About this Journal Submit a Manuscript Table of Contents
International Journal of Ecology
Volume 2012 (2012), Article ID 174813, 6 pages
http://dx.doi.org/10.1155/2012/174813
Research Article

Changes in Herbaceous Species Composition in the Absence of Disturbance in a Cenchrus biflorus Roxb. Invaded Area in Central Kalahari Game Reserve, Botswana

1Department of Basic Sciences, Botswana College of Agriculture, Private Bag 0027, Gaborone, Botswana
2Department of Wildlife and National Parks, Ministry of Environment, Wild life and Tourism, P.O. Box 4, Tsabong, Botswana

Received 8 December 2010; Revised 20 March 2011; Accepted 12 April 2011

Academic Editor: Herman Shugart

Copyright © 2012 Shimane W. Makhabu and Balisana Marotsi. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

A-nine year study was carried out to investigate changes in herbaceous species composition in an area invaded by Cenchrus biflorus Roxb, an exotic invader grass species. The study ensued termination of livestock and human activities in the area when residents of the area were relocated to another area. Vegetation characteristics from the disturbed sites (previous occupied areas) and undisturbed sites (previously unoccupied areas) were determined. The results show that C. biflorus has high tolerance to disturbance. It comprised the larger proportion of grasses in disturbed sites at the inception of the study. However, it decreased in abundance with time in disturbed areas and was absent in the undisturbed areas, suggesting that its ability to invade undisturbed sites is limited. Perennial species successfully reestablished on the third year after termination of disturbance. The study reveals that C. biflorus invasion in the Kalahari ecosystem can be controlled by termination of disturbances.

1. Introduction

The slow and uncertain rate of return of herbaceous species composition to the original status after severe disturbance is a major problem in wildlife management areas and on rangelands [1]. It is a function of the range condition at any given time, which is assessed as a function of grass composition, biomass, and cover [2, 3]. The dynamics of range condition is affected by environmental factors and land use such as animal grazing pressure. Heavy grazing pushes the range condition to degraded status dominated by weedy or unpalatable grass and forb species typical of disturbed areas [3, 4]. This has been supported by various studies done in semiarid Southern Africa [57]. The studies showed that basal cover, proportion of annuals and proportion of unpalatable species were higher in heavily grazed areas than in light grazed areas [57]. Vegetation outcrops on a disturbed area and surroundings are a function of availability of seeds or propagules of species or species groups.

Debates have ensued on the relative importance of both the biotic and abiotic factors on range conditions. Explanations for these are based on the equilibrium and nonequilibrium models [3, 811]. The equilibrium model stresses the importance of biotic feedbacks between herbivores and their resource, while the nonequilibrium model sees stochastic abiotic factors as the primary drivers of vegetation and livestock dynamics [3]. However, in most arid and semiarid rangelands both equilibrium and nonequilibrium models are at play, though at different scales.

It takes a considerable time for a species composition in a disturbed area to return to its previous condition. Samuel and Hart [1] reported that after 61 years of monitoring a disturbed area, its species composition did not reach the point of that of an undisturbed area it was compared with since there were visible differences between the two areas. The differences were attributed to various factors that included increase of invasive and alien species that often dominate in disturbed sites out competing native vegetation or disrupting native plant communities and nutrient cycling. On cessation of disturbance, expectations are that vegetation development follows a definite sequence of functional groups where early stages of recovery are dominated by species of light seeds, producing seeds in great numbers, fast growing and dispersing over vast areas [12]. These are hardened, annual plants that grow in very unfavorable conditions and improve the growth conditions resulting in plant communities that are adapted to the new, improved growth conditions, and replacing the existing plant communities. Species with heavy seeds, slower growth, and long life span dominate in later stages [13]. Studies by O’Connor and Roux [14] on the Kalahari landscape have shown that vegetation is able to reestablish after grazing pressure is lessened. Seitshiro [15] and Perkins [16] also observed that areas in the Kalahari sandveld vegetation communities, denuded of grass during dry years, had some rapid return of perennial grass cover in subsequent years of average or better rainfall or cessation of grazing. Indications are that, during prime growing conditions, it can take 8–10 years for disturbed vegetation to reestablish in the Kalahari landscape [17]. The Kalahari is a substantial part of Southern Africa interior covering countries among them Botswana, Namibia, and South Africa [18]. It is extensively elevated, flat and covered by Kalahari sand soils. Kalahari sand soils typically consist of over 95% fine sand-sized, aeolian-deposited sediment and are predominantly deep, structureless, and lacking in N, P, and organic matter [18].

Within the Kalahari ecosystem in Botswana lies Central Kalahari Game Reserve, the largest game reserve in the world that covers an area of some 52 800 km2. Some pockets of the Central Kalahari Game Reserve (CKGR) are inhabited by people. When people inhabited the area in the 1960s, their main source of food was hunting and gathering. However, with the decrease of natural resources, people started keeping domestic animals such as horses, donkey, and goats. Over time, it was realized that the area had been invaded by a noxious weed, Cenchrus biflorus Roxb. It is common in the savannas of sub-Sahara Africa. Cenchrus biflorus Roxb. is an annual grass of the Poaceae family with synonyms: C. annularis Andersson, C. barbatus Schumach, C. catharticus Delile, C. leptacanthus A.Camus, and C. perinvolucratus Stapf & C.E.Hubb. Its culms are 4–90 cm high. The leaf: lamina is 2–25 cm long, 2–6 mm wide and the panicle 2–15 cm long. Its spikelets are 1–3 per bur and 3.5–6 mm long. Molecular analyses suggest that the species invading the Kalahari originates from India hence the common name “Indian sandbur”. The grass was first observed in Botswana in the 1940 s growing around Nata. It is currently spreading rapidly in the western and central part of the country. The grass species is regarded as a noxious weed because of its objectionable burs, which adhere to animal skin, clothes, shoes and machinery. These burs are the likely vehicle to its rapid spread. The burs are harmful to grazers and may cause ulcers in the mouths of animals. The grass has also spread to arable fields making weeding difficult. As a result, some farmers in Botswana have opted to abandon their C. biflorus invaded arable fields. Intensive land use increases the potential to C. biflorus invasion as this increases the area of disturbed soil, thus creating conditions under which the invasive grass thrives. Control of this grass species is achievable in arable fields through weeding, cultivation, and herbicides application before seed formation. However, controlling this grass species in range and pasture areas still remains a challenge since there is lack of a selective herbicide that will kill only this grass species. As such other control measures have to be explored. Environmental friendly control measures will be more appropriate to be used in Game Reserves were minimal interventions are allowed.

In 1997, Central Kalahari Game Reserve (CKGR) residents at Xade were relocated outside the reserve as their activities were no longer compatible with wildlife conservation. This abandoned Xade area has been under disturbance for four decades. The nature of the disturbance ranged from simple cumulative trampling on the herbaceous layer to large-scale clearing of trees and shrubs. Establishment of modern and traditional homesteads and institutions like the “Kgotla” (traditional gathering place headed by a Chief) contributed to land disturbance. The keeping of livestock (horses, donkeys, sheep, and goats) exerted more pressure on the soil and vegetation leading to the disappearance of nutritious grass species and the increase of annuals [19]. Concerns were then raised that this would negatively affect wildlife populations and eventually deteriorate the biodiversity of the area.

The relocation of Xade residents prompted this study as an opportunity to determine how herbaceous species composition and cover changes overtime in the absence of livestock and human disturbance. Wildlife activities that had previously been observed to be absent or minimal were left to go on as usual. Thus, this study aimed at determining whether a disturbed area invaded by an exotic invader grass species can recover if human and livestock activities were terminated.

2. Methods and Materials

2.1. Study Site

Xade is in Central Kalahari Game Reserve (CKGR) and lies at Latitude 23.01 degrees south, Longitude 22.33 degrees east. The vegetation is characterized by medium to open bush savannah with occasional isolated tress. The vegetation has mainly been kept in shape by frequent bush fires. The climatic conditions of the area can be summarized as semiarid. Generally, the CKGR lies between 350 mm and 400 mm isohyets [19]. The Xade rainy season is between November and April with long-term annual rainfall of 350 mm. There is however considerable variation in the amount and pattern of rainfall between years 1998 and 2006 (Figure 1). The area is covered mainly by deep Kalahari sands. Its topography is flat to undulating with low immobile sand dunes.

174813.fig.001
Figure 1: Annual rainfall amounts in Xade during the study period.
2.2. Experimental Design

The vegetation data was collected from seven (10) strategically located plots at Xade. Strategically here means that it was made sure that five of the plots were in disturbed sites while two were in undisturbed sites (reference sites). The disturbance arose from livestock and human activities in the area before relocation in 1997. The plots measured 30 m × 20 m in size. Sampling was done from April 1998 to April 2006. Sampling was carried out in April of each year, the peak plant growing period. During April, herbaceous species are mature and easy to identify since most are either flowering or in seed. Within each of the 30 m × 20 m plots, five transects each measuring 30 m long and being 5 m apart were sampled. The herbaceous species frequency was recorded every two (2) meters along each transect using the wheel point method [20]. Numbers of individuals of each species by life form (forb, native grass and exotic) were recorded. The data was used to calculate species richness, composition, and herbaceous cover. Grass species nomenclature is according to Field [21] and Van Wyk and Van Oudtshoor [22].

Rainfall data was obtained from Botswana Department of Meteorological Service, Ghanzi offices.

2.3. Statistical Analysis

STATISTICA program, version Kernel 5.5 A, was used to calculate species richness, composition, and herbaceous cover. Species richness here refers to the number of individual plants of each species in each life form group (forbs, native, and invasive). Composition refers to relative abundance of each group. Herbaceous cover means the cover contributed by each group excluding that covered by litter and bareground. Excel was implored to draw charts showing species richness, composition, and cover. Spearman’s rank correlation coefficient (rs) was used to calculate the correlation between herbaceous species richness and site age (time). Spearman’s rank correlation was also applied to determine correlation between the native grass species, invasive species, and native forbs over time. The Spearman’s rank correlation coefficient values were tested for significance using the t statistic. The student t-test was also used to test if there was some significant difference in species richness and cover and functionality composition between sites (disturbed and undisturbed). STATISTICA [23] was used to perform these calculations.

The Czekanowski coefficient [24] was used to assess the similarity or dissimilarity in herbaceous species composition between disturbed and undisturbed plots (reference sites). The coefficient values range from zero (0), complete dissimilarity, to one (1), total similarity.

3. Results

3.1. Herbaceous Species Richness

The herbaceous species richness varied during the study period (Figures 2 and 3). Mean species richness was higher in disturbed sites than in undisturbed sites. Initially an average of fourteen (14) herbaceous species was recorded in the disturbed sites and six (6) in undisturbed sites. These included forbs, native grass species, and the exotic grass species Cenchrus biflorus. The native grass species included Stipagrostis uniplumis, Pogonarthria squarrosa, Schmidtia pappophoroides, Aristida meridionalis, Eragrostis lehmanniana, Aristida congesta, Melinis repens, Urochloa trichopus, Enneapogon cenchroides, and Schmidtia kalihariensis. The forbs included Tribulus terrestris, Amaranthus thunbergii, Heliotropium steudneri, Xenostegia tridentata, Harpagophytum procumbens, Hermbstaedtia odorata, and Cucumis myriocarpus.

174813.fig.002
Figure 2: Mean species richness at disturbed sites (Xade) from 1998 to 2006. Native species refers to all indigenous species while invasive refers to Cenchrus biflorus.
174813.fig.003
Figure 3: Mean species richness at undisturbed sites (Xade) from 1998 to 2006.

Cenchrus biflorus disappeared after four years of termination of disturbance but some reappeared during the sixth and ninth years after termination of disturbance (Figures 2 and 3). Species continued to recruit to reach a peak mean of nineteen (19) species and eighteen (18) for disturbed and undisturbed sites in 2000 and 2006, respectively (Figures 2 and 3). The lowest herbaceous species mean number of five (5) was recorded in 2005 for disturbed sites and of four (4) in 2003 for undisturbed sites. Species richness ranged from 5 to 18 (mean ± SE of 13 ± 1.26) plant species for disturbed sites and 4–18 (mean ± SE of 10 ± 1.26) for undisturbed sites. There was no significant difference in species richness between undisturbed and disturbed sites (t value = 1.71 P-value  =  .126).

3.2. Herb Layer Composition

The herbaceous layer composition differed substantially between sampling periods in both disturbed plots and undisturbed plots (Figure 4). The herbaceous layer in disturbed plots was composed of exotic grass species (C. biflorus), native grass species, and forbs and in undisturbed plots only native grass species and forbs were recorded. No exotic grass species were recorded in undisturbed sites. The Spearman’s rank correlation coefficient indicated that there were some negative relationships between percentage of forbs and time postdisturbance for disturbed sites ( , ). A similar pattern was also observed for C. biflorus ( , ). In contrast, Spearman’s correlation coefficient showed a strong positive correlation between native grass species and time postdisturbance for disturbed sites ( , ).

174813.fig.004
Figure 4: Herbaceous species composition at the disturbed (distbd) and undisturbed (undistbd) sites in CKGR: 1998 to 2006.

During the first two years of sampling, forbs dominated in disturbed plots but gradually declined with time. In the first year, forbs contributed 80.56%, C. biflorus 9.88%, while native grass species contributed 9.56%. Cenchrus biflorus gradually recruited reaching an average score of 21.94% in 2000 coinciding with the year of high rainfall (Figure 1) but disappeared in the fourth year with its traces being noted on the 6th and 9th years postdisturbance. Similarly, forbs decreased gradually reaching a minimum of 5.10% on the 8th year of study but went up to 27.23% on the subsequent 9th year.

On the contrary, native grass species recruited gradually in disturbed plots dominating in the third year of sampling. The native species reached a peak average of 89% on the 7th year of sampling. In disturbed plots, both forbs and native grass species fluctuated between years but native grass species were much higher throughout the study period except in 2000, 2004, and 2006 (Figure 4). In comparison, forbs dominated in disturbed plots while native grass species dominated in undisturbed plots during the first two years of sampling. Native grass species recruited gradually in the disturbed plots, exceeding that of undisturbed plots in 2004.

3.3. Perenniality

Annual grass species were relatively high during the first three (3) years in disturbed sites but gradually declined (Figure 5). In contrast, from 2001 until 2003 perennial species recruited gradually to a species composition of about 80% exceeding annuals but subsequently declined during the last three years of sampling. Nonetheless, the decline was not significant. The fluctuations indicate that the vegetation had not yet fully recovered. In undisturbed sites, perennial species dominated throughout the study period.

174813.fig.005
Figure 5: Composition of annuals and perennial grass species in disturbed (distbd) and undisturbed (undistbd) sites in CKGR:1998 to 2006.
3.4. Herbaceous Cover

Generally the disturbed sites showed some relatively higher percentage cover compared to undisturbed sites (Figure 6) but the difference was not significant ( value = −0.37, ). The highest herbaceous cover was recorded in 2000 and the least in 2005 for both disturbed sites and undisturbed sites, respectively. The peak score was 94.46% while the least score was 41%.

174813.fig.006
Figure 6: Herbaceous cover at disturbed and undisturbed sites in CKGR: 1998 to 2006. Error bars signify standard errors.
3.5. Similarity

There were great dissimilarities in species composition during the first two years of sampling (Table 1). However, as the range rested following termination of disturbance, the similarity index increased gradually reaching 0.48 in 2001. Nonetheless, the similarity index continued to fluctuate indicating that the range condition had not stabilized.

tab1
Table 1: Similarity Czekanowski coefficients of grass and forbs species between the disturbed and undisturbed sites.

4. Discussion

The study showed that forbs were the first to colonize the disturbed area and continued to dominate within the first two years posttermination disturbance. Dominance of forbs on disturbed area may be attributed to their hardness, ability to grow in very unfavorable conditions, adaptation to disturbed sites, and ability to grow vigorously to out-compete other plants [22].

Native grass species recruited gradually in disturbed plots dominating in the third year of sampling. Initially pioneer grass species that include Schmidtia kalihariensis, Enneapogon cenchroides, and Melinis repens dominated during the first five years of sampling but climax species which included Schmidtia pappophoroides and Eragrostis lehmanniana dominated during the last two years of sampling. This transition pattern of dominance by forbs and pioneer grass species in the early years of abandonment followed by an increase in climax species after some years postdisturbance are consistent with succession patterns observed by Holl [25].

Cenchrus biflorus recruited substantially as well in the first three years postdisturbance but declined significantly in the fourth year coinciding with the dominance of native grass species. This suggests that C. biflorus is out-competed by native grass species if disturbance is reduced or minimized. Of significance is that C. biflorus did not spread to undisturbed plots suggesting that the species is not competitive, and when factors promoting range deterioration are minimized, the species may not spread in the Central Kalahari Game Reserve. Also, the removal of livestock which degraded the area might have reduced the proportion of disturbance hence limiting the spread of C. biflorus to undisturbed areas. These results concur with previous studies that reported that reducing intensity of use of affected land is an efficient tool to control C. biflorus.

Annual grass species were relatively high during the first three (3) years in disturbed sites but gradually declined. In contrast, perennial species recruited gradually exceeding annuals from 2001 until 2003 but subsequently declined during the last three years of sampling. Nonetheless, the decline was not significant as the species continued to dominate. This pattern of vegetation succession growth was observed by Seitshiro [15] and Perkins [16] who concluded that perennial grass cover return rapidly to areas denuded of grass during dry years in subsequent years of average or better rainfall or cessation of grazing in the Kalahari sandveld vegetation communities. The presence of perennial grass species is vital since it protects the soil from erosion. It also changes the morphology of the landscape which would otherwise be bare during the dry season if it was dominated by annuals.

The results of this study concur with Wiggs [17] who concluded that it takes 8–10 years for a disturbed Kalahari biome to fully recover. However, studies by Samuel and Hart [1] have shown that it takes a longer time for the disturbed areas to fully recover if the sites were severely disturbed. This could be the situation here considering that the study area was under immense pressure of disturbance from human activities that included decades of small scale cultivation, livestock production, harvesting of grass and wood, as well as general infrastructure developments.

Generally the disturbed sites showed some relatively high percentage cover compared to undisturbed sites. The high cover in disturbed sites could be attributed to invasive species ability to vigorously regenerate and recruit rapidly in abandoned disturbed sites [22]. Cenchrus biflorus, though classified as a weed in Botswana, has potential of being a nutritious food plant for herbivores. The National Research Council [26] quotes it has “potential to improve nutrition, boost food security, foster rural development and support sustainable landcare”.

In conclusion, the results show that Cenchrus biflorus has some high tolerance to disturbance as it comprised the larger proportion of grasses in disturbed sites at the inception of the study. However, it decreased in abundance and was absent in the undisturbed area, suggesting that its ability to invade undisturbed sites is limited. The results indicate that resting an area invaded by C. biflorus might rehabilitate it, and this takes several years depending on rainfall variability. It is thus important to mitigate and prevent factors which promote range deterioration such as human activities (ploughing, trampling, developments, harvesting of grasses for thatching, and firewood collection) and overgrazing in wildlife protected areas. This action will reduce the probabilities of range deterioration and intrusion of undesirable invasive exotic species in protected areas. This will promote viable wildlife populations and its habitat, thus conserving the biodiversity in such a protected area.

Acknowledgments

Many thanks to N. Tsopito, B. Batsile, B. Seretse, L. Maswena, Morake Moetsabatho, and B. Thibedi of Department of Wildlife and National Parks, Botswana who assisted in the study at its different stages and times. Thanks also go to Dr. D. P. Lebatha for reviewing an earlier draft of this paper. This work was supported by the Department of Wildlife and National Parks, Botswana.

References

  1. J. M. Samuel and R. H. Hart, “Sixty one years of secondary succession on rangelands of Wyoming high plains,” Journal of Range Management, vol. 47, pp. 184–191, 1994.
  2. E. J. Dyksterhuis, “Condition and management based on quantitative ecology,” Journal of Range Management, vol. 2, pp. 104–115, 1949.
  3. S. Vetter, “Rangelands at equilibrium and non-equilibrium: recent developments in the debate,” Journal of Arid Environments, vol. 62, no. 2, pp. 321–341, 2005. View at Publisher · View at Google Scholar · View at Scopus
  4. S. W. Todd and M. T. Hoffman, “A fence-line contrast reveals effects of heavy grazing on plant diversity and community composition in Namaqualand, South Africa,” Plant Ecology, vol. 142, no. 1-2, pp. 169–178, 1999. View at Scopus
  5. D. A. B. Parsons, C. M. Shackleton, and R. J. Scholes, “Changes in herbaceous layer condition under contrasting land use systems in the semi-arid lowveld, South Africa,” Journal of Arid Environments, vol. 37, no. 2, pp. 319–329, 1997. View at Publisher · View at Google Scholar · View at Scopus
  6. A. Hoshino, Y. Yoshihara, T. Sasaki et al., “Comparison of vegetation changes along grazing gradients with different numbers of livestock,” Journal of Arid Environments, vol. 73, no. 6-7, pp. 687–690, 2009. View at Publisher · View at Google Scholar · View at Scopus
  7. C. Skarpe, “Decertification, no-change or alternative states: can we trust simple models on livestock impact in dry rangelands?” Applied Vegetation Science, vol. 3, no. 2, pp. 261–268, 2000. View at Scopus
  8. Y. Pueyo and C. L. Alados, “Abiotic factors determining vegetation patterns in semi-arid mediterranean landscapes: different responses in gypsum and non-gypsum substrates,” Journal of Arid Environments, vol. 69, pp. 490–505, 2006.
  9. J. E. Ellis and D. M. Swift, “Stability of African pastoral ecosystems: alternate paradigms and implications for development,” Journal of Range Management, vol. 41, pp. 450–459, 1988.
  10. M. Westoby, B. Walker, and I. Noy-Meir, “Opportunistic management for rangelands not at equilibrium,” Journal of Range Management, vol. 42, no. 4, pp. 266–274, 1989. View at Scopus
  11. M. E. Fernandez-Gimenez and B. Allen-Diaz, “Testing a non-equilibrium model of rangeland vegetation dynamics in Mongolia,” Journal of Applied Ecology, vol. 36, no. 6, pp. 871–885, 1999. View at Publisher · View at Google Scholar · View at Scopus
  12. Z. Bernacki, “Secondary succession of the vegetation in the young shelterbelt (Turew area, Western Poland),” Polish Journal of Ecology, vol. 52, no. 4, pp. 391–404, 2004.
  13. M. Rees, S. Pacala, D. Tilman, R. Condit, and M. Crawley, “Long-term studies of vegetation dynamics,” Science, vol. 293, no. 5530, pp. 650–655, 2001. View at Scopus
  14. T. G. O'Connor and P. W. Roux, “Vegetation changes (1949-71) in a semi-arid, grassy dwarf shrubland in the Karoo, South Africa: influence of rainfall variability and grazing by sheep,” Journal of Applied Ecology, vol. 32, no. 3, pp. 612–626, 1995. View at Scopus
  15. G. Seitshiro, “Gradient analysis of five non-operational boreholes in the Kweneng District,” Unpublished report, Land Utilisation Division, Ministry of Agriculture, Government of Botswana, Gaborone, Botswana, 1978.
  16. J. S. Perkins, The impact of borehole dependant cattle grazing on the environment and society of the eastern Kalahari sandveld, Central District, Botswana, Ph.D. thesis, University of Sheffield, Sheffield, UK, 1991.
  17. G. Wiggs, Geomorphic Thresholds: Aeolian Dune Activity in the Southwest Kalahari, Geographic Research 9, Centre for the Environment, University of Oxford, Oxford, UK, 2007.
  18. A. J. Dougill and A. D. Thomas, “Kalahari sand soils: Spatial heterogeneity, biological soil crusts and land degradation,” Land Degradation and Development, vol. 15, no. 3, pp. 233–242, 2004. View at Publisher · View at Google Scholar · View at Scopus
  19. Bonifica, Initial Measures for the Conservation of the Kalahari Ecosystem, Technical Assistance to the Department of Wildlife and National Parks, Gaborone, Botswana, 1992, Project N.
  20. C. E. M. Tidmarsh and C. M. Havenga, “The wheel-point method of survey and measurement of semi-open grasslands and karoo vegetation in South Africa,” Memoirs of the Botanical Survey of South Africa, vol. 29, p. iv-149, 1955.
  21. D. Field, A Handbook of Common Grasses in Botswana, Ministry of Agriculture, Gaborone, Botswana, 1976.
  22. E. Van Wyk and F. Van Oudtshoorn, Guide to Grasses of Southern Africa, Briza Publications, Pretoria, South Africa, 1st edition, 1999.
  23. Statistica, ““STATISTICA program, version Kernel 5.5 A (Stat Soft, Inc (2000),” STATISTICA for Windows [Computer program manual],” Tulsa, Okla, USA, 2000.
  24. M. Kent and P. Coker, Vegetation Description and Analysis. A Practical Approach, Belhaven Press, London, UK, 1992.
  25. K. D. Holl, “Long-term vegetation recovery on reclaimed coal surface mines in the eastern USA,” Journal of Applied Ecology, vol. 39, pp. 960–970, 2002.
  26. National Research Council, Wild Grains, vol. 1, The National Academies Press, Washington, DC, USA, 1996.