Abstract
Employing rooftops for the cultivation of crops in limited urban space has garnered interest in densely populated cities in the United States, where there is a growing demand for locally sourced vegetable products. Fertility management recommendations for rooftop farming, however, are scant. With insufficient research on nutrient cycling within rooftop farming systems, which tend to use soilless substrates with low organic matter content, the potential tradeoffs between the negative impacts (e.g., nutrient runoff) and the benefits (e.g., increased locally produced vegetables, stormwater retention, etc.) associated with rooftop farms are unclear. The objective of this study was to evaluate the effects of organic and inorganic nitrogen (N) inputs on the N dynamics within substrate typically used on rooftop farms. Substrate without added N inputs (control) was compared with substrates receiving N sources that are both realistic for and/or reflective of amendments currently applied on urban rooftop farms: a synthetic fertilizer (Osmocote® 14N–4.2P–11.6K), and three organic N inputs—composted poultry manure, municipal green waste (MGW) compost, and vermicompost. Aboveground crop biomass and yields of Beta vulgaris (swiss chard), along with inorganic N availability (ammonium:
Expanding agricultural capacity within or close to urban centers has the potential to supplement the growing demand for locally grown produce, while also reducing food transportation costs, lowering greenhouse gas emissions associated with food transportation, increasing food and economic security, and making a positive change in our food systems (Norberg-Hodge et al., 2002). Because land of suitable agricultural quality is often in high demand, and other forms of development are usually more economically viable, land availability is a major obstacle for urban farmers (Whittinghill and Rowe, 2012). Employing rooftops for the cultivation of crops in limited urban space has garnered interest and popularity in densely populated cities in the United States (e.g., Brooklyn Grange Farm in Queens and Brooklyn, NY; Eagle Street Rooftop Farm in Brooklyn, NY; Uncommon Ground, Chicago, IL). Experience with fertility management (e.g., basic N recommendations) and understanding of the environmental impacts associated with rooftop farms, however, is lacking.
Studies have shown that green roofs can provide multiple environmental benefits to urban areas, including attenuating surface temperatures, mitigating the urban heat island (UHI) effect, managing stormwater retention and flow, and restoring ecological habitat and biodiversity (Brenneisen, 2006; Oberndorfer et al., 2007; Solecki et al., 2005; Teemusk and Mander, 2007; VanWoert et al., 2005). While a rooftop farm essentially applies green roof technology for food production on a building roof, the aforementioned benefits associated with landscaped green roofs may not necessarily be associated with rooftop farms. A key question regarding the environmental impacts of rooftop farms is how does the fertility management of rooftop farms impact stormwater retention, detention, and pollution attenuation? One reason to expect a reduction in environmental benefits is rooftop farms operate with seasonal vegetation, i.e., biomass is low at planting and is then harvested at the end of the season; whereas, landscaped green roofs have perennial biomass cover. It has been suggested that plant establishment and the reduction of labile organic matter in the substrate via decomposition and assimilation would decrease the
A number of factors (e.g., substrate composition, type of vegetation, and fertilizer additions) can potentially influence the quantity and quality of runoff from vegetated roofs (Czemiel Berndtsson, 2010; Hathaway et al., 2008). Studies evaluating the properties of the substrate material (Moran et al., 2005) and fertilizer application rate (Emilsson et al., 2007) have found that green roof systems can act as sources of
The type and form of N inputs play a major role in determining the fate and the quantity of N that will be available for plant uptake as well as N loss from an agricultural system (e.g., Goh and Haynes, 1986) and from green roofs (e.g., Clark and Zheng, 2013). Compared with synthetic fertilizers, which readily release plant-available N, organic amendments must be mineralized and therefore release mineral N to field crops slower and more gradually throughout the growing season (e.g., Burger and Jackson, 2003; Sikora and Szmidt, 2001). Mineralization of added organic matter is of particular interest to systems such as rooftop farms that apply organic amendments as the primary source of plant-available N. Composts, particularly those made from food scraps and yard trimmings, will likely be more accessible for urban agriculture as cities and countries recognize the impact of diverting organic waste from landfills on greenhouse gas emissions (Materials Management & Product Stewardship Workgroup–West Coast Climate and Materials Management Forum, 2011). While the gradual release of inorganic N from composts could also reduce N losses and substantial evidence indicates that compost applications can improve soil conditions (e.g., Roe, 2001; Smith, 1996), N availability from composts needs to be synchronized with crop demand to optimize crop yield. A better understanding of N dynamics and cycling in rooftop farm substrate would elucidate how to best manage rooftop farm systems for optimal productivity and environmental benefits, while minimizing N losses via runoff.
This study is one of the first investigations of N dynamics in substrate used on rooftop farms. The composts used in this experiment represent feasible N input options for both conventionally and organically managed urban rooftop farms in New York City, where this study took place. Three hypotheses were tested: 1) synthetic fertilizer will provide the highest N availability for crop growth, thereby leading to the highest yields, but the high levels of available N in excess of plant demand will also lead to the greatest loss of N in leachates compared with the other systems, 2) the organic N input with the lowest C:N ratio will release plant available N at a rate most closely synchronized with crop N demand, thereby leading to the smallest N loss among the systems and the highest yields among the organic N treatments, and 3) N release from the organic amendment with the highest C:N will not meet plant N demand and, therefore, lead to reduced yields and the highest leachate-N among the organic N treatments.
Materials and Methods
Experimental design.
Starting in June 2012, an 8-week container experiment was conducted in a climate-controlled aluminum and glass rooftop greenhouse at Barnard College, Columbia University (New York, NY). Beta vulgaris (swiss chard) was selected as the test crop because of its high economic value to local rooftop farm businesses (B. Flanner, personal communication). rooflite® intensive ag (Skyland USA LLC, Avondale, PA), a soilless substrate designed for rooftop farms, consisting of lightweight mineral aggregates and an organic composted component [HydRocks® (Big River Industries, Inc., Livingston, AL and Erwinville, LA) and mushroom compost, respectively], was used as the base substrate (select characteristics shown in Table 1). rooflite® intensive ag, alone, served as the control to be evaluated against blends consisting of rooflite® intensive ag amended with four different N inputs (Table 2): “composted poultry manure” from the Stone Barns Center for Food and Agriculture (Pocantico Hills, NY); MGW compost from the New York City Department of Sanitation, which is available to urban farmers in New York City at no cost; “vermicompost” (compost produced by red wiggler worms, Eisenia foetida) from the Lower East Side Ecology Center (New York, NY), and “synthetic fertilizer” [Osmocote® Smart-Release 14N–4.2P–11.6K; The Scotts Company, Marysville, OH; a resin-coated, granular slow-release fertilizer (The Scott Company, 2009)]. An average of 125, 110, and 170 g of composted poultry manure, vermicompost, and MGW compost, respectively, were added to an average of 730 g base substrate per container; these application rates were based on the 1 compost:4 substrate ratio (by volume) used at the Brooklyn Grange Rooftop Farm (B. Flanner, personal communication). The synthetic fertilizer was added to the substrate according to the manufacturer’s recommendation for vegetable plantings [2.5 g·L−1 substrate (The Scott Company, 2009)]. Estimated rates of N application on an area basis are reported for each N addition in Table 2. The containers for both the control and fertilizer addition treatments contained ≈775 g substrate, which was equivalent in volume to the treatments receiving the organic N inputs (e.g., compost plus base substrate; ≈11 cm total substrate depth).
Select specifications (particle size distribution and water measurements) of the rooftop farm substrate (rooflite® intensive ag, Skyland USA, LLC) used in this study. These details were provided by B. Flanner (personal communication).
Total nitrogen (N) content and carbon-to-nitrogen (C:N) ratio of the organic N input additions and the substrate (control), along with the amount of N applied via the N additions to the different treatments. Mean separation in the same column, among organic N inputs only, by Tukey’s honestly significant difference test at α = 0.05.
The experiment consisted of the control and four N input treatments, replicated five times per 2-week destructive sampling interval [5 treatments × 5 replicates × 5 sampling events = 125 total containers (experimental units)]. The containers (round, white, polypropylene pots; 15 cm diameter; 182 cm2 area; 14 cm height; ≈2 L volume; ITML-Meyers Industries Inc., Canada) were randomly arranged ≈15 cm apart on greenhouse benches. Two B. vulgaris seeds per container were seeded on 19 June 2012 (week 0). Germination was observed on 22 June 2012 and swiss chard plants were subsequently thinned to one shoot per container on 3 July 2012 (week 2). The crop was watered via drip irrigation as needed to prevent wilting and grown to maturity (≈56 d).
Measurements
Crop and substrate sampling and analyses.
Samples of the crop aboveground biomass and substrate were collected biweekly: at the start of the experiment (week 0) and weeks 2, 4, 6, and 8. Crop aboveground biomass (referred to as “yield” for the final harvest during week 8) was collected, weighed for fresh weight, dried at 60 °C until the mass no longer changed, weighed for dry weight, and then ground. From each experimental unit, one subsample of the moist substrate was collected (≈30 g), weighed, passed through a 2-mm sieve, dried at 60 °C, and then ground [in accordance with the standard methodology used for determination of chemical characteristics of green roof substrates (FLL, 2002)]; another subsample of the moist substrate was collected for inorganic N measurements via the Saturated Media Extract (SME) method (described below). Both aboveground crop biomass and substrate samples were analyzed for C and N content using a FLASH EA 1112 elemental analyzer (Thermo Scientific, Cambridge, U.K.).
Plant-available inorganic N (
Measurements of pH, electrical conductivity (EC), total nitrogen (N), carbon-to-nitrogen ratio (C:N), and inorganic N levels (
Leachate collection and analyses.
At the onset of the experiment (week 0) and each following week (weeks 1, 2, 3, 4, 5, 6, 7, and 8), leachate was collected following the PourThru program, adapted from Whipker et al., (2001). Before collecting leachate, all containers were irrigated to saturation. After draining for 1 h, deionized water was added to designated containers (e.g., during weeks 1 and 2, leachate was collected from pots assigned for destructive sampling in week 2) to collect ≈70 mL leachate from each container (≈50% of total volume of deionized water applied). The leachate was subsequently filtered through Whatman no. 1 filter paper and EC and pH were measured. About 30 mL of the leachate was subsampled, 0.22-μm-filtered, and then stored at −20 °C until colorimetric analysis for inorganic N content (in the same manner as that described above for the SME extracts).
PMN measurements.
Potentially mineralizable N estimates have been widely used to assess the effects of various management practices, such as tillage, crop rotations, and fertilization, on N availability (e.g., Campbell and Souster, 1982; Doran, 1987; Franzluebbers et al., 1995). Potentially mineralizable N was measured using a 28-day aerobic incubation study, following the method outlined in Curtin and Campbell for mineral soils (Curtin and Campbell, 2006), which was similar to the method used to measure PMN in a soilless potting mix (Boydston et al., 2008). About 48 h before the start of the incubation, 10 g of 4-mm-sieved moist substrate from each of the treatments (five replicates per treatment) was brought to 20% water holding capacity (WHC) with deionized water. This equilibration or “priming” step is necessary to avoid the period of stimulated microbial activity upon rewetting the substrate. After priming, samples were brought to 60% WHC and sealed within canning jars (946 mL). On day 0 of the incubation, 3 h after being placed in the canning jars, one set of samples was extracted with 100 mL 0.0125 M CaCl2 at a rate of 1 sample:10 extractant (w/v) [according to the guidelines for soil analyses of green roof substrates (FLL, 2002)] and then shaken on a reciprocating shaker at ≈150 rpm for 1 h. The slurries were filtered through a Whatman no. 1 filter, and EC and pH were measured on the extracts. Extracts were 0.22-µm-syringe filtered and stored at −20 °C. The second set of samples was incubated at ≈25 °C for 28 d, and aerated every three days for ≈10 min. On the final day of the incubation (day 28), the samples were removed from the jars and extracted in the same manner as the first set of samples. Frozen extracts were thawed and then analyzed for
Statistical analyses
Analysis of variance (ANOVA) was used to compare the sensitivity of substrate C and N content, yields and crop-N, inorganic N in leachate, pH and EC values, and PMN concentrations with regard to experimental factors (sampling week and N input) using the function aov() of the R software package (R, 2013). Means separation was found using Tukey’s honestly significant difference, and significance was determined at α = 0.05 unless otherwise stated. Because treatments destructively sampled at different sampling periods were represented by distinct experimental units, ANOVA was also used to compare treatment means across sampling periods, rather than apply a multivariate ANOVA repeated analysis. Homogeneity of variances could not be achieved via transformation for:
Results and Discussion
Nitrogen dynamics
Plant-available nitrogen.
With the exception of week 0, substrate
Substrate depths on rooftop farms are similar to intensive green roof depths (>17 cm). To the best of the authors’ knowledge, nutrient application recommendations are, however, not easily obtainable for intensive green roofs planted with vegetable crops. The common practice regarding intensive green roof fertilization is one that is endorsed by green roof expert, Charlie Miller, P.E., and that is to apply care and maintenance as prescribed to a similar garden situation (Roofmeadow, 2015). The total N application via the synthetic fertilizer treatment was similar to an N fertilization recommendation of 136 N kg·ha−1 for field-grown swiss chard (Masabni, 2011); whereas, the total N applied from the composted poultry manure, MGW compost, and vermicompost were 13, 58, and 42% higher than the recomemendation for field-grown swiss chard. Despite the lower N application rate, the synthetic fertilizer treatment provided a more consistent supply of both
Potentially mineralizable nitrogen.
Although the C:N ratios of the composts were expected to be strong predictors of the size of the PMN pools [i.e., the lower the C:N, the larger the PMN pool (Nicolardot et al., 2001)] and there were differences in C:N among the N input treatments at week 0 (Table 3), no significant differences in total PMN were found among treatments (Fig. 1). This suggests that the contribution of mineralization to plant-available N is not different among the treatments. While the lack of differences in PMN among the N input treatments corresponds with the similarities in substrate and inorganic N availability (
Potential nitrogen mineralization (µg N/g moist substrate) measured by 28-day aerobic laboratory incubation. Treatments include control, no added N input to rooftop farm substrate; poultry manure, composted poultry manure; vermicompost; municipal green waste (MGW) compost, composted municipal green waste; and synthetic fertilizer, Osmocote® Smart-Release 14N–4.2P–11.6K. Error bars represent standard error of the mean (n = 5).
Citation: HortScience 50, 6; 10.21273/HORTSCI.50.6.806
Assuming that N mineralization under optimum temperature and moisture in the substrate follows first-order kinetics (Stanford and Smith, 1972), potential N mineralization rates in the control, composted poultry manure, and vermicompost treatments were calculated to be 0.04 mg N/L/d, while the MGW compost and fertilizer treatments had PMN rates of 0.05 and 0.06 mg N/L/d, respectively. No significant differences for PMN rates among the treatments were found (data not shown). In systems with slower N mineralization rates, we expect the total mineralized N to be made available to the crop later but potentially over a longer period than systems with faster PMN rates. Moreover, lower N mineralization rates could correspond to a reduced risk for N leaching.
As an N availability index, PMN, as measured here with a 28-d aerobic incubation, does not correlate with inorganic N measurements or yields (data shown below). Given the importance of estimating N availability for fertility recommendations for rooftop farm substrate, it would be in the interest of urban growers to identify techniques for predicting N availability and its relationships to total substrate N, yields, etc. (Schomberg et al., 2009).
Inorganic N in leachate.
Rooftop farmers face the challenge of applying N in sufficient amounts to maintain crop yields, while not contributing to nutrient loading of stormwater runoff, which would diminish the benefits associated with a vegetated roof. Concentrations of
Total nitrate concentration (mg
Citation: HortScience 50, 6; 10.21273/HORTSCI.50.6.806
Ammonium-N in the leachate was highest in the synthetic fertilizer treatment at all sampling weeks (Fig. 3). Concentrations of
Total ammonium concentration (mg
Citation: HortScience 50, 6; 10.21273/HORTSCI.50.6.806
The Environmental Protection Agency (Repository of Documents: New York State, 2000) has established enforceable standards for
Several factors (similar to those for landscaped green roofs) potentially influence runoff dynamics from rooftop farms: substrate composition, thickness, and drainage; maintenance and amendments used; type of vegetation; precipitation volume and patterns; local pollution sources and properties; and age of the growth substrate. The porous nature of the rooflite® intensive ag substrate might create a well-oxygenated environment that could encourage nitrification, which could potentially deplete
Yields and crop N.
Swiss chard biomass (fresh weight of aboveground biomass) was similar among the treatments at week 2 (Fig. 4). In weeks 4 and 6, the synthetic fertilizer treatment showed a significant increase in swiss chard biomass compared with the MGW compost treatment and the control. At harvest (week 8), the yield in the treatment receiving synthetic fertilizer (fresh weight = 44.4 g) was nearly 67% higher than the yield from the control (Fig. 4). Although biomass in the MGW compost treatment was lower than the synthetic fertilizer at weeks 4 and 6, yields between these two treatments were similar at harvest. Throughout the experiment, biomass and yields among the compost treatments were not different and did not differ from the control (Fig. 4).
Swiss chard (Beta vulgaris) aboveground biomass and yields of during an 8-week greenhouse experiment. Treatments include control, no added N input to rooftop farm substrate; poultry manure, composted poultry manure; vermicompost; municipal green waste (MGW) compost, composted municipal green waste; and synthetic fertilizer, Osmocote® Smart-Release 14N–4.2P–11.6K. Bars (means representing n = 5) within a sampling week bearing the same letter are not significantly different by Tukey’s honestly significant difference test at α = 0.05.
Citation: HortScience 50, 6; 10.21273/HORTSCI.50.6.806
In a greenhouse study, Echer et al. (2012) evaluated the impact of a range of urea-N fertilization doses on swiss chard and found linear increases in total and marketable yields (5.4 t·ha−1 increase for each 40 kg·ha−1 N applied). Engelbrecht et al. (2010) found the fresh biomass weight of swiss chard increased with increasing N doses and did not show a decrease even at an application of 800 kg N/ha. The N application rates in our study (126–189 kg N/ha) approached the highest N dose in the Echer et al. (2012) study (160 kg N/ha), but were less than 25% of the highest N dose used in Engelbrecht et al. (2010). Therefore, applying more N to the rooftop farm substrate might still lead to a linear increase in swiss chard yields. However, other factors, such as phosphorus concentration, planting density, salinity, etc., could become limiting to growth at higher N input levels.
At harvest, N content of the swiss chard yields among the four treatments were not different from the control (Table 4). Despite the generally higher inorganic N availability in the the synthetic fertilizer treatment, yield N (33 g N/kg biomass) in the synthetic fertilizer treatment was similar to the yield N from the organic N input treatments and was only higher than N content in swiss chard from the composted poultry manure treatment (21 g N/kg biomass). The higher C:N ratio in the swiss chard from the composted poultry manure (C:N = 15.2) vs. the synthetic fertilizer addition treatment (C:N = 10.9; P = 0.06; Table 4) was consistent with the differences observed in yield N.
Total nitrogen (N) content, carbon-to-nitrogen ratio (C:N), and partial factor of productivity of N (PFPN) of swiss chard (Beta vulgaris) yields at harvest. Mean (n = 5) separation in the same column, by Tukey’s honestly significant difference test at α = 0.05 (lower case) and α = 0.1 (upper case).
Fertility management
Substrate conditions: pH and EC.
Along with organic matter content, green roof substrate characteristics that are important to nutrient release include salinity, pH, particle size, porosity, and WHC (FLL, 2002). We observed pH values within the range of 6.97 to 7.77 for all treatments across the season (Table 3). Values of pH were generally highest at weeks 4 and 6 and did not change significantly within treatments across the season, suggesting that these systems are well buffered (statistical significance among weeks not shown in Table 3). With the exception of week 6, the pH of the substrate of the synthetic fertilizer treatment was significantly lower than that of the control and organic N input treatments, which were generally not different. Although the measured pH values for the treatments fall within the range of pH 5.5–8.0 that is recommended in the FLL (2002) guidelines for actively growing intensive green roofs, these values are higher than the preferred pH range of 6.75 for swiss chard grown in mineral soils (Jett, 2005). The lower pH of the substrate in the synthetic fertilizer treatment (Table 3) may have, along with other factors, contributed to higher yields in comparison with the control and organic N input treatments.
Substrate EC values across all treatments ranged from 2.46 to 7.66 dS·m−1 (Table 3), with EC values gradually decreasing from the start of the experiment toward harvest. This was to be expected as salts from the substrate and N inputs will gradually leach out with each irrigation event that exceeds substrate moisture holding capacity and plant water uptake. The substrate from the synthetic fertilizer and MGW compost treatments showed the highest EC values across the season (5.68 and 5.15 dS·m−1, respectively). According to Warncke and Krauskopf (1983), substrate EC values exceeding 5.00 dS·m−1 can result in plant wilting and leaf burn. Shannon and Grieve (1998) report that swiss chard growth diminishes in sand cultures with EC exceeding 11 dS·m−1 (≈3.67 dS·m−1 for saturated extract EC). While neither leaf burn nor wilting were observed in any of the treatments, rooftop farms using substrates similar to those in this study should monitor salinity to optimize crop productivity and substrate health.
Nutrient management for maximizing crop yields and reducing N loss.
Like most agricultural operations, rooftop farms must maximize their crop production with minimal inputs to be viable. In the case of N inputs, cropping systems with high nitrogen use efficiency (NUE) are key to optimizing the trade-offs among production, profit, and environmental impact. Cropping system NUE can be increased by achieving greater uptake efficiency of applied N inputs, by reducing the amount of N lost from organic and inorganic N pools in the growth medium, or both. Nitrogen applied via synthetic fertilizer (126 kg N/ha) and MGW compost (189 kg N/ha) were the lowest and highest, respectively, among the treatments. The partial productivity factor of N applied (PFPN), calculated as the ratio of the swiss chard yield (g dry weight per container) to the total amount of N applied to the treatment (g N per container), was highest for the synthetic fertilizer treatment (6.37 g yield/g N applied) and did not differ among the organic N input treatments (2.97–4.39 g yield/g N applied; Table 4). The substrate
The
Of the four N input treatments, the treatment receiving MGW compost-N best achieved the balance between higher yields and reduced N losses to potential roof runoff. It is not evident from our measurements, however, that inorganic N availability, total substrate N, and potential N mineralization concentrations contributed to the higher yields of the MGW compost treatment and the synthetic fertilizer treatments. Phosphorus and other macronutrients and micronutrients, which were not measured in this study, are also critical to crop growth (Fageria et al., 2010) and may have played a potential role in the differences observed in the yields. While the moisture holding capacities of the MGW compost, vermicompost, and composted poultry manure treatments (12%, 11%, and 15%, respectively, including base substrate) were similar, differences in porosity (not measured) may have also been influential in creating conditions that promoted crop growth and N loss.
Nitrate- and ammonium-N availability in the organic N input treatments and the control were lower at the end of the 8-week experiment than at the beginning of the season (Table 3; significance not shown). Williams and Nelson (1992) tested several organic N sources, including sewage sludge, poultry manure sludge, bonemeal, pine needles, and poultry feathers, and found that all sources ceased releasing sufficient N after 6–7 weeks. To replenish the inorganic N pools and sustain crops grown in soilless rooftop farm substrate, applying N inputs throughout the growing season (which can be longer than the 8-week period studied in this experiment) and/or supplementing organic amendments with synthetic fertilizers might improve N availability, thereby maximizing yields, providing N for subsequent crops, and promoting long-term productivity. Cover cropping with leguminous, N-fixing crops (e.g., vetch, clover, and/or beans) have been shown to replenish N pools and accumulate organic matter in traditional cropping systems (Dakora and Keya, 1997; Kuo et al., 1997; Stivers and Shennan, 1991), and should be considered for fertility management on rooftop farms.
Conclusions
Despite the lower total N applied, the higher inorganic N (
Literature Cited
Berghage, R.D., Beattie, D., Jarrett, A.R., Thuring, C., Razaei, F. & O'connor, T. 2009 Green roofs for stormwater runoff control. In: National Risk Management Research Laboratory (ed.). EPA/600/R-09/026. Office of Research and Development, U.S. Environmental Protection Agency, Cincinnati, OH. 81 p.
Boydston, R.A., Collins, H.P. & Vaughn, S.F. 2008 Response of weeds and ornamental plants to potting soil amended with dried distillers grains HortScience 43 191 195
Brenneisen, S. 2006 Space for urban wildlife: Designing green roofs as habitats in Switzerland Urban Habitats 4 27 36
Burger, M. & Jackson, L.E. 2003 Microbial immobilization of ammonium and nirate in relation to ammonification and nitrification rates in organic and conventional cropping systems Soil Biol. Biochem. 35 29 36
Campbell, C.A. & Souster, W. 1982 Loss of organic matter and potentially mineralizable nitrogen from Saskatchewan soils due to cropping Can. J. Soil Sci. 62 651 656
Carpenter, S.R., Caraco, N.F., Correll, D.L., Howarth, R.W., Sharpley, A.N. & Smith, V.H. 1998 Nonpoint pollution of surface waters with phosphorus and nitrogen Ecol. Appl. 8 559 568
Chapin, F.S. III, Vitousek, P.M. & Cleve, K.V. 1986 The nature of nutrient limitation in plant communities Amer. Nat. 127 48 58
Clark, M.J. & Zheng, Y. 2013 Plant nutrition requirements for an installed sedum-vegetated green roof module system: Effects of fertilizer rate and type on plant growth and leachate nutrient content HortScience 48 1173 1180
Curtin, D. & Campbell, C.A. 2006 Mineralizable nitrogen. In: M.R. Carter and E.G. Gregorich (eds.). Soil sampling and methods of analysis. 2nd ed. CRC Press, Boca Raton, FL
Czemiel Berndtsson, J. 2010 Green roof performance towards management of runoff water quantity and quality: A review Ecol. Eng. 36 351 360
Dakora, F.D. & Keya, S.O. 1997 Contribution of legume nitrogen fixation to sustainable agriculture in Sub-Saharan Africa Soil Biol. Biochem. 29 809 817
Doran, J.W. 1987 Microbial biomass and mineralizable nitrogen distributions in no-tillage and plowed soils Biol. Fert. Soils 5 68 75
Echer, M.D.M., Zoz, T., Rossol, C.D., Steiner, F., Castagnara, D.D. & Lana, M.D.C. 2012 Plant density and nitrogen fertilization in Swiss chard Horticultura Brasileira 30 703 707
Emilsson, T., Czemiel Berndtsson, J., Mattsson, J.E. & Rolf, K. 2007 Effect of using conventional and controlled release fertiliser on nutrient runoff from various vegetated roof systems Ecol. Eng. 29 260 271
Engelbrecht, G.M., Ceronio, G.M. & Motseki, P.C. 2010 Effect of nitrogen levels and sources on production of Swiss Chard (Beta vulgaris var. Cicla) SA J. Plant Soil 27 229 234
Fageria, N.K., Baligar, V.C. & Jones, C.A. 2010 Growth and mineral nutrition of field crops. CRC Press, Boca Raton, FL
FLL. Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau (Landscape, R., Development & Construction Society) 2002 Guideline for the Planning, Execution, and Upkeep of Green-roof Sites
Franzluebbers, A.J., Hons, F.M. & Zuberer, D.A. 1995 Soil organic carbon, microbial biomass, and mineralizable carbon and nitrogen in sorghum Soil Sci. Soc. Amer. J. 59 460 466
Gobel, P., Dierkes, C. & Coldewey, W.G. 2007 Storm water runoff concentration matrix for urban areas J. Contam. Hydrol. 91 26 42
Goh, K.M. & Haynes, R.J. 1986 Nitrogen and agronomic practice. In: R.J. Haynes (ed.). Mineral nitrogen in the plant-soil system. Elsevier Science, London, U.K
Groffman, P.M., Law, N.L., Belt, K.T., Band, L.E. & Fisher, G.T. 2004 Nitrogen fluxes and retention in urban watershed ecosystems Ecosystems 7 393 403
Hathaway, A.M., Hunt, W.F. & Jennings, G.D. 2008 A field study of green roof hydrologic and water quality performance Trans. ASABE 51 37 44
Jett, J.W. 2005 Plant pH preferences [Online]. West Virginia University extension service. 30 Jan. 2015. <http://www.wvu.edu/∼agexten/hortcult/homegard/pHpref.pdf>
Kuo, S., Sainju, U.M. & Jellum, E.J. 1997 Winter cover crop effects on soil organic carbon and carbohydrate in soil Soil Sci. Soc. Amer. J. 61 145 152
Masabni, J. 2011 Swiss Chard | Commercial and Specialty Crop Guides [Online]. Texas Agri Life Extension Service: Texas A&M. 29 Jan. 2015. <Aggie-horticulture.tamu.edu/vegetable/files/2011/10/swisschard.pdf>
Materials Management & Product Stewardship Workgroup–West Coast Climate and Materials Management Forum 2011 Reducing Greenhouse Gas Emissions through Recycling and Composting [Online]. Seattle, WA. 7 Oct. 2014. <http://www.epa.gov/region10/pdf/climate/wccmmf/Reducing_GHGs_through_Recycling_and_Composting.pdf>
Michael Beman, J., Arrigo, K.R. & Matson, P.A. 2005 Agricultural runoff fuels large phytoplankton blooms in vulnerable areas of the ocean Nature 434 211 214
Moran, A., Hunt, W.F. & Smith, J. 2005 Hydrologic and water quality performance from green roofs in Goldsboro and Raleigh, North Carolina. Proceedings of the Green Roofs for Healthy Cities Conference. Washington, DC
Nicolardot, B., Recous, S. & Mary, B. 2001 Simulation of C and N mineralisation during crop residue decomposition: A simple dynamic model based on the C:N ratio of the residues Plant Soil 228 83 103
Norberg-Hodge, H., Merrifield, T. & Gorelick, S. 2002 Bringing the food economy home: Local alternatives to global agribusiness. Zed Books, London, U.K
Norton, J.M. 1999 Nitrification. In: M.E. Sumner (ed.). Handbook of soil science. CRC Press, Boca Raton, FL
Oberndorfer, E., Lundholm, J., Bass, B., Coffman, R.R., Doshi, H., Dunnett, N., Gaffin, S., Köhler, M., Liu, K.K.Y. & Rowe, B. 2007 Green roofs as urban ecosystems: Ecological structures, functions, and services Bioscience 57 823 833
R 2013. R: A language and environment for statistical computing. In: R.C. Team (ed.). Foundation for Statistical Computing, Vienna, Austria
Repository of Documents. New York State, Tribal & Territorial Standards (US) 2000 Federal Regulations, 40 CFR 131.41, Section 1: Water Quality Regulation [Online]. In: Water Quality Standards Repository of Documents (ed.). U.S. Environmental Protection Agency. Washington, DC. 24 Sept. 2013. <http://water.epa.gov/scitech/swguidance/standards/wqslibrary/ny_index.cfm>
Roe, N. 2001 Compost effects on crop growth and yield in commercial vegetable cropping systems. In: P.J. Stoffella and B.A. Kahn (eds.). Compost utilization in horticultural cropping systems. CRC Press, Boca Raton, FL
Roofmeadow 2015 Horticultural Questions – Roofmeadow – Green roofs. For good. [Online]. 29 Jan. 2015. <http://www.roofmeadow.com/contact-us/ask-charlie/horticultural-questions/>
Schomberg, H.H., Wietholter, S., Griffin, T.S., Reeves, D.W., Cabrera, M.L., Fisher, D.S., Endale, D.M., Novak, J.M., Balkcom, K.S., Raper, R.L., Kitchen, N.R., Locke, M.A., Potter, K.N., Schwartz, R.C., Truman, C.C. & Tyler, D.D. 2009 Assessing indices for predicting potential nitrogen mineralization in soils under different management systems Soil Sci. Soc. Amer. J. 73 1575 1586
Shannon, M.C. & Grieve, C.M. 1998 Tolerance of vegetable crops to salinity Sci. Hort. 78 5 38
Sikora, L.J. & Szmidt, R.A.K. 2001 Nitrogen sources, mineralization rates, and nitrogen nutrition benefits to plants from composts. In: P.J. Stoffella and B.A. Kahn (eds.) Compost utilization in horticultural cropping systems. CRC Press, Boca Raton, FL
Smith, W.H. 1996 Utilizing composts in land management to recycle organics. In: M. De Bertoldi, P. Sequi, B. Lemmes, and T. Papi (eds.). The science of composting. Springer, Netherlands
Solecki, W.D., Rosenzweig, C., Parshall, L., Pope, G., Clark, M., Cox, J. & Wiencke, M. 2005 Mitigation of the heat island effect in urban New Jersey Environ. Hazards 6 39 49
Stanford, G. & Smith, S.J. 1972 Nitrogen mineralization potentials of soils Proc. Soil Sci. Soc. Amer. 36 465 472
Stivers, L.J. & Shennan, C. 1991 Meeting the nitrogen needs of processing tomatoes through winter cover cropping J. Prod. Agr. 4 330 335
Teemusk, A. & Mander, Ü. 2007 Rainwater runoff quantity and quality performance from a greenroof: The effects of short-term events Ecol. Eng. 30 271 277
The Scott Company 2009 Osmocote® Plus Multipurpose Plant [Online]. 22 Jan. 2015. <http://www.scotts.com/smg/products/osmocote/PDF/Osmocote%20-%20O%20_%20I%20Product%20Page.pdf>
VanWoert, N.D., Rowe, D.B., Andresen, J.A., Rugh, C.L., Fernandez, R.T. & Xiao, L. 2005 Green roof stormwater retention: Effects of roof surface, slope, and media depth J. Environ. Qual. 34 1036 1044
Warncke, D. 2011 Greenhouse root media recommended chemical soil test procedures. In: The Northeast Coordinating Committee for Soil Testing (NEC-1012) (eds.). Recommended soil testing procedures for the northeastern United States. Northeastern Regional Publication No. 493, 3rd ed. Agricultural Experiment Stations of Connecticut, Delaware, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont and West Virginia. Univ. Delaware Fact Sheets & Pubs., Newark, DE
Warncke, D. & Krauskopf, D.M. 1983 Greenhouse growth media: Testing and nutrition guidelines. Extension Bulletin E 1736. Michigan State University Ag Facts. East Lansing, MI
Whipker, B.E., Cavins, T.J. & Fonteno, W.C. 2001 1, 2, 3’s of PourThru [Online]. North Carolina State University Floriculture Research. 16 Oct. 2013. <http://www.ces.ncsu.edu/depts/hort/floriculture/Florex/PourThru Handout 123s.pdf>
Whittinghill, L.J. & Rowe, D.B. 2012 The role of green roof technology in urban agriculture Renew. Agr. Food Syst. 27 314 322
Williams, K.A. & Nelson, P.V. 1992 Low, controlled nutrient availability provided by organic waste materials for chrysanthemum J. Amer. Soc. Hort. Sci. 117 422 429