What is the impact of Impatiens glandulifera on species diversity of invaded riparian vegetation?

Summary

Himalayan balsam Impatiens glandulifera, native to the western Himalayas, has become naturalized in northern and Central Europe where it is one of the most widespread invasive species. In the Czech Republic it was first recorded in 1893, is now found along most large rivers and is considered invasive. I.glandulifera is competitively superior to some native species through attracting pollinators by a massive production of nectar and its tall growth.

In the Czech Republic, effects of I.glandulifera on riparian plant communities was studied along invaded banks of six rivers by removing balsam in some areas and comparing it with 'uninvaded' communities. Uninvaded plots harboured up to 23% more species but other effects were not apparent. Once I.glandulifera is removed, communities recover without any consequences for species diversity. It is concluded that I. glandulifera exerts negligible effect on the plant characteristics of invaded riparian communities, and hence represents little threat to the plant diversity of invaded areas investigated during this study.

Study area: The study was carried out in riparian areas of the valleys of the Berounka, Sázava, Vltava and Kamenice rivers, central Czech Republic. The sampled river banks were nutrient-rich and regularly disturbed by flooding. To evaluate the effect of Himalayan balsam Impatiens glandulifera on plant community characteristics of invaded stands, comparative and experimental approaches were used.

Comparative study of invaded & uninvaded plots: In 2004–2005, 30 pairs of 4 × 4 m plots were selected in the six river valleys. Each pair consisted of invaded (at least 60% balsam cover) and nearby uninvaded, vegetation. Uninvaded plots were selected so as to represent as close as possible the same habitat conditions as the corresponding invaded plot. A little balsam (1–5% cover) was present in 23 of the 30 uninvaded plots as it was difficult to find vegetated, totally balsam-free areas. This low cover was considered not to affect the established vegetation or exert impact on species diversity. The presence of I.glandulifera served as evidence that habitat conditions in uninvaded plots were suitable for the invasion. Vegetation in the plots was sampled in July and August 2004 and the percentage cover of each species was estimated.

Removal experiment: At the end of April 2003, 10 plots (each 2 × 1 m) were established along the Berounka river valley and in the Křivoklátsko Protected Landscape Area. Each plot was divided into two 1 m² subplots. Seedlings of I.glandulifera were removed from one of the subplots, the other served as a control (or 'experimental invaded'). The plots were monitored from May to July to remove any I.glandulifera growing in the 'uninvaded' subplots. At the end of July, when the vegetation was fully developed and no further shifts in the dominance of individual species were observed, species composition was sampled and cover recorded using a Braun–Blanquet scale.

Effect of invasion on species richness: Uninvaded comparative plots had on average 9.2 ± 3.1 (mean ± SD, n = 10) species. This was one fifth more than invaded plots but the difference was not significant. In total, 80 and 85 plant species were recorded in invaded and uninvaded plots respectively (see Table 1, attached). In uninvaded experimental subplots there were on average 11.9 (± 3.0) species, while invaded plots held 1.3 (± 4.6) fewer species (Table 2). This difference was not statistically significant. The total number of species recorded in invaded and uninvaded plots was 34 and 35, respectively (Table 1).

Effect of Himalayan balsam on species composition: Invasion had no effect on species composition as demonstrated by non-significant results. Common hemp-nettle Galeopsis tetrahit, large-flowered hemp-nettle G.speciosa and touch-me-not balsam Impatiens noli-tangere showed the strongest association with invaded plots, while reed canarygrass Phalaris arundinacea and greater celandine Chelidonium majus were most frequent in uninvaded plots. In experimental plots, small balsam Impatiens parviflora, rough meadow-grass Poa trivialis and creeping bent Agrostis stolonifera tended to be more represented in invaded plots, greater dodder Cuscuta europaea, white campion Silene alba, and burdock Arctium lappa in those from which I.glandulifera was removed.

Discussion: One problem with experiments such as this is that there will always be some uncertainty about the character of invaded plots prior to invasion i.e. to what extent they are comparable with control uninvaded plots. In this study, the uninvaded plots were selected close to invaded stands in habitat conditions matching as closely as possible. An ideal situation would be an ongoing spread of I. glandulifera over a site, with part already heavily infested, while the other still free or almost free of it. Visits over several years indicated that some comparative study sites were very close to such an ideal situation. Thus it is considered that the data provides a sound basis for the assessment of the effects of himalayan balsam on established plant communities.
In the UK, Hulme and Bremner (2005) conducted an experiment at a scale similar to this study in terms of the number of plots and their size, and found a highly significant increase in species richness and diversity following the removal of I. glandulifera. That the same treatment did not result in a significant effect in the Czech Republic, can be attributed to the difference in cover of the invading species. While in the British study, the cover varied from 80% to 100%, it only reached on average 43% at the Czech sites.

Conclusions: The effect of I. glandulifera on species composition of invaded communities was minimal. Invasion did not alter species composition in terms of presence and absence of species, only proportional covers, especially those of dominant species, had changed slightly. It appears that I. glandulifera merely takes over the role of native tall nitrophilous dominants e.g. stinging nettle Urtica dioica, root chervil Chaerophyllum bulbosum, fat hen Chenopodium album and welted thistle Carduus crispus. Riparian communities in the study area are often supplied with excessive amount of nutrients. Such conditions support the strong dominance of such tall competitive nitrophilous species. It may be that, unlike in UK (Hulme & Bremner 2005) where the nutrient load may be generally lower, species less resistant to the invasion of I.glandulifera do not naturally occur in the type of communities considered in this study.

The results suggest that the invasion by I. glandulifera does not represent a major problem in the study areas for conservation of native flora. In the light of this, the worth of occasional eradication attempts to preserve biodiversity in affected riparian areas is open to question, especially if the effect is rather limited and short-term. In addition, such control efforts may give way to invasions of other more robust and harmful alien riparian invasives, such as giant hogweed Heracleum mantegazzianum.

Reference
Hulme P.E. & Bremner E.T. (2005) Assessing the impact of Impatiens glandulifera on riparian habitats: partitioning diversity components following species removal. Journal of Applied Ecology, 43, 43–50.


Note: If using or referring to this published study, please read and quote the original paper. Please do not quote as a www.conservationevidence.com case as this is for previously unpublished work only.