Alien plant invasions are among the major threats to natural and semi-natural ecosystems in South Africa on approximately 18 million hectares of land. Much of the available data are not suitable for planning of local scale management because it is presented at a quarter degree grid square scale, which makes accurate location and estimates of invaded areas difficult.
The aim was to identify the dominant alien plant species and quantify their areal extent along a 479 km railway corridor in the Mpumalanga province.
The extent of the invaded area was obtained by manual digitising of alien plant distribution and density from Satellite Pour l’Observation de la Terre 5 imagery and by further applying an Iterative Self-Organising Data Analysis technique of the unsupervised classification method. Species’ occurrences were located and identified in the field using a Global Positioning System.
The most dominant invaders in terms of the number of individual polygons and the infested area were
Knowledge of the species’ occurrence and their extent will assist landowners and relevant authorities to control the spread of alien plants, which impact rail safety, agricultural production, water availability and biodiversity.
Invasive alien plants are a major threat to biodiversity because of their effects on the population dynamics of native species, effects on community dynamics (e.g. species richness, diversity and trophic structure) and disruption of ecosystem processes and functioning (Van Wilgen et al.
It was previously estimated that approximately 10 000 000 ha (8%) of land in South Africa was invaded by various taxa of alien species with the majority in the Western Cape followed by the Mpumalanga, KwaZulu-Natal and Limpopo provinces (Le Maitre et al.
An inventory of alien invasive species, an understanding of invasion processes and the management history of a region would constitute the baseline data necessary for effective management (Masubelele, Foxcroft & Milton
In Mpumalanga province, our initial field observations showed that woody alien plants along railway sides are increasing and are spreading to and/or from the neighbouring land because of the lack of appropriate control. This is a major concern to landowners who are bound by conservation laws and regulations to manage these populations (Department of Agriculture
In this context, we aimed to identify the dominant alien plant species and quantify their distribution and density along 479 km of railway corridors through endangered ecosystems and other adjacent areas in the Mpumalanga province, South Africa. To achieve this aim, the study had the following objectives: (1) to map all visible woody plant species from satellite imagery; (2) to survey and identify different alien plant species within each mapping unit including other visible herbaceous, shrub, grass and succulent alien species; and (3) to compare the results with other databases available for the study area. The data will be the first step towards developing proper control measures for implementation by landowners in order to manage and reduce the spread and impact of alien vegetation, which is a problem for railway managers and also for ecosystem functioning.
In this study, Satellite Pour l’Observation de la Terre (SPOT) 5 satellite imagery acquired in 2014 and 2015 was used. This imagery has four spectral bands (green, red, near infrared and shortwave infrared) and was accessed from the Mpumalanga provincial Department of Agriculture, Rural Development, Land and Environmental Affairs. It was preferred over Landsat (
The study area is within the grassland biome of the Mpumalanga province, South Africa (
Location of the selected railway line in the study area and vegetation type boundaries from a database edited by Mucina and Rutherford (
The Vryheid Formation is characterised by grey micaceous shale, course-grained sandstone and subordinate grit and coal beds found at the basin margin (Johnson
The first step of the mapping phase was to separate the woody vegetation from herbaceous and grass vegetation. This was based on visual interpretation of the SPOT 5 satellite imagery and manual digitising of the background woody vegetation boundary in ArcMap 10.3.1 (ArcGIS ESRI
An example of the colour contrast between woody vegetation (red outline) and the surrounding non-woody vegetation using the three selected colour bands and their combination (true colour) of the Satellite Pour l’Observation de la Terre 5 image in the area between Maviristad and Ermelo.
To solve the problems of island polygons and subjectivity in the output map associated with manual digitising, the unsupervised classification technique was implemented in ArcMap using the digitised polygons as a mask. Unsupervised classification uses the Iterative Self-Organising Data Analysis technique clustering algorithm that aggregates unknown pixels in an image, based on their natural groupings (Lillesand, Kiefer & Chipman
A rapid reconnaissance survey along the service road running parallel to the railway lines in our study area was carried out from November 2015 to February 2016 to identify the species occurring in each polygon of the unsupervised thematic maps. Only the species visible from a moving vehicle were recorded within a distance of up to 100 m on either side of the road travelled. A Garmin nüvi Global Positioning System (GPS) was used to record the co-ordinates for each woody vegetation stand or single tree. The co-ordinates were later loaded into the ArcGIS 10.3.1 (ArcGIS ESRI
Interpretation of the satellite imagery was biased towards woody plant species while grasses, succulents, shrubs and herbaceous plants were difficult to detect. However, field observations revealed some alien grass, succulent, shrub and herbaceous species in the study area. An attempt was made to record their location by estimating the extent of the invaded area in the field and by maximising the use of colour and texture contrast during manual digitising to delineate the invaded area. Field data also revealed the occurrence of indigenous woody species, mainly
Spatial distribution of alien vegetation along the railway line in Mpumalanga province.
The species observed belong to 12 different families (
Observed alien plant species organised by family, including their growth form and localities in which they were found along a 479 km railway line and on adjacent farms.
Family | Species | Common name | Growth form | Locality |
---|---|---|---|---|
Agavaceae | Sisal | Succulent | TD, OB | |
Cactaceae | Prickly pear | Succulent | OG, GW, TD, BL, EM | |
Fabaceae | Mauritius thorn | Tree | OB, GW, DG, TD, EM | |
Wattle | Tree | OB, OG, GW, DG, TD, ME, BL, EM | ||
Meliaceae | Seringa | Tree | OG | |
Myrtaceae | Gum tree | Tree | OB, OG, GW, DG, TD, ME, BL, EM, ED | |
Papaveraceae | Mexican poppy | Herb | OB | |
Pinaceae | Pine | Tree | OB, OG, GW, TD, ED, ME, BL, EM | |
Poaceae | Water reeds | Grass | OB, OG, TD, EM | |
Rosaceae | Peach | Tree | OG | |
Wild berries | Shrub | OG, GW | ||
Salicaceae | Poplar tree | Tree | OB, OG, GW, DG, TD, BL, EM | |
Weeping willow | Tree | OG, TD, ED, ME, BL, EM | ||
Solanaceae | Bugweed | Shrub | OB, OG | |
Verbenaceae | Desel | Herb | GW |
EM, Ermelo to Machadodorp; BL, Buhrmanskop to Lothair; TD, Trichardt to Davel; ED, Ermelo to Davel; ME, Maviristad to Ermelo; DG, Davel to Gelukplaas; GW, Gelukplaas to Wonderfontein; OG, Ogies to Gelukplaas; OB, Ogies to Blackhill.
It was difficult to quantify precisely the area invaded by each individual species from the satellite images because some polygons have more than one woody species and SPOT 5 spectral bands cannot be used to separate individual species. However,
Species observed and the extent of the area invaded calculated using the ArcMap calculate areas tool after converting the spectral classes derived from unsupervised classification of the satellite imagery to polygons.
Species (scientific name) | Occurrence (no. of individual polygons) | Area (hectares) |
---|---|---|
125 458 | 813.62 | |
46 081 | 456.73 | |
9666 | 46.46 | |
2534 | 10.29 | |
494 | 9.00 | |
19 | 7.78 | |
443 | 3.10 | |
8 | 2.80 | |
1 | 2.28 | |
15 | 1.79 | |
23 | 1.66 | |
90 | 1.43 | |
19 | 0.90 | |
332 | 0.60 | |
1 | 0.57 | |
2 | 0.42 | |
1 | 0.28 | |
8 | 0.25 | |
44 | 0.22 | |
5 | 0.21 | |
1 | 0.07 | |
1 | 0.07 | |
1 | 0.03 | |
15 | 0.02 | |
2 | 0.02 | |
2 | 0.01 |
The grouped species indicate polygons where more than one species occur and where species were not separated because of the lack of spatial and spectral resolutions of Satellite Pour l’Observation de la Terre 5 imagery utilised.
Other species are not prominent and only occur in isolated stands of trees, or they are succulents, shrubs and herbaceous plants.
The most dominant alien plant species in terms of the number of times they were observed are the
The spatial distribution map is limited to large trees because of the limitation of spectral bands and spatial resolution of SPOT 5 satellite imagery to detect smaller trees. Field observations revealed the occurrence of young trees in other parts of the study area. These young trees, mainly
An example of young or resprouting
Some of the land parcels next to the railway line have been planted with
Our study was not intended to provide a complete list of alien plant species occurring in the area and we did not record all species identified in other databases. This is mainly because the current inventory was only obtained along the railway line and adjacent farms, whereas other studies have included observations in environments such as watercourses, grassland, savannah and human-modified habitats (e.g. Henderson
Comparison of the current species observation and alien plant species records in Southern African Plant Invaders Atlas and National Herbarium’s Pretoria Computerised Information System databases.
Grid number | Species observed in this study | Species recorded in the SAPIA database ( |
Alien plant species recorded in the PRECIS database ( |
---|---|---|---|
2529CC | |||
2529DC | |||
2529DD | |||
2530CA | |||
2530CB | |||
2530CC | |||
2629AA | |||
2629AB | |||
2629AC | |||
2629AD | |||
2629BA | |||
2629BC | |||
2629BD | |||
2629CA | |||
2629CB | |||
2629DB | |||
2630AA | |||
2630AC | |||
2630AD | |||
2630CA |
In bold are category 1 (1a and 1b) species that must be eradicated or controlled in terms of the
SAPIA, Southern African Plant Invaders Atlas; PRECIS, National Herbarium’s Pretoria Computerised Information System.
The current study and SAPIA have more similar species recorded per quarter grid than the PRECIS database. The most prominent species in the current study list and SAPIA database include the
The inventory of alien plants in this study is notably different to the list from the PRECIS database. This is probably because PRECIS has limited data for alien species, and it has a bias towards herbaceous and shrub species, whereas this study primarily focused on woody plants. Similar species were only observed in grid 2529DD (
By combining the maps from remote sensing with field observations, it helped to improve our knowledge of alien plant species’ distribution and the extent of the invaded area along the railway line in the Mpumalanga province. This approach can be used in other regions where woody plants dominate. This study found that just like many other disturbed environments which include roadsides, plantations, heavily grazed and cultivated lands, railway lines are invaded by a variety of alien plant species in the Mpumalanga province. The biggest threat is posed by
This study should be regarded as a first step towards understanding alien plant invasion along the railway line in Mpumalanga province. While the current study only focused on the mapping and identification of species, future studies should look into the role of railway lines in dispersal of alien plants. It is well understood that commercial forestry and other means of introduction are the primary sources of alien plants in South Africa, but railway lines may have played a role as driver pathway for spread of alien plants from one locality to the other because not all invaded areas have a history of commercial forestry (Nyoka
The authors would like to thank the Department of Agriculture, Rural Development, Land and Environmental Affairs for the provision of software, data and imagery used in this study. This article benefited from comments by two anonymous reviewers and those of the editor, Dr Michelle Hamer.
The authors would also like to thank DEA – Working for Water Programme for their willingness to facilitate and fund the clearing of alien plants in the study area.
The authors declare that they have no financial or personal relationship(s) that may have inappropriately influenced them in writing this article.
N.M. was the project leader and was responsible for drafting the manuscript as well as geospatial interpretation. M.N.M., M.C.R., N.N.M. and S.R.N. were responsible for both field data collection and revising the scope and content of the draft manuscript.