The Effect of Human Trampling on Three Vegetation Communities in Mount Elgon National Park
 
 

An increase in the number of visitors to Mount Elgon National Park (MENP) is anticipated as a result of the Ugandan Government’s tourism strategy as described in the Summary.  Recreation ecology is the study of the impact of outdoor recreation on natural or semi-natural environments (Liddle 1991).  In particular, the effect of  human trampling on vegetation and soil has been the subject of many recreation ecology studies (Boucher et al 1991,Cole 1993, 1995a, 1995b, de Gouvenain 1996,  Liddle 1975, 1991, Sun and Liddle 1991, 1993, Taylor et al 1993).  Such investigations have been stimulated by the increasing demand for information on which to base management decisions (Liddle 1975) concerning environments open to human impact.  Boucher et al (1991) point out, ‘nearly all the published papers were done in the temperate zone’, and ‘with the development of ecotourism leading to increased visitation of parks in tropical rain forest areas, this lack of information is a handicap to managers wishing to control the impacts of trail use’.  However, Boucher et al (1991) found that their data on tropical Costa Rican rain forest vegetation, ‘tended to confirm the work of previous researchers in the temperate zone’.  Particular concern rests on the effects of trampling in tropical rain forests because ‘most tropical rain forest species combine several of the properties predicted to be associated with sensitivity to damage’ namely ‘they are vegetatively active all year round, have soft, delicate leaves and are adapted to wet habitats with easily compacted soils.’ (Boucher et al 1991).

A simulated human trampling study was therefore carried out in MENP, using the standard experimental procedure of Cole and Bayfield (1993), with the aim of assessing the effect of various intensities of human trampling on different vegetation communities contained within MENP.  The results of this study are available to the Ugandan Wildlife Authority, the MENP Conservation and Tourism Development Officer, and the MENP Conservation and Development Project to consider for further decision making regarding human trampling on MENP.

It was not possible to study the impact of livestock trampling as the movement of livestock in the area of study was uncontrolled.
 

Given the time available to me I was able to study the effects of human trampling on three vegetation communities of Mount Elgon.  In addition, as the expedition base camp was located at Piswa Patrol Hut, from where ‘Piswa Trail’; the trekking trail to attractions including Hunters’ Cave, Jackson’s Peak, Wagagai Peak and Mudi Cave, commences, I chose three vegetation communities along this trail.  The vegetation communities were also chosen for their proximity to the expedition base camp; for time and safety considerations.  Many alternative vegetation communities found along ‘Piswa Trail’ remain to be studied, including those found above 3,000 m altitude, classified by van Heist (1994) as ‘Afroalpine and Ericaceous communities’.

The altitude of the vegetation communities included in my study is 2,850 m.  The three vegetation communities respectively comprise:

Plot A: Transition from grassland to shrubland with dominant tree species Juniperus sp, Afrocranea volkensii and Dombeya goetzeni

Plot B: Disturbed grass clearing in shrubland with dominant tree species Hagenia abyssinica.

Plot C: Ground level forest canopy with dominant tree species Afrocranea volkensii

The methodology used is that of Cole and Bayfield (1993) which is a standard experimental procedure to study the recreational trampling of vegetation, thus allowing results generated by different studies and different observers to be readily comparable.
 

 Layout of treatment lanes Each Plot contained five delineated lanes (subplots), measuring 0.5 m wide and 5 m long.  Each of the five lanes should  have been replicated a minimum of four times but this was not possible due to the limited size of appropriate vegetation sites available nor the time constraints.  Each vegetation Plot was chosen for its plant community homogeneity and for being on flat ground.

 Trampling treatments Each subplot was randomly assigned one of five trampling treatments: control (no trampling), 25, 75, 200 and 500 passes.  A pass being a one-way walk along the subplot.

 Measurements  These are aimed at assessing the effect of the various trampling treatments on both vegetation cover and height (vegetation structure).  The results of the following measurements are given as Appendix (1).

1) Visual estimates of the canopy coverage of each vascular plant species; only green photosynthetic material should be included.  Cover was recorded as 0 for no cover then as 0.5%, 1, 5, 10, 15, 20, 30, 40, 50, 60, 70, 80, 90 or 100%.

2) Visual estimates of the cover of bare ground; defined as ‘ground not covered by live vegetation’ (Cole and Bayfield 1993) and therefore includes the decomposing litter, eg fallen tree leaves, and litter of recently trampled plants.  This latter form of litter can however be misleading if such ‘litter’ appears to be dead but after time does recover, ie reshoots.

3) Measures of vegetation height are meant to be taken using a quadrat frame.  However, as I did not have such a frame I measured the height of the species within in each structure of vegetation canopy at a distance of 40 cm intervals along the length of the subplot.

Measurements aim to assess both the damage and recovery responses of vegetation.  Damage refers to the amount of vegetation change that occurs as a result of trampling disturbance (Cole and Bayfield 1993), ie changes in cover and height.  As Liddle (1975) explains, ‘direct mechanical damage to plants is shown most clearly by the persistent reduction in height that occurs in trampled areas.’.  Recovery refers to the rate at which the vegetation reverts to pre-disturbance conditions once trampling ceases (Cole and Bayfield 1993).  Vegetation height  measurements were repeated immediately after trampling as this is when the greatest reduction in height occurs (Cole and Bayfield 1993).  Vegetation cover measurements were repeated two weeks after trampling as cover loss often continues to increase after the trampling treatment as trampled vegetation dies (Cole and Bayfield 1993).  I also used this two week post-trampling period as an indication of the recovery rate for those species that were able to re-shoot in this period.
 

Cole and Bayfield (1993) use the following methods of data analysis, which have since been used in several investigations (eg Cole 1993, 1995a and 1995b).  The two primary measurements of vegetation change are relative cover (RC) and relative height (RH). For both RC and RH, conditions after trampling are expressed as a proportion of initial conditions, with a correction factor (cf) applied to account for spontaneous changes on the control plot (in my study trampling by livestock was an example of this).  Additional information on the effect on vegetation by trampling can be derived from data for individual species and bare ground.

Relative Cover (RC) is based on the sum of the coverages of all species, not on a single estimate of total vegetation cover.

RC = surviving vegetation cover on trampled subplots _{x cf x 100%}
          initial vegetation cover on trampled subplots

where cf = initial vegetation cover on control subplot
                surviving vegetation cover on control subplot

Relative Height (RH) is also a sum of each height measurements taken for individual subplots (each species at 40 cm distance intervals in this case).

RH = surviving vegetation height on trampled supbplots _{x cf x 100%}
          initial vegetation height on trampled subplots

where cf = initial vegetation height on control subplot
                surviving vegetation height on control subplot

The results of the RC analysis are given in Table 1 and Figure 3.1 at the end of this chapter.  The data set for RC of Plot A is given as Table 1 also at the end of this chapter.

Figure 3.1 gives the percentage RC of each subplot in each plot two weeks after the respective number of tramples.  Liddle (1973 in 1975b) first proposed the methodology of comparing the vulnerability of different habitats to trampling by using the number of passes necessary to cause 50% reduction in RC as an index of vulnerability.  50% reduction in RC could also be taken as an index of the number of passes necessary to leave an evident path; because as Cole (1993) points out, ‘an evident path encourages additional use’.  From Figure 3.1 it can be seen that RC falls below 50% for each Plot at between 100 and 150 tramples; specifically 140, 130 and 115 tramples for Plot A, B and C respectively.

The variation in the number of tramples necessary for RC to fall below 50% can be accounted for by:

  The morphology and density of the vegetation at each canopy level within the  communities;  particularly;
 (i) the species type at ground level vegetation cover; and
 (ii) the % of ground level vegetation cover; and

 The soil type underlying the vegetation as affected by trampling and rainfall.

Liddle (1975a) has put together a theoretical relationship between the primary productivity of vegetation and its ability to tolerate trampling.  This comprises seven statements on the effect of human trampling on plant communities with three main factors affecting tolerance to trampling identified as:

  plant structure (eg Statement (4), ‘Tall grasses will at first give way to lower growing dicotyledonous species but as the amount of trampling increases the monocotyledonous species will replace the broadleaved plants before they, in turn are eliminated’);

 potential for regrowth (eg Statement (6), ‘ Trampling will at first stimulate and then reduce primary production as its (trampling) severity rises;); and

 environmental conditions (eg  Statement (2), ‘Plant communities are less tolerant to trampling when the ground is wet than when it is dry.’.
 

Plot A comprises several taller growing species eg Crassocephalum montuosum and Sambucas africana described as ‘weak’ and ‘soft’ respectively by Agnew and Agnew (1994) and as such these are vulnerable to trampling. In addition, Sun and Liddle (1993b) conclude that species with erect herbaceous forms, which Crassocephalum montuosum and Sambucas africana both have, are the least resistant species to trampling. Indeed, the ‘disadvantage of tall stems for survival’ with regard to the vegetation of footpaths has been recognised since as early as 1917 (Jeffreys in Liddle 1975a) and as shown above, is used as one of Liddle’s seven statements of his theoretical relationship between vegetation and trampling. However, the lower canopy of Plot A is dense and abundant in species such as Achyranthus aspera, Stellaria sennii and Galium sp which collectively form an initial resistance to trampling. Thus the effect of the interaction of plant species within a community in relation to trampling (Sun and Liddle 1993b) also becomes apparent. The ground level canopy of Plot A is comprised of Plantago palmata which proved to be highly resistant to trampling treatments. Although Plantago palmata is not particularly abundant in Plot A, where it does appear it remained after all trampling treatments. The morphology of Plantago palmata accounts for its high resistance; it has a basal rosette form of stalked, ovate leaves. Sun and Liddle (1993b) list several trampling studies which indicate that plants with such a rosette form are frequently identified as having trampling resistance. Liddle (1991) goes so far as to state that ‘There is little doubt that smaller plants survive better in trampled areas; plants that are not able to grow in rosette, creeping or other low-growing form do not survive long on new pathways and rarely appear in the trampled flora’.

Plot A has underlying soil with a 2cm thick litter layer (decomposing vegetation).  From 2-28 cm depth the soil is dark brown-black in colour which can be interpreted as it having a high organic content.  It could also be interpreted that this soil has been cultivated relatively recently, due to its stable structure and lack of clays both of which would be expected to be the opposite case, given the soils of Plots B and C.  Roots are fine and abundant throughout the profile to 28 cm, at which point I came across much thicker, woody roots (below).  I carried out the ‘feel’ method as defined by Brady and Weil (1996) to distinguish between a sand, a silt loam or a clay whereby moist samples are rubbed between the thumb and forefinger.  I also assessed the criteria for the field method of determining soil texture classes.  From these criteria I concluded that the soil of Plot A is a silt loam.

Plot A soil profile: roots at 30 cm

Moist samples are rubbed between thumb and forefinger. Note shiny appearance of clay (c), lack of cohesion in sand (b), and intermediate status of silt loam (a). Estimates of % clay are often a useful first step in determining textural class by feel. % clay is best estimated by squeezing a ribbon of moist soil (d).
 

Plot B has a ground level canopy comprising abundant Grass sp A which forms a prostrate mat of grass cover over the soil, ie turf-forming. Among this mat grows abundant Geranium arabicum which appeared to be particularly tolerant of trampling. This tolerance may be related to the interaction of the Grass sp A mat with the Geranium arabicum as the latter is only approximately 1 cm higher than the mat and the mat creates a protective cushion against trampling; Cole (1993) describes turf-forming graminoids (grasses) as the ‘most resistant growth forms’. Oxalis corniculata also grows among the Grass sp A mat, but the thinner stemmed and smaller leaved Oxalis corniculata appeared to be less tolerant of trampling than the relatively thicker stemmed and much broader leaved Geranium arabicum.

Plot B also has several tussock ‘clumps’ of Grass sp B.  Although the height of these clumps was not directly reduced by trampling, it was indirectly reduced as a result of the clumps splitting and the tussock dividing and falling over.  This had the added effect of bringing loose soil to the ground surface; increasing the % of bare ground cover.  While Sun and Liddle suggest (1993a) and conclude (1993b) that a tussock form may have higher resistance to trampling than prostrate species, I did not find this necessarily to be the case.  I found that the relative resistance to trampling of these two forms also depended on, for example, the thickness of the ‘mat’ formed by the prostrate species.  Cole (1993) also concludes that the most resistant plants are tuft-forming and mat-forming graminoids.

    Plot B (subplot 2) after 75 tramples - flattened Polygonum sp.

The apparent low resistance to trampling of Polygonum sp, including both stem-less and tall-stemmed forms (above) substantiates Sun and Liddle’s (1993b) conclusion that erect herbaceous forms are more vulnerable to trampling than prostrate or tussock forms, and their conclusion (1993b) that broad leaves may also be less resistant than tussock species. However, the unidentified Species A in Plot B, when stem-less is of similar form to the stem-less Polygonum sp form, ie clump of large stalked oblong- ovate leaves emerging directly from the soil surface (below), yet Species A appeared to be less tolerant to the same treatment of trampling (75) than the Polygonum sp. My suggestion here is that the Polygonum sp leaves have a different chemical structure to that of the leaves of Species A eg a waxy cuticle, making the Polygonum sp relatively more resistant to trampling.

         Unidentified Species A in plot B

The soil underlying Plot B appears to contain less organic matter and more clay than that of Plot A’s soil (below).  There is no substantial layer of litter or fermentation as in Plot A which can account for the apparent lack of organic matter in the soil itself. The surface of the soil is covered by the mat of Grass sp A  and low-height plant species growing among the mat.  Roots are most abundant at 0-10 cm depth and are mostly fine with some thicker grass roots.  Roots continue throughout the profile but are less abundant beyond 10 cm and continue to decrease with increased depth,  The A horizon from 0-20 cm depth appears to be a silty loam, the B horizon from 20-50 cm depth has increasing amounts of clay appearing and from 50-70 cm the soil is more compact and apparently has a higher clay content with a shiny, (Brady and Weil 1996) orange-brown appearance.

    60 cm soil profile in Plot B

The increased clay content and decreased organic matter content may account for Plot B’s soil’s weaker structure than that of Plot A.  The higher clay and lower organic matter also makes the soil more slippery underfoot when vegetation cover has been removed through trampling.  This slipperiness is exacerberated by rainfall.
 

Plot C comprises forest understory with a ground vegetation layer this time of trailing stems of species including Didymodoxa caffra and the Species A of Plot C (different from Plot B Species A). This ground cover is not turf-forming and this appears to be one of the reasons that this plot is less tolerant of trampling than Plot B. However, Plot C does have a higher abundance of Plantago palmata and the individuals of this species also have larger leaves than those found in Plot A. In addition to the high resistance of Plantago palmata to trampling, two Ranunculus spp in Plot C also appear to have a high resistance to trampling. Indeed, often on the established trails throughout the forest the only remaining vegetation to be found would be Plantago palmata and these two Ranunculus spp (below). A number of impact studies have found closed forest understory vegetation (such as is Plot C) to be more severely impacted than vegetation types found in open areas (see Cole 1993 for list of studies). Although Cole (1993) explains that ‘direct sunlight does not confer greater durability; rather, the growth forms of plants adapted to shade tend to make them susceptible to damage from trampling’.

Evident footpath through forest with only Plantago palmata and Ranunculus sp. remaining on path

Plot C also contains Mimulopsis sp which appears highly resistant to trampling in both its trailing and erect-branched woody form. This disputes Sun and Liddle’s (1993b) conclusion that erect woody forms are less tolerant of trampling. Cole (1995b) states that ‘woody species have been variously reported as more or less resistant than herbaceous species’.

   70 cm soil profile in Plot C

The soil underlying Plot C has a thin 3 cm litter layer.  The soil still appears to be a silty loam down to approximately 60 cm with less clay at this depth than the soil of Plot B (above).  However, with heavy rainfall and trampling the soil has a very weak and slippery structure; as with Plot B and is more apparent due to the lack of the protective mat of Grass sp A found in Plot B.  The Grass sp A of Plot C is not prostate and does not therefore form such a protection against trampling nor rainfall.  Thus the required number of tramples to reach 50% relative cover is less for Plot C than for Plot B; despite the abundance of Plantago palmata and Ranunculus spp as mentioned above.
 

It can be seen from this study that:

  different plant species react differently to human trampling; and it follows that

  the composition of the plant community is important to that community’s survival against human trampling.

  100 consecutive passes was necessary to reduce the three vegetation types to 50% relative cover.

  The effects of trampling eg loss of vegetation cover, soil compaction and soil erosion, are increased under conditions of heavy rainfall on clay soils without protection from a prostate graminoid mat cover.

  An evident path attracts further trampling from visitors.
 

  Further human trampling studies on different vegetation communities within MENP should be carried out;

  footpaths on MENP should be restricted wherever possible to the most tolerant vegetation communities ie those that have highest resistance to vegetation loss from trampling.  Wherever possible, these communities should also have high recovery rates, ie able to recover vegetation production after losses from trampling;

  such footpaths should be clearly signed for visitor use and wherever possible access to other ‘evident paths’ on MENP should be closed to avoid further attraction to these less tolerant paths;

  closed footpaths should remain closed for a period of at least one year in order to allow the lost vegetation to recover and/or a new vegetation community to grow in the place of the original (a secondary community);

  stabilisation of footpaths, particularly on clay soils; however this appears to be a difficult problem to solve, due to the apparent widepsread absence of rocks on MENP.  Bamboo or timber are fragile resources and could also prove to be as slippery underfoot in wet weather as the soil itself.
 
 

Project Co-ordinator: Jaqueline Pratt
Field Work Team: Jaqueline Pratt and Maganyi Olivia
Plant Identification: Maganyi Olivia
Report Writer: Jaqueline Pratt