Introduction Across southern Africa, nature-based tourism reportedly now generates roughly the same revenue as farming, forestry, and fisheries combined [1]. Worldwide, tourism as a whole has been estimated to account for roughly 10% of gross domestic product (GDP) [2], with wildlife viewing and outdoor recreation (much of it centred on protected areas [PAs]) reportedly making up one of its fastest growing sectors [3]–[5]. Though statistics like these are rarely supported by detailed data, they underpin widespread recognition that nature-based tourism is an important ecosystem service [6], capable of generating substantial resources for both conservation and local economic development [3],[7],[8]. This is particularly significant given that PAs are under increasing pressure to provide economic justification for their existence [9]–[12]. This positive perspective stands in sharp contrast to growing concerns about an emerging disconnect between people and their natural environments. Increasing urbanisation and the rise of sedentary, indoor pastimes (such as television, the Internet, and video games) have been linked to a reduction in informal, outdoor recreation (Pyle's “Extinction of Experience” [13]), with potentially serious consequences for childhood development, mental and physical wellbeing, and environmental knowledge and concern [14]–[21]. Many see this as a major challenge for biodiversity conservation [13],[14],[21],[22]: if people no longer experience and know their natural environments, how can they be expected to care about them? These worries have been further fuelled by a recent and widely publicised paper examining trends in 16 measures of outdoor recreation (14 from the United States, plus one each from Japan and Spain [23]). This analysis showed that, expressed per head of population, visits to natural areas in the United States and Japan (as well as participation in duck-hunting and fishing in the United States, but not hiking, camping, or other hunting) have declined since the late 1980s (though for contrasting US figures, see [24]). From these per capita trends the authors conclude there has been “…a fundamental and pervasive shift away from nature-based recreation” [23; see also 21,25]. However, the paper produced no evidence of declines outside the United States and Japan (and per capita national park attendance in Spain, the only other country sampled, has not declined), raising the possibility that the reported shift may not be universal. To date, lack of data has meant no study has looked at trends in nature-based tourism across more than a handful of countries. Here, we use newly compiled information on visitor numbers to 280 PAs in 20 countries (Australia, Canada, Chile, China, Ecuador, Ghana, India, Indonesia, Japan, Korea, Madagascar, Peru, Philippines, Rwanda, South Africa, Sri Lanka, Tanzania, Uganda, UK, United States) between 1992 and 2006 to explore the generality of the United States and Japan results and to understand the apparent mismatch with the claim that globally, nature-based tourism is on the rise. Importantly, because we are interested in trends in nature tourism as a whole as well as individual interest in nature, we analyse changes in both total visit numbers and visit numbers corrected for national population size. The latter are a better reflection of per capita interest in a country's PAs [23], but the former are a more sensible proxy for trends in the overall benefit derived from nature tourism as an ecosystem service. Results Our analysis of standardised rates of change in PA visit numbers provides limited support for the previously reported declines in nature-based activities in the United States. Using total visit numbers, only 14 out of 51 US PAs for which we could get data showed significant decreases in visit number (at p 1.5× interquartile range), respectively. (A) Changes in total visit number; (B) changes in per capita visit number. These patterns of spatial heterogeneity were confirmed when data were analysed by country (Table S1). Total visit numbers to PAs on average grew in 15 out of the 20 countries sampled and fell in four (with Uganda showing no change). Even allowing for population growth, per capita visit numbers rose in 14 countries (with Uganda and Australia added to the list of countries showing falling visitation). The only country we sampled outside the Organisation for Economic and Co-Operation Development (OECD) with consistently falling PA visitation was Indonesia. National rates of change are closely associated with wealth. In contrast to the United States and Japan, poorer countries typically had increasing numbers of PA visits, with median standardised rates of growth in total visit numbers showing a clear negative relationship with per capita GDP (Figure 2A; regression weighted by number of PAs sampled per country: adjusted r 2 = 0.52, n = 20 countries, F 1,18 = 21.8, p 70% of visitors are nationals, so that errors caused through adjusting by national population size are relatively limited. The exception is Africa, where on average only ∼30% visitors are nationals. For this continent, adjustment by national population growth (which is also generally higher than elsewhere) is probably excessive and so negatively biases estimates of trends in per capita visit rates. For each PA we next performed linear regressions of total visit number and per capita visit number on year, and derived standardised measures of rates of change (ranging from +1 to −1) as (slope/maximum total [or per capita] visit number predicted by the regression during the 15-y range). We explored geographical variation in trends in our two measures of visit numbers by calculating median standardised rates of change across continents, and across countries (Table S1). We compared the latter with per capita GDP adjusted for purchasing power parity (PPP) (for 2005, from [30]), using linear regression weighted by the number of PAs sampled in each country. As an additional check to see whether our results for total visit number were confounded by changes in national population size, we performed an equivalent weighted regression of national median change in total visit number versus annual population growth (for 1990–2006, from [30]). Last, to see whether our findings were specific to nature-related tourism, we also obtained data on trends in all foreign arrivals between 1995 and 2005 (again from [30]), and compared standardised national rates of change (calculated in the same way as for PA visits) with per capita GDP and with median standardised rates of change in total visit numbers. Supporting Information Figure S1 Median national rates of change in total numbers of PA visits in relation to annual population growth (1990–2006); the number of PAs sampled per country is reflected in point size. (0.53 MB EPS) Click here for additional data file. Figure S2 Standardised annual change in foreign arrivals (1995–2005) in relation to per capita GDP (in 2005), adjusted for PPP; solid line represents the best model, dashed lines represent ±1 standard error (SE). (0.58 MB EPS) Click here for additional data file. Table S1 National values of annual rates of change in total and per capita visits to PAs, per capita GDP, number of PAs sampled, and annual rates of change in foreign arrivals. (0.05 MB DOC) Click here for additional data file.
