The value of wildlife-viewing tourism as an incentive for conservation of biodiversity in the Okavango Delta, Botswana

Abstract

The Okavango Delta is a large wetland safari destination in north-western Botswana. Given that future threats to its water supply may affect the biodiversity of this ecosystem, it was important to elicit a value of the Delta from the tourism sector and assess how biodiversity influences that value. This paper presents a valuation of visitors' preferences for the preservation of the Okavango Delta in 2001 and 2002 using contingent valuation (CV) and travel cost (TC) approaches. The results showed that the quality of wildlife viewing was significantly correlated with willingness-to-pay (WTP) for preservation and suggested that impaired biodiversity would negatively affect the value of this ecosystem. The combined CV and TC values totalled US$285/visitor/annum. Extrapolated to the annual pool of visitors to the Delta in 2002, this translates to US$23 million, a large reservoir of funds from the tourism sector that could be used for preservation.

1. INTRODUCTION

Wildlife-viewing tourism contributes to the Gross Domestic Product (GDP) of many southern African countries, yet the environment that supports wildlife is quickly diminishing. This may soon be the case for the Okavango Delta, a large pristine wetland in north-western Botswana. The Delta is one of the most frequently visited tourist destinations in Botswana and home to many rare and endangered species of animals. With an area of between 6000 and 13 000 km² (Gieske, 1997), it is the largest Ramsar Wetland of International Importance (Ramsar Convention on Wetlands, 1996). It is also part of the Zambezian Flooded Savannahs Ecoregion, one of the World Wildlife Fund's top 200 ecoregions of global significance. The Okavango Delta has been recognised by the Government of Botswana as a source of floral and faunal biodiversity.

Situated just north-west of the Kalahari Desert, the Okavango Delta is not unlike other arid zone wetlands that are threatened by the growing need for water. It is important to recognise that the Delta and its entire basin system, which has its origins in Angola and is shared by Namibia, is under threat from the socio-economic activities taking place in these two countries and Botswana. Mbaiwa (2004a) argues that for the Delta to remain a natural ecosystem that also supports wildlife resources for the tourism industry, sustainable management measures need to be adopted at both the national and the international level by the basin states of Angola, Botswana and Namibia. Mbaiwa (2004a) also notes that at the international level protocols such as the 2001 Revised Southern African Development Community Protocol for Shared Watercourses encourage the development of Integrated Management Plans for the basin and the formation of a Secretariat for the Basin States to ensure their sustainable management.

Potential diminished water supply because of upstream water withdrawals (Andersson et al., 2003; Mukumbira, 2003), such as the proposed hydropower project by Namibia (Water Transfer Consultants, 2003), and other indirect causes, such as climate change (Tyson, 1991; Arnell et al., 2003), may cause significant ‘shrinking’ of the Delta (Andersson et al., 2003). It is likely that seasonally flooded areas of the Delta, where most large mammals are found, will be severely affected if the Delta shrinks. Also, given that safari camps and public campsites are located in seasonally flooded areas, reduced water supply could have non-linear impacts on wildlife-viewing opportunities. Such impacts could have ripple effects on Botswana's tourism industry, which contributes 4.5 per cent to Botswana's GDP, or 7 per cent of the non-mining sector GDP (Mbaiwa, 2003).

While the biodiversity of the Delta clearly has intrinsic value, which is difficult to quantify, it is important to demonstrate that these environmental assets are also of economic value, which can be measured. Two methods that can be used to accomplish this are the contingent valuation (CV) method and the travel cost (TC) method. A CV model measures the amount a respondent is willing to pay for a good or for the maintenance of that resource, for example, willingness-to-pay (WTP) for the preservation of the Okavango River water rights. The contingent valuation method is one of the most popular methods employed to determine non-use values (Folmer et al., 1997; Blomquist & Whitehead, 1998; Oglethorpe & Miliadou, 2000) and has been used successfully in developing countries (Navrud & Mungatana, 1994).

The amount that tourists are willing to spend in addition to the cost of the trip, or the consumer surplus, can be measured using the travel cost method. The TC method is a revealed preference economic valuation technique that uses information on travel costs to recreation sites as a proxy for the benefit of those recreation sites (Kahn, 1997; Bateman et al., 1999; Garrod & Willis, 2000). The consumer surplus, the area under the demand curve and above the price that visitors pay, represents the value of the site to the individual (Kahn, 1997). According to the theory of decreasing marginal utility of consumers, the marginal consumer surplus for subsequent visits is expected to be smaller than that for the first visit (Garrod & Willis, 2000).

Previous studies on tourism in the Okavango Delta have dealt with the socio-economic and environmental impacts of tourism development Mbaiwa (2004b, 2004c; 2005) and regional protocols and programmes that can lead to the conservation of the Okavango River Basin (Mbaiwa, 2004a). To advance our understanding of the value of wetland resources and of their conservation, this study seeks to contribute to a wider and important literature on tourism and biodiversity and, in particular, on issues of costing environmental assets. It is assumed that this study will be of interest not only to scholars or researchers of the Okavango Delta but also to a wider set of tourism scholarship. The study uses ecosystem valuation tools to better understand how changes to the biodiversity of the ecosystem may affect the value of this safari destination and, ultimately, its sustainable management. We begin by establishing socio-economic profiles and analysing Delta visitors' trip preferences. We hypothesise that the quality of the trip affects the value of the Delta and that those tourists who take high-end safari trips will have a higher WTP than budget safari tourists. We also hypothesise that diminished biodiversity and decreased wildlife-viewing opportunities will have a negative impact on the value of the Delta, which can be elicited through both CV and TC methods. To evaluate our hypotheses, we present the results of a tourism survey that employs CV and TC methods combined with socio-economic and trip characteristics questions to elicit a value for the Okavango Delta as a safari destination. Therefore the goals of this study are to:

• 1) provide socio-economic data, which can be used in a larger societal context to inform management decisions;

• 2) estimate the value of the Delta for the tourism sector, which represents a reservoir of funds that can be used for preservation;

• 3) determine whether decreased biodiversity affects the value of the Delta;

• 4) evaluate the differences in WTP from visitors taking high-end and budget high-end and budget safari trips; and

• 5) use these results to inform sustainable management of the Okavango River Basin.

2. METHODS

2.1 The study site

Safari camps are mostly located in the seasonally flooded fringe areas of the Delta (Figure 1), where the wildlife-viewing opportunities are greatest. Tourists are generally transported to private safari camps from Maun International Airport by small plane or four-wheel drive vehicle. Visitors to the Panhandle, at the upstream end of the Okavango Delta (Figure 1), are often self-driven or part of an overland tour. From there, tourists can taking fishing trips or be transported deeper into the Delta by motorboat or dugout canoe (mokoro). Some safari camps and public campsites are located in Moremi Game Reserve, a government-protected park within the Okavango Delta, where tourists can enter via four-wheel drive vehicle or overland truck after paying the appropriate national park and camping fees. In 1998, a survey by the Department of Tourism (2001a) estimated that there were 52 512 visitors to the Okavango Delta, and a sizable portion of these represented travel to Moremi Game Reserve (49 556 visitors).

Figure 1. Map of the Okavango Delta, showing safari camps and Moremi Game Reserve. Source: Adapted from Department of Tourism of Botswana, Okavango Delta map, 2001b.

2.2 Survey design

The survey instrument used in this study was a two-page questionnaire (Mladenov, 2004) and all the tourists taking the survey were given an explanation of its purpose and background information on the international water rights conflict involving the Okavango River and Delta. They were informed that Botswana's wildlife is ‘under threat from competing water uses’ and that the size of the Delta could be significantly reduced should there be a large-scale water withdrawal upstream of the Delta. The CV and TC questions were presented first, followed by socio-economic questions, and finally open-ended questions to elicit further comments. The CV question was structured as an open-ended, hypothetical payment request. The first part of this question provided information about the proposed upstream water withdrawals. Two possible uses of hypothetical donations were specified: 1) to invest money toward the preservation of upstream water rights and 2) to fund research to identify the impacts from upstream water withdrawals. Respondents were asked whether they were willing to contribute to this preservation fund for the Delta and what amount they would be willing to contribute. They could choose not to contribute or they could specify amounts for one-off contributions, contributions per visit, or membership subscriptions.

To arrive at a direct valuation, the TC method generally estimates consumer surplus from repeat visits to a site. Because some visitors to the Delta indicated that they would not make repeat visits (not because they did not benefit from it but because it is an exotic, isolated place that they were unlikely to visit more than once), we determined the consumer surplus by asking each respondent: ‘What is the maximum amount of money you would have paid to take the trip?’ The choices were from the present cost to three times the present cost in increments of 0.5. Visitors were also asked questions about trip characteristics, such as trip cost, transport costs, countries visited, type of accommodation and mode of travel into the Delta.

Socio-economic data were collected to determine a profile of the type of tourist drawn to the Delta and the effects of income, age, gender, and nationality on WTP. Because of the already extensive length of the survey, socio-economic variables such as education and size of household were omitted.

2.3 Survey administration and analysis

In our sampling strategy, we attempted to engage tourists taking all types of trips into the Delta: high-end safaris, budget overland tours, boat tours and self-drive safaris. Many of the respondents were intercepted at the Maun International Airport because this location offered the opportunity to sample high-end and budget tourists, as they waited for their flights to the next destination (high-end) or for their overland truck pick-up (budget). Any persons who appeared to be waiting while at Maun Airport were approached and asked if they had recently visited the Okavango Delta. Because one of the study goals was to assess the effects of wildlife-viewing opportunities on WTP, only tourists who had already visited the Delta were surveyed. Tourists were also surveyed at fishing camps in the Panhandle and safari camps near Maun, and those who were self-driven were surveyed at their campsites in Moremi Game Reserve (Figure 1). Translation in German and Spanish was provided for visitors who did not speak English.

The surveys were administered during the tourist seasons of 2001 and 2002 (July to September) to 201 visitors. The number of respondents was constrained by the project budget, interviewer work load, and difficulty in intercepting visitors before they departed from Maun.

2.4 Analytical methods

Between 1994 and 1998, a 13 per cent annual growth rate in tourist arrivals was observed in Botswana (Department of Tourism of Botswana, 2001c). Using the 13 per cent growth rate and a 1998 estimate of 52 512 visitors to the Delta (Department of Tourism of Botswana, 2001a), we projected the number of visitors to the Delta for the study period of 2001–2002 to be 80 700.

Total willingness-to-pay (DONATE) was computed as the amount specified multiplied by the number of times specified, where applicable. In certain instances where an ongoing contribution structure was selected but no frequency of contribution was given, a default of two trips or years was assumed. ‘No’ responses were counted as zeros, and missing responses were excluded.

Consumer surplus, CS _i , computed from the travel cost portion of the 2002 revised surveys, was calculated as:

where CDELTA is the amount spent only during the Delta portion of the trip (many visitors spent money travelling to other destinations in Africa or elsewhere), CTRANSPORT is the amount spent on transport to the Delta from the point of departure, COUNTRIES represents the total number of countries visited during the trip, and MULT is the proportion of the price paid that would have prevented the trip (ranging from 1 to 3, in increments of 0.5), all for the ith individual.

Statistical analyses for the CV method were performed in SAS and Eviews, and for the TC method in SPSS and Excel. T-tests and bivariate correlations were run to establish initial patterns of WTP for preservation among visitors.

3. RESULTS

3.1 Descriptive statistics

Descriptive statistics (Table 1) indicate that most of the respondents were Europeans. The respondents' average age was 36 and their mean annual income was high, at over US$43 000, which is just above the US median household income for 2001 (US$42 228) and 2002 (US$42 409, from the US Census Bureau, 2005). A more realistic distribution of visitor profiles is shown by histograms of age and income (Figures 2 and 3). The 17–30 age group was the largest group (about 41 per cent) of respondents, and incomes in the US$25 000–30 000 range were the most common (about 45 per cent).

Figure 2. Histogram of visitors' age (years) with normal distribution superimposed

Figure 3. Histogram of visitors' income (US$) with normal distribution superimposed

Table 1: Descriptive statistics in the contingent valuation and travel cost models (Abbreviations are shown in parentheses for specific variables used in subsequent tables)

	Variable	Mean	δ	N
Socio-economic characteristics	Age	36	14.51	195
	Gender (0 = female, 1 = male)	0.50	0.50	196
	Income	43 456	52 691	127
	European nationality	0.53	0.50	148
	African nationality	0.20	0.40	148
	North American nationality	0.19	0.39	148
Trip characteristics*	Total cost of trip (US$)	3 616	4 210	184
	Cost of Okavango Delta portion (OKACOST) (US$)	1 122	2 162	163
	Transport cost (US$)	896	1 463	120
	Number of days away	41	60	195
	Number of days lost from work	18	25	89
	Number of days spent wildlife viewing in the Delta (DAYSWILD)	4	4	179
	Also visited South Africa as part of this trip	0.57	0.50	157
	Also visited Zimbabwe as part of this trip	0.57	0.50	157
	Also visited Namibia as part of this trip	0.57	0.50	157
	Visited other countries as part of this trip	0.34	0.47	157
	Also visited Chobe National Park as part of this trip	0.19	0.39	162
	Also visited Makgadikgadi Pans as part of this trip	0.75	0.43	162
	Also visited Tsodillo Hills as part of this trip	0.10	0.31	162
	Lodge accommodation while in the Delta (ACLODGE)	0.20	0.40	196
	Campground accommodation while in the Delta (ACCAMP)	0.74	0.44	196
	Travelled by plane into the Delta (TRPLANE)	0.40	0.49	197
	Travelled by overland truck into the Delta	0.11	0.32	197
	Travelled by boat into the Delta	0.20	0.40	197
	Self-driven into the Delta (TRSELF)	0.23	0.42	197
	Quality ratings**	Rating of camp/lodge employees	3.54	1.49	159
Rating of guides	3.60	1.58	183
Rating of park officials	2.26	1.84	165
Rating of wildlife-viewing opportunities (RTWILD)	4.19	0.90	193
Would pay the same if wildlife-viewing opportunities were diminished (PAYSAME)	0.29	0.46	152
Concerned with environmental impacts on the Delta	0.91	0.28	147
Would recommend the trip to others	0.98	0.14	194
Felt the trip was worth the money spent	0.86	0.35	163
Maximum amount (MAXPAY) visitor would have spent on Delta trip (# of times the present cost)	1.17	0.34	171
Comment (0 = negative, 1 = positive)	0.58	0.50	60
Values	Amount visitor is willing to donate (DONATE) to a preservation fund for the Delta (US$/annum)	60	172	174
	Consumer surplus (SURPLUS): difference between maximum price and actual price paid (US$/annum)	225	744	168
	*0 = no and 1 = yes, where value not specified.
** For ratings, 0 = lowest and 5 = highest. For all other responses, 0 = no and 1 = yes.

Most of the respondents (97 per cent) combined their Botswana trip with trips to other countries, such as South Africa, Zimbabwe and Namibia. While they were in Botswana they also travelled to the neighbouring Makgadikgadi Pans (75 per cent), Chobe National Park (19 per cent) and Tsodillo Hills (10 per cent). While in the Okavango Delta, most stayed in campground accommodation (74 per cent), though a significant number (20 per cent) stayed in lodges. About 40 per cent travelled into the Delta by plane, and the number of self-driven tourists was also high (at 23 per cent).

Average trip costs were also high (Table 1), with the cost of the Delta portion of the trip (OKACOST) representing almost one-third of all trip costs. This portion included only the safari costs (the cost of transport to the Delta from the departure city was not included), and modes of travel ranged from budget self-driven trips to high-cost safaris. The average annual tourist spending on safaris in the Delta, calculated by multiplying OKACOST (US$1122; Table 1) by the projected number of visitors to the Delta for the study period of 2001–2002 (80 700), was US$90.5 million.

Overall, the respondents rated wildlife-viewing opportunities higher than the service they received from the employees who staff the lodges and parks (Table 1). Only 29 per cent said they would pay the same amount for their trip if wildlife-viewing opportunities were diminished and most (91 per cent) were concerned about potential negative impacts on the Delta. Most (86 per cent) felt the trip was worth the money they had spent and 98 per cent would recommend the trip to others. Approximately a third (60) of the respondents provided written comments, and of these 58 per cent were positive. Generally, the wording of the negative comments reflected satisfaction with the Delta as a whole, but displeasure with some aspect of the Delta visit. For example, there were specific negative comments about the camping facilities in Moremi Game Reserve, but positive ones about wildlife viewing in this reserve.

3.2 Socio-economic analysis

Income may be an important determinant for both DONATE and SURPLUS, as indicated by the low ‘p’ values (p < 0.05) in Table 2. However, the lower r² values suggest that the correlations are weak. As expected, higher income was significantly correlated with total cost of the Delta trip (Table 2). Accommodation at lodges was significantly correlated with income (Table 2), and visitors who stayed at lodges had higher incomes than those who stayed in campgrounds (Table 3). Visitors who travelled into the Delta by plane had, on average, higher incomes than those who were self-driven (Table 3).

Table 2: Significant bivariate correlations between income and key variables

	Correlation	P value	N
DONATE	0.208*	0.028	112
SURPLUS	0.284**	0.003	109
OKACOST	0.708**	0.000	109
ACLODGE	0.404**	0.000	127
*Correlation is significant at the 10% level (p < 0.05).
**Correlation is very significant (p < 0.001).

Table 3: Mean WTP for preservation (DONATE), consumer surplus (SURPLUS), income for visitors who stayed in lodges (ACLODGE) and campsites (ACCAMP) and visitors who travelled by plane (TRPLANE) and were self-driven (TRSELF)

		Mean ($)	δ ($)	N
ACLODGE	DONATE	99	204	41
	SURPLUS	609	1 470	35
	INCOME	77 700	73 299	32
ACCAMP	DONATE	50	161	130
	SURPLUS	129	312	128
	INCOME	37 513	52 932	94
TRPLANE	DONATE	68	161	70
	SURPLUS	374	1 075	69
	INCOME	56 987	65 162	58
TRSELF	DONATE	119	277	42
	SURPLUS	314	1 200	40
	INCOME	39 784	52 163	32

3.3 Contingent valuation analysis

The average annual WTP for preservation (DONATE) was calculated from 174 valid surveys at US$60 per person (Table 1). Multiplying the projected number of visitors to the Delta for the study period 2001–2002 (80 700) by DONATE, we calculated a substantial annual donation amount of US$4.8 million. Independent sample tests (Table 4) indicated that DONATE was most highly correlated with rating of wildlife viewing (RTWILD), cost of the Delta trip (OKACOST), and days spent viewing wildlife (DAYSWILD). However, each of these correlations only explained approximately 20 per cent of the variation. DONATE was higher among respondents who had higher responses for wildlife-viewing contributions to trip enjoyment (RTWILD) and also among those who had spent more days in the Delta and those who had paid more for their trip (Table 4).

Table 4: Significant bivariate correlations between amount willing to pay (DONATE) and key variables

	Correlation	P value	N
RTWILD	0.223**	0.004	167
OKACOST	0.231**	0.006	141
DAYSWILD	0.165*	0.032	168
*Correlation is significant at the 10% level (p < 0.05).
**Correlation is very significant (p < 0.001).

3.4 Travel cost analysis

The average annual consumer surplus (SURPLUS) was calculated from Equation 1 for 168 valid surveys at US$225 per person (Table 1). Multiplying the projected number of visitors to the Delta for the study period of 2001–2002 (80 700) by the SURPLUS value, we calculated a total surplus value of US$18.2 million from the tourism sector. The correlation (p = 0.001) between SURPLUS and DONATE (Table 5) was highly significant and suggests consistency between the two valuation methods. However, the strength of the correlation explained less than 30 per cent of the variation. The bivariate correlation between SURPLUS and RTWILDH was just under significant (p = .07). The number of days spent viewing wildlife was significantly correlated with SURPLUS and the R value explained just under 50 per cent of the variation (Table 5). Compared to visitors who stayed in campground accommodation, those who stayed in lodges had a higher average consumer surplus (US$609/person for lodge stays vs US$129/person for camping) and a higher WTP for preservation (US$99/person vs US$50/person; Table 3).

Table 5: Significant bivariate correlations between consumer surplus (SURPLUS) and key variables

	Correlation	P value	N
DONATE	0.283**	0.001	146
RTWILD	0.139	0.076	164
MAXPAY	0.480**	< 0.001	166
OKACOST	0.301**	< 0.001	148
DAYSWILD	0.492**	< 0.001	153
*Correlation is significant at the 10% level (p < 0.05).
**Correlation is very significant (p < 0.001).

4. DISCUSSION

4.1 Results from socio-economic profiles

A generalisation that can be made from the results of our survey is that the average visitor to the Okavango Delta is a middle-aged, middle-income European, South African or North American in his or her mid-thirties. However, the histograms showed that age and, moreover, income were not best represented with a normal curve and that tourists in the 20 to 30 age group and those with annual incomes of about US$27 000/annum displayed the more common socio-economic profile. However, all the socio-economic results should be interpreted cautiously because our study was restricted to the boreal summer when many northern hemisphere residents are out of school or taking vacations. This may influence age, income and nationality estimates.

Correlations between socio-economic data and trip characteristics suggested that higher income visitors could be expected to contribute more to preservation and pay more for trips than budget travellers. Generally, the higher income visitors travelled by plane and stayed at lodges. Therefore, it is no surprise that visitors who stayed in lodges also had a higher WTP for preservation and higher consumer surplus values than visitors who camped (Table 3). The average trip cost had a high standard deviation because of the varying length of stay and the wide range of travel packages, from budget self-driven trips to high-cost safaris. Nevertheless, the average annual amount (US$90.5 million) spent by tourists on safaris in the Okavango Delta during the study period 2001–2002 indicates that tourism generates significant revenue.

The relationship between annual income and DONATE is consistent with the findings of other studies that show a positive correlation between income and WTP (Hadker et al., 1997; Muriithi & Kenyon, 2002; Horton et al., 2003). In agreement with the findings of the CV analysis, there is also a relationship between annual income and consumer surplus, although it is weaker. This is consistent with findings that household income does influence participation patterns for natural resource based trips (Leeworthy & Bowker, 1997; Park et al., 2002).

4.2 Implications of CV and TC values for management and preservation of the Delta

The combined value of US$285 per person per annum (or US$23 million/annum) represents a significant sum. The WTP for preservation elicited from the CV method (US$60 per person per annum) is significantly higher than the range of WTP values reported by Turpie (2003) for the conservation of biodiversity in South Africa in 2001 (from R5 to R100). However, the consumer surplus estimate (US$225 per person per annum) is in the range reported by Turpie and Joubert (2001) for the Kruger National Park (ranging from R554 to R9819 per person among foreign visitors in 2000). These results indicate that the tourism sector may offer a large reservoir of funds that could be tapped for managing and preserving the Delta.

Moreover, the value of the Delta could be extrapolated to a much larger global population that may have preservation and bequest values for this ecosystem without ever having seen it. Although only Delta visitors were questioned in this survey so that the effects of biodiversity on WTP could be assessed, in principle non-visitors also would be able to reveal their maximum WTP for the trip within the same survey structure. Non-visitors deciding to visit the Delta could make the decision only on the basis of information not experience, while the visitors we surveyed had the advantage or bias of experience. A sampling of both visitors and non-visitors would be recommended in a future study to quantify and compare 1) WTP on the basis of expectation and uncorroborated by experience and 2) WTP after experiencing the recreational value of the site.

Consistent with previous research (Sellar et al., 1985; Haneman, 1994), direct use valuations that were elicited using the TC method (US$225 per person per annum) exceeded donation amounts elicited using the CV method (US$60 per person per annum). The high average trip cost may have contributed to higher TC values. In addition, using an open-ended CV question, which typically generates a lower WTP than an iterative bidding question or a dichotomous choice question (Gren et al., 1994), may have contributed to the lower average CV value. A preferred payment structure for future contingent valuation studies is a payment ladder approach, which avoids starting point bias and anchoring effects (Horton et al., 2003). Given that using an open-ended payment structure generates a low WTP, it is likely that our findings of visitors' economic value for the Delta are conservative.

4.3 Effects of biodiversity on ecosystem value

The significant correlation between DONATE and RTWILD (Table 4) indicated that biodiversity positively affects the preservation value of the Delta. Higher DONATE amounts from visitors who spent more days in the Delta viewing wildlife (Table 3) may also confirm the importance of biodiversity, especially if more time viewing wildlife allows for more successful wildlife-viewing opportunities. The significant correlation between DONATE and SURPLUS supports our view that the respondents provided consistent use and non-use values.

Positive correlations between SURPLUS and DAYSWILD (Table 5) further support the importance of biodiversity in the Delta. Taken as a proxy for recreation site quality (i.e. more days viewing wildlife may correspond to greater wildlife-viewing opportunities and more time to take in the safari experience), these results revealed that diminished wildlife-viewing opportunities could decrease the recreational value of the Delta.

5. CONCLUSION

The results of this study highlight a compelling financial incentive to invest in water rights protection and preservation of the Okavango Delta's biodiversity. Income was a determinant both of WTP for preservation and of consumer surplus, suggesting that preservation value and the participation patterns for natural resource based trips are both influenced by personal income. However, the strength of correlation was low and further study would be necessary to verify this preliminary result. The other important finding of this study, that WTP was influenced by the quality of the wildlife-viewing opportunities, has important hydro-political implications. Diminished wildlife-viewing opportunities, a likely consequence of decreased water supply to the Delta, will probably reduce the amount tourists are willing to contribute for preserving the Delta.

The high value of the Delta, as shown in this study, prompts questions about natural resource conservation and sustainable use of the wetland. Tourism activities in the Delta have triggered a number of debates, suggesting the need to re-evaluate the benefits of wildlife-viewing tourism (Mbaiwa, 2003, 2004b, 2004c, 2005). For example, Mbaiwa (2003) shows that the use of the Okavango Delta particularly by the wildlife based tourism industry may not strictly adhere to conservation practices and may in turn lead to unsustainable use of the natural resource. This is the case despite Botswana's high-cost, low-volume tourism policy, which was designed to reduce high tourist numbers and activities so as to conserve the wetland. Further, in the light of our findings of the high annual revenue generated from tourism spending on Delta safaris alone, the leakage of tourism revenue to developed countries by foreign safari companies (Mbaiwa, 2005) also becomes an important issue. At a national level, ongoing plans, such as the Okavango Delta Management Plan, should be developed so that overuse of natural resources and resource conflicts, as observed by Darkoh and Mbaiwa (2005), are minimised to promote sustainability. An approach that minimises resource conflicts has the potential to sustain this high value of the Delta derived from the tourism sector. Such an approach could also promote sustainable economic activities such as wildlife based tourism in the Delta.

In order to draw further policy implications from valuations of the Delta, a rigorous study with increased sample size (including visitors and non-visitors) and more effort to minimise biases (using an improved payment structure and a more tangible payment vehicle, such as a tax) would be needed. The socio-economic profiles and important correlations determined in this study can be used to guide future survey design.

The authors wish to thank May Lehmensiek of the University of Florida and Ariel Esposito and Anthony Lane of the University of Colorado for voluntarily administering the visitor surveys; May Lehmensiek and Ariel Esposito for translating the surveys for foreign visitors; Marnie Norris for data entry; Sepopa Swamp Stop, Audi Camp, Crocodile Camp, Sitatunga Camp, Maun International Airport and tourist respondents for their cooperation; and anonymous reviewers for constructive comments on the manuscript. Funding was made possible by an international supplement to the NSF IGERT Grant # DGE 9987607. The research was also funded in part by the Undergraduate Research Opportunities Program at the University of Colorado.