A threat to the natural World Heritage site rarely happens alone

This study contributes an empirical investigation of the likelihood that different external threats to a UNESCO Natural World Heritage Site occur in combination with each other when site characteristics and location are controlled for. For the purpose of the analysis, the World Heritage database and the UNESCO State of Conservation Reports are used and the nine most frequently appearing external threats are identified. These databases include 6852 site-year observations and 3316 threats over the period 1979-2023. The most commonly identified external threats are illegal activities, with eleven percent of all observations and mining with six percent. Transport infrastructure, tourism and visitor pressure are also common threats. Estimation results based on the multivariate Probit (equation system) model demonstrate that there are strong positive correlations between many pairs of the nine external threats. Most apparent are the links between illegal activities and loss of identity, social cohesion, changes in local population and community, water infrastructure (dams) and farming, as well as illegal activities and land conversion. There are also clear links between tourism and infrastructure. This emphasises that the various threats seldom appear in isolation from each other. Results also highlight that the threats have different drivers. Among the determinants, site characteristics and location are the most important ones. The likelihood of threats is highest for Natural Heritage Sites covered by forests or those in marine and coastal areas, Africa as well as the Arab region. It is also possible to identify a general increase in threats over time, although with a diminishing rate of growth towards the period 2015-2019. Contrary to this development and the general downturn in threats during the Covid-19 pandemic period of 2020-2023, pressure from tourism continues to grow. Methodologically, the results emphasize the need for multivariate Probit models when research goes beyond analyses of descriptive statistics and single equation approaches.