Roof gardens are becoming an increasingly important issue in urban developments worldwide. This is due to the densification of cities and the complexity of intertwined infrastructure and different functions.
They are also a response to the increasing hardening and petrifaction of cities.

It is well known that greenery has a cooling effect and that roof gardens also buffer water. Roof gardens also contribute to increasing biodiversity. They also form stepping stones in cities for natural hotspots.

It is a well-known fact that biodiversity is the result of sometimes a very long process whereby living creatures become dependent on each other for food or shelter in one specific place. That is why in the low countries (Belgium and the Netherlands), governments very often insist on the use of native plants, as these usually have the highest biodiversity linkage.

However, the complexity of today’s urbanisation creates conditions in which native plants find it increasingly difficult to survive. A Belgian city, for example, is deliberately switching to trees of Mediterranean origin in order to have them planted on its territory, because they are more resistant to the ever-increasing heat.

The native species are less and less resistant to these conditions.

This is how we have been dealing with roof gardens for a number of years now. Roof gardens are actually steppes. They are very hot and dry in summer and very cold in winter. So the temperature curve is very extreme.

Plants that can cope with such conditions have to be very strong and survive this winter humidity and that temperature curve.

That is why we make roof gardens that contain both native and non-native plants.

This means that the palette of usable plants is very large, and we do select them on the basis of the following criteria:

First of all, drought resistance. It goes without saying that this is an important criterion. The Korys roof garden published  is only irrigated sporadically in the event of extreme drought. As the weather is becoming extreme, we consider it advisable to provide emergency irrigation for all roof gardens. This is also the case here.

Then there is the construction. We are now making more and more roof gardens based on the retention principle. An artificial water table with polypropylene crates is placed on the flat roof in which rainwater is buffered. This is fed via capillary cylinders to the substrate from which the plants can draw. This gives them a much greater chance of survival. The roof published here does not have this because of the heights of the substrate on this roof vary from usually only 15cm to 80cm at the edges. Also, because this roof is sloping; it was constructed above a parking garage.

The substrate is a high-quality mixture of sand, compost, bentonite and volcanic materials.

Then there is the planting and the planting system.
We are opting more and more for a naturalistic dynamic. The same principle we apply as well on rain gardens or bioswales. We mix plants on the basis of a specific recipe. We do this on the basis of aesthetic criteria of leaf contrasts, a certain degree of colour compatibility, an amount of evergreen species (in favour of the winter-image) and height structure.

These are criteria that are also used for perennial borders, for example. The big difference, however, is that we set a process in motion here. This is totally different with regard to a system where you want to see the planting unchanged over the years.

This is not the case here. The process is assessed to such an extent that even wild plants can spread spontaneously on the roof. This helps the naturalistic picture and also helps biodiversity. The management of ‘intelligent and selective weeding’ then helps in the selection process to remove what takes over or that which compromises diversity.

This requires gardeners who have a good knowledge of plants. We therefore call for gardening education to make a strong commitment to this knowledge, because we are convinced that the cities of the future will have more and more naturalistic plantations.

The main aim is to achieve a diversified picture with an upmost groundcovering aspect. The process being set in motion is Darwinian. It is a ‘Survival of the Fittest’. That is why we are also implying risk spreading. By using many species, there is a very good chance that the soil will be covered. The chosen plants sow themselves out or run by stolons. Every year the picture changes in this way, which makes the process exciting and represents a negotiation with nature. This is philosophically also an important attitude. You don’t force nature, but let nature guide its process partly by itself. So you only set something in motion. This is a major turnaround in the approach to classical parks, as we knew them in the 20th century.

Along with Henk Gerritsen, Piet Oudolf and Rob Leopold, we were already intensively exchanging ideas about such dynamic systems in the early 1990s. Meanwhile, over the last three decades, we have been applying these “patchwork systems”, as we call them, more and more in various plantations.

This is also the case in the forest area under mature trees in the same park where we designed this roof garden.

The plants themselves are chosen on the basis of their drought resistance and their appearance, as mentioned. The selection is made as wide as possible in order to maximise the groundcovering aspect. Of course, this is also reflected in the great diversity in terms of both shape and colour.

There are actually two types of plants used in this system. Groundcovering and self-sowing species. The first take care of the lateral covering structure, the letter shape the vertical image. This withered vertical image is also present in the winter and makes the image at that time of the year interesting in a very specific way.

The self-sowing species are also the ones that cause the unpredictable and dynamic aspect.
Species such as Euphorbia oblongata, Oenothera versicolor and Verbena bonariensis seek their own way and opportunity and make the image change every year.
Others, such as Stachys byzantina, Sedum takesimense and Silene maritima are groundcovering, creating a horizontal layer that complements all those vertical plants in their structure.

Of all the species that we have planted, there is actually not one species that has disappeared. All of them have been present since the initial planting, although sometimes in more numbers and sometimes in smaller numbers as well. This is, of course, due to the process that is going on.
In the meantime, Linaria vulgaris has spread between them, which is actually already the first native plant to take its place among the vegetation. This gives us an extra flowering beautiful endemic plant.

The most important flowering period is between May and July, but we have also ensured that there is a spread of flowering, which means that various plants continue to flower after the summer.

The Sysirinchium, the Stachys, the Sesleri and the Sedum takesimense are particularly visible in terms of their structure in winter and thus form the backbone of the image.
Sedum takesimense is a very important plant for our rooftop concepts. Hans Simon blessed us with this plant. A man with fantastic plant knowledge of whom we still have very good memories.

What makes the overall picture come true is that a sea of flowers extends behind the castle in which offices are located. And from spring onwards, there are many insects and butterflies visiting this sea of flowers. Birds also come to this piece of naturalistic vegetation.

Meanwhile, we are experimenting with many more species on various new roofs to test what grows well and what does not. Always with the aim of getting as much image quality and biodiversity as possible in one concept. It goes without saying that a good knowledge of plants is indispensable for this.

Project Data

Landscape architecture: Denis Dujardin

Location: Halle, Belgium

Project date: 2016

Photo credits: drone photos – David Carpentier, other images – Denis Dujardin