Engineered and Environmental Controls of ... - ACS Publications

Subscriber access provided by - Access paid by the | UCSB Libraries

Remediation and Control Technologies

Engineered and Environmental Controls of Microbial Denitrification in Established Bioretention Cells Lucas J Waller, Gregory K. Evanylo, Leigh-Anne Henry Krometis, Michael Strickland, Tess M Wynn-Thompson, and Brian D. Badgley Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b06704 • Publication Date (Web): 10 Apr 2018 Downloaded from http://pubs.acs.org on April 10, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 36

Environmental Science & Technology

1

Engineered and environmental controls of

2

microbial denitrification in established

3

bioretention cells

4

Lucas J. Waller1, Gregory K. Evanylo1, Leigh-Anne H. Krometis2, Michael S. Strickland3,

5

Tess Wynn-Thompson2, Brian D. Badgley1*

6 7

1

Department of Crop and Soil Environmental Sciences, 2Department of Biological

8

Systems Engineering, Virginia Polytechnic Institute and State University, Blacksburg,

9

Virginia 24060, United States

10

3

11

United States

Department of Soil and Water Systems, University of Idaho, Moscow, Idaho 83844,

12 13 14

Abstract Bioretention cells (BRCs) are effective tools for treating urban stormwater, but

15

nitrogen removal by these systems is highly variable. Improvements in nitrogen removal

16

are hampered by a lack of data directly quantifying the abundance or activity of

17

denitrifying microorganisms in BRCs and how they are controlled by original BRC

18

design characteristics. We analyzed denitrifiers in twenty-three BRCs of different

19

designs across three Mid-Atlantic states (MD, VA, and NC) by quantifying two bacterial

20

denitrification genes (nirK and nosZ) and potential enzymatic denitrification rates within

21

the soil medium. Overall, we found that BRC design factors, rather than local 1 ACS Paragon Plus Environment


Page 2 of 36

22

environmental variables, had the greatest effects on variation in denitrifier abundance

23

and activity. Specifically, denitrifying populations and denitrification potential increased

24

with organic carbon and inorganic nitrogen concentrations in the soil media and

25

decreased in BRCs planted with grass compared to other types of vegetation.

26

Furthermore, the top layers of BRCs consistently contained greater concentrations and

27

activity of denitrifying bacteria than bottom layers, despite longer periods of saturation

28

and the presence of permanently saturated zones designed to promote denitrification at

29

lower depths. These findings suggest that there is still considerable potential for design

30

improvements when constructing BRCs that could increase denitrification and mitigate

31

nitrogen export to receiving waters.

32 33 34

Introduction Bioretention cells (BRCs) are engineered soil systems designed to intercept and

35

treat stormwater runoff before it infiltrates underlying soil or is discharged to nearby

36

surface waters. BRCs successfully reduce many stormwater contaminants1-6 and are

37

rapidly becoming one of the most popular strategies used in urban stormwater

38

management. However, observed nitrogen (N) removal rates have been highly

39

variable,7-9 which is likely due to the many possible N transformations that occur in

40

these systems. Stormwater contains nitrogen in many forms, including particulate

41

organic nitrogen (PON), dissolved organic nitrogen (DON), and dissolved inorganic

42

forms such as ammonium (NH4+), nitrite (NO2-), and nitrate (NO3-).10-12 This diverse N

43

speciation contributes to a high degree of complexity in N cycling. For example,

44

negatively charged NOx ions (NO2- and NO3-) are highly mobile in soil and therefore

2 ACS Paragon Plus Environment

Page 3 of 36


45

difficult to retain. In contrast, other forms such as PON can be physically captured by

46

the soil medium, but then undergo mineralization, ammonification, and nitrification

47

during aerobic conditions that can further contribute NOx species during subsequent

48

storms.9, 13

49

Denitrification is a microbial form of anaerobic respiration that is valuable for N

50

management because, when carried out to completion, it completely removes mobile

51

NOx forms from soil or water by transforming them into inert nitrogen gas (N2). However,

52

several underlying conditions must be present including: 1) the presence of denitrifying

53

organisms; 2) an organic carbon source as an electron donor; and 3) an oxygen

54

deficient environment coupled with oxidized N species to serve as alternate electron

55

acceptors.14 BRC design has evolved in attempt to facilitate bacterial denitrification,

56

most commonly by incorporating an internally saturated zone (ISZ) in the bottom layer

57

to promote a sustained anaerobic environment. While the incorporation of an ISZ into

58

BRCs can promote total nitrogen removal rates ranging from 60-80% in some studies,15-

59

17

60

significant difference in the reduction of N when comparing several BRCs with and

61

without ISZs.

others have shown only minor reductions in N removal,18 and Hunt et al.7 found no

Bioretention research has focused on design options that optimize vegetative

62 63

cover,12, 19, 20 soil medium composition,13, 21, 22 system capacity,23-25 and the addition of

64

carbon sources15, 17, 26 to increase N removal efficiency. For instance, multiple studies

65

have demonstrated enhanced N removal in BRC mesocosms containing vegetation,19,

66

27, 28

67

medium composition is more complicated because labile organic matter (OM) tends to

particularly with fast growing, high biomass plant species.12, 29 In contrast, soil



Page 4 of 36

68

leach N and phosphorus.7, 22 However, an organic carbon source for microbial activity is

69

also important, so the C:N ratio may be a critical factor to consider when selecting OM

70

additions.26 By contrast, fewer studies have examined the effects of environmental

71

factors on N removal in bioretention systems. In those studies, N removal rates have

72

been positively correlated with antecedent precipation,30 but effects of temperature can

73

vary by design. While conventionally drained BRCs generally export N as temperature

74

increases, BRCs containing a saturated zone show an inverse correlation between N

75

export and temperature.30-32

76

Despite the substantial amount of research focused on promoting nitrogen

77

removal in BRCs, important knowledge gaps remain. Firstly, there are few published

78

studies that evaluate broad patterns across numerous established BRCs. Second, while

79

denitrification has been studied in other stormwater systems,33-35 we are aware of only

80

two published studies that have quantified bacterial denitrifiers in a BRC.36, 37 Both

81

reported concentrations of denitrifying genes in a single BRC but the designs were

82

different, particularly considering the presence of an ISZ in the research by Willard et

83

al.37 Both also reported higher abundances of denitrifying genes in the surface media

84

layers compared to deeper layers, and hypothesized that saturation time and C content

85

influenced the microbial denitrification population.

86

In this study, we address both of these knowledge gaps – uncertainty about

87

denitrification controls in established systems and the lack of data that directly quantify

88

denitrifiers – by investigating generalizable relationships between BRC design factors

89

and microbial denitrification. Specific research questions were: 1) What are the most

90

important factors controlling the abundance and activity of denitrifying bacteria in


Page 5 of 36


91

established BRCs and do they relate more to design or to environment? And, 2) among

92

the most important factors controlling denitrifiers in BRCs, how can those relationships

93

be leveraged for improved design? We conducted a comprehensive field-based survey

94

of twenty-three BRCs across the mid-Atlantic region that represent different design

95

specifications, climates, physiographic regions, and ages. From each BRC we analyzed

96

a variety of design aspects, soil and climate characteristics, and denitrification indicators

97

to identify the factors that correlated most strongly with denitrification capacity to inform

98

future improvements and management strategies.

99 100

Materials and Methods

101 102 103

Site Selection & Sampling Design specifications were originally acquired for approximately 50 BRCs in the

104

eastern mid-Atlantic region (Maryland, Virginia, & North Carolina) from a variety of

105

published articles, public databases, and personal communications. These were

106

narrowed down to twenty-three BRCs (Table S1 and Figure S1) that represent a range

107

of features related to design (e.g. presence of ISZ, vegetation type, media mix

108

composition, etc.) and also characteristics related to environment and locale (e.g. age,

109

geographic region, mean temperature, precipitation, etc.) that we hypothesized could

110

influence denitrifier abundance and activity (Table 1).

111

BRCs were sampled during a one-month period in Nov/Dec 2014 during dry

112

weather (i.e., following at least one week of no rainfall) to compare BRCs in a common

113

functional state. Soil samples were collected as cores extending to the full depth of the



Page 6 of 36

114

soil medium. Triplicate cores were collected from both the front (inlet) and the rear

115

(outlet) portions of each cell. To assess vertical stratification, the top 10 cm and bottom

116

10 cm of media from each core were separated aseptically. The three subsamples from

117

each location (i.e., top/inlet, bottom/inlet, top/outlet, bottom/outlet) were homogenized to

118

provide a total of four soil samples per cell (Figure S2). The samples were stored on ice

119

during transport to the laboratory and stored at -80° C for further analysis. A portion of

120

the sample was refrigerated at 4° C for < 7 d for analyzing denitrification potential.

121 122 123

Denitrifier Abundance and Activity For each soil medium sample, DNA was extracted using the PowerSoil DNA

124

Isolation-Kit (MOBIO Laboratories INC, CA, USA) following the manufacturer’s protocol.

125

Concentrations of extracted DNA were measured using the Qubit 2.0 fluorometer

126

(Invitrogen, USA) and stored at -20° C. Abundances of bacterial denitrifiers were

127

measured by quantifying the copy number of two key functional genes present in the

128

soil medium using quantitative polymerase chain reaction (qPCR). Target genes

129

included nirK,38 which codes for a nitrite reductase that converts nitrite (NO2-) to nitric

130

oxide (NO),39 and nosZ,40 which codes for a nitrous oxide reductase that converts

131

nitrous oxide (N2O) to inert dinitrogen gas (N2).41 These were chosen to represent two

132

key steps in the N removal process. Firstly, the conversion of NO2 to NO (nirK)

133

represents the beginning of the denitrification pathway where N typically becomes

134

unavailable to most organisms and will likely be removed from the system. Second, the

135

conversion of N2O to N2 (nosZ) is critical because it represents the final step where

136

nitrous oxide is converted to inert N2, thus preventing the production of a potent


Page 7 of 36


137

greenhouse gas. Gene copies for each target were estimated by comparing cycle

138

threshold values to known standards of plasmids containing the target gene and

139

populations corresponding to each functional gene were computed as copies per gram

140

of soil medium. Standard curve R2 values for all genes were 99% or greater and

141

efficiencies were similar to or higher than previous studies ranging from 90.9 – 106.4%

142

for nirK 40, 42 and 83.9 – 96.5% for nosZ.43, 44 Reaction mixtures, primers, volumes, and

143

thermal profiles used for qPCR are included in Table S2.

144

While the number of denitrification gene copies indicates the abundance of

145

denitrifying bacteria, it does not quantify actual gene expression and/or cellular activity.

146

Denitrification potential, a direct measurement of denitrification enzyme activity, was

147

measured by the acetylene blockage technique.45 The protocol described by Drury et

148

al.46 was followed with the exception that headspace was flushed with N2. Briefly, soil

149

samples were incubated in an anoxic environment after adding glucose, nitrate, and

150

acetylene, and the amount of nitrous oxide was measured regularly over a 5 h period to

151

determine the rate of N2O production and estimate the rate at which denitrification could

152

potentially occur within a soil sample under optimal conditions. Gas samples were

153

stored less than two weeks and analyzed on a gas chromatograph (Shimadzu, Kyoto,

154

Japan). Due to instrument difficulties, estimates of denitrification potential are only

155

presented for 48 of the 86 total BRC media samples.

156 157 158 159

Total Organic Carbon, Ammonium, and Nitrite-Nitrate Total organic carbon (TOC) and N concentrations in the BRC medium were determined from potassium sulfate (K2SO4) extracts. Extractions consisted of the



Page 8 of 36

160

addition of BRC medium to 0.5M K2SO4 at a 1:7 ratio, agitation using a side-arm shaker

161

for four hours, and gravity filtration through a 2.5 micron, 15 cm diameter, cotton fiber

162

Whatman No. 42 filter (GE Healthcare, Germany). Extracts were frozen at -20 °C prior

163

to analysis. Total organic C was analyzed on an OI Model 1010 total organic carbon

164

analyzer using the standard method 5301c.47 NH4+-N and NOx-N concentrations were

165

determined using a Lachat QuikChem 8500 Flow Injection Analyzer following the

166

QuikChem Method 10-107-04-1-L and APHA Method 4500-NO3- I.47, 48 Given that

167

dissolved inorganic nitrogen (DIN) was measured as multiple species, a principle

168

component analysis (PCA) was conducted on NH4+-N and NOx-N concentrations to

169

calculate a single response variable that would provide simpler representation of the

170

effect of soil media inorganic N. The PCA analysis explained 83% of the variation

171

between nitrite-nitrate and ammonium concentrations – the principle component values

172

were then used for further analysis.

173 174 175

Data Analysis Vegetation in the BRCs surveyed in this study was broadly classified into three

176

types: grassed, landscaped, and overgrown. Grassed BRCs (n=4) were planted with

177

sod, did not contain a mulch layer, and did not contain forb, shrub, or woody plant

178

growth. Landscaped BRCs (n=12) contained a combination of herbaceous plants,

179

shrubs, woody plant species, were typically mulched, and were well-maintained to

180

prevent volunteer species from overgrowing the planted vegetation. Overgrown BRCs

181

(n=7) had not been well maintained and, although they contained some of the original

182

planted species, were primarily dominated by dense volunteer vegetation. To


Page 9 of 36


183

investigate the effect of the original composition of the soil medium, BRCs were divided

184

into two categories: mixes comprised of ≤ 50% sand (n=8) and those with ≥ 80% sand

185

(n=11). The original media mixes of the remaining four BRCs were unknown. It is

186

important to note that denitrification potential estimates were available from only one

187

BRC with a medium ≤ 50% sand, so only four samples from one BRC were available for

188

that comparison. Measurements of denitrifying gene abundances, however, were

189

available for all BRCs.

190

Linear regression models were used to investigate the relationship between

191

denitrification response variables and the design and environmental data collected on

192

BRCs. Multi-model inference using Akaike’s Information Criterion (AIC), a model

193

selection approach, was used to identify which models were most important in

194

predicting the denitrifying response variables.49 Akaike’s Information Criterion

195

incorporates a likelihood function that is corrected for the number of variables included

196

in a particular model, which allows for the determination of variables that most influence

197

the observed data and prevent continual addition of an increasing number of variables

198

to a model. We used a corrected variation of the AIC (AICc), which is a second order

199

formula recommended for analyses when the number of parameters is not large in

200

comparison to the number of samples.49 The result is an AICc value, with lower values

201

identifying more probable models. This approach has become increasingly popular in

202

understanding ecological processes and dynamics among datasets from which specific

203

variables cannot be isolated,50-52 which is a common problem in field-based surveys.53

204

The final model selection results were then produced by averaging the top

205

models for the denitrification response variables. Top models included in the averaging



Page 10 of 36

206

were identified as having ∆AICc of < 10, with ∆AICc equaling the difference between the

207

AICc value for the model in question and AICc value of the best model for a given

208

response variable, as proposed by Burnham and Anderson.49 The reported values for

209

each factor represent a relative variable importance (RVI), which are calculated by

210

summing the AIC weights across the top models and range from 0 to 1. Variables with

211

RVI approaching one have a greater model weight or appear more often in the top

212

models or both.

213

To further reduce the risk of identifying spurious correlations that can result from

214

random data dredging, the model selection approach was used in a structured

215

hypothetical manner to evaluate the relative importance of design and environmental

216

variables. Variables that are environmentally influenced and those that are fixed at the

217

time of design and construction were initially separated and analyzed independently to

218

reduce complexity (Table 1). After identifying the top models for each category

219

(environmental and design), the predictor variables from only the top models in each

220

category were then combined and analyzed together. The most important variables

221

were determined by those with the greatest RVI value and variables that had an RVI of

222

> 80% of the top RVI value. The intent of this approach was to eliminate elements which

223

indicated a marginal effect on the predictor variable and ultimately determine if the most

224

influential variables were environmental in nature or could be manipulated by design.

225

Predictor variables were transformed using a log or square root transformation to

226

reduce non-normal distributions in the dataset. This provided a better representation of

227

the gradients within our data and allowed for the identification of fewer top models with

228

higher confidences. Following the model selection process, Wilcoxon non-parametric


Page 11 of 36


229

tests (P < 0.05) and continuous variable regression ANOVA (P < 0.05) were used to

230

identify significant differences and examine relationships between categorical and

231

continuous variables for each denitrification response variable. Model selection

232

analyses were carried out using the MuMIn package54 in R55 and statistical significance

233

tests were determined using JMP.56

234 235

Results and Discussion

236

Model Selection Results

237

The primary objective of this research was to identify which factors or

238

combination of factors most strongly correlated with denitrification in established BRCs

239

and to determine whether those factors can be manipulated in future BRC designs to

240

potentially increase nitrogen removal rates. Given that numerous potential predictor

241

variables were measured (Table 1), it is not possible to present relationships between

242

denitrifiers and all measured predictor variables. Therefore, we used the model

243

selection process to identify which factors were the best predictors of denitrification

244

gene abundance and denitrification potential as response variables. This initial analysis

245

allowed us to focus the remainder of the discussion on the factors with the most

246

potential for facilitating increased denitrification.

247

The results of the complete multi-model inference analysis of all possible

248

measured predictor variables (Table 1) identified soil inorganic N, vegetation type, soil

249

TOC, medium composition, soil medium depth, and sampling depth to be the most

250

important variables affecting denitrification in the surveyed BRCs (Table 2). While this

251

still encompasses several factors that could have an impact, it is important to note that,



Page 12 of 36

252

from a water quality management and engineering perspective, each of these variables

253

can be manipulated by BRC design. This is a key finding because it results from a

254

uniquely comprehensive examination of many established BRCs to show that original

255

design factors have lasting and dominant effects on BRC function compared to local

256

environmental variables.

257

While denitrification potential is representative of complete denitrification, the

258

denitrification genes quantified in this study represent different steps in the

259

denitrification pathway. It’s important to recognize that the denitrification response

260

variables correlated most strongly with different factors (Table 2), suggesting that

261

individual steps in the denitrification pathway are affected differently. For example,

262

inorganic N, sampling depth, and vegetation type significantly affect all of the measured

263

response variables, suggesting they have broad control over the complete denitrification

264

pathway. In contrast, TOC was a strong predictor of nirK abundance and denitrification

265

potential, but it did not correlate well with nosZ abundance, which seemed to respond

266

more strongly to total BRC depth and medium composition. This distinction is critical

267

because nosZ codes for the enzyme that converts nitrous oxide (a potent greenhouse

268

gas) to inert N2, suggesting that some design factors may affect two key impacts of

269

bioretention systems – dissolved nitrogen removal and greenhouse gas emission – in

270

different ways. In the following sections, we further describe each of these important

271

factors independently and discuss how they could potentially be used to improve BRC

272

design.

273 274

Vertical Stratification as a Denitrification Control


Page 13 of 36


275

Overall, indicators of denitrification activity were significantly elevated in the

276

topmost 10 cm of BRCs with nirK abundance, nosZ abundance, and denitrification

277

potential rates respectively averaging 5.7, 3.6, and 23 times higher than in the

278

bottommost 10 cm (Figure 1). These results are similar to recent studies that also found

279

lower concentrations of denitrifying bacteria in the deeper layers of single BRCs,36, 37

280

suggesting that this is a widespread phenomenon. The current paradigm is that lower

281

layers of BRCs, particularly those with ISZs, facilitate greater denitrification because

282

they stay saturated (and thus anaerobic) for longer periods of time. However,

283

measurements of proportionally greater denitrification in the upper layers suggests that

284

large fractions of many BRCs may not be performing as designed.

285

We hypothesize that, over time, particulate organic matter carried in stormwater

286

runoff accumulates primarily in the surface layer of the medium. This results in limited

287

amounts of available N and C in the deeper layers, which restrict denitrifying

288

populations. In our samples, mean N and C concentrations were 5.0 and 2.9 times

289

higher, respectively, in the top 10 cm compared to the bottom 10 cm of soil medium and

290

only one of the 5 BRCs with an ISZ documented the addition of a carbon source to the

291

saturated zone. Furthermore, denitrification gene abundances and denitrification

292

potential were actually lower in the bottom 10 cm samples of BRCs that included an ISZ

293

(Figure S3) compared to conventionally drained designs, casting doubt on whether

294

ISZs, which are currently recommended in some states,57, 58 promote denitrification in

295

the lower layers as they are currently designed.

296 297

Higher abundances of denitrification genes and rates of denitrification activity in the surface layers, where aerobic conditions are expected to dominate, suggest that



Page 14 of 36

298

substantial amounts of denitrification may occur either in anaerobic microsites or during

299

very brief inundation periods following each storm event. Additionally, degradation of

300

plant matter and the accumulation of particulate OM in the upper layers of BRCs may

301

result in the persistence of denitrification “hot spots” within the system that could explain

302

the higher denitrification rates in the upper layers of BRCs, as has been seen in other

303

soil environments.59 Further research on how denitrifier abundance and activity are

304

affected by C, N, and oxygen at different time scales and in different soil

305

microenvironments will help determine how to best promote nitrogen removal

306

throughout the entire volume of the BRC. Particularly in these systems, due to common

307

size and volume constraints and the mobile nature of NOx species, exploiting the entire

308

vertical profile of BRCs must be a vital strategy for reducing the export of stormwater

309

nutrients.

310 311 312

Vegetation as a Denitrification Control Relative to other factors, vegetation type was a strong predictor of microbial

313

denitrification. Specifically, denitrifying populations and activity were significantly

314

decreased in grassed BRCs (Figure 2), where mean nirK abundance, nosZ abundance,

315

and denitrification potential were respectively 5.5, 2.5, and 27 times lower than in BRCs

316

that were categorized as landscaped or overgrown. An important caveat exists in our

317

data because many BRCs that were classified as grassed were also constructed with

318

media containing a high sand content. As a result, there is some confounding of these

319

two effects in this study, making it difficult to separate causation from correlation.

320

However, there are clear biological mechanism that could be responsible for such an


Page 15 of 36


321

effect. First, grassed BRCs do not contain a mulch layer that could capture organic N.

322

Using BRC mesocosms containing a 2.5 cm mulch layer and shrub vegetation, Davis et

323

al.13 found that the majority (> 50%) of total Kjeldahl nitrogen (TKN) removal occurred in

324

the top 20 cm of the soil medium, suggesting significant sorption potential in the mulch

325

layer and captured organic matter. In our data, ammonium (NH4-N) was 3.0 and 4.3

326

times higher while nitrite-nitrate (NO2-NO3-N) was 13.7 and 13.6 times higher in the soil

327

medium of landscaped and overgrown BRCs, respectively, than in grassed BRCs. Once

328

sorbed onto mulch and organic matter, mineralization of organic N can fuel

329

denitrification and potentially a greater abundance of denitrifiers in BRCs containing

330

herbaceous and woody vegetation types. However, given that our data represent only a

331

single point in time, additional work is required to understand the balance among the

332

amount of nitrogen available in the soil medium, denitrifier abundance, and the rate at

333

which N may be leached, denitrified, or otherwise lost from the system.

334

A second important aspect of grassed BRCs is the dense rooting systems that

335

may compete with bacteria for available N between saturation events. For example,

336

Passeport et al.31 found grassed BRCs could remove N as efficiently as BRCs

337

containing “traditional” vegetative cover (trees, shrubs, & mulch) and Payne et al.60

338

observed that > 97% of nitrate in BRC soil columns was plant assimilated. Considering

339

nitrification occurs in the shallow aerobic zones of BRCs, vegetation with high nutrient

340

accumulation potential and dense rooting systems in these shallow layers may capture

341

and reduce the export of nitrified products. Multiple studies have confirmed the ability of

342

high biomass plant species, particularly those with deep rooting structures, to assimilate

343

a significant portion of stormwater N in BRC mesocosms.61-63 This is a critical



Page 16 of 36

344

consideration especially if, as discussed in the previous section, BRCs are

345

underperforming with regard to microbial denitrification in deeper layers. In that case, in

346

addition to deep rooting plant species, engineering improvements that facilitate

347

denitrification from the deeper layers could be an important factor for maximizing N

348

removal in bioretention systems. It is also important to remember that plant assimilation

349

only works as a N removal mechanism when plant biomass is captured and removed,

350

so improved denitrification would be a significant advantage in BRCs that are not well

351

maintained. Overall, both plant and microbial processes can clearly be important for N

352

removal and increased ecological understanding of these systems is required to

353

facilitate complimentary mechanisms to achieve significant improvements in total

354

reductions.

355 356 357

Soil Nutrients as a Denitrification Control Extractable TOC concentrations in the BRC soil medium samples ranged from

358

below detection to 1.02 mg/g dry soil with a mean value of 0.16 mg/g dry soil. All

359

denitrification response variables had a statistically significant linear correlation with

360

total organic carbon (Figure 3). These findings are different than those of Chen et al.,36

361

who found significant effects of organic matter on a different nitrite reductase gene, nirS,

362

but not on nirK or nosZ gene abundances. Biologically, a relationship between TOC and

363

denitrifiers is expected considering that most are heterotrophs that utilize organic

364

carbon as an electron donor. Although our results indicate that increased organic matter

365

could facilitate more denitrifiers, from a water treatment perspective, the benefits of

366

increased denitrification could be negated if the added organic matter is also a source of


Page 17 of 36


367

leachable carbon and nutrients.11 However, there is evidence that some organic carbon

368

sources can fuel the microbial community but not contribute to nutrient leaching. For

369

example, Peterson et al.26 had success with the addition of woodchips and found

370

optimal N removal rates in gravel bioretention soil columns using 5 mm woodchips at a

371

quantity of 4.5% of the soil medium by mass. Presumably, the high C:N ratio of

372

woodchips resulted in minimal N leaching and a slow decomposition rate that

373

continuously supplied the microbial community with organic carbon.

374

Denitrification potential and denitrifying gene abundances also all had significant

375

correlations with inorganic soil N (Figure 3). Extractable ammonium concentrations

376

ranged from below detection to 136 µg/g dry soil with a mean of 12.9 µg/g dry soil.

377

Extractable nitrite – nitrate (NO2--NO3-) concentrations ranged from below detection to

378

38.2 µg/g dry soil with a mean of 6.27 µg/g dry soil. This particular effect represents

379

something of a design paradox since stimulating denitrification by adding N would

380

clearly be counterproductive. However, the N cycle is biologically complex and the

381

amount of inorganic N available to denitrifiers is a function of not just total nitrogen in

382

the medium but also other factors, including how quickly N is mineralized from

383

stormwater inputs, decomposition rates of organic N in the soil medium, and turnover

384

from plant and microbial biomass. Furthermore, after mineralization, the mobility of N

385

species and their availability to denitrifiers is affected by whether they are subsequently

386

nitrified during aerobic conditions. Relating these complex dynamics to BRC design is

387

particularly challenging because stormwater N generally cannot be controlled, so

388

increasing N removal efficiency in these systems will most benefit from optimization of



Page 18 of 36

389

organic matter in the starting soil media, promoting beneficial microbial transformations,

390

and increasing vegetative uptake.

391

On average, denitrifying gene abundances and potential denitrification rates were

392

higher in the BRCs containing low sand media (i.e., higher starting amounts of organic

393

matter), where they respectively averaged 1.1, 2.9, and 1.5 times higher nirK, nosZ, and

394

denitrification potential than in BRCs with high sand contents (Figure 4). Due to the low

395

sample size available for denitrification potential within BRCs with low sand content, the

396

relationship between denitrification potential and media mix serve only as a preliminary

397

assessment and will benefit from future studies. Organic carbon and nitrite-nitrate

398

concentrations also had higher averages in the low sand media mixes, but these

399

differences were not statistically significant in our data set. While increasing nutrient

400

availability via organic matter in the media mix may have the ability to enhance the

401

denitrifying microbial community, there are practical limitations to this strategy. For

402

example, low sand mixes are prone to compaction and clogging,64 are typically more

403

expensive,65 and have been shown to leach DON and phosphorus.9, 22 Constructing

404

BRCs with high sand mixtures and supplementing with organic carbon sources of more

405

favorable C:N ratios or N forms, particularly in specific underperforming areas of the

406

BRC, may be the most feasible design options that facilitate infiltration and optimize N

407

removal through denitrification.

408 409

Denitrifying groups nosZ vs. nirK

410

Differences between abundances of nirK and nosZ genes are particularly

411

important because a lack of nosZ activity is an indicator of incomplete denitrification that


Page 19 of 36


412

could result in the emission of nitrous oxide, a potent greenhouse gas.66 As

413

demonstrated by the RVI data based on the model selection analysis (Table 2), the

414

abundance of the nirK nitrite reductase gene was most affected by sample depth and

415

vegetation type. TOC and inorganic N were also included in the top models but by

416

comparison were less important. For nosZ abundance, however, inorganic N and

417

sample depth were the most important variables followed by total soil medium depth,

418

vegetation type, and medium composition, respectively. Interestingly, unlike nirK and

419

denitrification potential, the model selection results identified medium composition and

420

soil medium depth to be important factors affecting nosZ gene populations specifically.

421

Relative to each other, mean nirK and nosZ gene abundances were weakly

422

correlated (Figure 5A), but on a per sample basis, nirK abundances were over an order

423

of magnitude higher than nosZ (Figure 5B). To our knowledge, relationships between

424

nitrous oxide emissions and denitrifying gene abundances have not been studied in

425

BRCs. However, the nirK/nosZ gene ratio is positively correlated with nitrous oxide

426

emissions in other soil environments.67-69 Additionally, nosZ gene transcription can be

427

suppressed at low oxygen concentrations, while nirK expression generally exhibits a

428

higher oxygen tolerance.70, 71 Given that we found the highest denitrification activity in

429

the upper (i.e., frequently aerobic) layers of soil medium, these relationships suggest

430

incomplete denitrification is highly likely and that nitrous oxide is potentially being

431

emitted from many BRCs. And while N2O emissions in BRCs are estimated to be

432

equivalent to that of other urban landscape features,72, 73 future design

433

recommendations should seek to minimize nitrous oxide emissions. To address this



Page 20 of 36

434

topic, quantifying nitrous oxide flux, rather than simply quantifying N removal from these

435

systems, should be included when possible in future research and monitoring.

436 437 438

Implications for Future Design Nitrogen removal continues to be a challenging facet of bioretention

439

performance. This research suggests that denitrifying bacteria are largely influenced by

440

the design of BRCs rather than local environmental factors, and that nitrogen removal

441

efficiency can potentially be enhanced by simple design alterations. Specifically, these

442

alterations include constructing BRCs with media mixes that provide the microbial

443

community with non-leachable nutrients, creating synergic relationships between high

444

biomass vegetation and denitrifiers, and by encouraging denitrification in the lower

445

layers by supplementing with C sources. Further research on denitrification in BRCs is

446

needed to determine if these findings can be used reliably to reduce net N export from

447

these systems at a large scale.

448 449

Acknowledgments

450

The authors would like to thank the Science Museum of Virginia, Dr. Allen Davis, the

451

North Carolina Department of Transportation, and the Montgomery County, MD DEP for

452

supplying BRC design information that was critical to this research. The authors would

453

also like to thank the Virginia Tech Institute for Critical Technology and Applied Science

454

for the funding that made this research possible.

455 456

Supporting Information


Page 21 of 36


457

Supporting information is available for this manuscript which includes bioretention cell

458

locations and sources of design information, qPCR thermal profiles and reaction

459

mixtures, sampling map, sampling locations within bioretention cells, gene abundances

460

in bottom 10cm of BRCs with and without ISZs, number of BRCs classified within

461

categorical variables, and mean and range data for continuous variables collected on

462

BRCs. This information is available free of charge via the Internet at http://pubs.acs.org.

463 464

Corresponding Author Contact

465 466 467 468 469 470

Address: RB1880 Suite 1129, Room 1121 Blacksburg, Virginia 24061 E-Mail: [email protected]

471 472 473



474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518

Page 22 of 36

References (1) Li, H.; Davis, A. P., Water quality improvement through reductions of pollutant loads using bioretention. J. Environ. Eng. 2009, 135, (8), 567-576. (2) Sun, X.; Davis, A. P., Heavy metal fates in laboratory bioretention systems. Chemosphere 2007, 66, (9), 1601-1609. (3) Randall, M. T.; Bradford, A., Bioretention gardens for improved nutrient removal. Water Qual. Res. J. Can. 2013, 48, (4), 372-386. (4) DeBusk, K. M.; Wynn, T. M., Storm-water bioretention for runoff quality and quantity mitigation. J. Environ. Eng. 2011, 137, (9), 800-808. (5) DiBlasi, C. J.; Li, H.; Davis, A. P.; Ghosh, U., Removal and fate of polycyclic aromatic hydrocarbon pollutants in an urban stormwater bioretention facility. Environ. Sci. Technol. 2009, 43, (2), 494-502. (6) Davis, A. P., Field performance of bioretention: Water quality. Environ. Eng. Sci. 2007, 24, (8), 1048-1064. (7) Hunt, W. F.; Jarrett, A. R.; Smith, J. T.; Sharkey, L. J., Evaluating bioretention hydrology and nutrient removal at three field sites in North Carolina. J. Irrig. Drain. Eng. 2006, 132, (6), 600-608. (8) Hsieh, C. H.; Davis, A. P., Evaluation and optimization of bioretention media for treatment of urban storm water runoff. J. Environ. Eng. 2005, 131, (11), 1521-1531. (9) Hatt, B. E.; Fletcher, T. D.; Deletic, A., Hydrologic and pollutant removal performance of stormwater biofiltration systems at the field scale. J. Hydrol. 2009, 365, (3-4), 310-321. (10) Taylor, G. D.; Fletcher, T. D.; Wong, T. H. F.; Breen, P. F.; Duncan, H. P., Nitrogen composition in urban runoff - implications for stormwater management. Water Res. 2005, 39, (10), 1982-1989. (11) Li, L.; Davis, A. P., Urban stormwater runoff nitrogen composition and fate in bioretention systems. Environ. Sci. Technol. 2014, 48, (6), 3403-3410. (12) Read, J.; Wevill, T.; Fletcher, T.; Deletic, A., Variation among plant species in pollutant removal from stormwater in biofiltration systems. Water Res. 2008, 42, (4), 893-902. (13) Davis, A. P.; Shokouhian, M.; Sharma, H.; Minami, C., Water quality improvement through bioretention media: Nitrogen and phosphorus removal. Water Environ. Res. 2006, 78, (3), 284-293. (14) Philippot, L.; Hallin, S.; Schloter, M., Ecology of denitrifying prokaryotes in agricultural soil. Adv. Agron. 2007, 96, 249-305. (15) Kim, H.; Seagren, E. A.; Davis, A. P., Engineered bioretention for removal of nitrate from stormwater runoff. Water Environ. Res. 2003, 75, (4), 355-367. (16) Lucas, W. C.; Greenway, M., Hydraulic response and nitrogen retention in bioretention mesocosms with regulated outlets: part II--nitrogen retention. Water Environ. Res. 2011, 83, (8), 703-713. (17) Zinger, Y.; Fletcher, T. D.; Deletic, A.; Blecken, G. T.; Viklander, M. In Optimisation of the nitrogen retention capacity of stormwater biofiltration systems. Novatech 2007: 6th International Conference on Sustainable Techniques and Strategies in Urban Water Management, Lyon, France, June 25-28, 2007; Groupe de Recherche Rhone-Alpes sur les Infrastructures et l’Eau: Villeurbanne, France, 2007.


Page 23 of 36

519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563


(18) Dietz, M. E.; Clausen, J. C., Saturation to improve pollutant retention in a rain garden. Environ. Sci. Technol. 2006, 40, (4), 1335-1340. (19) Lucas, W. C.; Greenway, M., Nutrient retention in vegetated and nonvegetated bioretention mesocosms. J. Irrig. Drain. Eng. 2008, 134, (5), 613-623. (20) Turk, R. P.; Kraus, H. T.; Hunt, W. F.; Carmen, N. B.; Bilderback, T. E., Nutrient sequestration by vegetation in bioretention cells receiving high nutrient loads. J. Environ. Eng. 2017, 143, (2), 06016009. (21) Liu, J.; Sample, D. J.; Owen, J. S.; Li, J.; Evanylo, G., Assessment of selected bioretention blends for nutrient retention using mesocosm experiments. J. Environ. Qual. 2014, 43, (5), 1754-1763. (22) Clark, S.; Pitt, R., Storm-water filter media pollutant retention under aerobic versus anaerobic conditions. J. Environ. Eng. 2009, 135, (5), 367-371. (23) Luell, S. K.; Hunt, W. F.; Winston, R. J., Evaluation of undersized bioretention stormwater control measures for treatment of highway bridge deck runoff. Water Sci. Technol. 2011, 64, (4), 974-979. (24) Jones, M. P. Effect of urban stormwater BMPs on runoff temperature in trout sensitive regions. Dissertation, North Carolina State University, Raleigh, N.C., 2008. (25) Brown, R. A.; Hunt, W. F., Impacts of media depth on effluent water quality and hydrologic performance of undersized bioretention cells. J. Irrig. Drain. Eng. 2011, 137, (3), 132-143. (26) Peterson, I. J.; Igielski, S.; Davis, A. P., Enhanced denitrification in bioretention using woodchips as an organic carbon source. J. Sust. Water in the Built Environ. 2015, 1, (4), 04015004. (27) Henderson, C.; Greenway, M.; Phillips, I., Removal of dissolved nitrogen, phosphorus and carbon from stormwater by biofiltration mesocosms. Water Sci. Technol. 2007, 55, (4), 183-191. (28) Barrett, M. E.; Limouzin, M.; Lawler, D. F., Effects of media and plant selection on biofiltration performance. J. Environ. Eng. 2013, 139, (4), 462-470. (29) Gautam, D. N.; Greenway, M., Nutrient accumulation in five plant species grown in bioretention systems dosed with wastewater. Australas. J. Env. Manage. 2014, 21, (4), 453-462. (30) Manka, B. N.; Hathaway, J. M.; Tirpak, R. A.; He, Q.; Hunt, W. F., Driving forces of effluent nutrient variability in field scale bioretention. Ecol. Eng. 2016, 94, 622-628. (31) Passeport, E.; Hunt, W. F.; Line, D. E.; Smith, R. A.; Brown, R. A., Field study of the ability of two grassed bioretention cells to reduce storm-water runoff pollution. J. Irrig. Drain. Eng. 2009, 135, (4), 505-510. (32) Blecken, G. T.; Zinger, Y.; Deletic, A.; Fletcher, T. D.; Hedstrom, A.; Viklander, M., Laboratory study on stormwater biofiltration nutrient and sediment removal in cold temperatures. J. Hydrol. 2010, 394, (3-4), 507-514. (33) Morse, N. R.; McPhillips, L. E.; Shapleigh, J. P.; Walter, M. T., The Role of Denitrification in Stormwater Detention Basin Treatment of Nitrogen. Environ. Sci. Technol. 2017, 51, (14), 7928-7935. (34) Baker, B. H.; Kroger, R.; Brooks, J. P.; Smith, R. K.; Czarnecki, J. M. P., Investigation of denitrifying microbial communities within an agricultural drainage system fitted with low-grade weirs. Water Res. 2015, 87, 193-201.



564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608

Page 24 of 36

(35) Bettez, N. D.; Groffman, P. M., Denitrification Potential in Stormwater Control Structures and Natural Riparian Zones in an Urban Landscape. Environ. Sci. Technol. 2012, 46, (20), 10909-10917. (36) Chen, X.; Peltier, E.; Sturm, B. S.; Young, C. B., Nitrogen removal and nitrifying and denitrifying bacteria quantification in a stormwater bioretention system. Water Res. 2013, 47, (4), 1691-700. (37) Willard, L. L.; Wynn-Thompson, T.; Krometis, L. H.; Neher, T. P.; Badgley, B. D., Does it pay to be mature? Evaluation of bioretention cell performance seven years postconstruction. J. Environ. Eng. 2017, 143, (9), 04017041. (38) Henry, S.; Baudoin, E.; Lopez-Gutierrez, J. C.; Martin-Laurent, F.; Brauman, A.; Philippot, L., Quantification of denitrifying bacteria in soils by nirK gene targeted realtime PCR. J. Microbiol. Methods 2004, 59, (3), 327-335. (39) Braker, G.; Zhou, J. Z.; Wu, L. Y.; Devol, A. H.; Tiedje, J. M., Nitrite reductase genes (nirK and nirS) as functional markers to investigate diversity of denitrifying bacteria in Pacific northwest marine sediment communities. Appl. Environ. Microbiol. 2000, 66, (5), 2096-2104. (40) Henry, S.; Bru, D.; Stres, B.; Hallet, S.; Philippot, L., Quantitative detection of the nosZ gene, encoding nitrous oxide reductase, and comparison of the abundances of 16S rRNA, narG, nirK, and nosZ genes in soils. Appl. Environ. Microbiol. 2006, 72, (8), 5181-5189. (41) Canfield, D. E.; Glazer, A. N.; Falkowski, P. G., The evolution and future of Earth's nitrogen cycle. Science 2010, 330, (6001), 192-196. (42) Regan, K.; Kammann, C.; Hartung, K.; Lenhart, K.; Muller, C.; Philippot, L.; Kandeler, E.; Marhan, S., Can differences in microbial abundances help explain enhanced N2O emissions in a permanent grassland under elevated atmospheric CO2? Glob. Change Biol. 2011, 17, (10), 3176-3186. (43) Towe, S.; Albert, A.; Kleineidam, K.; Brankatschk, R.; Dumig, A.; Welzl, G.; Munch, J. C.; Zeyer, J.; Schloter, M., Abundance of microbes involved in nitrogen transformation in the rhizosphere of Leucanthemopsis alpina (L.) Heywood grown in soils from different sites of the Damma glacier forefield. Microb. Ecol. 2010, 60, (4), 762770. (44) Harter, J.; Krause, H. M.; Schuettler, S.; Ruser, R.; Fromme, M.; Scholten, T.; Kappler, A.; Behrens, S., Linking N2O emissions from biochar-amended soil to the structure and function of the N-cycling microbial community. ISME J. 2014, 8, (3), 660674. (45) Smith, M. S.; Tiedje, J. M., Phases of denitrification following oxygen depletion in soil. Soil Biol. Biochem. 1979, 11, (3), 261-267. (46) Drury, C. F.; Myrold, D. D.; Beauchamp, E. G.; Reynolds, W. D., Denitrification Techniques for Soils. In Soil Sampling and Methods of Analysis, Second Edition; Carter, M. R.; Gregorich, E. G., Eds. CRC Press: Boca Raton, FL, 2007; pp 471-493. (47) Federation, W. E., Standard Methods for the Examination of Water and Wastewater. American Public Health Association (APHA): Washington, DC, 2005. (48) Lachat Instruments. Determiniation of Nitrate/Nitrite by Flow Injection Analysis. Low Flow Method. QuikChem Method 10-107-04-1-L. Lachat Instruments: Loveland, CO, 2007.


Page 25 of 36

609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653


(49) Burnham, K. P.; Anderson, D. R., Model selection and multimodel inference: a practical information-theoretic approach. Springer Science & Business Media: 2003. (50) Symonds, M. R. E.; Moussalli, A., A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Akaike's information criterion. Behav. Ecol. Sociobiol. 2011, 65, (1), 13-21. (51) Richards, S. A.; Whittingham, M. J.; Stephens, P. A., Model selection and model averaging in behavioural ecology: the utility of the IT-AIC framework. Behav. Ecol. Sociobiol. 2011, 65, (1), 77-89. (52) Johnson, J. B.; Omland, K. S., Model selection in ecology and evolution. Trends Ecol. Evol. 2004, 19, (2), 101-108. (53) Richards, S. A., Testing ecological theory using the information-theoretic approach: Examples and cautionary results. Ecology 2005, 86, (10), 2805-2814. (54) Bartoń, K. MuMIn: Multi-Model Inference, R package version 1.40.4; 2018. (55) R Core Team. R: A language and environment for statistical computing., R Foundation for Statistical Computing: Vienna, Austria, 2017. (56) JMP, Version 13; SAS Institute Inc: Cary, NC, 2007. (57) NCDENR. Stormwater Best Management Practices Manual. North Carolina Department of Environment and Natural Resources: Raleigh, NC, 2007. (58) ARC. Georgia Stormwater Management Manual – Vol. 2: Technical Handbook. Atlanta Regional Commission: Atlanta, GA, 2016. (59) Parkin, T. B., Soil microsites as a source of denitrification variability. Soil Sci. Soc. Am. J. 1987, 51, (5), 1194-1199. (60) Payne, E. G.; Fletcher, T. D.; Russell, D. G.; Grace, M. R.; Cavagnaro, T. R.; Evrard, V.; Deletic, A.; Hatt, B. E.; Cook, P. L., Temporary storage or permanent removal? The division of nitrogen between biotic assimilation and denitrification in stormwater biofiltration systems. PLoS One 2014, 9, (3), e90890. (61) Cording, A. Evaluating stormwater pollutant removal mechanisms by bioretention in the context of climate change. Dissertation, The University of Vermont, Burlington, VT, 2016. (62) Glaister, B. J.; Fletcher, T. D.; Cook, P. L. M.; Hatt, B. E., Interactions between design, plant growth and the treatment performance of stormwater biofilters. Ecol. Eng. 2017, 105, 21-31. (63) Rycewicz-Borecki, M.; McLean, J. E.; Dupont, R. R., Nitrogen and phosphorus mass balance, retention and uptake in six plant species grown in stormwater bioretention microcosms. Ecol. Eng. 2017, 99, 409-416. (64) Hatt, B. E.; Fletcher, T. D.; Deletic, A., Hydraulic and pollutant removal performance of fine media stormwater filtration systems. Environ. Sci. Technol. 2008, 42, (7), 2535-2541. (65) Davis, A. P.; Hunt, W. F.; Traver, R. G.; Clar, M., Bioretention technology: Overview of current practice and future needs. J. Environ. Eng. 2009, 135, (3), 109-117. (66) Zumft, W. G., The biological role of nitric oxide in bacteria. Arch. Microbiol. 1993, 160, (4), 253-264. (67) Philippot, L.; Andert, J.; Jones, C. M.; Bru, D.; Hallin, S., Importance of denitrifiers lacking the genes encoding the nitrous oxide reductase for N2O emissions from soil. Glob. Change Biol. 2011, 17, (3), 1497-1504.



654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673

Page 26 of 36

(68) Li, S.; Song, L.; Jin, Y.; Liu, S.; Shen, Q.; Zou, J., Linking N2O emission from biochar-amended composting process to the abundance of denitrify (nirK and nosZ) bacteria community. AMB Express 2016, 6, (37), 1-9. (69) Li, S.; Song, L.; Gao, X.; Jin, Y.; Liu, S.; Shen, Q.; Zou, J., Microbial abundances predict methane and nitrous oxide fluxes from a windrow composting system. Front. Microbiol. 2017, 8, 409. (70) Bergaust, L.; Shapleigh, J.; Frostegard, A.; Bakken, L., Transcription and activities of NOx reductases in Agrobacterium tumefaciens: the influence of nitrate, nitrite and oxygen availability. Environ. Microbiol. 2008, 10, (11), 3070-81. (71) Dalsgaard, T.; Stewart, F. J.; Thamdrup, B.; De Brabandere, L.; Revsbech, N. P.; Ulloa, O.; Canfield, D. E.; DeLong, E. F., Oxygen at nanomolar levels reversibly suppresses process rates and gene expression in anammox and denitrification in the oxygen minimum zone off northern Chile. Mbio 2014, 5, (6), e01966-14. (72) McPhillips, L.; Goodale, C.; Walter, M. T., Nutrient leaching and greenhouse gas emissions in grassed detention and bioretention stormwater basins. J. Sust. Water in the Built Environ. 2018, 4, (1), 04017014. (73) Grover, S. P. P.; Cohan, A.; Sen Chan, H.; Livesley, S. J.; Beringer, J.; Daly, E., Occasional large emissions of nitrous oxide and methane observed in stormwater biofiltration systems. Sci. Total Environ. 2013, 465, 64-71.


Page 27 of 36


674

Figure and Table Captions

675

Table 1. Designation of design and environmental factors used as predictor variables in

676

the model selection analyses to predict denitrifier abundance and activity (response

677

variables).

678 679

Table 2. Model averaging results for predicting nirK abundance, nosZ abundance, and

680

denitrification potential. RVI values are presented on a scale from 0 to 1. Variables with

681

higher values (i.e. closest to 1) are considered to have a greater effect on the response

682

variable. Cells that do not contain an RVI value indicates that the explanatory variable

683

was not included in the top models for the denitrification response variable.

684 685

Figure 1. Mean denitrifying gene abundances and denitrification potential in top 10 cm

686

vs. the bottom 10 cm of BRC soil media samples.

687 688

Figure 2. Mean values for nirK abundance, nosZ abundance, and denitrification

689

potential for all soil medium samples from BRCs with grassed (n=15), landscaped

690

(n=48), and overgrown (n=24) vegetation schemes.

691 692

Figure 3. Linear regressions for total organic carbon (TOC) and inorganic N

693

concentrations in the soil medium vs. nirK abundance, nosZ abundance, and

694

denitrification potential across all BRC medium samples. All regressions are statistically

695

significant (p < 0.05).

696



697

Figure 4. Mean values for nirK abundance, nosZ abundance, and denitrification

698

potential in all soil medium samples from BRCs with known media mixes containing ≤

699

50% sand (n=32) and ≥ 80% sand (n=40).

Page 28 of 36

700 701

Figure 5. Linear regression of nosZ vs nirK gene abundances in each sample. Panel A

702

represents correlation between the abundances of each gene and Panel B presents the

703

same data on equally scaled axes to compare relative abundance. The dotted line

704

represents a 1:1 ratio of nosZ to nirK that has been presented as an ‘ideal’ ratio to

705

prevent nitrous oxide emission.


Page 29 of 36

706


Table 1. Design Factors

Environmental Factors

Vegetation (Grassed vs. Landscaped vs. Overgrown)

Soil Medium Total Organic Carbon

BRC:Catchment Surface Area Ratio

Region (Piedmont vs. Coastal Plain)

Presence/Absence of Saturated Zone

BRC Age

Soil Medium Composition

Soil Medium pH

(≤ 50% sand vs. ≥ 80% sand) Soil Medium Depth

Mean Annual Temperature & Precipitation

Sample Depth (Top vs Bottom)

Soil Medium Ammonium & Nitrite-Nitrate

707



708

Page 30 of 36

Table 2.

DIN

Sampling Depth

Vegetation Type

TOC

nirK

0.46

0.96

0.78

0.64

Denitrification Potential

1.00

0.82

1.00

1.00

nosZ

1.00

1.00

0.91

Soil Medium Composition

Soil Medium Depth

0.85

0.97

709


Page 31 of 36

710


Figure 1.

711



712

Page 32 of 36

Figure 2.

713


Page 33 of 36

714


Figure 3.

715



716

Page 34 of 36

Figure 4.

717


Page 35 of 36

718


Figure 5.

719 720



TOC Art 84x47mm (225 x 225 DPI)

ACS Paragon Plus Environment

Page 36 of 36

Engineered and Environmental Controls of ... - ACS Publications

Recommend Documents