Index‐based large carnivore population density and abundance estimates for the Ruaha‐Rungwa conservation complex in Tanzania

• DOI: 10.1111/aje.13092
• Publication Date: 2022-12-02T10:38:51Z
• Authors: Paolo Strampelli; Philipp Henschel; Charlotte E. Searle; Alex Loiruk Lobora; Halima Kiwango; David W. Macdonald; Amy J. Dickman
• Keywords: Crocuta crocuta; Carnivore; Acinonyx jubatus; Game reserve

The Ruaha-Rungwa conservation landscape is a ~45,000 km2 Protected Area (PA) complex in south-central Tanzania. area believed to harbour globally important populations of lion (Panthera leo; Bauer et al., 2016), cheetah (Acinonyx jubatus; Durant 2017), and African wild dog (Lycaon pictus; Woodroffe & Sillero-Zubiri, 2020), as well leopard pardus), spotted hyaena (Crocuta crocuta), striped (Hyaena hyaena) (TAWIRI, 2009). However, no empirical estimates landscape-wide population abundance exist for these populations, with the only available being based on expert opinion or extrapolations from densities elsewhere (Mesochina 2010; Riggio 2013). Here, we present landscape-level density five large carnivore species Ruaha-Rungwa. We calculated an index spoor (track) data, while acknowledging technique's low precision (as revealed by recent research) providing recommendations future monitoring. includes Ruaha National Park (NP), one largest NPs eastern Africa at 20,226 km2, which used photographic tourism; three Game Reserves (GRs – Rungwa, 9175 km2; Kizigo, 5140 Muhesi, 2720 km2), where trophy hunting tourism primary revenue generation mechanism; number multiple-use areas, including Lunda-Mkwambi Controlled (GCA), Rungwa South Open (OA), MBOMIPA Waga Wildlife Management Areas (WMAs) (Figure 1). Vegetation cover primarily comprises mosaic Acacia-Commiphora open savannah/bushland central Zambezian Eastern Brachystegia-dominated miombo woodland, complemented riverine forests floodplain grasslands (Olson 2001). carried out vehicle-based surveys over two dry seasons, between July November 2017 June 2018. study was divided into 225-km2 grid cells, minimum 6 km maximum 20 were surveyed within all cells sufficient road access (Henschel 2020). transects sites this possible. For data collection, employed protocols presented Henschel al. (2020). counted fresh track observations each transect standardised ‘track densities’ (Table 1), equivalent tracks per 100 (Funston 2010). Population then estimated using general equation developed Winterbach (2016), modified Funston (2010). confidence intervals (CIs) method (2017) Dröge (2020); although results wider than (2010), latter has been shown overestimate (Dröge As result ongoing land disputes, portion southwestern NP (~2400 km2) hosts permanent settlements industrial agriculture activities found be largely devoid wildlife Transects excluded analyses, refer without such (17,827 km2). total 2393 transects. Overall, ~85% sampling landscape, ensuring similar both vegetation types, regard distance rivers, PA boundaries type, other natural anthropogenic features detected 151 independent lion, 149 leopard, 9 cheetah, 54 dog, 572 hyaena. Table 2 presents overall PA-specific (where possible) estimates. See Appendix S1 detection maps carnivores, S2 calculations. Our indicate that home hyaena, dog. Although exhibit very precision, detections suggest relatively sparse 2). Results confirm suggesting Great Mzombe river valleys are particularly season (Appendix S1). findings also what likely continent's remaining dogs, standard errors estimate especially high due variability transects, should kept mind when interpreting results. GR appears species, possibly consisting woodlands, have suggested suitable (Creel, Spotted highest complex, comparable those studies employing same methodology (Bauer 2015; Finally, appreciate research 2020) showed lower previously thought. Indeed, CIs through amended formula (2020) 2) unlikely monitor changes time. Nevertheless, believe there value sharing our estimates, given complete lack landscape-scale Ruaha-Rungwa's carnivores. Going forward, however, recommend exploring instead scalability methods can provide greater levels spatially explicit capture-recapture (SECR) models applied camera traps (e.g., Strampelli, Henschel, Dickman 2022) direct sightings Elliot Gopalaswamy, typically smaller areas. Alternatively, proves challenging too resource-intensive, considering estimation alternative robust status parameters occupancy; Searle scales. Fieldwork under permits 2017-210-NA-2017-107 2018-367-NA-2017-107, granted Tanzania Commission Science Technology (COSTECH) Research Institute (TAWIRI). would like thank Government Tanzania, TAWIRI, Parks Authority (TANAPA), (TAWA), Idodi-Pawaga WMA their support research. TANAPA Rangers TAWA Scouts, Mr. Stivin G. Pangamwene Hamis S. Dongo, assistance during fieldwork. Mdonya Old River Camp Essential Destinations, Ikuka Safari Nomad Ruaha, Big Safaris Ikiri. Scholarship funding PS, some fieldwork costs, provided University Oxford's NERC Environmental DTP, Queen's College Oxford. Additional grants Geographic Society, Columbus Zoological Association, Inc., Chicago Panthera Royal Geographical Society (with IBG), who generosity. AD funded Recanati-Kaplan Fellowship. authors conflicts interest declare. Data freely accessed at: https://github.com/pstrampelli/RuahaRungwaLargeCarnivoreDensities. AppendixS1 AppendixS2 Please note: publisher not responsible content functionality any supporting information supplied authors. Any queries (other missing content) directed corresponding author article.