key: cord-0838625-s8xje1mu
authors: Muylaert, Renata L.; Kingston, Tigga; Luo, Jinhong; Vancine, Maurício Humberto; Galli, Nikolas; Carlson, Colin J.; John, Reju Sam; Rulli, Maria Cristina; Hayman, David T. S.
title: Present and future distribution of bat hosts of sarbecoviruses: implications for conservation and public health
date: 2021-12-13
journal: bioRxiv
DOI: 10.1101/2021.12.09.471691
sha: f08e0411397e6977b2ac0a56af6e51719e127627
doc_id: 838625
cord_uid: s8xje1mu

Global changes in response to human encroachment into natural habitats and carbon emissions are driving the biodiversity extinction crisis and increasing disease emergence risk. Host distributions are one critical component to identify areas at risk of spillover, and bats act as reservoirs of diverse viruses. We developed a reproducible ecological niche modelling pipeline for bat hosts of SARS-like viruses (subgenus Sarbecovirus), given that since SARS-CoV-2 emergence several closely-related viruses have been discovered and sarbecovirus-host interactions have gained attention. We assess sampling biases and model bats’ current distributions based on climate and landscape relationships and project future scenarios. The most important predictors of species distribution were temperature seasonality and cave availability. We identified concentrated host hotspots in Myanmar and projected range contractions for most species by 2100. Our projections indicate hotspots will shift east in Southeast Asia in >2 °C hotter locations in a fossil-fueled development future. Hotspot shifts have implications for conservation and public health, as loss of population connectivity can lead to local extinctions, and remaining hotspots may concentrate near human populations.

Major current and future global changes pose a severe risk to biodiversity and human 48 survival [1] . Global climate change and human encroachment into natural habitats is 49 simultaneously driving the biodiversity extinction crisis and increasing disease emergence risk 50 [2] . Climate and land cover change will alter the range distribution of species [3] , an important, 51 poorly defined predictor of zoonotic (animal to human) disease risk [4, 5] , and the direction and 52 magnitude of range shifts are not estimated for many species, leaving the impacts on their 53 viral interactions uncertain [6, 7] .

In early 2020, genomic analysis identified the severe acute respiratory syndrome 

The rapid increase in bat data after the COVID-19 pandemic provides opportunities to 89 better understand bats' distributional ecology, but may bring sampling biases in areas where 90 surveillance has been greatest [11, 29, 30] . Avoiding misprediction is essential, and we have 91 the opportunity to update ecological niche models with the help of big data, reproducible tools 92 and open science. We need adequate inferences regarding bat species distributions from the 93 current period projected to proximate future scenarios, so we can establish guidelines for how 94 to transition from the current trajectory of biodiversity loss and pandemic risk to a more 95 sustainable future [1] .

Here, we use ecological niche models to assess the potential distribution of bat hosts we compared all projected ranges within the accessible area in future scenarios as being the 169 maximum limit for dispersal. We define range as the area where the species most likely occurs 170 (estimated occupied area) driven by the environmental covariates used.

We defined environmental variables that are important drivers of target species 173 distributions in the current geographic space, which can also be projected into the future.

Based on the focal species' ecology, we chose selected climatic (annual precipitation, 175 precipitation seasonality, annual mean temperature, temperature seasonality) and landscape 176 variables (karst and primary forest cover) as covariates (Table S3 ).

Habitats used by each species were extracted using rredlist v0.7 ( Figure S2 ). We 

Sarbecoviruses were reported from 35 bat species (Table S1 ). We could model 17 240 species using IUCN-intersected data, and 23 with non-intersected data ( Figure S4 ). From the 241 23 species, six could only be modelled without intersecting their points with IUCN data; nine 242 species did not show improvement in TSS values after intersecting them with IUCN ranges, 243 while eight had small improvements after cropping occurrences within IUCN range limits 244 (Table S4) .

The maps show three focal areas of suitability across species; one each in Western 246 Europe, Indochina, and Central Africa ( Figure S5 , S6). IUCN delimitation for occurrence 247 inclusion does not improve model performance for more than 10% added value in TSS in most 248 cases (Table S4 ). Richness maps for the two datasets were highly positively correlated (|r| 

The highest number of bat species (i.e. host hotspots) in the present and future 258 projections occurred in Southeast Asia ( Figure 1 ; Table 1 in Southeast Asia are also well sampled, especially eastern coastal areas (purple in Figure 2 ).

Highest values for richness were estimated for areas with low sampling rates ( Figure S11 ). 

Highest host richness values decline in future SSP585 projections (Figure 3 ).

Average temperature of current hotspots -the few where 13 species are present -is 20.6 °C, Potential ranges for all periods, scenarios, and GCMs used are in 

However, some species with large ranges ( Figure S12 ), such as Rhinolophus 306 ferrumequinum, and Rhinolophus affinis, are predicted to suffer range contractions (Table   307 S5, Table S6 ). Population trend data showed that many species do not have a current 308 evaluation (Table S7) . Considering the most extreme global warming scenario (SSP585), 309 most species will suffer range contractions (N=17, 74%), while six may gain area (N=6), with 310 species overlap decreasing ( Figure S13 , Table S6 ). For SSP245, fewer, but still more than 311 half the species will suffer range contraction (N=14, 61%). In terms of total area overlaps by (Table S5) . 

Our focal species are insectivorous bats with varying geographical ranges and 332 sensitivity to habitat disturbance [58] . Species responses to climate change can be complex 333 [59, 60] . Though some species are resilient ( Figure S12) 

We chose a pessimistic carbon emissions SSP585 scenario as an example here, but 355 it is a likely scenario [64] and considered a possible future, though less likely according to 356 recent reports [65] . However, there is high convergence between SSP245 and SSP585 357 hotspot projections (Figure 3) , though SSP585 concentrates more species. In fact SSP245 is,

to some extent, less extreme with fewer range contractions, whereas species are projected to 359 become more spatially concentrated, especially in SSP585, probably due to a refugia effect 360 [66] since there will be less suitable habitat. We predict slight range gains for 2015-2040, 

More intensive sampling in species-rich effort gap areas can reduce biases. Despite these 375 challenges, after data curation to reduce sampling bias and autocorrelation, we could still 376 model most species but identified important classic shortfalls [74] . The smallest-ranged bats 377 in our dataset did not reach our modelling criteria ( Figure S7 ) because of data gaps, and these China hotspots (areas highlighted in red, Figure 2 ) [77] .

There are conservation implications of our findings. Range contractions are predicted 389 for several species, even for species using variable habitats, such as Rhinolophus pearsonii 390 (Table S4 , Figure S5 ). Most species' populations are currently declining (N=8) or have 391 unknown population trends (N=20 , Table S6 ). Local species loss predictions were almost 

Our results identify broad regions where bats reported positive for sarbecoviruses most 398 probably occur and co-occur. These hotspots coincide, but are not restricted only to

Rhinolophidae diversity hotspots previously reported [81] and to hotspots of mammal 400 vulnerability to climate change [82] . Projections suggest that hundreds of new future viral 401 sharing events may occur in Southeast Asia [7] . Novel interactions may be of concern for 402 species survival as pathogens could spread more easily in vulnerable wild populations, which 

Our future projections assume models using present data will perform adequately.

However, our models do not account for biotic components that also interfere with suitability, 413 so we are limited to inferences of distribution derived from landscape and climate drivers. 

IPBES. 2020 Workshop Report on Biodiversity and Pandemics of the Intergovernmental 445 Platform on Biodiversity and Ecosystem Services

Global hotspots and correlates of emerging zoonotic 448 diseases

2020 A review of the major threats and challenges to 450 global bat conservation

Host 452 and viral traits predict zoonotic spillover from mammals

The future of zoonotic risk prediction

Bat Research Networks and Viral Surveillance: Gaps and 457 Opportunities in Western Asia

Climate change will drive novel cross-species viral transmission

A pneumonia outbreak associated with a new coronavirus of 462 probable bat origin

Coronaviruses with a SARS-CoV-2-like receptor-binding domain 464 allowing ACE2-mediated entry into human cells isolated from bats of Indochinese 465 peninsula

Predicting wildlife hosts of betacoronaviruses for SARS-CoV-2 467 sampling prioritization. bioRxiv

2021 Land-use change and the livestock 469 revolution increase the risk of zoonotic coronavirus transmission from rhinolophid bats

472 2021 A strategy to assess spillover risk of bat SARS-related coronaviruses in Southeast 473

2021 Bat Species of the World: A taxonomic and 475 geographic database. Bat Species of the World: A taxonomic and geographic database

Bats as Viral Reservoirs

2020 Bat-borne virus 480 diversity, spillover and emergence

No silver bullets in correlative ecological niche 564 modelling: Insights from testing among many potential algorithms for niche estimation

The Predictive 567 Performance and Stability of Six Species Distribution Models

Evaluating resource 570 selection functions

Prevalence, thresholds and 572 the performance of presence-absence models

Selecting thresholds for the prediction of species 574 occurrence with presence-only data

Climate Change 2021: The Physical Science Basis

Parasite and host assemblages: 579 embracing the reality will improve our knowledge of parasite transmission and virulence

The science of the host-virus network

Parasites in food webs: the ultimate missing links

Microchiropteran bats: global status 586 survey and conservation action plan

Species Associated with High Levels of Climate Warming

2021 Drivers of change in the realised 591 climatic niche of terrestrial mammals

Habitat Suitability 593 Evaluation of the Chinese Horseshoe Bat ( R. sinicus ) in the Wuling Mountain Area 594 Based on MAXENT Modelling

Surge in global metal mining threatens vulnerable ecosystems

The projected 598 effects of climatic and vegetation changes on the distribution and diversity of Southeast 599 Asian bats

RCP8.5 tracks cumulative CO2emissions. 601 Proceedings of the National Academy of Sciences

2020 Emissions -the 'business as usual' story is misleading

Managing Climate Change Refugia for Climate Adaptation. PLoS 606 One 11

The future of Southeast Asia's forests

Challenges and considerations of applying nature-based 610 solutions in low-and middle-income countries in Southeast and East Asia. Blue-Green 611 Systems

Geographical Variation of Rhinolophus affinis

Rhinolophidae) in the Sundaic Subregion of Southeast Asia, including the Malay 614 Peninsula, Borneo and Sumatra

Rhinolophidae): expanded geographic and taxonomic sampling of the Afrotropics

Variation of 620 mitochondrial DNA in the Hipposideros caffer complex (Chiroptera: Hipposideridae) and 621 its taxonomic implications

Carlson 623 CJ. 2021 Mammal virus diversity estimates are unstable due to accelerating discovery 624 effort. bioRxiv

A comprehensive survey of bat sarbecoviruses across China for the 626 origin tracing of SARS-CoV and SARS-CoV-2

Seven 629 Shortfalls that Beset Large-Scale Knowledge of Biodiversity

Want to model a species niche? A step-by-step guideline on correlative ecological 632 niche modelling

Ranking the risk of animal-to-human spillover for newly 634 discovered viruses

Ecology, evolution and spillover of coronaviruses from bats

Individual dispersal, 638 landscape connectivity and ecological networks

Landscape resistance influences effective dispersal of endangered golden lion 641 tamarins within the Atlantic Forest

Modelling the functional connectivity of landscapes for greater horseshoe bats 644

Rhinolophus ferrumequinum at a local scale

Tracing the origins of SARS-CoV-2: lessons learned from the past

A framework for the identification of hotspots of 649 climate change risk for mammals

Setting the Terms for Zoonotic Diseases: Effective 653 Communication for Research, Conservation, and Public Policy

Serological Evidence of Bat SARS-Related Coronavirus Infection in 656

SARS-related Virus Predating SARS Outbreak, Hong Kong. Emerg. Infect. Dis. 10, 176. 659 87. WHO. 2021 WHO-convened global study of origins of SARS-CoV-2: China Part. See

2020 Ecosystem perspectives are 663 needed to manage zoonotic risks in a changing climate

2021 Does public 666 fear that bats spread COVID-19 jeopardize bat conservation?

The role of land use and land cover change in climate 670 change vulnerability assessments of biodiversity: a systematic review

Understanding Risk Perceptions to Enhance Communication about 674

Human-Wildlife Interactions and the Impacts of Zoonotic Disease