiful and high in both nutrients and organic matter. A common practice has been to use them in reclaiming mining sites and other soils disturbed by industrial applications (Li et al. 2000). The use of biosoilds as a soil amendment for the fertilization of grasses and vegetables is not a recent innovation Harper ( (Wakefield and Sawyer, 1986 In a review of the potential benefits of compost for restoring soils disturbed by urban development, Cogger (2005) found that most research regarding the use of compost and its associated guidelines had been generated for agriculture. He noted a lack of literature regarding the use of composts as amendments for soils affected by urban development, recommending that more research take place. While extensive studies have been conducted on the use of biosolids in mine site reclamation (Jenness 2001;Daniels et al. 2002;Evanylo et al. 2005;Stehouwer et al. 2006), research concerning their use in engineered soils is less common. However, relevant to the establishment of grass on roadsides, he did find two studies (Tester 1990;Sullivan et al. 2003) that looked at the persistent benefits of a single compost application. Both studies, conducted in humid, temperate climates, used a single application of compost and found that the benefits of compost lasted for at least several years after application. He also examined the issue of over-application of compost. He describes studies that found potential problems with elevated salt content within composts, excessive wetness and anaerobic conditions in poorly drained soils, and high rates of nitrate leaching in nitrogen-rich composts. These are all potential issues in roadside applications, where salt is often applied to roads, drainage must remain at a high level and, due to drainage patterns, leachate must be monitored for excessive nutrients. He also notes a study that found that high rates of application (50% or greater by volume) led to excessive settling and waterlogging in an urban soil. Craul (1999) recommends an application rate for degraded landscapes of 5 cm to 8 cm incorporated to 20 to 25cm deep (20%-40% v:v) for long-term improvement of the soil environment. Loschinkohl and Boehm (2001) and Dunifon et al. (2011) both looked at the use of composts as a means of restoring fertility to disturbed soils. Loschinkohl and Boehm (2001) focused on the restoration of turf in a controlled plot scenario, while Dunifon et al. (2011) investigated the effectiveness of different compost application methods on existing roadsides. Loschinkohl and Boehm (2001)  applied at a rate of 235m 3 ha -1 . The hydromulch, a wood cellulose fiber, was applied at a rate of 112 kg ha -1 . The plots were mowed once per year to a height of 10 cm by Virginia DOT maintenance contractors using rotary mowers. Dunifon et al. (2011) concluded that a surface application of poultry litter and woody waste compost (C:N 17:1) increased important plant nutrients in disturbed roadside soils and significantly improved coverage of the intentionally planted tall and chewing's fescues relative to the hydromulched plots. However, they also concluded that these gains would be short-lived. They state that "any beneficial effects of the organic amendments would largely have been restricted to the soil surface". This was due to a lack of evidence of compost incorporation beyond 5 cm into the root zone after 2 years. While the compost amended plots outperformed the hydromulched plots in all regards, it was observed that by the second year of the study the fescues in the compost amended plots had declined and stands were dominated by spotted knapweed and horsenettle.

Nitrate Mineralization
Nitrogen is often the limiting nutrient in turf production. It is commonly the nutrient that fertilizer application rates are based upon in landscape and athletic field management (Henry et al. 2002;Henderson et al. 2012). Nitrogen is commonly added to soil as synthetic fertilizers such as nitrate and ammonium salts or urea. In this form it is either plant available or quickly mineralized, but also very mobile and can be lost from soil through leaching, volatilization and denitrification. Synthetic nitrogen fertilizers allow for precise and timely application, but requires that regular additions of fertilizer take place.
Mineralization of the nitrogen in organic amendments such as biosolids and composts into plant available forms depends on a number of microbial processes that are controlled by temperature, moisture, pH, soil texture and other physical and chemical soil properties. This creates a lack of precision in the application of nitrogen via organic amendments and requires that the amendments be applied at rates of nitrogen greater than those at which inorganic fertilizers are applied ).
The current guidelines for the application of fertilizer along Rhode Island highways calls for the use of 10-10-10 (N-P-K) fertilizer, with fifty percent of nitrogen in slow release form (RIDOT 2013). These requirements are easy to meet when using synthetic fertilizers but almost impossible to control for when using organic amendments. A rate of application of 850 pounds of 10-10-10 per acre is specified for general highway seeding with 500 pounds per acre following the installation of sod (RIDOT 2013). Only recommendations and no specifications are available for the application of compost along Rhode Island highways (RIDOT 2013, RISCC 2014.
Compost is suggested for use in increasing soil organic matter and not as a fertilizer.
Because biosolids and composts are comprised of complex organic compounds that require breakdown by microbes, the speed at which nutrients are released varies depending on factors such as the community of microbes present, the size and type of compounds present within the biosolids and composts, soil particle size, temperature, moisture, pH and other factors specific to a given location (Sylvia et al, 2005). All of these variables impact the mineralization of the nitrogen present within the biosolids and composts and subsequently the availability of that nitrogen to vegetation. Garau et al. (1986) looked at the problem of nitrogen mineralization rates from a multivariate perspective. Their study observed the impact that soil type, stabilization method (aerobic and anaerobic), amendment incorporation rate, leaching and time had on the rate of nitrogen mineralization for sewage sludge. They determined that soil was the most important factor in determining nitrogen mineralization rate, followed by application rate and then the properties of the sewage sludge. They also determined that organic nitrogen mineralized more quickly from aerobically treated sludges than from anaerobically treated sludges. contained what was thought to be the greatest proportion of woody materials, had an extended period of nitrogen immobilization and the lowest cumulative N mineralization. Far exceeding the N mineralization rates of the yard waste composts, the co-composted biosolids released twice as much N as the granite topsoil (1.17 g N/kg), and two-thirds that of the nutrient rich, sedimentary topsoil (2.14 g N/kg) used as a reference. The rate at which total available N was mineralized differed greatly between the yard waste composts and the biosolid co-composts. According to the study, 1% to 7% of the total N contained in the yard waste composts was released during the incubation period, while the co-composts released about 27% of total N.
The yard waste composts provided a more gradual, long-term supply of N, though those with high levels of uncured and fibrous materials could initially slow plant establishment due to N immobilization, while the co-composts provided a more rapid source of N that might have less potential for continued N release.
To better understand the causes of these different mineralization rates it helps to view organic sources of nitrogen as parts of different "pools" that are released at different rates over time. The ratio of these pools in different sources of biosolids and composts will significantly impact their short-term and long-term effectiveness as sources of nitrogen for roadside turf. Smith et al. (1998) state that stabilized (composted or air-dried) biosolids have up to five pools of mineralizable nitrogen to draw from, though these can be condensed into two pools for modeling and predictive purposes. The relative proportions of these "fast release" and "slow release" pools are determined by the process used to treat the biosolids. Digested biosolids had up to 60% of total N converted to nitrate within 4000 d°C (thermal time units), while dewatered undigested biosolids had only 15% mineralization to nitrate over the same 4000 d°C. Composted biosolids were the most stable, with only 10% of the N mineralizing to nitrate after 160 d of incubation at 25°C (4000 d°C). This stability indicates a minimal chance of leaching, though it appears to limit the utility of composted biosolids as a primary source of nitrogen. The study emphasizes the need for incubation tests for individual products and the importance of a centralized processing facility for biosolids so that product uniformity can be achieved. Furthermore, the thermal-time relationship between products and mineralization rates brings up the importance of considering climate when determining application rates.
An additional method of characterizing nitrogen (as well as carbon) pools in composted biosolids is presented by Doublet et al. (2010). Their study examined the role particle size plays in the mineralization of nitrogen and the sequestration of carbon. They used a composite material comprised of a biosolids co-compost composed of aerobically-digested sewage sludge (13%), refuse from green waste compost screening (47%), stored yard trimmings (20%) and crushed wood pallets (21%). They then separated the finished material into seven particle size fractions: Mineralization rate is additionally impacted by the moisture present within soil. Hudson (1994) draws a link between increased soil organic matter and increased soil moisture. The addition of organic amendments should increase soil organic matter and lead to increased soil moisture which should then lead to increased mineralization of nitrogen containing compounds within the soil (Orchard and Cook 1983).

Potential for nitrogen surplus and leaching
Mineralized nitrogen in the soil that is in excess of what plants can immediately utilize is made unavailable to plants through either immobilization, volatilization, leaching, or denitrification. This over-application of a nitrogen is not only inefficient and costly, but leaching into groundwater can cause eutrophication and contaminated drinking water. Over-application of nitrogen in landscaping applications can also cause overly vigorous growth that can create the need for more frequent mowing. Petrovic (1990) and Di and Cameron (2002) identified factors that influence the fate of nitrogen when it is applied as a turf fertilizer as well as in a variety of natural and agricultural systems. These factors include soil texture, nitrogen release rate, nitrogen storage, climate, vegetation harvesting, soil management and the type of vegetation being fertilized. The actual amount leached can vary widely and was determined to be controllable through management techniques.
While all fertilization runs the risk of leaching nutrients, Easton and Petrovic (2004) contend that the greatest risks of N and P leaching and runoff into ground water from turfgrass occur within the first 20 weeks following seeding. By the second year of the study there was no significant difference between plots fertilized with biosolids and the control in terms of nutrient leaching. The authors suggest, based on fertilized plots showing less leaching than the unfertilized control plots, that fertilization of turf that leads to greater rooting and tissue growth can provide a greater benefit to ground water than no fertilization.
This theory that turf serves as a sink for plant available nitrogen is further supported by the findings of Bushoven et al. (2000) and Jiang et al. (2000). They found that sudden turf death leads to increased nitrate leaching and that turf reestablishment significantly reduces the losses.
The source of the nutrients should be considered as well. King and Torbert (2007) state that land-applied organic sources of nutrients (composted animal manures) pose less of a runoff threat for nitrate and ammonium than inorganic sources. This is attributed to micro-organisms stabilizing nutrients in complex organic compounds during the composting process. These compounds lead to a slower and steadier release of nutrients, more in sync with plant uptake. Further, they determined that ammonium is a much smaller runoff threat than nitrate. This is predicated, however, on NH4 + -N binding to clay particles in the soil. Very sandy soils and those soils with low cation exchange capacity are at a greater risk of leaching NH4 + -N, though most should volatilize. Tian et al. (2008) demonstrate that while compost, is superior to peat for restoring soil microbial biomass in the sandy soil of golf course putting greens, biosolids offer greater potential for restoring N mineralizing bacteria that provide long-term plant mineralizable nitrogen. A caveat to this is that biosolids need to be applied at greater rates of total N than inorganic fertilizer in order to achieve similar levels of mineralized nitrogen. Since the nutrients are contained within organic compounds and decompose at a rate of approximately 10% per year (Tian et al. 2009), applying the product at the same N rate as inorganic fertilizers is likely to lead to N deficiency. The authors suggest that application rates for the biosolids products on turf be based on the annual mineralized nitrogen levels for the individual product.

Roadsides and engineered verges as grasslands, their vegetative composition and their suitability for perennial cool season grasses
As mentioned earlier, the roadside environment, including other transportation corridors such as verges adjacent to railroad tracks, has functional properties that create conditions which are less than optimal for the growth of perennial cool season grasses. These areas have long been thought to have little ecological value and their vegetative communities have been little studied. What has been found, however, is that these areas are home to a diverse array of annual and perennial grasses and forbs, including rare native species. (Tikka et al. 2001) evaluated the role that roadside and railway verges play as hosts and dispersal corridors for grassland plant species. They found 78 grassland species within these unmanaged verges. Only three of the species found were of the sort that would have been sown. They conclude that these unmanaged verges could be important resources in the preservation of grassland species as natural grasslands disappear.
In their study of Rhode Island highway roadside vegetation, Brown and Sawyer (2012) identified 80 graminoid and forb species. Of these, 35 were native plant species and 4 were rare species. With little management beyond occasional mowing, these native and rare species can proliferate as niches develop within the roadside environment. Highways as habitats for both native and invasive species is further supported by Rentch et al. (2005). In their study of West Virginian highway roadsides, they identified a total of 538 vascular plant species. While they emphasize the impact of invasive species and the need to actively manage the threat they pose to native species, they do note that native species were often well-represented in the study plots.
While there is evidence to support these roadside areas as havens for native grass species there is also evidence that they can act as dispersal corridors for invasive and opportunistic species that colonize disturbed habitats. Hansen and Clevenger (2005) state that highway and railway corridors play host to a number of invasive species due to their high levels of disturbance and light. This impacts floral communities well beyond the corridors, leading to the spread of invasive species and the decline of native species. They recommend reseeding corridors with native plants to prevent this.
There is also the theory that by applying composts and other amendments to roadsides the nutrients released will fertilize the weedy and annual species as much as they will the sown and perennial species. This would give the weedy species that are better adapted to disturbed environments a further advantage. This is stated succinctly by Dunifon et al. (2011) when they say, "the high nutrient rates can promote the growth of competing broadleaf weeds as well as that of the sown species." If this holds true, then we should take actions that provide the greatest benefit to those species we would like to see dominate and persist. For instance, with crabgrass, a dominate species along Rhode Island roadsides, Turner and Van Acker (2014) contend that neither fall nor spring fertilization have an impact on the germination rates of smooth or large crabgrass. This means that if crabgrass is an issue, the fertilization regime most beneficial to cool season perennials should be the one used so as to allow them to compete with crab grass, without concern being given to the effect of fertilizing the crabgrass.
An additional concern regarding the suitability of roadsides as host environments for perennial grasses has been the perception that the use of road salt over the winter months leads to soil salinity levels that would damage common turf grasses. Rodgers and Anderson (1995) studied the impacts on soil salinity by the application of sewage sludge as an amendment and found that even though salinity increased to levels thought inhibit vegetation there was a linear increase in plant biomass. Research has shown that various factors influence soil salinity and that sensitivity to salinity varies between plant species (Bernstein 1975). In the case of Rhode Island roadsides, the large particle size, low CEC and high levels of drainage that characterize the roadside soils do not allow salts to accumulate to levels detrimental to vegetation (Brown and Gorres 2012). Karim and Mallik (2008) propose that by identifying the characteristics of substrates associated with the different microhabitats that comprise the roadside environment that native species can be selected for their ability to persist within those microhabitats. This adopts the strategy that rather than altering the conditions present on the roadside we should seed those native species best suited for the microhabitats created during road construction. The study further identifies particular native species suited for the four identified microhabitats. While they do not identify individual grasses, they did identify two low-growing herbaceous species that are native to both practices. Without adherence to those codified best practices and attention to current research, no effort will achieve consistent success in attaining the goal of safe, effective and attractive roadsides. The goal of this paper is to add to the body of knowledge that will inform future management guidelines and regulations for Rhode Island highways by investigating the potential of biosolids and composts to act as soil amendments that will support the physical and chemical needs of perennial species within roadside habitats. The study site was divided into 66 plots, each 2.1 m wide by 4.5 m long located within the shoulder slope of both sides of the median. Treatments were assigned using a randomized block design with three replications. The 66 individual plots were divided between the two sides of the median, with the second block split between the two sides. All amendment treatments utilized the same three unamended control plots ( mounted rototiller was used to till the soil to a depth of 7.6 cm, incorporating the existing vegetation and loosening the soil. Second, the amendments were applied by hand and raked evenly across the appropriate plots. Large rocks were removed at the same time. Third, the amendments were incorporated to a depth of 15 cm using a tractor-mounted rototiller. This step was done in three passes in the same direction in order to keep displacement of soil and amendments consistent across the plots. Soil profile analysis was conducted after the incorporation of the amendments. All plots (amended and control) were then hydroseeded using the RIDOT's "Park Mix" (70% Festuca rubra , 15% Poa pratensis and 15% Lolium perenne) at a rate of 168 kg/ha.
The hydroseed mixture also contained 19-19-19 fertilizer (applied at a rate of 95.2 kg product/ha, equivalent to the 385.5 kg product/ha of 10-10-10 prescribed by RIDOT), hydrostraw and binder. No further fertilization or irrigation took place over the course of the study. RIDOT maintenance crews mowed the site occasionally using flail mowers.

Description of Amendments
The amendments included five stabilized biosolids and two compost products.

Boston Beans (BB)
"Boston Beans" is a heat-treated biosolids product similar to Milorganite; it is produced by passing anaerobically digested biosolids through a rotary drum drying and granulation process. The product is sold as a 4-3-0 +Iron fertilizer and is distributed by Casella Organics in both bulk and 40-lb bags. The product is listed as

US EPA Class A, Exceptional Quality, has Type I Approval of Suitability in
Massachusetts and is fully approved for use in Massachusetts, Connecticut and Vermont (Casella Organics 2009).

RMI Biosolids (RMI)
RMI Biosolids is a wood-ash-stabilized biosolid product. It is produced and distributed by Resource Management, Inc. The product is manufactured by mixing "raw cake" (biosolids that have been dewatered at the wastewater treatment plant) with a biomass fly ash (wood ash) at a 1:1 v:v ratio. It is marketed as Heart + Soil Complete pH+Plus with an N-P-K of .008-.003-.0155 and 171 lbs of lime per ton.
Analysis for the material used for the study is: Total Solids -38.6%, TKN -27 g/kg, Ammonium-N -5.4 g/kg, Total K -9.57 g/kg, Total P -5.38 g/kg. The product is recognized as a Class A product under New Hampshire and federal regulations. Class A pathogen standards mean that the material is effectively sterile relative to pathogens and can be used on any type of crop with no waiting period necessary after application. Due to its lack of friability and objectionable appearance, it is only sold as a bulk commercial product. It is recommended that it be applied via a manure spreader or other commercial equipment (Charlie Hanson, pers. comm).

Rates of application
The RIDOT Standard Specifications for Road and Bridge Construction (2013) lists the specifications that an approved fertilizer must meet and the guidelines for applying approved fertilizers. The specification are for synthetic fertilizers, which is to be formulated to provide 10% available N (50% of which is slow-release), 10% available phosphate, and 10% available potash (10-10-10), with significant trace elements and a salt index not exceeding 35 (M.18.06). RIDOT does not currently have specifications for the use of organic fertilizers, including biosolids. Biosolids do not contain equal amounts of nitrogen, phosphorous, and potassium, so the application rate in this study was based on nitrogen, as is standard in turfgrass management (Henry et al. 2002). The biosolids were applied at rates of first-year available N of 48 kg N/ha (1 lb N/1000 ft 2 ), 144 kg N/ha (3 lb N/1000 ft 2 ) and 288 kg N/ha (6 lb N/1000 ft 2 ). These rates are based upon a range of what is considered to be the low and high end of annual N turf fertilization rates (RISCC 2014;Henderson 2012, Henry et al. 2002. RIDOT has specifications for compost approved for use as a mulch but does not provide guidelines for its use as a soil amendment. Compost was applied based upon a percentage volume (v:v) of the first 15 cm (6 in) of soil, a common practice in the landscaping industry. The composts were applied at the rates of 15%, 30% and 45% v:v of the first 15 cm of the soil without consideration of first-year mineralizable nitrogen. The control plots did not receive any amendments or fertilizers beyond the equivalent of 385.5 kg 10-10-10 fertilizer/ha applied to all in the hydroseeding mixture.
Because the biosolids products and compost products were applied based on different parameters the total amount of material applied differed vastly between the two types of amendments. Total applied material ranged from 1 Mg product/ha to 48 Mg product/ha for biosolids and from 103 Mg product/ha to 523 Mg product/ha for composts. Carbon application ranged from <1 Mg C/ha to 14 Mg C/ha for biosolids and from 32 Mg C/ha to 109 Mg C/ha for composts (Table 2). Turf data collection methods  September, October and November of 2014. Samples were collected using a 3.2 cmdiameter soil core sampler. During each sampling period each plot was sampled in five locations distributed in an "X" pattern. Individual cores were taken down to 15cm below the soil surface. Depth was often limited by compacted and gravelly soil.

Turf quality was assessed in
Fresh samples were placed in resealable plastic bags and stored at 4°C until analysis.
Samples were prepared for ammonium and nitrate analysis within 5 days of collection.
Soil samples were sieved at field moisture conditions through a 2 mm-mesh screen prior to chemical and physical analysis. Nitrate and ammonium were determined using the method described by Gugino et al. (2009)
Soil pH and EC were determined using a 1:1 soil water mixture and analyzed with a Denver Instruments UB-10 Ultra Basic pH/mV Meter (Denver Instruments Inc., Bohemia, NY). Soil moisture content was determined by drying and reweighing a 20g samples of field-condition sieved soil. Organic matter was determined by loss-onignition at 550°C for 5 hours. The C/N ratio of amendments was determined by the URI Graduate School of Oceanography using the protocol set by the manufactures of the Costech 4010 Elemental Analyzer (Costech Analytical Tecnologies Inc., Valencia, CA).

Vegetation survey methods
We conducted the vegetation survey in the early summer over a 14-day period (June 11, 2014to June 24, 2014, and again in late summer over a 13-day period (September 5, 2014to September 17, 2014. A 1-m 2 quadrat, subdivided into a grid of 100 sections, each 100 cm 2 , was randomly placed within each of three distinct areas of each plot. The areas were designated as Location 1 (0 m -1.5 m from the road), Location 2 (1.5 m -3 m from the road), and Location 3 (3 m -4.5 m from the road).
Species coverage was determined by counting the number of sections that contained above-ground vegetation of a particular species utilizing the methods of Brown and Sawyer (2102). All plants were identified to genus, and to species when possible.
In June the eight most common non-planted species (minimum coverage of 500 units) and in September the nine most common non-planted species (minimum coverage 250 units) from the respective vegetation surveys were compared.

Statistical analysis
Statistical analysis was conducted in SAS v.9.2. Repeated measures ANOVA was utilized for analyzing vegetation quality ratings, soil moisture, and soil organic matter data. Factorial ANOVA was utilized for soil nitrate, soil ammonium, vegetation survey, pH, and electrical conductivity data. Least Squared Means testing within the Factorial ANOVA function was conducted using Fischer's LSD.
Main effects for all analyses were the specific amendment and the rate of application. Interaction terms were tested and further analyzed if they were significant. For the analysis of the vegetation survey, location was also used as a main effect. Biosolids and compost amendments were analyzed separately to account for their differing application rates.

Vegetation Quality
Over the 26 months following hydroseeding, the turf quality in plots amended at any rate of biosolids products (no significant difference between products) and the turf quality in plots amended with either compost product (no significant difference between rates) was significantly better than the turf quality in the unamended control plots. The generally superior performance of the vegetation within the amended plots over the two-year study agrees with the results reported by numerous other studies and management guides that have investigated the effects of amending disturbed soil with biosolids and compost (Byju et al. 2015;RISCC 2014;Dunifon et al. 2011;Wright et al. 2008;Cogger 2005;Claassen and Carey 2004;Loschinkohl et al. 2001;Landschoot 1996;Sikora and Yakovchenko 1996;Wakefield and Sawyer 1984;Wakefield et al. 1981;Wakefield et al. 1974).

Turf Quality in Biosolids-Amended Plots
Turf in plots amended with biosolids, as a class, performed well at all application rates in the 2013 season (year one), only to steadily degrade in quality over the 2014 season (year two) ( Figure 4). In year one, the turf in plots at all three rates of application of the biosolids amendments outperformed the control (P<0.0001) while having an average quality score within the acceptable range. The method of biosolids stabilization did not have an effect on turf performance. Overall quality increased with N application rate (48 kg N/ha -2.44, 144 kg N/ha -2.96, 288 kg N/ha -3.40).
The control averaged 1.29 over the same time period, which was below the acceptable range. In year two, there was a substantial difference in performance in the biosolids- significantly (P<0.05) better than the turf within the plots amended at 144 kg N/ha (average score 2.15), 48 kg N/ha (average score 2.00) and the control (average score 1.67). There were no significant differences between the two lower rates and the control. From July 2014 to November 2014, no significant differences (all P for LSM >0.1) were observed between either the individual biosolids amendments or the individual application rates and the control. Over that time period the turf had quality score that averaged 1.89 (44 kg N/ha), 1.95 (144 kg N/ha) and 1.91 (288 kg N/ha) while the control averaged 1.60. While the turf within the plots amended with biosolids at all three application rates had acceptable average quality scores (between 2 and 4) throughout the first year, the turf quality of all rates fell below 2 during the end of the year two. These results are contrary to those reported by Wakefield et al. (1986), who over three years of annual observations saw that performance of vegetation in their turf plots amended with composted biosolids exceeded that of their control plots. This difference is likely due to Wakefield et al. (1986) applying much greater rates of product (145 Mg/ha, 291 Mg/ha and 437 Mg/ha) than used in this study (between 1 Mg/ha and 44 Mg/ha).

Figure 4. Year 1 and Year 2 Turf Visual Ratings for biosolids amendments by rate of application.
Individual products are not shown because no significant differences were detected between them. The rate of application had a significant effect on the quality of the turf in the first year of the study, with all three rates producing results better than the control. In the second year, these differences disappear with little significant difference between any of the rates and the control. By the end of the second year, all of the rates of biosolids application dropped below the lower limit of acceptable turf quality. Error bars represent standard error of the control.

Compost Amended Plots
The quality of the turf in the plots amended with BBCC was significantly better (P<0.05) than the plots amended with YWC ( Figure 6) while also demonstrating the best performance among all amendments. In the first year of the study the turf within the compost-amended plots failed to demonstrate consistently acceptable performance due to an over-application of quickly-mineralizable nitrogen in the case of the BBCC, and an under-application of quickly-mineralizable nitrogen in the case of YWC. The high level of woody material within the YWC likely led to nitrogen immobilization within the first year, a process that did not occur within the BBCC amended plots due to the presence of easily mineralized biosolids. On average, the amendments did produce quality scores within the acceptable range of between 2 and 4. The BBCC turf within the amended plots averaged a quality score of 2.51 which was significantly greater (P=0.0185) than the turf within the YWC amended plots that averaged 2.01.
Both amendments performed significantly better (P<0.001) than the control, which averaged a quality score of 1.29. In year two, both products produced more consistently acceptable results. This is likely due to nitrogen mineralization rates stabilizing as the slower-to-mineralize forms of nitrogen in the organic matter fraction of both products began to be utilized by the soil microbes. Significant differences were not identified between the two products, with the turf within the BBC-amended plots averaging a score of 2.65 and the turf within the YWC-amended plots averaging a score of 2.60. The turf in the plots amended with either products performed significantly better (P<0.01) than the control, which averaged a score of 1.63. There was no significant difference between rates of compost application in either year. In   (Smith et al. 1998). Claassen and Carey (2004) reported that biosolid/yard waste cocomposts, such as BBCC, initially mineralize nitrogen at a greater rate than yardwaste composts such as YWC, but the long-term mineralization rates of the two products are similar, and both provide steady sources of mineralized N over time. The early season vigor of the BBCC-amended plots in the first and second year, as well as the early season vigor of the YWC in the second year, agree with the findings of Sullivan et al. (2003), who observed strong, early season biomass production over a seven-year study of tall fescue (Festuca arundinacea Schreb. 'A.U. Triumph') grown in plots that received a heavy one-time application of food waste compost. Additionally, the results observed within the compost-amended plots more closely agree with those of Li et al. (2000), Tester (1990), and Wakefield et al. (1986) who saw improvements in vegetation using composted amendments. Rates are not show due to no significant difference appearing between them. Despite an early spike followed by a massive drop in quality by the BBCC, the two compost products largely produced similar results, especially in the second year. They consistently remained within the acceptable limits for turf quality while performing significantly better than the control. Error bars represent standard error of the control.  (Table 3 and Table 4). Only type of amendment had a significant effect on total nitrogen in August 2014 (Table 3). Neither rate nor type of amendment was significant in any other month and no significant interactions between rate and amendment occurred.
Compost-amended plots had significant differences between type of amendment and between rates in August and October of 2013 (Table 5 and Table 6). Only the type of amendment was significant in early and late May and July 2014 (Table 5).
Significant interactions between rate and type of amendment occurred in May, June and September 2013 and October 2014 (Tables 7-10). Neither rate nor type of amendment was significant in any other month.

Potentially Mineralizable Nitrogen
The amount of ammonium (NH4 + ) produced by a soil sample over a 7-day waterlogged incubation represents the amount of potentially mineralizable nitrogen (PMN) (ammonium + nitrite + nitrate) within a soil (Waring and Bremner 1964). All of the plots associated with the study were sampled and subsequently tested for anaerobic ammonium production.
For the biosolid-amended plots, rate of application significantly affected potentially mineralizable nitrogen (P<0.05) during six sampling periods (Table 12).
During those months the PMN in the amended plots never significantly exceeded that of the control plots, suggesting that differences were due to background variation in the soil. In all other months, differences based on rate were not significant for the plots amended with biosolids.
Among the biosolids-amended plots, amendment significantly affected levels of PMN during seven of the sampling periods (Table 11). Only in November 2014 did any of the amendments have a PMN significantly greater than the control. Significant differences were present between individual amendments, with certain amendments often producing greater PMN levels than others. The soil amended with BB and WRB most often produced the greatest PMN levels while soil amended with WW, RMI and CRD most often produced the least PMN. The BB and WRB amendments, both anaerobically digested products, were applied at the lowest volumes, among all amendments, and therefore would have had the least impact on the physical characteristics of the soil. The RMI and CRD products were alkaline stabilized with either wood ash or lime. The application of such highly alkaline amendments may have negatively impacted the soil microbial community, which likely is adapted to the acidic soils of the region. The minimal impact of the BB and WRB products may have preserved or had little effect on that microbial community, allowing PMN to remain within the range of the control.
Among the compost-amended plots, amendment application rate significantly affected PMN during only two months (Table 14) and product significantly affected potentially mineralizable nitrogen during 5 months (Table 13), with a significant interaction effect between rate and product occurring in November 2013 (Table 15).
Consistently greater levels of PMN in the control plots than in the biosolidsamended plots indicate that the addition of biosolids as an amendment at the three rates utilized either has no impact or an inhibitory effect on nitrogen mineralization in soil. Tian et al. (2008) found the opposite was true in their multiyear study of sandy putting greens that had been amended with biosolids, compost and peat. The biosolids and compost-amended soils in their study increased both soil microbial activity and PMN in comparison to the control. The difference in our results can be attributed to the range of conditions that impact the presence of soil microbial life, as well as the differences between individual composts and biosolids. Putting greens are highlymanaged and regularly-watered environments that would have very different microbial dynamics from a minimally-managed roadside.
The differences in both total soil nitrogen and PMN between the different amendments and rates, despite being significant (P<.05) in some months, do not explain the differences in turf performance. Furthermore, we were not able to draw a relationship between the mineralized nitrogen in the soil and the quality of the vegetation observed in the plots. The fate of the nitrogen within turf stands is difficult to determine since nitrogen can leach, be taken up by plants, be microbially immobilized, or be volatilized (Petrovic 1990). Nitrogen is also deposited via precipitation in substantial amounts over the eastern United States (Keene et al. 2002) in amounts substantial enough to impact carbon cycling in forest systems (Townsend et al. 1996). In crops such as corn, pre-planting and pre-sidedress soil nitrate tests have been shown to be a significant indicator of crop performance (Ma and Wu 2008).
These nitrogen tests, however, have been established in monocropped agricultural soils with improved and standardized crop varieties and the ability to control for factors such as soil moisture and competition from weeds. The variability and diversity of the soil and flora of the roadside makes it impractical to forecast future turf quality based only on the current and potential mineralized nitrogen within the soil. The mineralization of nitrogen within soil is impacted by a range of variables (e.g. organic matter, soil moisture, temperature, disturbance, particle size), as is the quality of vegetation growing within those soils (e.g. vegetation management, sunlight, competition). Simply having the potential to mineralize nitrogen does not mean it will mineralize at rates that will positively impact vegetation. Measurements other than monthly soil nitrate extractions appear to be better indicators of future vegetation quality in roadside soils.

Comparisons between Biosolids and Composts
I cannot make direct comparisons between the compost and biosolids products used in this study because different parameters were used to determine application rates. Application rates were based on expected first-year mineralizable nitrogen for the biosolids and on product volume for the composts. Furthermore, due to the differing kinetics of nitrogen mineralization between stabilized biosolids and composts (Smith et al. 1998), and the role of particle size in the release and immobilization of both carbon and nitrogen (Doublet et al. 2010), we can only compare these two products within their own classifications.
The effects of the different application rates on the quality of the vegetation quality within the study plots for the biosolids and composts were observed over the course of the two years of the study. Applying the organic biosolids amendments at the same rates of first-year plant-available nitrogen that inorganic fertilizer is typically applied appears to have been insufficient for sustaining vegetation quality over the second year of growth. First-year results for the biosolids-amended plots were generally acceptable throughout the year, with manageable growth throughout.
Though this method provided sufficient first-year nitrogen to the vegetation, it did not address other physical properties of the soil that are critical for sustained perennial vegetation, such as soil bulk density, cation exchange capacity, organic carbon and plant-available soil moisture. This became evident in year two when the overall quality of the biosolids-amended plots begin to decline. The rates at which the biosolids amendments were applied did not account for second-year availability of nitrogen and were determined based upon the methods recommended for standard inorganic fertilizers. Tian et al. (2009) states that applying biosolids at the same rate of nitrogen as inorganic fertilizers is likely to lead to nitrogen deficiency. Moreover, considering the coarse, gravelly texture of the roadside soil, along with its low organic matter content, the sustained release of nitrogen from the biosolids remains unlikely.
In contrast to the first-year available nitrogen metric used to determine application rates for the biosolids amendments, the compost amendments were applied based on a percent volume of the existing soil. Applying organic amendments based upon either volume (v:v of soil to a certain depth) or mass (Mg/ha) has been the method often utilized by studies investigating their impact in agricultural, remediation, and revegetation. Three studies that investigated the impacts of the one-time application of amendments, the method desired by the RIDOT, used application rates of 60 Mg/ha, 120 Mg/ha, and 240 Mg/ha (Tester 1990), 155 Mg/ha (Sullivan et al. 2003), and 145 Mg/ha, 291 Mg/ha and 437 Mg/ha (Wakefield and Sawyer 1986). For comparison, the biosolids amendments were applied at rates that ranged from 1 Mg/ha to 44 Mg/ha, while the composts were applied at rates that ranged from 103 Mg/ha to 523 Mg/ha ( Table 2). The far greater volume of compost material applied, in comparison to the biosolids, along with the amount of total carbon present within the composts (35% for BBCC and 18% for YWC) resulted in a far greater addition of organic matter to the soil and a much larger reserve of organic nutrients that could be mineralized over time.

Organic Matter
Soil organic matter levels ranged from 5.15% to 6.09% for plots amended with biosolids. No significant differences were observed between either products or rates.
Plots amended with composts had organic matter percentages that averaged 9.29% in BBCC amended plots and 7.50% in YWC amended plots. Organic matter level increased significantly with increasing application rate for both composts. Control plots averaged 5.68%.

Soil Moisture
Over the course of the entire study, none of the biosolids amended soils had a soil moisture percentage significantly different from the control, while soils amended with either compost amendment had soil moisture percentages significantly greater than the control ( Figure 17). This trend also occurred during the period of low rainfall from The significantly greater levels of soil moisture help to explain the superior performance of the turf within the compost-amended soil during the second year, especially during the period of below-average rainfall. The WW product is an aerobically composted biosolid product, giving it physical characteristics similar to the composts used in the study. This is potentially why the soil it was amended with had moisture levels not significantly different than the compost-amended soils.
Organic matter and soil moisture data obtained from both the biosolids-amended and compost-amended plots had strong correlations to each other (biosolids r = 0.655 P<0.001, composts r = 0.774 P<0.001). This is in line with the determination of Hudson (1994) that a soil's available water content increases with its organic matter content. Following this reasoning, increased soil moisture likely led to increased microbial nitrogen mineralization in the compost amended plots (Orchard and Cook 1983), further increasing turf quality.  (Rodgers and Anderson 1995), and in some climates repeated applications can lead to soil salinity levels that can negatively impact vegetation.
The soil pH of samples collected in September 2012 and October 2014 from all treatments were within the pH range of 6.0 to 8.0 recommended by Landschoot (1996) for good turf root growth (Tables 16 -19). In September 2012, the soil amended with CRD (7.44), RMI (7.17) and WW (7.58) biosolids products and the YWC (7.02) compost product had soil pH levels that differed significantly from the control (6.11) (   (Bernstein 1975). The low initial levels for all biosolids-amended soils are probably due to the small quantities that were applied. In this study no more than 48 Mg/ha of product were applied. For comparison, when Rodgers and Anderson (1995) applied biosolids to soil at 56 Mg/ha they observed a soil an EC of 900 μS/cm 2 .
The BBCC product, applied at rates of 103, 207 and 310 Mg/ha, had an average initial EC of 2519 μS/cm 2 ± 837 μS/cm 2 . Those values are similar to the ones obtained by Rodgers and Anderson (1995) when they amended soil with sewage sludge at rates of 111, 222 and 333 Mg/ha (1400, 2900 and 3400 μS/cm 2 , respectively). Despite these initial elevated concentrations, all treatments dropped to levels well below those considered harmful to most plants. This is due to the combined factors of the generally mesic climate and the low cation exchange capacity of the soil providing few exchange sites to retain Ca + and Na + , and the gravelly soil providing high levels of drainage that leach out the soluble Ca + and Na + .

Vegetation Survey
The roadside is host to wide array of flora, ranging from noxious and invasive annual weeds to rare and endangered native perennials. The reasons for the presence or absence of individual species can include soil conditions, slope orientation, traffic patterns and microclimates (Forman and Deblinger 2000).

Amendment and rate -Planted vs. Weed
Both the biosolids and the compost amendments significantly increased coverage by planted species and decreased weedy/volunteer species (all non-planted species). Relative coverage was calculated using the count of instances of planted species as a percentage of the total instances of all species. In June (Table 24), all plots amended with biosolids had significantly higher relative coverage of planted species than the control, with the product averages ranging from 49.1% (WRB) to 60.5% (RMI). Compost-amended plots also had significantly more planted species coverage than the control, averaging 56.0% for BBCC and 39.9% for YWC. Control plots averaged 27.8% (Table 25). In September (Table 24), the range of averages for relative coverage of planted species within plots amended with biosolids was from 33.1% (WRB) to 43.8% (CRD). Compost-amended plots averaged 44.7% for BBCC and 39.7% for YWC (Table 25). Control plots averaged 24.5%, significantly less relative coverage of planted species than all products except WRB (Table 24 and   Table 25).
In general percent coverage of planted species demonstrated very little difference between rates in both June and September. Significant differences between rates occurred only for the biosolids amendments in June where the 288 kg N/ha rate produced significantly greater coverage with planted species than the 48 kg N/ha rate and 144 kg N/ha rate (  (Table   27). Control plots in June averaged 27.8% and in September averaged 24.5% (Table   26 and Table 27).

Amendment and rate -Perennial vs Annual
In addition to observing the coverage of planted species relative to weed species, I analyzed the plant community based upon the broader categories of perennial and annual species. This acknowledges that many of the naturally occurring species are potentially desirable perennials. While some perennial species are considered noxious weeds, they will still provide greater benefits to the roadside than annuals which will die over the winter.
During the June survey of the biosolds amended plots, the type of product had a significant effect on the coverage of perennial species (Table 28). In that survey, RMI at 83.3% and WW at 81.9% had a percentage of perennial species coverage significantly different than the control at 68.3%. Type of product was not significant for the compost-amended plots (Table 29). In September, none of the plots amended with any of the biosolids or compost products had perennial species coverage significantly different than the control at 50.3% (Table 28 and Table 29). Application rate did not significantly impact the coverage of perennial species relative to annual species within any of the amended plots in either June or September. In both June and September, neither product nor rate had a significant effect on either the total annual or perennial counts in either the biosolids-amended or compostamended plots (Tables 30 -33). Control counts for June were 68.1 for total annuals and 154.0 for total perennials. In September, control counts were 73.1 for total annuals and 109.9 for total perennials. In June, while product was not significant to p<0.05 across the overall main effect, individual treatments were significant in certain biosolids-amended plots. Both RMI at 32.3 and WW at 36.9 had significantly fewer annuals than the control ad WRB at 121.7 had significantly fewer perennials than the should be noted that while counts of annuals within amended plots in June were only significantly different from the control for two of the amendments, the control plots had 30 -50% more annuals than most of the plots. While not statistically significant in most cases, it could be practically significant as the relationship between earlyseason perennial to annual species is explored.
Other perennial species, in addition to the planted species, that were affected by type of amendment were Trifolium repens (white clover), in June (Table 34 and   Table 35) and September (Table 37 and Table 39), and Juncus tenuis (path rush), in June (Table 34). White clover counts in June were significantly greater in the control plots at 32.4 than in all biosoilds amended plots except WW, ranging from WRB at 4.4 to CRD at 15.6. Within compost-amended plots, YWC-amended plots had significantly more clover than the control at 52.9 while BBCC-amended plots had significantly less at 7.0. In September, the control, at 31.6, had significantly more clover than all biosolids amended plots except WW, ranging from WRB at 1.2 to RMI at 13.1. Among compost-amended plots, the control had significantly more clover than BBCC at 4.2, but not YWC. Path rush counts in June were significantly greater in the control plots at 34.7 than in all biosolids-amended plots, ranging from CRD at 4.8 to BB at 12.9, and all compost-amended plots from BBCC at 2.7 to YWC at 8.7.
Type of amendment also affected the most common annual species, Digitaria sp. (crabgrass). Crabgrass counts in June (Table 34 and Table 35), at 42.4 for the control, were significantly higher than in plots amended with the biosolids products RMI at 16.5, CRD at 20.4 and WW at 22.9, and the compost products BBCC at 15.3 and YWC at 21.9. In September the control, at 68.1, was not significantly different from any biosolids or compost products (Table 37 and Table 39).
It appears that the perennial plants that would have naturally established were replaced with the seeded species. No substantial increase in the soils' favorability to perennial species was observed, aside from the RMI and CRD plots in June. The turf within the plot amended with these products had the lowest crabgrass counts and the greatest planted species counts among the biosoilds-amended plots. These impact, however did not carry over into the September survey. While perennial species in general did not appear to benefit from the addition of the amendments, the planted species did. Future observations will establish if this trend allows for the long-term establishment and spread of the planted species within the amended plots.   Location relative to the road I analyzed the areas 0 -1.5 m from the paved surface (Location 1), 1.5-3 m from the paved surface (Location 2), and 3-4.5 m from the paved surface (Location 3) independently of each other to determine if coverage of planted and perennial species changed significantly, as compared to the control, as distance from the road changed.
Different coverage based on distance from the road could suggest that different management techniques would be needed for the different areas.

Planted coverage by location
In Location 1, in both June and September (Tables 41 -44), I observed no significant difference in the coverage of planted species between the turf in the control plots and the plots amended with either biosolids or compost. In June, the turf within the control plots had planted coverage of 31.5% and in September had planted coverage of 30%. In June, coverage of planted species within all amended plots ranged from YWC at 29.7% to BBCC at 48.2%. In September, the coverage of planted species within all amended plots ranged from WRB at 28.4% to CRD at 35.0%.
In Location 2, for all amended plots in June (Table 41 and Table 42) and all amended plots, except those amended with WRB, in September (Table 43 and

Perennial coverage by location
In Location 1, in both June and September (Tables 45 -48) In Location 2, in June (Table 45 and Table 46) for plots amended with either biosolids or compost, only the turf within the plots amended with RMI, at 88.3%, and WW, at 89.9%, were significantly different than the control at 73.0%. In September (Table 47 ad Table 48 I also analyzed the total counts of both perennial and annual species within plots, as opposed to their relative coverage within those plots. I did this to determine if amending the soil increased total plant coverage and not just the ratio of perennials to annuals. When all locations were included in the models for June and September of 2014 for turf amended with either biosolids or compost, neither product nor rate had a significant (P<0.05) effect on total counts of annual or perennial species, but location did. When only Locations 2 and 3 were included in the model, product had a significant effect in some of the models, but location did not. Rate was not significant in any model and no significant interaction were detected.
In June 2014, excluding Location 1 from the model, type of product had a significant effect on the total count of annual species within the turf of biosolids and compost amended plots (Table 49 and Table 50 In all cases the area nearest the road was less able to support perennial species. This area was at the greatest risk of disturbance from vehicles leaving the paved travel surface and causing soil compaction, erosion and destruction of vegetation. Perennial species require resources (water, nutrients, stored energy) to recover from damage where annual species may regenerate from large soil seedbanks, giving them an advantage is highly disturbed soils such as those nearest the road.
The increase in perennial species as distance from the road increased is in keeping with the finds of Brown and Sawyer (2012). Their study looked at mowed, yet unamended areas along Rhode Island highways. They noticed this trend on a larger scale, with their three zones (Zone 1 0-3.3m from the pavement, Zone 2 3.3-6.6 m from the pavement, Zone 3 6.6-10 m from the pavement) being the same size as an entire plot in this study. Karim and Mallik (2008) observed that all four of the microhabitats that they surveyed along the Trans Canada highway were dominated by perennial species. They, however, did not identify individual grass species, making their study and this one difficult to compare. The regions they surveyed were relatively undisturbed for 10 to 15 years, allowing for a degree of succession to take place and trees and shrubs to establish. This natural succession may not be an option along Rhode Island highways where woody vegetation poses unacceptable safety risks.

CONCLUSIONS
The primary conclusion drawn from this study is one that was not addressed in the initial hypotheses. Two years after amending roadside soils with biosolids and compost products, the most significant factor in predicting the quality of the turf within those plots is the volume at which these organic amendments were applied.
Applying the products at greater volume increased the soil organic matter and the soil's ability to retain moisture.  (Wakefield and Sawyer 1986, Tester 1990, Sullivan et al. 2003.
The large amounts of organic matter present within the composts likely changed the mineralization dynamics and long-term biology of the soil rather than just providing immediate fertilization. This increase in soil organic matter led to significantly greater soil moisture levels within the compost amended plots. The increased soil moisture likely led to increased nitrogen mineralization, greater buffering of soil temperature levels and the prevention of drought-induced senescence.
Despite the biosolids products appearing to providing sufficient first-year mineralizable nitrogen, they did not significantly increase soil organic matter or soil moisture. The difference in organic matter was likely the key factor in the declining performance of the biosolid-amended plots in the second year of the study as either. The quality of the turf, as signified by quality scores, within the biosolidsamended soils were significantly better than those given to the control in the first year of growth while in the second year the quality steadily declined until it was no longer significantly different from that of the turf within the unamended plots.
These trends in quality are attributed to two factors. First, the biosolids products were applied at rates that controlled for expected first-year mineralizable nitrogen and were applied at rates commonly associated with synthetic fertilizers. The nitrogen mineralization rates of the following years were not addressed. Second, the relatively small amounts of biosolids amendments that were applied did not improve the soil organic matter and its ability to retain plaint-available moisture, leaving it prone to droughty conditions and low rates of nitrogen mineralization. The rates at which the biosolids amendments were applied are not recommended for a one-time application if the goal is to promote the persistence of perennial species. The rate of amendment application did, however, appear to effectively promote the early establishment of planted species and could potentially be used to amend annually fertilized soils.
The primary conclusion also provides reasoning for the support of the second hypothesis: The addition of composts to roadside soil at a rate of 30% v:v prior to hydroseeding will significantly increase vegetation quality scores as compared to an unamended control two years after seeding because of increased plant available nitrogen in the soil. The addition of composts at a rate of 30% v:v to the first 15 cm of soil did significantly improve the quality of vegetation as compared to the control.
This improvement was also observed at the 15% and 45% v:v rates. The turf grown in the soil amended by the both of the composts performed well over the two year study, though differences existed in their performance in the first year. The BBCC appeared to provide sufficient first and second year mineralized nitrogen while both products improved the soil organic matter and water holding capacity of the soil. The 15% v:v application rate of the BBCC can be recommended due to its ability produce turf of acceptable quality throughout the study. The YWC demonstrated the ability to consistently produce turf of acceptable quality in the second year and could benefit from the addition of a more quickly mineralizing source of nitrogen such as one of the biosolids products. Creating conditions conducive to the long-term mineralization of organic nitrogen appears to be an effective strategy for improving the persistence of perennial species within the roadside environment.
Overall, the products utilized in the study appeared to impact the vegetation in a manner consistent with the findings of previous studies. Smith et al. (1998) observed digested biosolids, such as BB, mineralizing nitrogen at a rate six times faster than composted biosolids, such as the WW product. This is reflected in the application rates by mass of the two products, with WW being applied at approximately seven times the rate of the BB. Furthermore, the compost products reacted in a manner generally consistent with the findings of Claassen and Carey (2004) who observed far greater release of nitrogen from biosolid co-compost than from yard waste composts.
That same study, however, also demonstrated the variability in nitrogen release rates between similar organic products. Four different yard waste compost (short curing time, typical curing time, 18 month curing time, and highly processed) had initial nitrogen rates that varied by as much as 48% between products and nitrogen release rates that ranged from 1% to 7%. Just basing application rates on broad categories such as biosolids and composts is not sufficient for effectively predicting the mineralization rates of organic amendments.
This complexity is compounded when we add in the findings of Doublet et al. (2010) and Garau et al. (1986) who demonstrated that a range of factors that include particle size, soil type, and incorporation rate impact the rate of nitrogen mineralization. By acknowledging the influence of these variables, it becomes imperative that individual products be bench tested in soils and conditions (e.g. moisture, temperature, pH, EC) similar to those they will be used in in order to effectively model their impacts on vegetation. It will then fall on regulators to ensure that guidelines are established and adhered to in order to ensure the successful utilization of these products. Also, most turf in that same region of the plot, whether in soil amended by biosolids or composts, had perennial coverage greater than the control, just not significantly so.
This brings us to an additional conclusion that was not previously hypothesized as part of this research. The factor that had the greatest effect on the coverage of planted and perennial species was the location in the plot relative to the roadside.
When the plots were divided into three distinct areas, differences in the composition of the turf emerged. Within Location 1 (0 -1.5 m from the road) no amendment had the ability to significantly change the percent coverage of either planted or perennial species or the total instances of either annual or perennial species in the turf.
As distance increased from the road, so did significant differences between the amended plots and the control. Within Location 2 (1.5 -3 m from the road) and location 3 (3 -4.5 m from the road), product significantly affected the coverage of planted species within the turf. In both June and September 2014, all amended plots, except for the Location 2 WRB-amended plots in September, had coverage of planted species within the turf significantly greater than that of the unamended control plots.
In June 2014, only plots amended with RMI had turf with relative coverage of perennial species that was significantly greater than the control in both Location 2 and 3. By September 2014, relative perennial coverage within Location 3 for all biosolids and compost amended plots, except for those amended with WRB, was significantly greater than the control. Total perennial counts in September within Locations 2 and 3 were only significantly greater in plots amended with WW and YWC.
Since this study did not track or classify stresses on the turf, either on the whole plot or on specific locations within the plot, we can only hypothesize as to the cause of the differences between locations. The location closest to the road would have been more likely to have been damaged by vehicles, be more exposed to increased heat radiated from the road surface and intercept more runoff and particulates. As distance increased, the risk of disturbances should have decreased, allowing for better survival of planted and perennial species.
The stress applied by the proximity to the travel surface and the proximity to vehicles was observed during data collection, but not classified. This stress was especially notable when the two sides of study area were compared. The side with vegetation closer to vehicle traffic, not just a paved surface, had noticeably poorer vegetation and greater coverage of Digitaria sp. within one meter of the paved surface.
Not only was the region within a meter of the paved surface significantly more disturbed than the areas further away, but the level of disturbance was different based on the soil's proximity to vehicle traffic. This is also the region of greatest importance to the structural integrity of the roadside where gullying and the erosion of soil would have the greatest negative impact. I encourage further study of this region of the roadside, with consideration being given to both vegetative and non-vegetative options for construction and management practices that preserve the services of the roadside while excluding invasive and noxious plant species.
A further observation is that Trifolium repens should be considered as an addition to the RIDOT Park mix of seeds due to its success in establishing in plots with no or little added nitrogen. Within otherwise poor plots, the areas with populations of T.
repens supported stands of seeded species while maintaining noticeably greater soil moisture. The design of this study did not allow for the consideration of the effects of smaller stands of vegetation within the plots, but the observation of the investigators leads us to recommend the effects of this species, or another nitrogen fixing species, for further study within non-slope roadside environments.

Recommendations for future studies
With the expectation that the results from this study will justify the future investigation of the use of organic amendments within roadside soils, the following recommendations are presented.
 Future amendment studies should control for a single property among the amendments used. It is recommended that organic matter be explored as a critical factor in the amending of roadside soils. The effectiveness of different compost/biosolids combinations in improving vegetation and soil quality would be a worth investigation.
 Use fewer amendments with larger study plots. By using 21 different treatments along with a control, the ability to make specific observations was limited due to there being only three replications of each treatment.
Furthermore, the great deal of variability within individual plots, along with the regular damage caused by vehicles disturbing large sections of soil, made the small plots poor indicators of larger trends. The use of larger plots would allow for more randomized sampling procedures as well as minimizing the potential impacts of variability.
 Perform random soil profiles prior to amending and seeding and then again at the conclusion of the study in order to assess the impacts of amendments on the physical properties of existing soils while also documenting variability within the soils.
 Future studies should refrain from splitting replications between physically distinct areas. By splitting a replication in this study between two sides of the study area, a direct comparison of the effects of the two sides was not possible.
 Future roadside studies should control for the distance not only from the paved surface, but also from vehicle traffic. Utilizing an area that is less prone to vehicular damage is recommended as well. Further study of the impact of paved surface and vehicle distance from vegetation and methods to mitigate any negative effects they cause is highly recommended.
In observing the amended plots, the area closest to the road on the southern side of the study area (plots 34-66) was more heavily and uniformly dominated by Digitaria sp. and appeared to be under much more stress than the equivalent area on the northern side of the study area (plots 1-33). When viewed from above ( Figure 1) it can be seen that the southern side directly abuts the traffic lanes while on the northern side there is a paved shoulder that ranges in width from 2 m to 5 m and is guarded with posts. We conducted a two sample t-test for means on the coverage of both

Amendment effects on adjacent environments
The swale was surveyed using a 10 x 100 cm quadrat with a 100 cm 2 grid. The quadrat was placed every 1.8 meters starting at the point where the swale met the drain for the median and continuing for 54 m until the swale no longer had treatment plots on both sides of it. As a control for the swale survey, we surveyed 54 m of swale from a nearby unamended region of the median that utilized the same drain as that of the amended swale.
Fertilizer can impact flora beyond that area which it was directly applied to (Di and Cameron 2002). In order to observe the effects that applying organic amendments would potentially have on downslope vegetation we conducted a vegetation survey of the swale adjacent to the two blocks of study plots. A nearby swale was surveyed in the same manner and used as a control. All measurements were taken based upon distance from the drain utilized by the swales.
Thirty-eight species (some only identified to genus due to age or mowing) were identified between the two regions. Twenty-one species were identified in the study area while 32 species were identified in the control area. Of these species, 17 had counts that varied significantly between the two areas (Table 3). There was no significant difference in the presence of annual or perennial species between the two areas.
Soil moisture appears to have the greatest impact on the speciation within the two swales. Significantly greater coverage of Cyperus esculentus and Persicaria maculosa, two species common within cultivated soils indicate greater soil moisture.
The USDA also classifies Cyperus esculentus as a facultative wetland species in the Northeast. It is difficult to ascribe the significantly greater coverage of Elymus repens and Plantago major to moisture differences since they were both common within the amended plots. The greater coverage of Poa pratensis is likely due to it having been seeded within the amended plots. The significantly greater presence of Aster sp.,

Andropogon virginicus, Dichanthelium acuminatum, Dichanthelium sphaerocarpon
and Digitaria sanguinalis within the control swale indicate a lower soil moisture percentage. These species are associated with dry and disturbed soils (Uva et al. 1997). Other species with significantly greater coverage within the control swale are Plantago aristata, described as "found in sandy drought-prone sites" (Uva et al. 1997), Rumex acetosella, described as "often found on, but not limited to, acid soils and area with poor drainage, low nitrogen, and little competition" (Uva et al. 1997), and Potentilla recta which is generally found on dry soil in the Northeast (Uva et al. 1997). While both areas were predominately populated with native and naturalized weed species, the addition of organic amendments appears to have altered the moisture dynamics of the soil and led to the differences in individual species coverage. We theorize that differences in moisture dynamics between the two swales can be attributed to both greater organic matter within the swale adjacent to the amended soils, due to the movement of OM as runoff from the amended plots, and greater biomass due to the leaching of nitrate contributing to increased fertility within that swale. How this would impact the overall flora of the roadside is not yet evident.
However, greater levels of moisture within the swale could lead to increased denitrification and carbon sequestration, which are a services often provided by wetlands. Increased vegetation would also help serve as a buffer to take up and filter heavy metals and contaminates that might occur in elevated levels within biosolids products (Wakefield et al. 1981) and road surface runoff.