Mexican Hillside in London

“It doesn’t look very maintenance-free.” With a dubious, sidelong look.

It’s a non-question I get asked a lot. Along with “why did you build a green roof?”, that one often accompanied by a “where’s the straitjacket?” look.

I have a standard answer for the second question – and it answers the first one, too.

Three reasons to build an extensive green roof:

  1. The first is the relatively obvious complex of ecological reasons: restoring the carbon sink, increasing biodiversity and reducing flooding by slowing down storm run-off.
  2. The second is that it made it significantly easier to get planning consent to build my modern timber house in a Conservation Area that otherwise consists of large, brick-built Victorian villas. When the neighbours objected (and they did), I was able to stand up in the planning committee meeting and point out that they wouldn’t even see my house – all they would see were the plants on the roof – and did everyone seriously prefer to look at a cracked concrete slab scattered with used car parts?
  3. Then there’s the real reason, which is that I’m a gardener and it doubles the size of my garden. What distinguishes my roof from most of the other green roofs in Britain is that, although it is an “extensive” green roof with a modest depth of soil, it is intended as an ongoing horticultural experiment and as a gardener’s garden. The object is to find out what one can grow in 100mm depth of soil on a roof in London, to find the envelope and then push it.

There is a wide variety of plants that will grow on an extensive green roof in London’s climate. Not just Sedum.

First, some technical and design details

  • The house was designed with architect Jon Broome, following his general specification for sustainable buildings, and was constructed in 1999 and 2000.
  • I planted it out gradually over the course of 2000 and 2001, since when it has been in a continuing state of development and redesign, as the “envelope” gets bigger.
  • The green roof covers the whole house. The planted area is 15m x 5.5m and the roof is a curved mono-pitch, with a flat area at the top and an increasing slope toward the bottom, with its steepest point at 22º. It slopes down toward the Northeast.
  • The top (flat) end is shady, the middle is very sun-baked and the bottom end is moderately sunny.
  • The planting medium or substrate is 100mm depth of sieved loamy garden soil, which was specified to be pre-mixed with 30% sharp sand. Unfortunately, what arrived was only about 10% sharp sand and we didn’t realise this until it was already laid on the roof, at which point I wasn’t realistically going to scrape it all off and send it back. As a result, I’ve spent a great deal of time amending the soil locally when planting. This has mainly been by adding large quantities or grit and sharp sand when planting xerophytes (drought-loving plants) and succulents. Furthermore small quantities of organic material have been added when planting other plants.
  • The soil is laid over a welded PVC membrane 1.5mm thick, which so far has not leaked. In theory, the presence of the soil should extend the life of the roof membrane by protecting it from the weather.
  • Timber rails, 50mm x 25mm, were laid across the slope to prevent the soil from slumping down the slope in the early stages (latterly, the plant roots stabilise the soil) and a biodegradable coir mat designed for sand-dune erosion-control was tied on top.
  • There is a built-in vertical access ladder at the lower end of the slope – this is the only way up.
  • I installed an under-soil “leaky pipe” irrigation system fed from the mains, but it was used mainly to get plants established and for “death control” during prolonged droughts. As the roof has been replanted with a wider range of more drought-resistant species, the irrigation has been used less and less – this year only twice, for 15 minutes at a time.

Planting the green roof

Planting out is an elaborate business involving a rucksack (every one of the 500+ plants on the roof had to be carried up on my back) containing a round-ended trowel, a bag of grit, a bag of sharp sand, a kneeling pad and, until the coir mat finally rotted away, a pair of shears. Now that the roof is covered with plants, planting out resembles a one-man game of Twister – a knee goes here, dodging an Agave, and an elbow goes there, avoiding a Cistus, leaving one hand free to dig, also ensuring that the mouth is free, as frequent full-blooded cursing is an intrinsic part of the process, especially when planting Agaves or Cacti.

Having been encouraged by tales of rampant Sedum plants colonising acres of green roof in mere minutes, I installed a few patches of various species of Sedum and stood back. And waited. Winter 2000/2001 was the wettest in living memory. My Sedum plants survived, but by Spring 2001 it was evident that my plan to plant out 80m2 of green roof myself, using individual plants, constituted a major attack of hubris. I bit the bullet and bought a job-lot of 220 Sedum plugs from a well known green roof specialist ( Evenings during Summer 2001 were largely spent up on the roof, doggedly planting out eight to ten plants at a time, to the extent that my new neighbours introduced themselves to me by putting a note in my letterbox addressed to “The Man on the Roof”.

A presentable green roof

By Spring 2002 I had a presentable green roof at last. As well as installing hundreds of Sedum plants, I experimented with various herbaceous plants and small shrubs that had proven themselves to be drought-resistant in my previous garden and that I knew to grow, in habitat, in locations with a restricted root-run. These were not plants that, to my knowledge, had been tried on a green roof before and inevitably there were unsuccessful experiments, but a remarkable number of them not only survived, but thrived and flowered profusely, especially:

  • Armeria maritima (the British native Thrift, a cushion-forming Alpine plant that is a familiar sight on the cliffs of the Cornish coast)
  • Cistus x pulverulentus “Sunset” (a free-flowering small Mediterranean shrub)
  • Sisyrinchium idahoense “E K Balls” (a dwarf variety of a small herbaceous perennial related to Iris) Culinary herbs such as Thyme and Rosemary (Sage failed)
  • Persicaria affinis (another low-growing, spreading herbaceous perennial)
  • Lysimachia nummularia “Aurea” (a creeping gold-leaved plant that, in theory, should thrive only in moist places, but spreads vigorously on the roof)

On the question of weeds

Over the next two or three years, this version of the roof grew and flowered well, but at the same time, it started to manifest the down-side of having secured soil that was excessively fertile and contained too little added sand. The advantage of a free-draining, nutrient-impoverished substrate, as is typical of an extensive Sedum roof, is that perennial weeds and especially grasses will not survive. Not so my roof. By 2005, I had collected the Compendium of Perennial Weeds and Grasses. I had the lot, with the exception of Ground Elder: Couch Grass, Creeping Ranunculus, Dandelion, Creeping Potentilla, Dock and, most prolific of all, millions of wind-blown lawn grass seeds, intent upon turning my green roof into a lawn.

The gardener side of me resented this intensely, but at the same time I was watching a bio-diverse ecosystem establishing itself, including, as well as garden weeds, mosses, fungi and a healthy population of invertebrates.

Native or non-native?

Let me digress for a moment on the subject of biodiversity and the “native plants tendency”. My experience, growing a high proportion of exotic species in a succession of “terrestrial” gardens as well as a green roof, is that invertebrates and soil micro-organisms really do not care where the plants in a garden originally hailed from. They just want something to eat and a place to breed. They’ll breed in a Banana or amongst Cistus and Agave as happily as they will in a garden rigorously limited to native plants such as Armeria. My decisions on what to plant on my roof are restricted quite enough by two factors – what will thrive and what will look good together – without adding the third, spurious limiting factor of asking how long it has been since the plant’s ancestors invaded Britain.

During the couple of years I was battling the invasion of the weeds I was also in frequent correspondence with a number of gardeners growing exotic and marginally-hardy plants in Britain, especially xerophytes in the relatively dry climate of the Southeast, so I decided in 2005 to rise above the weed problem. Literally. By planting taller plants. That way I could keep my almost-organic regime and my bio-diverse roof, without having it turn into an un-aesthetic mess. And I could experiment with even more unlikely plants.

Succulents and Xerophytes

So I started to re-plan the whole look of the roof by planting a framework of architectural succulents and xerophytes, mainly Mexican species of Agave and Dasylirion, but also South American members of the Bromeliad (Pineapple) family, creeping South African plants such as Delosperma and Mesembryanthemum, Echinopsis Cacti and two species of Aloe. The key to the whole plan is that all of these plants are marginally frost-hardy, especially when they are given a well-drained environment. Amateur gardeners and professional horticulturalists have had long-term success in Southeast England planting some of the species in virtually pure ballast and crushed rock. I am relying on the shallowness of the soil on my roof to function in the same way, so these plants from arid habitats do not sit with wet, rotting roots for weeks on end during Winter. I am particularly indebted to the garden designer Paul Spracklin, who has pioneered the use of these plants in a British environment in his garden in Essex and who has supplied me with a number of plants and advice (

Agave parrasana, Aloe Aristata and Armeria maritimaAbout 60% of the roof has now been replanted in this style and, so far, there have been no losses in two relatively mild Winters (one wet, one not) with overnight temperatures down to -5ºC.

One aspect of this change of approach has been the discovery that xerophytes from all over the world often harmonise well visually. I don’t think there’s probably another garden anywhere that has Armeria and Sisyrinchium interplanted with Aloe, Bromeliads, Cistus and fifteen varieties of Agave. But it works.


Unresolved Problems

There remains an unresolved problem area – the flat part of the roof at the top of the slope not only tends to pond after heavy rain, making planting with succulents impossible, but is also in shade. The soil conditions veer repeatedly between saturated and bone-dry over the course of a season. At present, it is dominated by those plants that I have been able to persuade to grow there, invasive horticultural thugs such as Vinca major (Periwinkle) and Phalaris arundinacea picta (Gardener’s Garters). A plan is required for its future, but first I have to replant the rest.

There’s also the lurking question of what will happen when we finally have that once-every-30-years winter. The lowest temperature in the eight years I have lived here has been about -6ºC and it has rarely stayed below freezing for more than 24 hours at a time. Both of these factors, along with the degree of moisture in the soil, the drying effect of the wind and the length and heat of the previous growing season, affect the cold-hardiness of plants. When we finally get a week of continuous deep frost, then I will find out which of my xerophytes are long-term prospects for growing in London.

It isn’t a roof terrace – the planning authority saw to that – but I’m up there doing some form of gardening about once a day in Summer. As a garden, as an ecological contribution and as an ongoing experiment, it’s a satisfying experience.

More photos of the roof on this forum.

You can listen to David talking about his roof on the Guardian Gardening Podcast, along with’s Dusty Gedge.

David Matzdorf, 30 August 2008