Green roofs use an engineered substrate, rather than natural soil, to support plant growth.  There are a number of commercial varieties available, but these are often shipped considerable distances from supplier to consumer and can be expensive.  To reduce the cost of our roofs, it would be best if we could produce our own green roof soil.  In this post I will examine information available from a variety of roof soil producers and independent studies to explore the feasibility of making our own green roof substrate.  (n.b. Substrate formula differs based on the type of roof being installed (i.e. extensive vs. intensive); since Eco Brooklyn focuses mainly on extensive roofs, I will concentrate on soil formulas for extensive roofs only.)

The planting substrate used in a green roof installation is crucial to the success and longevity of the roof itself.  If the soil mixture does not support plant growth while simultaneoulsy allowing the roof to function as a traditional roof would (i.e. shedding water), then the installation will fail.  Therefore, the standards to which green roof soil will be held are extremely high, and we need to do our best to get the formula right!

Steve Skinner, a green roof products manager with American Hydrotech and soil scientist, writes that the most important green roof substrate considerations include:

• grain size distribution
• density
• water & air management
• pH, lime & salt content
• organics
• nutrients
• cation exchange capacity (CEC)

Since the modern green roof movement began in Germany, the Germans have a head-start on establishing standards for successful green roofs.  The FLL is held as the authority on green roof planning, execution, and upkeep, and has established guidelines for individual components, including substrate; these guidelines can be downloaded here.

Before we dive into FLL guidelines, let’s check out some of the commercial soils and identify relevant attributes.

In our most recent green roof installation, we used Gaia Soil for Green Roofs, a locally-produced, lightweight (10 lbs./cu. ft. dry, 30 lbs./cu. ft. saturated) substrate made from expanded polystyrene (modified to improve water-holding and cation-exchange capacities), clay, and finshed compost.  We liked that it was made close to New York City, that it was light, that it retained water well (200% of dry weight!), and that it uses recycled content.  However, the soil was expensive and tricky to pick up and transport.

The makers of Live Roof, a modular green roof system, have developed their own brand of green roof substrate which aims to increase soil lifespan and reduce substrate shrinkage by incorporating a high percentage of inorganic material (94+%) and avoiding the use of peat moss, compost, and perlite, all of which can break down and erode after exposure to the elements.  In addition, the system uses clay to bind nutrients, a buffering agent to reduce the impact of acid rain, and a disease-supressing organic element.

G-Sky produces two blends of Roof Soil from pumice, well graded sand, and aged bark mulch compost.  The manufacturers claim that, given appropriate installation, these soils can be used with little or no structural support on roof slopes of up to 35 degrees.

Rooflite Extensive MC uses a blend of mineral lightweight aggregates like HydRocks, a calcined clay, and organic components to achieve FLL certification.  Detailed specifications for the product can be found here.

The most important commonality among these substrates is that they include a high percentage of low-density inorganic particles.  These elements, whether pectin-coated expanded polystyrene, pumice, or calcined clay, play an important role in air and water management: they allow the soil to drain quickly during heavy rains and, because they do not break down, maintain interstitial space over the long-term.  Each blend also incorporates an organic element to absorb water and provide a nutrient reservoir.  The organic component within roof soils, however, typically represents a lower proportion of the soil mass than would be found in a natural system and CEC (and nutrient availability) suffers as a result.  Some companies, therefore, use calcined clay or add clay to their soils to increase CEC and bind essential plant nutrients like potassium, calcium, magnesium, and ammonium.

Independent studies are a good source for information on substrate formula, as well.

A recent Michigan State University study looked at the ability of a variety of substrates to support Sedum and native plant growth on rooftops in Michigan.  In their study, Rowe et al. found that at a depth of 10 cm., a soil mixture of 80% heat-expanded slate (PermaTill, from Carolina Stalite), plus USGA grade sand, aged compost, and peat, amended by approximately 50 grams of controlled-release fertilizer per sq. m per year enabled a Sedum community to establish itself (in about a year) and thrive with minimal decomposition of organic material and nutrient-laden runoff.

Another study by Oberndorfer et al. suggests that organic components typically comprise about 10% (by weight) of green roof substrate and notes that crushed brick can be substituted for expanded (calcined) clay, depending on the load-bearing capacity of the roof.

In his 2007 dissertation for Ohio State University’s Horticulture and Crop Science program, Reid Coffman notes that that extensive green roof soil is typically engineered to be 75% inorganic and 25% organic and closely mimics the conditions found within the rock outcrop ecosystem.  In a 2006 article, Lundholm notes that the most popular plants used in green roofs are naturally found in these outcrop systems.  Though I will not do so in this post, study of productive rock outcrop ecosystem soil may provide additional insight into optimal green roof substrate composition.

Lastly, Skinner’s article provides very helpful information on green roof substrate composition.  He notes that extensive green roof soils are typically between 3% and 6% organic by mass, but that additional organic matter can be added initially to help to establish young plants.  The organic component, however, should have a low carbon to nitrogen ratio (i.e. be highly decomposed) so that plants can easily extract nutrients and to prevent the severe shrinkage and accumulation of toxic compounds that can accompany rapid decomposition.  Humus made from composted straw, sawdust, wood, leaves, grass clippings, agricultural waste, biosolids, and/or animal carcasses, separately or in combination, could be used as the organic component in green roof media and will initialy provide the macro- and micronutrients necessary for plant growth, though later it may be necessary to amend the soil with fertilizer.  The inorganic component, Skinner states, could be comprised of synthetic expanded slate, clay, or shale or of natural materials like pumice.  Sand, perlite, scoria, or pumice fines are often used as secondary inorganic components.  Skinner advises that soil pH be maintained between 6 and 7.5 and that the substrate contain carbonate levels sufficient for plants to readily absorb nutrients and water (though architect Allan Wingfield suggests that carbonate levels be kept in check since high levels (above 6 grams/liter) can lead to corrosion of roof drains and outlet pipes).

Based on the general formulas outlined here, it seems reasonable for us to develop our own Eco Brooklyn substrate blend for our extensive roof installations.  The next step would be to source materials, mix a trial batch and, if we choose to try to meet FLL guidelines, perform an empirical comparison to the FLL figures.  For example, FLL places emphasis on grain size, organic content, permeability, water storage, air content, pH, and salt and nutrient content, and all of these parameters must be within a specific range for the soil to be certified FLL compliant.  Most of these characteristics could be easily measured, and our formula tweaked to ensure compatibility with the FLL guide.

Without undertaking an effort to guarantee FLL certification, however, I believe that it is possible to mix a good quality green roof substrate.  To begin, I would advocate the use of calcined clay or expanded slate as our primary inorganic substance, as both have reasonably high CEC.  Pumice, repurposed polystyrene, and lava rock (with some clay mixed in) would work as well, and might lower the embodied energy of our substrate mix (although calcined clay is produced as a byproduct of cement production in Brazil).  Depending on the load-bearing capacity of the roof, crushed brick could be added to the mix, too.  Cost will be a major factor in determining our ultimate formula.  To the clay/slate/etc., I would recommend adding well graded sand and finished compost or humus.  I would aim for 10% organic material initially, with the expectation that the percentage will fall to closer to 5%.  I would plan to amend the substrate with fertilizer annually, and water the roof if it has not received precipitation for two weeks.

In an ideal world, we would work to get our soil to FLL spec, and periodically test the product to ensure that it is appropriately mixed.  For now, however, the time that it would take to test and reformulate the soil is not a luxury that we have.  Still, based on my analysis of the green roof studies and substrate formulas available, I feel reasonably confident that my recommendations will produce a soil that will green a roof and support plant growth for years to come.