The value of flower tourism on the Bokkeveld Plateau – a botanical hotspot

Abstract

The Bokkeveld Plateau is situated at the interface between the Cape Floristic Region and the Succulent Karoo and has high levels of endemism and plant biodiversity. During spring, the area attracts tourists to view the wildflowers in bloom. Wildflower tourism has been suggested as a means to protect the plant biodiversity of the region. A questionnaire survey was used to estimate the value of flower tourism and to explore visitors' preferences for different attributes of their experience. The travel–cost method, conjoint modelling and direct questions were used. It was estimated that R2,35 million is spent annually in the area. According to the travel–cost analysis, the average consumer surplus was R28 per person. On average, 37 per cent of visitor enjoyment came from seeing expanses or carpets of flowers and 26 per cent from seeing the variety of flowers. The average length of stay would be reduced to 37 per cent of its original length if there were no flowers present. The conjoint analysis showed that seeing carpets of flowers contributed by far the most to the scores given to scenarios, as compared with seeing a variety of species of flowers or bulbs. Biodiversity conservation will therefore depend on marketing the biodiversity and other attributes of the area rather than relying on the visual displays that are highly weather dependent.

1 INTRODUCTION

The Bokkeveld Plateau, located in the vicinity of Nieuwoudtville (31°23′S, 19°07′E) in the Northern Cape, South Africa, is recognised as a hotspot in terms of its high level of plant biodiversity (ca. 1 350 species) and endemism (6,5 per cent) (Van Wyk & Smith, 2001; Todd, 2003). This semi‐arid region lies at the interface between the Cape Floristic Region and the Succulent Karoo, both of which are recognised biodiversity hotspots. The Bokkeveld Plateau, which is approximately 3 million hectares in extent, comprises a variety of soil and vegetation types, of which the areas of dwyka tillite (under Renosterveld vegetation) and dolerite soils are of particular conservation interest. The area is particularly rich in bulbs (geophytes) and is a primary target for conservation. Indeed, the Bokkeveld Plateau is exceptional from a botanical point of view, with as much as 40 per cent of the flora being geophytes (Snijman & Perry, 1987).

The study area is almost entirely under private farmland. Sheep farming, together with some wheat and fodder cultivation, is the predominant activity (Turpie, 2003). Small areas are set aside in municipal and provincial reserves. However, these areas are not considered sufficient for the conservation of the area's unique biodiversity. The biodiversity on private farmlands is under threat due to grazing and other farming activities, and will only be conserved if appropriate management strategies are adopted. One option that has been mooted is the use of important biodiversity areas on farms for generating tourism revenue instead of traditional farming revenues.

One option that has been mooted is the use of important biodiversity areas on farms for generating tourism revenue instead of traditional farming revenues.

During spring, both the conserved and farming areas produce impressive displays of wildflowers in bloom, which attracts tourists to the area. Tourists are accommodated in the small towns, as well as on several farms that provide bed‐and‐breakfast facilities during the season. One farm in the area, Glen Lyon, has changed its grazing practices to encourage the conservation of wildflowers on certain parts of the farm, which generates income through organised flower tours during the spring season.

Could this be a viable means for conservation in the area on a broader scale? A key issue being debated is whether it is viable for more farmers to follow suit. In order to answer this question, it is necessary to understand the demand and supply aspects of the flower tourism industry on the Bokkeveld Plateau.

The main aim of this study was to quantify the demand for wildflower tourism on the Bokkeveld Plateau. The following key questions were addressed:

	What is the visitor profile (origin, interests, etc.)?
	What is the total number of visitors to the study area during spring?
	What is the total value of flower tourism on the Bokkeveld Plateau (expenditure and consumers' surplus)?
	How do different attributes of biodiversity contribute to tourism value?
	Can tourism contribute to the conservation of rare species on the Bokkeveld Plateau?

2 METHODS

Data were collected by means of a questionnaire survey of visitor groups to the area. A total of 164 visitors, each representing a visitor group, were interviewed in the vicinity of Niewoudtville in September 2001. They were interviewed in the town, at a nearby waterfall, at the wildflower reserve and on Glen Lyon, a farm offering flower tours. The questionnaire was brief, taking about 10–15 minutes to complete. The questions, which are described below, were ordered in the most conversational way possible, opening with visitor origins and the journey.

The study used a combination of three resource economic valuation methods to estimate the value of flower tourism on the Bokkeveld Plateau – the travel–cost method, conjoint or choice modelling, and direct questions.

2.1 Total visitor numbers

One part of the questionnaire served to gather data that could be used to estimate the total number of tourists to the area. Visitors were asked whether they were visiting the local waterfall or flower reserve, and whether they were going on the Glen Lyon wild flower tour. The proportions of visitors in the sample engaging in these activities were then used together with known visitor numbers to these attractions, to estimate total visitor numbers.

2.2 Willingness to pay for tourism services

In addition, the questionnaire investigated visitor reactions to the prices asked for the above‐mentioned attractions. If they did not consider the price to be too high, they were asked to estimate the most that they would have been willing to pay for that experience. Although akin to a contingent valuation question (Hanemann, 1994), this part of the questionnaire established willingness to pay by direct questioning rather than through posing any hypothetical scenario.

2.3 The value of flower tourism (travel–cost analysis)

Total visitor expenditure on the Bokkeveld Plateau was ascertained on the basis of estimated total visitor numbers and stated expenditure in the area. Expenditure only represents part of total recreational value, however, as visitors are often willing to pay more than they had to for the experience. The total value, including this extra willingness to pay, or consumers' surplus, is established by estimating the demand function for the tourism experience. This is usually estimated using the travel–cost method (Bockstael, 1995).

The first part of the questionnaire thus established the details about the visitors' trip required for the travel–cost analysis. Respondents were asked about their home town or country, the origin of the trip, type of trip (tour or self‐drive), and group size. They were also asked how often they had visited the area during the past five years, how long their stay in the area was in relation to their entire trip, and to what extent the reason for the entire trip was to visit the Bokkeveld Plateau. The budget for the entire trip was established, as well as the amount budgeted for the Bokkeveld Plateau. Visitors were also asked where they were staying in the area. The inclusion of a value for travel time is a contentious issue in the travel–cost method (Shaw, 1992). In this study, the opportunity cost of visitors' time in travelling to the area was conservatively assumed to be zero, based on the point of view that vacation time does not have an opportunity cost to the travellers.

A zonal travel–cost method was applied, which compares the frequency of visitors from zones of different distances to the reserve with the travel costs from each zone to estimate the visitors' demand curve and the consumer surplus (Hof & King, 1992). The origin zones were first taken as the different provinces in South Africa plus the rest of Africa, Europe and North America. However, numbers of foreign visitors in the sample turned out to be too small to include in the analysis, and the analysis was thus limited to South African visitors.

Estimated total visitor numbers from each zone, based on the proportion of visitors in the sample, were expressed as a proportion of the total zonal population to obtain visitation rates from each zone. Populations for 2001 were estimated on the basis of the 1996 census and an assumed constant growth rate of 2,5 per cent.

In order to avoid the problems of multiple‐site visits, the travel costs were not taken as full trip costs, but as the proportion of total trip costs that visitors ascribed to the Bokkeveld Plateau. Visitation rates were regressed against these adjusted trip costs with the following form of semi‐log travel–cost equation:

where Q is the quantity of visits per zonal inhabitant and TC is the average travel cost per zone adjusted by the average percentage contribution of the Bokkeveld Plateau. A demand curve was then estimated by imputing a series of hypothetical entrance fees to the Bokkeveld Plateau and plotting the effect of these price changes on the total number of visitors, as determined by the travel–cost equation.

Consumers' surplus for each zone was estimated from the derived demand curve using the following function:

Total consumers' surplus is equal to the sum of zonal surpluses. The total calculated is an estimate of consumers' surplus above actual travel costs (UNEP, 1995).

2.4 The contribution of botanical attributes to flower tourism value (conjoint analysis)

The third part of the questionnaire investigated visitor interests and how different botanical attributes of the area contributed to tourism value. Visitors were asked to estimate how each of the following attributes of the area contributed to the total enjoyment of their visit, in terms of percentage contribution:

	Seeing the expanses (‘carpets’) of flowers
	Seeing a variety of species (different colours, forms)
	Identifying rare and endemic species that cannot be found elsewhere
	Enjoying the scenery and ‘getting away from it all’
	The cultural experience of village or farm life, history, restaurants, etc.
	Other

Visitors were asked about their level of botanical interest – whether they distinguished between annuals and bulbs, whether they identified flowers and carried an identification guide, what they wanted to see most and how many species they thought they had seen so far. They were also asked how much time they would have spent in the area if there were no flowers, or if there were flowers but no bulbs.

The contribution of different attributes to tourism value was ascertained using conjoint analysis (cf. Green & Rao, 1971; Stevens et al., 2000; Turpie & Joubert, 2001). Conjoint methods, often used in marketing, seek to ascertain the relative value of different components of a ‘package’ of goods, or in this case the different attributes of the botanical diversity of the area. Respondents are usually asked to score a range of scenarios that differ in the state of their attributes. In this study, a multifactor evaluation was used whereby respondents were presented with a combination of all attributes at once. The number of attributes and attribute levels had to be kept to a minimum to minimise respondent burden. Three attributes of flowers were selected: abundance, variety and number of bulb species, and three levels were defined for each. Abundance levels were ‘none’, ‘some patches’ or ‘vast carpeted areas’. Levels of variety were defined as 10, 30 or 55 types of flowers. Numbers of bulb species were 5, 15 or 30. These are much lower than the overall numbers of species that one finds with intensive sampling – about 100 annual and 188 bulb species (Todd, 2003), but are closer to the numbers of species (in flower) that visitors are likely to encounter in a single visit.

With three attributes and three levels, there were 27 possible combinations. Seven of these were selected, and four were presented in each of three different questionnaire versions (Table 1). Each version included both the ‘middle’ and ‘best’ scenarios.

Questionnaire versions showing the levels of the three attributes presented to respondents

Questionnaire		View/scenery (V)	Level	Flower species (F)	Level	Bulb species (B)	Level
Version 1
Q11	Middle	Some patches	V12	30 spp.	F22	15 spp.	B32
Q12	Worst	None visible	V11	10 spp.	F21	5 spp.	B31
Q13		Some patches	V12	55 spp.	F23	15 spp.	B32
Q14	Ideal	Vast carpeted areas	V13	55 spp.	F23	30 spp.	B33
Version 2
Q21	Middle	Some patches	V12	30 spp.	F22	15 spp.	B32
Q22	Ideal	Vast carpeted areas	V13	55 spp.	F23	30 spp.	B33
Q23		Some patches	V12	55 spp.	F23	5 spp.	B31
Q24	Worst	None visible	V11	10 spp.	F21	5 spp.	B31
Version 3
Q31	Middle	Some patches	V12	30 spp.	F22	15 spp.	B32
Q32		Vast carpeted areas	V13	10 spp.	F21	5 spp.	B31
Q33		None visible	V11	30 spp.	F22	30 spp.	B33
Q34	Ideal	Vast carpeted areas	V13	55 spp.	F23	30 spp.	B33

In addition, provision was made to describe the current visit experience in terms of the same attributes and levels, to be scored as a fifth, ‘status quo’ scenario. The respondents were asked to score the set of scenarios as follows:

Imagine that you had a choice of areas to visit for the day, and this is what you could see:

Table

	Abundance of flowers	Variety of flowers	Number of bulb flowers species	Score
Area 1	Some patches	30 species	15
Area 2	None visible	10 species	5
Area 3	Some patches	55 species	15
Area 4	Vast carpeted areas	55 species	30
Bokkeveld

How would you score all these options (10=100 per cent satisfaction, 1=no satisfaction)? Finally, how does what you have seen on the Bokkeveld Plateau compare, and how would you score this in comparison to the rest?

The relationship between the different levels of the attributes and the response Z was examined using a generalised linear model. In all of the models tested, the attribute view V (abundance of flowers) was coded as a categorical variable. In other words, the three levels V₁₁ , V₁₂ , V₁₃ were either present or absent as indicated by an indicator variable I_1j . The attributes variety of flower species (F) and number of bulb species (B) were coded either as continuous or categorical variables. The regression model with V categorical and F and B continuous is represented as follows:

where I_ij is the indicator variable for the level j of relevant attribute i and is either 1 if the level of the attribute is ‘present’ in the scenario and 0 otherwise. If all attributes are coded as categorical, the equation is therefore:

where the lowest level of each attribute (V_i1) is subsumed in the constant α.

All combinations of these attribute codings were examined for the amount of variance explained and for the significance of the attribute coefficients.

3 RESULTS

3.1 Visitor origins and characteristics

The majority of visitors surveyed (94 per cent) were South Africans (Table 2). The Bokkeveld Plateau is on the southern boundary of the northern Cape Province, adjacent to the Western Cape province, and most visitors came from these two provinces (70 per cent). In addition, a significant proportion came from Gauteng (12 per cent).

Origins of the sample of visitors interviewed on the Bokkeveld Plateau

Origin	Number of groups	%	Number of people	%
Western Cape	71	43,3	222	45,2
Northern Cape	33	20,1	123	25,1
Gauteng	24	14,6	61	12,4
Eastern Cape	6	3,7	16	3,3
KwaZulu‐Natal	5	3,0	12	2,4
Free State	4	2,4	9	1,8
North‐West	3	1,8	7	1,4
Mpumalanga	2	1,2	6	1,2
Northern Province	3	1,8	6	1,2
Africa (Zimbabwe)	1	0,6	3	0,6
United States	2	1,2	4	0,8
Europe	10	6,1	22	4,5
Total	164		491

Less than half (43 per cent) of respondents were visiting the area for the first time. Of the 92 respondents that had been there before, the average visitation rate was approximately 2,8 to 3,1 times over the past five years.

A total of 91 per cent of the groups sampled were on self‐drive visits (Table 3). Several tour companies include the Bokkeveld Plateau on their tours, including specialised mini‐bus tours and more generalist coach tours. Groups on tour included two groups from the United States, one from the United Kingdom and seven South African groups from KwaZulu‐Natal, Gauteng, Northern and Western Cape.

Number of visitor groups sampled that were on tours or self‐drive visits

	On tour	Self‐drive	Total	%
International	3	12	15	9
South African	7	142	149	91
Total	10	154	164	100

Group size on tours (not the number of people on the tour bus) was similar to that on self‐drive visits (2,9±1,4 on tours vs 3,0±1,8 for self‐drive groups). Of the ten groups on tour, eight started their tour in Cape Town and one each in Durban and Johannesburg.

All visitors on organised tours were only visiting the area for the day. In contrast, less than half of self‐drive visitors (43 per cent) were day visitors, with the remainder mostly staying one or two nights, but up to a week (Figure 1). Self‐drive visitors stayed for an average of 1,8±1,1 days.

Number of nights spent on the Bokkeveld Plateau by self‐drive visitors to the area

Of those overnighting in the area, 97 per cent listed places within a 60‐km radius of Nieuwoudtville. A total of 68 per cent were overnighting in Nieuwoudtville or on guest farms in the area. Some 13 per cent were staying in Van Rhynsdorp, 11 per cent in Calvinia and 4 per cent in Loeriesfontein. Some 64 per cent were staying in guesthouses or hotels and 21 per cent were camping. The remaining 15 per cent were staying with friends or family.

The role of the Bokkeveld Plateau in motivating the trip was different for people on tour and those on self‐drive trips. For the former group, the Bokkeveld Plateau contributed an average of 17 per cent to the decision to take the tour, whereas for self‐drive visitors the average was 70 per cent (Table 4). This is not surprising, as tours generally include several destinations. Almost half (49 per cent) of self‐drive visitors cited visiting the Bokkeveld Plateau as the only reason for their trip away from home, and a further 32 per cent of visitors stated that it made up 50 per cent or more of the reason. Visitors from further afield gave a lower average percentage than those from nearby (Table 5).

Percentage that the Bokkeveld Plateau contributed to the reason for going on the overall trip

	Visitors on tour	Self‐drive visitors	Total
Average	17,2	70,2	66,9
Standard deviation	8,4	33,3	34,7

Average±standard deviation (s.d.) of the percentage that the Bokkeveld Plateau contributed to the reason for the trip, for visitors from different origin zones

Trip origin	Ave. % reason	s.d.	Trip origin	Ave. % reason	s.d.
Western Cape	78,5	29,4	Europe	40,7	34,1
Northern Cape	76,8	31,7	Mpumalanga	30,0	28,3
Eastern Cape	75,0	31,6	Northern Province	30,0	39,1
Free State	56,3	23,9	KwaZulu‐Natal	26,0	21,9
Gauteng	53,5	34,2	United States	7,5	3,5
North‐West	45,0	48,2

3.2 Total number of visitors

Some 96 per cent of visitors interviewed had been to or were going to visit the waterfall near Nieuwoudtville while in the area (84 per cent were interviewed there), and 55 per cent were visiting the Nieuwoudtville Wildflower Reserve. These are both municipal sites for which an entrance fee of R4 per person is levied. Some 20 per cent of groups had also been on the Glen Lyon flower tour (Figure 2).

Numbers of visitors recorded at the waterfall and the Nieuwoudtville Flower Reserve during 2001

Based on daily ticket revenues, it is estimated that a total of about 7 880 and 2 110 people bought tickets at the waterfall and the flower reserve, respectively, during the August–September 2001 flower season. However, ticket collectors were absent on 12 days. With the conservative assumption that half the weekday and weekend average number of visitors visited the waterfall on those days, the total estimated number of visitors was adjusted to 8 962. With the assumption that a third of weekday and weekend average visitors came on days that there were no ticket collectors at the wildflower reserve, the adjusted estimate was 3 114.

Of the visitors not interviewed at the waterfall, 77 per cent were planning to visit or had visited the waterfall. Thus, the total numbers of visitors to the study area were estimated as the total number visiting the waterfall×100/77=11 638 people.

3.3 Willingness to pay for facilities and services

Of the 90 respondents who had been to the Nieuwoudtville Wildflower Reserve, 52 responded to the question as to whether it was worth it, with all answering in the affirmative. A total of 82 responded to the question about their maximum willingness to pay for the Nieuwoudtville Wildflower Reserve, and 80 per cent were willing to pay more than the current R4 entry fee (R2 for pensioners and scholars), the average being R9,15 and the median R10 (Table 6). Some 20 per cent (32) of the respondents had been on the Glen Lyon flower tour and of these 29 responded to the question as to whether it was worth the price. Although only one reported that the price seemed high, margin comments indicated that many felt that R50 was about the right price and that more would be exploitative. The average price that people were willing to pay for the Glen Lyon tour was R60 and the median R50.

Responses to questions regarding visits to three different attractions in the study area, whether the areas were worth the price paid (for those that had already been), and their willingness to pay (WTP) to go to each attraction

	Been/going	Worth the price?	N	WTP range	WTP (mean±std. dev.)	WTP median	WTP mode
Waterfall	158/164 (96%)	–	–	–	–	–	–
Nieuwoudtville Flower Reserve	90/159 (57%)	52/52 (100%)	82	R4–R50	R9,15±R6,20	R10	R10
Glen Lyon tour	32/160 (20%)	9/10 (90%)	26	R25–R150	R60±R50	R50	R50
Notes: The first two are given as a number of affirmative responses out of a total of yes or no responses (i.e. excluding ‘maybe’). N is the number of respondents who answered the WTP question.

3.4 Visitor expenditure

Visitors to the Bokkeveld Plateau spent an average of R2 940 on their total trips, amounting to some R34 million in total. Not all of this is attributable to the study area, as many visitors were on multiple destination trips, and not all is spent in South Africa (e.g. international flight costs). Adjusting individuals' expenditure by the percentage reason that the study area contributed to the decision for their entire trip, the average expenditure that could be attributed to the Bokkeveld Plateau was R790 per person, or just over R9 million (2001 rands). Of this, visitors spent an average of R572 to R643 per group or R235 to R277 per person in the Bokkeveld Plateau area, amounting to approximately R2 735 000 for the season.

3.5 Total recreational use value

The average travel costs incurred per person visiting the Bokkeveld Plateau ranged from R243 to R5 702 for visitors from different zones (Table 7). This cost is calculated from the average total trip cost multiplied by the average percentage reason ascribed to the Bokkeveld Plateau for each zone.

Average travel cost per visitor per zone (a) for the whole trip and (b) attributed to visiting the Bokkeveld Plateau (BP)

Zone	Mean cost of trip per person	Portion attributable to BP	Zone	Mean cost of trip per person	Portion attributable to BP
Northern Cape	589	453	Gauteng	1 767	946
Western Cape	694	545	Mpumalanga	3 000	900
KwaZulu‐Natal	3 250	845	Northern Province	4 000	1 200
North‐West	1 889	850
Free State	1 425	802	Europe	17 464	7 108
Eastern Cape	1 521	1 141	United States	78 348	5 876

A travel–cost relationship was estimated on the basis of these data, but excluding foreign visitors (Figure 3). Thus, 69 per cent of the variance in visitation rate per zone was explained by the average travel cost attributed to the Bokkeveld Plateau. The relationship in Figure 3 was used to construct a hypothetical demand curve for visits to the Bokkeveld Plateau, based on a series of hypothetical entrance fees (Figure 4).

Relationship between visitation rate per zone (South African provinces) and average travel cost per zone attributed to the Bokkeveld Plateau

Demand curve for visits to the Bokkeveld Plateau

Using this demand curve, the total consumer surplus was estimated at R193 000. This translated to an average consumer surplus of R28 per person. In other words, the average visitor would have been willing to spend R28 more in the area than he or she had to.

3.6 Visitor interests and the contribution of botanical attributes to tourism value

Some 74 per cent of respondents said that they could distinguish between annuals and bulbs; 31 per cent had tried to identify flowers, and an additional 22 per cent had tried to identify a few, but only 33 per cent had identification books with them. Most were not aiming to see particular species – only 14 (8,5 per cent) responded to the question as to which species they were most interested in seeing. Among those that were, the species mentioned were babiana, bulbinella, cat's tail, daisies, gazanias, gladioli, kelkiewyn, kalkoentjies, kokerboom, lachenalia, perdeblom, sparaxis, viooltjies and vygies.

By far the majority of respondents (71 per cent) felt that seeing expanses or ‘carpets’ of flowers contributed most to their enjoyment. On average, this attribute contributed 37 per cent of enjoyment. Seeing the variety of species (contributing 26 per cent) and the general scenery and getting away from it all (contributing 18 per cent) were second and third most important (Figure 5).

Average percentage contribution of different attributes to the tourism experience

Most respondents (88 per cent) stated that the length of their stay would not be affected if there were no bulbs in the area, and only 4 per cent would be discouraged from coming altogether. The average duration of visits would be reduced by less than 10 per cent (Table 8). However, most respondents (69 per cent) would reduce the length of their visit if there were no flowers, and 49 per cent said they would not come to the area at all. Visitor days, based on the change in average length of stay, would be expected to be reduced to 37 per cent of the current number.

Impact of the presence of flowers or bulbs on the duration of tourist visits

	Current mean duration of visits (days)	Mean duration of visits if no bulbs	% of current stay if no bulbs	Mean duration of visits if no flowers	% of original stay if no flowers
Average	1,8	1,6	91,42	0,63	37,2
Standard deviation	1,1	1,1	24,85	0,92	44,8
Median	1	1	100	0,1	5
Mode	1	1	100	0	0

The average score for what the respondents had actually experienced was 8,7. Average scores given for the different scenarios (see Table 1) presented to respondents ranged from 4,5 for the ‘worst’ scenario to 9,7 for the best (average score, Table 9). The scenario with middle levels of all three attributes scored an average across all versions of 7,13. The difference between the overall ‘middle’ average and the ‘middle’ average for each version was used to adjust the scores to a common benchmark score by adding this difference to each score (average adjusted score, Table 9).

Average score, adjusted score and model predicted scores

		Average score	Standard deviation	Average adjusted score	Model predicted scores	Residual
	Q11 Middle	6,74	1,50	7,13	7,13	0,00
	Q12 Worst	4,45	2,41	4,85	4,48	0,37
n=53	Q13	7,47	1,61	7,87	7,80	0,07
	Q14 Ideal	9,49	0,99	9,89	9,49	0,4
	Q1 Actual	8,72	1,17	9,12	n/a
	Q21 Middle	7,22	1,78	7,13	7,13	0,00
	Q22 Ideal	9,56	1,11	9,47	9,49	−0,02
n=54	Q23	7,81	1,73	7,73	7,80	−0,07
	Q24 Worst	4,76	3,04	4,67	4,48	0,19
	Q2 Actual	8,72	1,49	8,64	n/a
	Q31 Middle	7,43	2,06	7,13	7,13	0,00
	Q32	8,34	1,77	8,05	8,29	−0,24
n=56	Q33	4,79	3,26	4,49	5,02	−0,53
	Q34 Ideal	9,68	0,64	9,38	9,49	−0,11
	Q3 Actual	8,65	1,54	8,36	n/a
Note: Qmn indicates questionnaire version m scenario n.

The respondents' evaluation of the scenarios (Z) was best modelled either with all variables categorical or with only V categorical. Scores for low‐scoring scenarios were more variable than those for high‐scoring scenarios. In all cases, only 44 per cent of the variation in the responses was explained by the model (r²=0,44; n=652). With V categorical and F and B continuous, all variables were significant except B, and so in the final model B was omitted. Then the intercept, the coefficients for level 2 and level 3 of V and for F were all significant (p<0,001). The parameters were as follows:

remembering that I_ij is either 0 or 1, depending whether that level was present in the scenario.

Note that increasing the levels of V adds the most value to the visitors' predicted utility (Z). This confirms the intuitive view that people are primarily interested in seeing ‘vast carpets’ or at least ‘some patches’ of flowers. Of interest is an assessment of the effects of changing the various levels of the attributes V and F. If V and F are at their best level, the predicted Z is 9,5. If F is dropped to its lowest level (only ten flower species), keeping V at its maximum, the score drops to 8,3 (87 per cent of the maximum score). If V is dropped to its lowest level, keeping F at its highest level, the score drops to 5,7 (60 per cent of the maximum score). Thus, enjoyment will be reduced if the variety of species is reduced, but it is not important whether the ‘vast carpets’ are made up of many flower species or just a few. If all attributes are at their worst level, the Z score is 4,5 (47 per cent of the maximum score).

4 DISCUSSION

4.1 General profile of visitors

Only a small percentage of respondents were foreign. This is a lower percentage than recorded in a previous survey in Namaqua National Park, where 27 per cent of visitors were foreign (Loubser et al., 2001). The difference is probably real, in that national parks are more likely to attract foreign visitors. In general, the unpredictability of the quality of floral displays may discourage foreign visitors (Loubser et al., 2001). Indeed, it is largely local people who can monitor weather conditions during spring and visit the area at short notice when conditions are good. As both the onset of the season and the good weather days are variable this is not something that can be planned in advance, unless the visitor is willing to take a risk or is visiting the area for other reasons as well. Thus, tourists in the study area are predominantly South Africans from the same and neighbouring provinces.

The fact that only a small proportion (9 per cent) of visitors arrive on bus tours is to be expected for a site that is mostly a local attraction, as well as being somewhat off the ‘beaten track’. Most visitors were self‐drive visitors. As expected, the Bokkeveld Plateau plays a fairly minor role in the decision for the journey in the case of bus tours, whereas it is the primary destination for self‐drive visitors, with almost half of self‐drive visitors being on a single‐destination trip to the area.

The majority of visitors, including those on tours, were day visitors, with most of the remainder spending only one or two nights away. Overnighting visitors were nearly all staying within a 60‐km radius of Nieuwoudtville. This pattern is to be expected, as it is relatively easy to take in the area's attractions in a day, and because the area is just within range of a weekend outing from Cape Town.

4.2 Total numbers of visitors

Tourism on the Bokkeveld Plateau, as for all Namaqualand tourism (Boonzaaier et al., 2002), is highly seasonal, with tourist numbers being high during August and September during the flower season and extremely low for the rest of the year. Indeed, a large proportion of the smaller tourist businesses in the study area are only operational during the short flower season.

Estimating numbers of visitors or visitor days in an area is always difficult, but is perhaps slightly easier when only a short season is considered. Fortunately, in the case of Nieuwoudtville, the limited number of attractions and the fact that three of these are paying attractions, made estimation possible without having to physically monitor visitor arrivals. It was decided to use the most commonly visited attraction, the waterfall, as the benchmark with which to estimate visitor numbers. This was achieved by applying the proportion of visitors interviewed away from the waterfall that had been or were going to the waterfall, to the numbers of visitors recorded at the waterfall. There was a slight conflict of interest in that the waterfall site was the best place to interview tourists, leaving a smaller sample size from which to calculate the above‐mentioned proportion. Nevertheless, we are comfortable with this estimate. The fact that more people (mostly flower tourists) visit the waterfall than the flower reserve is easily explained by two factors: first, the waterfall is still a good attraction on overcast and rainy days, whereas the flowers are not, and second, on good days, flowers can be seen throughout the area, so many visitors do not see the need to visit the reserve. The numbers recorded as having visited the reserve, calculated from daily revenues, was an underestimate of actual visitor numbers, as staff did not man the gate every day during the season and entry was possible after hours. The total visitor numbers were thus estimated under certain assumptions that were felt to be realistic, but may err slightly on the low side.

A more serious shortcoming in the estimation of total visitor numbers is that it does not capture inter‐annual variability. Visitor numbers are known to change with the quality of the floral display. The year in which this study was carried out was a good year for flowers.

4.3 The total value of flower tourism

‘Tourism value’ can be interpreted in more than one way. The ‘recreational value’ accruing to visitors is technically the total consumers' surplus experienced by the visitors. Consumers' surplus is the value that visitors would have been prepared to pay for the experience over and above what they did actually pay. This is a net value, in that the gross value is their expenditure plus consumers' surplus. The total expenditure by visitors is the value accruing to businesses on the Bokkeveld Plateau and beyond. This is effectively the ‘economic impact’ of tourism. It includes money spent en route as well as on site.

In terms of economic impact, flower tourism on the Bokkeveld Plateau was estimated to generate R9 million per annum in total, most of which would have been spent in South Africa, and R2,7 million of which was estimated to have been spent locally within the areas. This is a substantial boost to the local economy, considering that there are only about 19 000 people in the entire Hantam municipal district (2001 census), although it probably does not provide much increase in employment due to the extreme seasonality. Of the accommodation expenditure, much goes to the wives of farmers.

Consumers' surplus was relatively small, however, with a total estimated value of R193 000. The estimate is fairly rough, as the travel–cost model explained only 70 per cent of the variation in visitation rates in terms of travel costs. Nevertheless, it is a fairly good fit as far as such models go, and also given the vagaries of human behaviour. The result suggested that consumers, on average, would not be willing to spend much more than a further R28 per trip. Quite the opposite situation is found in areas such as the Kruger National Park, where the consumers' surplus far exceeds actual travel costs (Turpie & Joubert, 2001). The low consumer surplus experienced in the study area may be due partly to the fact that people from Cape Town wishing to view flowers will travel as far north as they can afford to go. There is a common perception that flower viewing increases in quality the further north one goes. Nevertheless, the almost total dominance of local visitors in this analysis, in contrast to the Kruger National Park, is probably the major reason.

4.4 How different attributes of biodiversity contribute to tourism value

The overall level of botanical knowledge amongst the respondents was fairly high, with 74 per cent being able to distinguish bulbs from annuals. Almost a third had identification guides and tried to identify flowers, although not many had come in search of particular species. Nevertheless, the most popular attribute of the area was the expanses of flowers that are visible throughout much of the Bokkeveld Plateau, as well as in many other areas such as Namaqualand to the north and further south in the Western Cape. Understandably, however, variety was also an important characteristic, implying that vast carpets of the same species would not be as appreciated as areas of different colours. General scenery and ‘getting away from it all’ were more important than seeing rare endemics or enjoying the cultural attractions of the area. It is important to note that the latter two aspects require knowledge and marketing in order to be appreciated widely. Indeed, Loubser et al. (2001) reported that visitors felt that better knowledge would have enhanced the value of their visits to the Namaqua National Park, also a flower tourism area. If people were more aware of the interesting biodiversity of the area, more people might be inclined to visit it. Furthermore, the area has much to offer in terms of a cultural experience, with interesting attractions such as historic trekpaaie (travel roads). These attractions, if they were more widely marketed, might provide a greater incentive for tourists to risk arriving on bad weather days.

The above findings were echoed in the more quantitative conjoint analysis, which found that the diversity of bulbs per se did not provide a significant drawcard for tourists visiting the area. This is in spite of Nieuwoudtville marketing itself (albeit only locally) as the ‘bulb capital of the world’. Although significant, the conjoint model that was derived from the data had a fairly poor fit. However, this could be expected, given the limited number of attributes that could be used in relation to the complexity of factors that motivate tourists. In the conjoint analysis, visitor satisfaction is measured as utility (Z). The model shows that numbers of flower species contribute much less to this satisfaction than does the extent of vistas of fields of flowers.

4.5 The potential impact of a loss in biodiversity on tourism

There is much concern about the vulnerability of the endemic flora of the Bokkeveld Plateau. The area's biodiversity of interest has always been concentrated on the fertile tillite and dolerite soils, and thus much of these areas have already been ploughed for agriculture. Situated on private lands, the fragmented remains of these vegetation areas are under threat from overgrazing and ploughing, and some species may die out due to the small size of their surviving populations (Turpie, 2003). Most of the threatened species are bulbs.

Tourists, however, appear to be fairly insensitive to bulbs, despite most being able to tell the difference between bulbs and annuals. On average, respondents said they would stay only 91 per cent of their original time in the area if there were no bulb species, but the median change in length of stay was actually zero. This corresponds with the conjoint model – the coefficient for bulb diversity was very small, not significant and in fact slightly negative.

If all attributes are at their worst level (no carpets of flowers, with only ten flower species and five bulb species visible), the utility (Z) score is 47 per cent of the maximum score for the situation of vast carpeted areas of flowers, with 55 flower species and 30 bulb species visible. This can be compared with the responses to Question 22, in which respondents were asked how much less time they would have spent in the area if there were no flowers present (i.e. worse than the worst scenario of the conjoint analysis). On average, they said they would spend only 37 per cent of their original amount of time in the area. Interestingly, this corresponds well with the amount that other attributes of the area (scenery/getting away, cultural and other) contributed to respondents' enjoyment. While the latter gives some idea of the sensitivity of the area to flower abundance in different rainfall years the overall level of conservation in the area is unlikely to have a major impact on flower tourism values, as the species at risk are mostly rare bulbs that would be overlooked by most visitors.

4.6 Supply and demand: the implications for conservation, farming and the local economy

Ecotourism is frequently considered to be a solution for conservation of rare species (King & Stewart, 1996). Indeed, ecotourism has been touted as a possible incentive to ensure the survival of some of the threatened species of the area. This idea has emerged in part because one farmer in the area conducts tours on areas of his farm that are protected from overgrazing or ploughing, and the income from these tours is assumed to cover the opportunity costs involved. However, this study suggests that there is little difference between the demand and supply of tourism services. While there is a high level of occupancy on the Glen Lyon tours, it is unlikely that more than one farm would be able to generate the same level of income from tours. Indeed, this is indicated by the fact that respondents' willingness to pay for the tour was close to the price currently charged. In other words, the tours are generating negligible consumers' surplus. Furthermore, overall consumer surplus among visitors to the area was low in relation to overall expenditure, suggesting that tourists did not have much excess willingness to pay for flower viewing in general. Tourists were willing to pay more for entry to the wildflower reserve, but this is because the entry fee is very low.

The above findings suggest that conservation action or the lack thereof will not have much of an impact on visitor behaviour under the current circumstances, as tourists are more interested in widespread floral displays than in their actual composition. While visitors appreciate diversity, they do not tend to appreciate rarity in a more global context, undoubtedly due to lack of knowledge and marketing.

This study describes the current demand for flower viewing on the Bokkeveld Plateau. It should be realised that several factors could lead to a decrease or increase in overall demand (a left or right shift of the demand curve, respectively). If demand increases, then the average willingness to pay for flower viewing can be expected to increase, and supply of services will be expected to follow suit. An increase in demand, generally, is brought about by a change in preferences, changes in the prices of complements or substitutes and increases in population. Increased prices of complements, such as the fuel price, would decrease demand and increased price of substitutes, such as the cost of visiting the Namaqua National Park, would increase demand for the Bokkeveld Plateau. Changing preferences is the objective of marketing campaigns, and is probably the critical element needed to make ecotourism a viable option for enhancing conservation in the area. However, it must be recognised that increasing the demand is only desirable up to a point, as high densities of tourists would lead to negative congestion effects and possibly also environmental damage.

Additional information

Notes on contributors

Alison Joubert 1

Respectively, Senior Lecturer, Percy FitzPatrick Institute; and Chief Scientific Officer, Department of Statistical Sciences, University of Cape Town, Rondebosch, South Africa.

Notes

Respectively, Senior Lecturer, Percy FitzPatrick Institute; and Chief Scientific Officer, Department of Statistical Sciences, University of Cape Town, Rondebosch, South Africa.