Negative interactions between people and tigers Panthera tigris on the island of Sumatra, Indonesia, have been documented for centuries. Such interactions include sightings of tigers, loss of livestock, injury or loss of lives of both people and tigers, and evacuation of tigers that have been injured near settlements in incidents unrelated to poaching (Nyhus et al., Reference Nyhus, Tilson and Sumianto2000; Boomgard, Reference Boomgard2001; Campbell-Smith et al., Reference Campbell-Smith, Sembiring and Linkie2012). The tiger is categorized as Endangered on the IUCN Red List (Goodrich et al., Reference Goodrich, Wibisono, Miquelle, Lynam, Sanderson and Chapman2022). Following the extinctions of the Javan tiger Panthera tigris sondaica and the Bali tiger Panthera tigris balica, the Sumatran tiger Panthera tigris sumatrae is the only remaining subspecies of tiger endemic to Indonesia. Its numbers have declined significantly and populations have become increasingly fragmented (Luskin et al., Reference Luskin, Albert and Tobler2017a). Rapid land-use change in Sumatra reduced the extent of forests from c. 161,000 to 121,000 km² during 2001–2019 (Gaveau et al., Reference Gaveau, Locatelli, Salim, Husnayaen, Manurung and Descals2022), forcing tigers into human-dominated landscapes (Wibisono et al., Reference Wibisono, Linkie, Guillera-Arroita, Smith, Sunarto and Pusparini2011; Sunarto et al., Reference Sunarto, Kelly, Klenzendorf, Vaughan, Zulfahmi, Hutajulu and Parakkasi2013). The threat of tiger attacks on livestock and people has resulted in persecution of the tiger, threatening the conservation of this species (Nyhus & Tilson, Reference Nyhus and Tilson2004).

There are currently no publicly available, published datasets of human–tiger interactions in Sumatra. Earlier studies compiled information on human–tiger interactions in Sumatra, including interactions in the province of Riau, but data from 2010 onwards are needed (Nugraha & Sugardjito, Reference Nugraha and Sugardjito2009; Lubis et al., Reference Lubis, Pusparini, Prabowo, Marthy, Tarmizi, Andayani and Linkie2020). Riau used to be a stronghold for the Sumatran tiger and had one of the highest percentages of remaining forest cover (52%) in Sumatra (Nyhus & Tilson, Reference Nyhus and Tilson2004). But Riau was also one of the provinces with the highest number of fatal tiger attacks in Sumatra during 1978–1997 (Nyhus & Tilson, Reference Nyhus and Tilson2004). To address the lack of up-to-date information on human–tiger interactions in this province, which is being rapidly deforested, we compiled information on human–tiger interactions in Riau, and assessed where these events occurred, for 2010–2020. Our study area is the 89,913 km² province of Riau, which comprises 12 districts, 163 subdistricts and 1,737 villages. The major land-cover classes in 2017 were monoculture plantations (45% of the total land area), forests (18%), croplands (15%), shrubs (9%) and bare land (6%; Indonesian Ministry of Environment and Forestry, 2018b).

We collated data from online news reports for 2010–2020 using eight sources that reported news at the international, national and provincial levels (Supplementary Material 1, Supplementary Tables 1 & 2). For each news source we conducted a comprehensive search using the keywords harimau for tiger and konflik manusia dan harimau for human–tiger interactions, downloading all relevant articles published during 1 January 2010–31 December 2020. We conducted web scraping and text mining to derive this list of online news reports, using the packages rvest (Wickham & RStudio, Reference Wickham2021), stringdist (van der Loo et al., Reference van der Loo, van der Laan, Logan, Muir and Gruber2021), corpus (Huang et al., Reference Huang, Perry, Nielsen, Porter and Boulton2021) and tidyverse (Wickham et al., Reference Wickham, Averick, Bryan, Chang, McGowan and François2019) in R 4.1.2 (R Core Team, 2021).

We obtained reports of a total of 454 individual human–tiger interactions from the eight news sources. After removing duplicates, we had information on 101 human–tiger interaction events, from which we obtained a total of 107 interactions, which we categorized into seven interaction types (one interaction event could consist of more than one interaction type). We based our definition of negative human–tiger interaction types on that of Goodrich (Reference Goodrich2010) and the Indonesian Government's definition of human–tiger conflict in Ministry of Environment and Forestry Law No. P.48/Menhut-II/2008 (Kholis et al., Reference Kholis, Faisal, Widodo, Musabine, Hasiholan and Kartika2017). The latter defines human–tiger conflict as all interactions between people and tigers that result in negative effects on people, including their social life, economy, culture, and on the conservation of tigers or their environment. These include tigers present in human settlements, tigers attacking livestock, tigers attacking people, and tigers being hunted in retaliation. Our seven interaction types are: people injured, people killed, livestock killed, sightings of tigers, tigers killed, tigers injured and tigers evacuated (Supplementary Material 1, Supplementary Tables 1 & 2). For each interaction we recorded date of the article that reported the interaction, number of tigers involved, interaction type and location. We based the location of the interaction on the land use described in the report. When the reported number of tigers was unclear (e.g. ‘a number of tigers’ or ‘several tigers’) we recorded the number of tigers as ‘NA’. We mapped human–tiger interactions to village administrative boundaries (Badan Pusat Statistik, 2017), the lowest-level administrative boundary in Indonesia, using ArcGIS 10.8 (Esri, Redlands, USA).

We obtained records of 101 human–tiger interaction events in a total of 107 interactions published during 2010–2020 in 78 villages (4.5% of the total number of villages in the province) in Riau (Fig. 1). The majority of the interactions were tiger sightings (56), followed by attacks (19) in which tigers killed 34 animals. Seven tigers were reported killed, three injured and evacuated, and three otherwise evacuated. Eleven people were killed and nine people injured as a result of tiger attacks. Of the 101 interaction events, 77 reported the number of tigers involved, with most (81%) reporting a single tiger. The largest number of tigers reported in an interaction was five. Of the unique reports of human–tiger interactions, 100 (99%) reported the location of the interaction. The highest number of interactions occurred close to settlements (35%) followed by in plantations (26%) and smallholdings (25%; Table 1). Human–tiger interactions in the forest or at the forest edge comprised 14% of the reported events. Tiger sightings dominated human–tiger interactions in settlements, plantations and smallholdings. However, more interactions that resulted in people being injured or killed by tigers were reported in plantations and smallholdings (13) compared to settlements (three). The second most common interaction type reported in settlements and smallholdings was the killing of livestock.

Fig. 1 The distribution of 101 human–tiger interactions in the province of Riau, Indonesia, during 2010–2020 (Table 1). Deforestation data for 2011–2020 and forest cover data for 2020 are from the Indonesian Ministry of Environment and Forestry (2018a). Tanjung Simpang village is identified in the figure by an arrow. (Readers of the printed journal are referred to the online article for a colour version of this figure.)

Table 1 Locations and numbers of seven human–tiger interaction types in the province of Riau, Indonesia, during 2010–2020 (Figs 1 & 2). Multiple locations could be reported for each interaction type and more than one interaction could occur at a location. Hence the combination of locations and interaction types results in a total number of 121.

The mean annual number of documented human–tiger interactions increased from 4.6 during 2011–2017 to 21.3 during 2017–2020, an increase of more than fourfold (Fig. 2). The village with the highest number of human–tiger interactions (n = 13, c. 13% of the total) was Tanjung Simpang, Pelangrian subdistrict, Indragiri Hilir district (Fig. 1). Interactions in this village resulted in six human deaths, one human injury, one livestock death, two evacuations of problem tigers and three tiger sightings.

Fig. 2 Reports of the number of human–tiger interactions by type and per year in the province of Riau, Indonesia, published in eight news sources during 2010–2020 (Table 1).

The number of human–tiger interactions reported in our study was comparable to that reported previously around the Leuser Ecosystem (n = 148), which spans Aceh and North Sumatra, during 2008–2018 (Lubis et al., Reference Lubis, Pusparini, Prabowo, Marthy, Tarmizi, Andayani and Linkie2020). Our study and this previous study used a similar definition for human–tiger interactions. In Riau, the majority of human–tiger interactions occurred in settlements, plantations and smallholdings, indicating that human–tiger interactions occur with greater frequency at sites of high human activity and disturbance compared to in forests or at forest edges. This aligns with the results of similar studies elsewhere in Sumatra that found the probability of human–tiger interactions was higher in areas closer to settlements (Lubis et al., Reference Lubis, Pusparini, Prabowo, Marthy, Tarmizi, Andayani and Linkie2020) or in sites with intermediate disturbance such as multiple-use forests, and were low in sites with less disturbance such as protected areas (Nyhus & Tilson, Reference Nyhus and Tilson2004). The frequencies of human–tiger interactions in settlements, plantations and smallholdings in our study indicate there is a sharing of space and resources between people and tigers in Riau (Nyhus & Tilson, Reference Nyhus and Tilson2004; Maddox, Reference Maddox, Tilson and Nyhus2010). The expansion of plantations and associated settlements into previously forested areas results in overlap between human land uses and the distribution of tigers (Wibisono et al., Reference Wibisono, Linkie, Guillera-Arroita, Smith, Sunarto and Pusparini2011). Although plantations are often dominated by monocultures such as oil palm, they can harbour generalist wildlife species such as wild pigs that are key prey species for tigers (Luskin et al., Reference Luskin, Brashares, Ickes, Sun, Fletcher, Wright and Potts2017b). Settlements may be sites with high numbers of human–tiger interactions because of the increased numbers of people in these areas, and potentially the availability of livestock as prey when wild prey populations have been depleted (Lubis et al., Reference Lubis, Pusparini, Prabowo, Marthy, Tarmizi, Andayani and Linkie2020).

Web scraping of online news sources is a cost-effective way to gather spatio-temporal information on human–wildlife interactions (Egri et al., Reference Egri, Han, Ma and Chakraborty2021). However, the limitations of this approach include the inconsistent ways in which human–tiger interactions are reported across news sources, the bias in reporting that could arise as a result of more people or better telecommunication access, and some interaction types being considered more newsworthy than others (e.g. tiger sightings may be reported less than tiger evacuations). Given that our findings align with results from previous research (Lubis et al., Reference Lubis, Pusparini, Prabowo, Marthy, Tarmizi, Andayani and Linkie2020), we nevertheless believe that our approach is useful and relevant to developing a spatio-temporal database of human–wildlife interactions.

The Sumatran tiger has been the subject of many conservation efforts (IUCN, 2021). It is important to document and monitor human–tiger interaction events to ensure these efforts do not unduly burden communities that live close to tigers (Nyhus et al., Reference Nyhus, Osofsky, Ferraro, Madden, Fischer, Woodroffe, Thirgood and Rabinowitz2005), and so that the recovery of tiger populations is not undermined by retaliatory killings for the loss of livestock or of human lives (Inskip & Zimmermann, Reference Inskip and Zimmermann2009). Although the greater number of human–tiger interactions in Riau during 2017–2020 comprises mostly tiger sightings, this is still a cause for concern given that every year at least one person is killed or injured and one tiger is killed, injured or evacuated. As with previous research (Lubis et al., Reference Lubis, Pusparini, Prabowo, Marthy, Tarmizi, Andayani and Linkie2020), we found that settlements, plantations and smallholdings are sites that should be targeted for prevention and mitigation efforts to reduce conflicts between people and tigers. These efforts include establishing conflict mitigation teams, improving animal husbandry practices, and providing training and education on human–tiger interactions. Our mapping of human–tiger interactions could contribute towards spatial prioritization for such mitigation and prevention efforts. Assessing the biophysical variables that drive human–tiger interactions and improving social tolerance for tigers in Riau (Struebig et al., Reference Struebig, Linkie, Deere, Martyr, Millyanawati and Faulkner2018) are important future steps towards understanding and mitigating these events.