Spontaneous Vegetation on Green Roofs = Weeds? How Can We Balance Urban Biodiversity and Maintenance Costs?

Green roofs, as a "Nature-Based Solution," have become a popular green building element in the past ten to twenty years. However, many extensive green roofs worldwide (including Hong Kong), which have shallower and lighter growing media (generally 15 cm or less), are manicured with neat and tidy patch plantings of short woody shrubs and monotonous herbaceous covers limited to a few plant species such as grass (like Zoysia and Bermuda grass). This practice is perceived as prioritizing maintenance efficiency over ecological complexity, failing to effectively enhance biodiversity. In tropical and subtropical regions, if regular maintenance tasks such as pruning and weeding are not performed, the roofs can quickly be overtaken and replaced by spontaneous vegetation. Roofs will be invaded by spontaneous vegetation if green roof owners cannot afford frequent maintenance or if managers experience obstacles in arranging timely maintenance. Spontaneous vegetations are not limited to herbaceous ones, those undesirable tree and shrub species can create long-term structural damage to green-roof engineering layers and the roofs themselves, leading to water leakage in some cases.

Spontaneous vegetation is often defined as weeds and many people perceived that they must be removed during regular maintenance. However, many studies have demonstrated that plant species diversity on extensive green roofs enhances efficiency overall, such as water retention, thermal insulation, reduced nutrient runoff, pollutant removal, and enhanced ecological value. How can we achieve a balance between urban biodiversity and green maintenance costs? To answer this question, we need local data and experience to establish and optimize planting designs and horticultural maintenance guidelines.

I conducted a study on a local extensive green roof involving naturalized planting: Over the course of 15 months, the research team removed seven species of ornamental plants, and replanted 15 native ferns and herbaceous plant species (over 1,000 in quantity), transforming a monotonous ornamental plant roof into a renaturalized biotope.

In the following year, the study recorded and analyzed the development and changes of spontaneous and planted vegetation species. The research combined field data collection (baseline plant surveys in the first month: plant specimen collection and species identification; three seasonal plant surveys: photographing, recording vegetation coverage, and measuring plant height) to analyse plant growth dynamics on green roofs, including seasonal growth trends of planted and spontaneous species and their species rank relationships. It also conducted hierarchical cluster analysis to illustrate the groupings of spontaneous species with differential seasonal change patterns. The automatic irrigation system operated as usual during the monitoring year, and no horticultural maintenance was performed.

During the first month of baseline survey following roof transformation planting, a total of 54 spontaneous species and 14 planted species were found, where seven native species planted at the time of replanting survived, and seven species planted before replanting came back to life.

• During each quarterly survey (three in total), planted species ranged from 14 to 16, and spontaneous species remained between 51 and 54. In each quarter, 11 to 12 species died, and 11 to 12 new ones sprouted or revived, indicating a constant turnover of species.

The majority of spontaneous species were annuals and perennials that grew on bare soil after roof transformation, resulting in an increase in biodiversity on this green roof. Biodiversity peaked in the early observation period, but later, certain fast-growing species expanded their coverage, leading to a decline in biodiversity index.

• The quarterly survey indicated that vegetation coverage of different plant forms fluctuated slightly with seasons (within a range below 7%); perennial herbaceous plants had the largest coverage, followed by ferns and annual herbaceous plants.
• Dominant and aggressive plants (such as Nephrolepis auriculata) were planted on the roof prior to the roof transformation and possess extensive rhizomes. Even if the above-ground parts are removed, the fern will still resprout from the rhizomes, making it nearly impossible to remove all. Therefore, planting such rapid-growing and difficult-to-remove species requires thoughtful deliberation. Their performance outstripped other plant forms (additional expansion of 12% of roof area over a year); the abundance of annual and perennial herbaceous plants is on a decline, indicating a slight decrease in both species richness and species evenness, with a noticeable disparity between the two.

The height of planted species was influenced by seasonal changes, showing significant fluctuations over the year. They were particularly susceptible to summer heat, leading to wilting and shortening, while spontaneous species remained unaffected, maintaining stable heights throughout the year, demonstrating resilience to high temperatures and intense sunlight. This also implies that spontaneous plants have lower water resource demands, making them a more sustainable option.

 The coverage of spontaneous tree seedlings increased from 0% to about 2%, indicating that strict monitoring and complete removal of seedlings are necessary during the initial stage of denaturalized planting, until continuous and taller vegetation cover discourages tree seed germination.

 Most planted species maintained stable coverage, while only a few spontaneous species exhibited stable coverage, with most experiencing seasonal fluctuations, declines, or revivals throughout the year. Species with stable and unstable growth patterns can help meet maintenance efficiency and enhance biodiversity needs respectively, providing guiding principles and options for plant species selection for naturalized planting.

• If landscape designers and managers aim to enhance biodiversity on green roofs, they should select and retain spontaneous species that are less stable. These species can diversify larvae and nectar plants, increasing visual variability on seasonal change.
• If maintaining a stable aesthetic appearance of the green roof is a priority, it is advisable to use planted species with stable coverage performance, and retain spontaneous species that also demonstrate stable coverage (e.g., Gomphrena celosioides).

 The differences in growth patterns between spontaneous and planted species provide comprehensive information and decision-making support for plant selection and maintenance. The study proposes management recommendations based on analytical results to realize both maintenance efficiency and ecological benefits of extensive green roofs. These are summarized and listed as follows:

• If frontline gardeners cannot accurately identify plant species, it may lead to delayed decision-making, causing irreversible damage to green roof structural stability. Hence, training for frontline gardeners and managers is crucial.
• It is recommended that green roof owners and managers establish long-term management goals for green roofs and discuss the maximum allowable height for plants, enabling managers to arrange for frontline gardeners to prune and remove plants based on height requirements.

Due to limited resources, this study did not collect further data to assess medium to long-term (e.g., 2-5 years, 5-10 years) species growth dynamics. It is hoped that systematic medium to long-term research will further optimize planting design and management in the future.

If you would like to learn more about this study, you can read the journal article "Short-Term Growth Dynamics of Spontaneous and Planted Vegetation on Subtropical Extensive Green Roof as Renaturalized Biotope" through the following link: https://doi.org/10.3390/su16198314