sharing for science: high-resolution trophic interactions ......facebook alone has more than 2.45...

Submitted 26 March 2020Accepted 15 June 2020Published 9 July 2020

Corresponding authorRobin A Maritzmaritzrobinagmailcom

Academic editorPetteri Muukkonen

Additional Information andDeclarations can be found onpage 16

DOI 107717peerj9485

Copyright2020 Maritz and Maritz

Distributed underCreative Commons CC-BY 40

OPEN ACCESS

Sharing for science high-resolutiontrophic interactions revealed rapidly bysocial mediaRobin A Maritz and Bryan MaritzDepartment of Biodiversity and Conservation Biology University of the Western Cape Bellville South Africa

ABSTRACTDiscrete ephemeral natural phenomena with low spatial or temporal predictability areincredibly challenging to study systematically In ecology species interactions whichconstitute the functional backbone of ecological communities can be notoriouslydifficult to characterise especially when taxa are inconspicuous and the interactionsof interest (eg trophic events) occur infrequently rapidly or variably in space andtime Overcoming such issues has historically required significant time and resourceinvestment to collect sufficient data precluding the answering of many ecologicaland evolutionary questions Here we show the utility of social media for rapidlycollecting observations of ephemeral ecological phenomena with low spatial andtemporal predictability by using a Facebook group dedicated to collecting predationevents involving reptiles and amphibians in sub-Saharan Africa We collected over1900 independent feeding observations using Facebook from 2015 to 2019 involving 83families of predators and 129 families of prey Feeding events by snakes were particularlywell-represented with close to 1100 feeding observations recorded Relative to anextensive literature review spanning 226 sources and 138 years we found that socialmedia has provided snake dietary records faster than ever before in history with preybeing identified to a finer taxonomic resolution and showing only modest concordancewith the literature due to the number of novel interactions that were detected Finallywe demonstrate that social media can outperform other citizen science image-basedapproaches (iNaturalist and Google Images) highlighting the versatility of social mediaand its ability to function as a citizen science platform

Subjects Biodiversity Conservation Biology Ecology ZoologyKeywords Citizen science Facebook Social media Crowd-sourcing Feeding interactionsTrophic ecology Natural history Herpetofauna Snake ecology Southern Africa

INTRODUCTIONMany ecological processes exist as the product of a large number of discrete ephemeralevents At fine spatial and temporal scales these events are often difficult to predict makingthem challenging to study systematically This challenge is particularly true for interspecificbiological interactions and is magnified when one or both interacting species are difficultto detect with important impacts on our understanding of the ecology of many systemsSuch challenges can be overcome with large investments of time andmoney but these costscan be prohibitive and are likely part of the reason for the remarkable absence of empirical

How to cite this article Maritz RA Maritz B 2020 Sharing for science high-resolution trophic interactions revealed rapidly by socialmedia PeerJ 8e9485 httpdoiorg107717peerj9485

datasets characterising species interactions in ecosystems (McCann 2007 Hegland et al2010 Jordano 2016)

Together the origin and development of social media and the concurrent advances inaccess to mobile cameras represent a disruptive innovation that has changed the mannerand rate at which modern events are recorded and communicated With over 380 billionpeople using social media worldwide (Kemp 2020) the synergy of social media and readilyaccessible mobile cameras has increased the observational effort of researchers by ordersof magnitude Facebook alone has more than 245 billion monthly active users (Facebook2019) making it the digital platform with the largest social networking potential (Kemp2020) Harnessing this power has far-reaching implications for understanding ecologicaland evolutionary processes characterised by difficult to detect discrete transient eventsthrough the resultant increase in observation coverage and depth

Trophic interactions defined as interspecific interactions in which one organismconsumes another form the basis for understanding processes and system characteristicsas diverse as energy flow population dynamics food web dynamics and the evolution ofbehavioural morphological and physiological adaptations by predators and prey (Garveyamp Whiles 2017) Moreover with a world experiencing climatic changes and worseningenvironmental conditions (Vitousek et al 1997) attention to species interactions willbe crucial for understanding ecosystem function and integrity (Tylianakis et al 2008Valiente-Banuet et al 2014) Despite their central position in ecological and evolutionarytheory the characterisation of trophic interactions between species and within food websparticularly those in which such interactions are difficult to study are often incompleteor absent (Paine 1988 Chacoff et al 2011 Miranda Parrini amp Dalerum 2013 Jordano2016) Moreover because certain organismal traits can reduce the detection likelihood of agiven trophic interaction trophic interactions involving terrestrial vertebrates (particularlynon-herbivorous interactions) are underrepresented in the literature (Miranda Parrini ampDalerum 2013) In cases where such datasets exist endotherms tend to be better representedthan ectotherms in trophic studies (Miranda Parrini amp Dalerum 2013)mdashpossibly becauseof ease of sampling or because endothermy often demands higher food intake ratesFinally of the interactions reported organisms involved in lower-trophic-level interactionsoften suffer from taxonomic aggregation (Polis 1991) which can mask complex speciesinteractions and influence metrics associated with food webs community assemblagesand interspecific competition (Greene amp Jaksić 1983 Paine 1988 Thompson amp Townsend2000) Thus a method for improving the quantity and quality of collected trophic data iswarranted and essential

Together reptiles and amphibians (hereafter herpetofauna) include more than 18000ectothermic vertebrate species and account for more than half of all global tetrapoddiversity (Pincheira-Donoso et al 2013) In many terrestrial ecosystems these animalscan make up a large proportion of the total abundance of vertebrates and contributesignificantly to the total biomass of a region (Western 1974 Iverson 1982 Jacobsen 1982Petranka amp Murray 2001) Moreover herpetofauna (mostly amphibians and squamates)often occupy intermediate trophic levels providing important trophic links between small-bodied invertebrate primary consumers and higher trophic levels occupied primarily by

Maritz and Maritz (2020) PeerJ DOI 107717peerj9485 221

endothermic predators (eg Polis 1991) Interestingly snakes a monophyletic lineage ofmore than 3700 species (approximately 10 of global tetrapod diversity Pincheira-Donosoet al 2013) are exclusively carnivorous and potentially occupy intermediate trophicpositions between many other herpetofauna and species residing at higher trophic levels(FitzSimons 1962 Greene 1997) However many species of herpetofauna are notoriouslydifficult to detect and observe in the wild (Steen 2010 Durso Willson amp Winne 2011Durso amp Seigel 2015 Lardner et al 2015 Rodda et al 2015) and individuals of manyspecies feed infrequently or discreetly (Greene 1997) making the systematic observationand quantification of trophic interactions incredibly challenging

In this article we demonstrate the utility of a method that uses social media specificallya dedicated Facebook group to collect images and videos of difficult to detect feedinginteractions involving herpetofauna in sub-Saharan Africa We hypothesised thatinformation regarding ecological phenomena specifically trophic interactions can becollected at large spatial scales and across diverse taxonomic clades by employing theassistance of a large network of potential observers (ie Facebook users) First wehighlight the remarkable diversity of predator and prey interactions identified over afive-year period and provide an overview of observer statistics Next because snake feedingevents are well-represented in our dataset and are notoriously difficult to observe in thewild we tested the hypothesis that an increased quantity of data on snake feeding wouldresult in the detection of novel species interactions either due to increased sampling effortor differences in detectability Finally we compared our dataset to data collected fromiNaturalist and Google Images to test whether other digital media platforms offer theobservational power required to detect difficult to record trophic interactions Togetherour findings emphasise that previous methods have left gaps in our understanding offeeding interactions involving southern African herpetofauna and to gain a more holisticunderstanding data collection must incorporate multiple lines of inquiry Ultimately ourapproach highlights the application of Facebook to rapidly improve trophic interactionsampling coverage and depth in many ecosystems and act as a model for utilising socialmedia to study rare and difficult to detect ecological events

MATERIALS amp METHODSFacebook data collectionThe idea for a dedicated predation records Facebook group arose amongst a small groupof South African reptile researchers and enthusiasts after noticing feeding observationsbeing shared across the platform and recognising the scientific value in having a placeto store and later record these observations Since then we have administrated andcurated the Predation Records - Reptiles and Frogs (Sub-Saharan Africa) Facebook group(httpswwwfacebookcomgroups888525291183325) from its creation in August 2015until December 2019 We requested that members include details such as predator andprey identity location date time and observer or photographerrsquos name when sharingan observation to the group When information was missing administrators or groupmembers asked for the post to be updated with the necessary details Predator and prey


identities were confirmed to the finest taxonomic-level possible using a combination oflocality information and key physical characteristics and with support from taxon expertgroup members In challenging cases persons with taxon-expertise were consulted usingFacebook or via email Observations that appeared on other social media groups wereincorporated in an ad hoc manner

Literature data collectionWe performed an extensive review of diet records for snake species in southern Africansnakes (the region where most Facebook observations occurred) We searched primaryand grey literature sources (museum bulletins society newsletters and bulletins wildlifemagazines and non-indexed journals) for substantiated feeding records Searches wereconducted in English and themain platforms usedwere Google Scholar and the BiodiversityHeritage Library Interactions published without supporting details (eg field guidedescriptions) were categorised as secondary records and were not included in our finalanalyses In all instances prey identity was recorded with modification based on updatedtaxonomy In instances where only a generic name was provided the most representativetaxonomic name was assigned based on geographic location Feeding interactions in whichmultiple prey items of the same type were ingested at once (eg lsquothree nestling chicksrsquo)were treated as a single record in the database Captive-fed observations were recorded butexcluded from this study A list of literature sources used (N = 226) and the snake specieswhich they provide data for can be found in the supporting information (Table S1)

Data management and curationData were recordedmanually and kept in local storage withmonthly back-ups to a personalcloud storage service The data files are accessible via Figshare (106084m9figshare11920128) Images and videos from all Facebook posts have been downloaded in case usersdelete or change the privacy settings on their uploaded media For each feeding interactionwe recorded predatorprey identity predatorprey life stage direction of ingestion (forsnake predators) interaction specifics (date time location) and any noteworthy detailsTaxonomic hierarchies were updated automatically for each predator and prey itemby referencing a local hierarchy database with information obtained from biodiversitydatabases (reptile-databaseorg sabap2aduorgza amphibiaweborg gbiforg)

For Facebook records additional information included microhabitat (eg treeshrubartificial surface) type of interaction (true predation or scavenging) type of event (eg insitu roadkill capturedndashregurgitated) share date person who shared the record person(s)who observed the record and post permalink Duplicates were excluded in a semi-automated manner using a photo comparison program (Duplicate Photo Cleaner v47WebMinds Inc) Additionally records were flagged and verified whenever an identicalcombination of predator prey and observer arose

For literature records additional information included predator and prey snout-vent-length predator and prey mass type of study (eg incidental museum) museum vouchernumbers (when available) and reference We treated any record in which a given authorhad published the same interaction previously and did not provide any information onlocality or date along with the most recent account as a duplicate


Data collection from other digital media sourcesWe retrieved relevant observations from the iNaturalist citizen science platform(iNaturalistorg) in December 2019 These included all records shared on the iSpotplatform (ispotorgza) for southern Africa that were migrated to the iNaturalistplatform during 2017 Currently there is no centralised method for reporting speciesinteractions on iNaturalist but a pre-existing iNaturalist project lsquoInteractions (s Afr)rsquo(inaturalistorgprojectsinteractions-s-afr) gathers feeding interaction data using theobservation field lsquolsquoEating (Interaction)rsquorsquo which we used to query (lsquolsquoampfieldEating(Interaction)=rsquorsquo) and retrieve all snake feeding records in southern Africa logged ontothe platform (N = 77) Uncatalogued observations were located using the followingindependent queries lsquofeedingrsquo lsquoeatingrsquo lsquomealrsquo lsquopredationrsquo lsquoswallowrsquo and lsquopreyrsquo specieswas set to lsquoSerpentesrsquo and location used was lsquosouthern Africarsquo (N = 25) Records wereexported using the download observations function Duplicate interactions were identifiedbased on iNaturalist observation numbers The crowd-sourced identification was usedwhen available Wemanually inspected images of the target species (the four most observedspecies in the Facebook dataset) including brown-house snake (Boaedon capensis) southernAfrican python (Python natalensis) boomslang (Dispholidus typus) and cape cobra (Najanivea) for additional instances of feeding that had been missed Prey could not be identifiedin nine of the records because the item had already been fully ingested (ie food bulgevisible)

Google Images results were retrieved in October 2019 Searches were performed for eachof the four target study species using the following query lsquolsquo(lsquolsquoscientific namersquorsquo | lsquolsquocommonnamersquorsquo) (eating | prey | predation | swallow | meal | feeding)rsquorsquo and all resulting imageswere inspected for evidence of feeding Only images of wild feeding observations wererecorded Photos documenting the same encounter were excluded manually Observationsderived from Google Images were more coarsely identified as the geographic locationwas frequently missing but prey were identified to the finest taxonomic-level wheneverpossible

Analytical approaches and comparisonsAll data manipulations graphical outputs and statistical analyses were conducted in Rversion 361 Code can be accessed via Figshare (106084m9figshare12287714) In allanalyses non-southern African snake species were excluded To compare the accumulationrates of Facebook and literature records the number of records in a given year was averagedwith the previous n years in 2ndash15ndashyear windows (moving average) To assess interactionnovelty duplicate interactions within the dataset for each approach were removed Thenfor interactions in which the prey was identified to the species-level (repeated at eachtaxonomic level) the presence of a given predatorndashprey interaction was assigned to eitherliterature digital media source or both (shared) To test for discordance between theFacebook and literature dataset Cohenrsquos kappa coefficient (κ) was calculated using thenumber of unique and shared interactions (repeated at each taxonomic level) (Cohen1960) For the comparison of prey-ratios derived from digital media sources prey itemscategorised as lsquolarge mammalsrsquo are species that typically exceed five kilograms


RESULTSBetween 2015 and 2019 we gathered a total of 1917 trophic interactions involvingherpetofauna using images and videos shared to the Predation Records - Reptiles and Frogs(Sub-Saharan Africa) Facebook group (Figs 1Andash1L) We detected trophic interactionsbetween 83 families of predators (across 30 orders and 9 classes) and 129 families ofprey (across 51 orders and 14 classes) (Figs 1A 1G Fig S1) Our data encompassesobservations from 18 African countries However most feeding interactions were observedwithin South Africa (755N = 1446) which is reflective of a geographic bias in Facebookgroup participation Observations were dominated by predation events involving reptilesas the predator representing 660 (N = 1266) of all trophic interactions in our dataset(Fig 1A) Remarkably snakes accounted for the majority of these observations (858N = 1086) We detected feeding events by 85 species of snakes including five of the eightfamilies that occur in Africa

In our study therewere at least 1369 unique observerswhouploadedmedia documentinga predation event However the actual number of observers may be larger because 48records did not explicitly state who photographed the event On average observers shared144 instances of predation (SD = 188 range = 32) Across all observations (ie anyclass of predator) 820 of observers reported only a single record whereas 180of observers reported at least two predation events Only 095 of observers reportedmore than 10 predation events There were 891 unique observers who had documented apredation event in which the predator was a snake with an average of 124 observations perperson (SD = 123 range = 19) For instances of snakes as predators 895 of observershad one observation and only 045 of observers have reported more than 10 snakepredation events Notably the records posted by the top four observers were dominated byobservations of prey in road-killed specimens or from snakes that regurgitated prey itemswhile being translocated following a request for the snake to be removed (ie to minimisehuman-wildlife conflict)

Our extensive literature review of southern African snake diets revealed a total of 2884feeding records covering 109 of southern Africarsquos 168 species collected over a period ofmore than 130 years (Fig 2) Contrastingly in five years wewere able to collect 1066 feedingrecords which equates to 270 of all documented observations When unsubstantiatedrecords are included as a conservative measure our observations from Facebook accountfor 243 of all feeding records

Overall feeding observations accrued at a significantly faster rate (Welchrsquos t-testt =minus394 p= 00163) by utilising Facebook (micro= 213 records yrminus1) compared to historicalcollection and reporting approaches (micro= 209 records yrminus1) To account for gaps inreporting we conducted a moving average analysis to test if there were any time periodsthat produced comparable rates of data accumulation There were no five-year periodsthat approached or exceeded the accumulation rate observed using Facebook The mostcomparable period was between 2006ndash2010 (micro = 139 records yrminus1) which produced 696records Only after expanding the time frame to a 10-year window did the number ofaccumulated records from the literature (N = 1187) exceed the number of records that


Reptilia Amphibia Aves

Mammalia Actinopterygii Insecta

Gastropoda

Other

Prey of Reptiles amp AmphibiansA

J

Predators of Reptiles amp Amphibians


Mammalia OtherArachnida

G

D

C

L

I

F

B

K

H

E

Figure 1 Diversity of predatorndashprey interactions detected using Facebook (A) Number of reptileand amphibian feeding records grouped by prey class Each square represents 10 records Prey classeswithlt10 records are included within lsquoOtherrsquo (Arachnida Chilopoda Clitellata Diplopoda LiliopsidaMagnoliopsida and Malacostraca) (BndashF) Examples of in situ feeding interactions with reptiles andamphibians as predators collected from Facebook (B) Cape cobra (Naja nivea)ndashTubercled gecko(Chondrodactylus sp) Photo credit Michele-Ann Nel (C) Brown house snake (Boaedon capensis)ndashAfrican yellow bat (Scotophilus dinganii) Photo credit Nick van de Wiel (D) Cape river frog (Amietiafuscigula)ndashCape white-eye (Zosterops virens) Photo credit Basil Boer (E) Waterberg flat lizard (Platysaurusminor)ndashAfrican thief ant (Carebara vidua) Photo credit Kobus Pienaar (F) Rhombic egg-eater (Dasypeltisscabra)ndashIndian peafowl (Pavo cristatus) Photo credit Sherry Woods (G) Number of feeding recordsinvolving reptiles or amphibians as prey grouped by predator class Each square represents 10 records(continued on next page )

Full-size DOI 107717peerj9485fig-1


Figure 1 ( continued)Predator classes withlt10 records are included within lsquoOtherrsquo (Actinopterygii Chilopoda Insecta andMalacostraca) (HndashL) Examples of in situ feeding interactions involving reptiles and amphibians asprey collected from Facebook (H) Many-horned adder (Bitis cornuta)ndashSpotted desert lizard (Merolessuborbitalis) Photo credit Jessica Kemper (I) Jumping spider (Salticidae)ndashCape dwarf gecko (Lygodactyluscapensis) Photo credit Anton Roberts (J) Brown water snake (Lycodonomorphus rufulus)ndashTable Mountainghost frog (Heleophryne purcelli) Photo credit Theo Busschau (K) Rock kestrel (Falco rupicolus)ndashBloubergstrand dwarf burrowing skink (Scelotes montispectus) Photo credit Paul Clinton (L) Black-backed jackal (Canis mesomelas)ndashMole snake (Pseudaspis cana) Photo credit Reg Lyons

lt10 records0

500

1000

1500

2000

2500

3000

3500

400018

8318

8718

9118

9518

9919

0319

0719

1119

1519

1919

2319

2719

3119

3519

3919

4319

4719

5119

5519

5919

6319

6719

7119

7519

7919

8319

8719

9119

9519

9920

0320

0720

1120

1520

19

Year

Cum

mul

ativ

e nu

mbe

r of

rec

ords Facebook

Literature

Figure 2 Accumulation of snake feeding records from Facebook and literature sourcesFull-size DOI 107717peerj9485fig-2

were collected using Facebook in half of the time This period of high reporting rate canbe attributed to the publication of several multi-taxa museum studies from 1998ndash2007(Shine et al 1998 Keogh Branch amp Shine 2000 Webb et al 2000 Webb Branch amp Shine2001 Shine et al 2006a Shine et al 2006b Shine et al 2007) Importantly the periods ofcomparable reporting rates are the result of decades of work that ultimately culminatedin the publication of the literature during those periods rather than actual rates of recordaccrual as represented by our Facebook dataset

We found important differences in the taxonomic resolution to which prey species wereidentified when comparing the two datasets Prey were identified to the species level in766 of Facebook records compared to only 504 of literature records (χ2

= 2169plt 00001) (Fig 3)mdashprobably because digestion of prey items in the gut of museumspecimens often eliminates diagnostic characteristics Similarly a significantly largerproportion of the Facebook records were identifiable to at least the level of genus familyand order than records in the literature dataset (χ2

=4913ndash2506 all plt 00001) (Fig 3)Broadly speaking the number of feeding observations for each snake species was

moderately correlated across the two approaches (Spearmanrsquos correlation ρ = 0490plt 00001) However this relationship obscures some dramatic differences in the overlapin trophic interactions detected via each approach For interactions with species-levelidentification of prey items we identified 441 and 781 distinct interactions within theFacebook and literature datasets respectively (Fig 4) Surprisingly only 114 of these


000

025

050

075

100

Class Order Family Genus Species

Taxonomic level

Pro

port

ion

of r

ecor

ds

FacebookLiterature

Figure 3 Taxonomic resolution of snake prey items identified from Facebook and literature sources p-valuelt 00001


interactions were shared between the two datasets and notably 327 interactions (of the441 interactions detected 741) were unique to the Facebook dataset Cohenrsquos kappacoefficient (κ =minus0652) confirmed that the approaches were highly discordant in anon-random manner Given the bias toward higher-level taxonomic resolution for prey inthe literature dataset (Fig 3) we recalculated Cohenrsquos kappa coefficient with interactionsaggregated at the level of genus family and order and found low concordance across alltaxonomic levels of prey identification (κ = minus0652ndashminus0289) with order-level taxonomicassignment showing the greatest but still poor level of concordance Depending on level ofprey identification (ie taxonomic aggregation) our analyses revealed that 284ndash741 ofthe interactions detected via Facebookwere previously undocumented (Fig 4) Remarkablyeven at the coarse taxonomic aggregation level of order 284 of interactions detectedusing Facebook were novel

On iNaturalist we found 102 snake feeding observations by querying the databaseThe earliest upload date of a feeding observations was in 2011 with an average of 113observations added per year since thenmdasha rate that is significantly slower (Welchrsquos t-testt = 415 p= 00139) than Facebook (micro= 213 records yrminus1) The greatest number offeeding observations reported on iNaturalist occurred in 2019 (N = 39) and representsfewer observations than the number of uploads to our Facebook group during its firstyear (N = 54) Eight of the top-ten snake species recorded in the iNaturalist dataset werealso in the top-ten in the Facebook dataset Notably the brown house snake (Boaedoncapensis) had the most observations on both platforms Interestingly of the 57 distinctinteractions with prey identified to the species-level that were reported on iNaturalist 19species interactions were not detected using Facebook and 13 interactions were not presentin either the Facebook or the literature dataset


Species

Genus

Family

Order

000 025 050 075 100

Proportion of interactions

Taxo

nom

ic le

vel

Facebook unique Shared Literature unique

Figure 4 Proportion of unique snake feeding interactions identified from Facebook and literaturesources An interaction is defined as any instance of a specific snake species consuming a specific preyitem Interactions were included only if the prey were identified to the taxonomic level under analysis Du-plicate interactions within an approach were removed


We gathered an additional seven feeding records for four target species (B capensisPython natalensis Dispholidus typus and Naja nivea) by visually searching through speciesobservations for records that were missed by our querying methods Across all four targetspecies Facebook outperformed iNaturalist (Fig 5A) However the number of recordsobtained for each of the target species was proportionally similar across the two platformsThe maximum difference in proportions equated to 368 (D typus Facebook 732vs iNaturalist 110) Finally at least 92 of all iNaturalist records were duplicates ofrecords found using Facebook (ie exact photo)

Targeted Google Images searches for the four target species returned 13ndash25 recordsper species (N = 72) which exceeds the number of records posted to iNaturalist for eachtarget species but still underperformed relative to Facebook (Fig 5A) 194 of recordswere duplicates of feeding events found using Facebook Importantly the ratio of preytypes for several of the target species were heavily skewed depending on the source ofthe observation (Figs 5Bndash5E) In particular 842 of records for N nivea depictedophiophagy (ie snake-eating) particularly involving puff adders (Bitis arietans) andmole snakes (Pseudaspis cana) (Fig 5E) Additionally 846 of P natalensis observationsinvolved animals feeding on large mammals (eg antelope) and our search failed toproduce any instances of bird-eating (Fig 5C) Remarkably this method did not produceany novel species interactions

DISCUSSIONOur study demonstrates the utilisation of Facebook as a crowdsourcing tool to gather ageographically and taxonomically diverse dataset of difficult to observe trophic interactions


0

20

40

60

80

100

120

Boaedon capensis Python natalensis Dispholidus typus Naja nivea

Snake species

Num

ber

of o

bser

vatio

ns

Facebook iNaturalist Google ImagesA

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Pro

port

ion

of o

bser

vatio

ns

Boaedon capensisB

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Python natalensisC

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Pro

port

ion

of o

bser

vatio

ns

Dispholidus typusD

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Naja niveaE

Figure 5 Comparison of feeding observations for four target snake species collected from FacebookiNaturalist and Google Images (A) The number of observations collected (BndashE) The proportion ofrecords with prey belonging to a given prey category



involving southern African herpetofauna as predators and as prey Despite these types ofinteractions being difficult to observe our approach has yielded observations faster at finertaxonomic resolution and that differ significantly from what is currently known within138 years of herpetological literature Taken together these findings provide a powerfulexample of the potential application of social media to gather discrete ephemeral ecologicalinteractions

Importantly our work is part of a growing recognition of the remarkable power ofsocial media and citizen science to gather biological information (reviewed by Toivonenet al 2019 and Jarić et al 2020) Although a number of studies have made use of digitalmedia platforms (ie not specifically designed for citizen science) to better understand thegeographic and temporal distribution of biological traits or organisms (Leighton et al 2016Jimeacutenez-Valverde et al 2019 Marshall amp Strine 2019) other studies have started to detailecological and evolutionary processes explicitly Google Images has been used to quantifyinsect-pollinator relationships (Bahlai amp Landis 2016) commensalism-like relationshipsbetween birds and large mammals (Mikula et al 2018) to assess the diets of predatorybirds (Mikula et al 2016Naude et al 2019) and the diets of predatory insects (HernandezMasonick amp Weirauch 2019) Similarly Facebook has been used to quantify co-grazingpatterns between two deer species (Mori Bari amp Coraglia 2018) and ad hoc observationshave revealed a fascinating foraging strategy in skunks (Pesendorfer Dickerson amp Dragoo2018) Importantly many of these taxa are often conspicuous due to their size colourationmicrohabitat usage or duration spent in one location and the resources in several of thestudies are conspicuous (for the same reasons) or spatially restricted Ultimately thesecharacteristics improve detection probability and reporting rates Conversely our studyhas demonstrated that social media (specifically Facebook) draws observational powerfrom such a large network that even elusive ecological interactions with low temporal andspatial predictability can be gathered rapidly

Our approach has several strengths that make its application in ecological andevolutionary studies appealing Firstly the ease of reporting means that observers aremore likely to share their observations A dedicated actively managed public groupallows for photos to be funnelled from across Facebook and many of our observers hadalready shared their observations to Facebook in some other context before those postswere shared to our dedicated predation records group Importantly the group acts as anoutlet for observations that would never otherwise have been documented formally nowthose records can be incorporated into a growing database Secondly while citizen scienceprojects like iNaturalist and iSpot attract many users citizen science platforms are mainlypopulated by a few very active users (Sauermann amp Franzoni 2015) Facebook does notrequire an inherent interest in a particular topic which allows for a diverse range of mediato be posted and shared publicly Together with the low probability of encountering feedingeventsmdashas indicated by the number of single observations in our datasetmdashdedicated floraand fauna platforms do not attract enough observers to gather sizeable datasets especiallyoutside of major populated areas Thirdly the interactive nature of Facebook facilitatesdirect communication with observers which can result in more photos or details if neededInformation such as locality data can be requested directly from observers thus reducing


the reliance on geo-tagging functions of social media platforms which can be incorrector missing from posts altogether (Di Minin Tenkanen amp Toivonen 2015) Fourth theFacebook group format provides an ideal platform to discuss identification of species withinterested experts thereby facilitating expert-crowdsourcing of species identificationsAusten et al (2018) proposed that the identification of species in digital natural historyobservations should be based on more than one photo and verified by more than oneexpert Thus Facebook groups offer effective mechanisms to meet these criteria Finallythe community of observers receive informed feedback from researchers regarding theirobservation Unlike passive data collection methods (ie media and data scraping) activeengagement with observers and other members acts as an opportunity to educate thepublic about the importance of an observation and active engagement and feedback hasthe potential to incentivise continued participation

The data gathered via our approach is not without its context-specific challengesPrimarily our approach does not offer an obvious mechanism for quantifying samplingeffort prohibiting rate- or density-dependent analyses of these processes Secondly ourapproach as with nearly all sampling approaches may overrepresent certain interactionsin important ways (Glaudas Kearney amp Alexander 2017) For one our approach is likelyto include events that happen (1) frequently (2) near humans (either urban areas orwell-trafficked nature reserves) (3) near humans with the means to access the internet(which shows socioeconomic and regional bias Kemp 2020) and (4) over longer periodsof time Third the permanence of posts and their associated media which appear onsocial media platforms like Facebook is not guaranteed and images may be removed ortheir visibility settings may be changed by the owner at any time As a result there is aneed to store images and data outside of the platform in a timely manner Finally wehave adopted to manually curate and log observations into a database rather than seekautomated approaches in part due to the loss of API function in April 2018 associatedwith a change in Facebookrsquos terms of service (Freelon 2018) This manual approach hasworked well at the scale of our analysis but will become problematic at the scale of someof the data that social media has the potential to gather Advances in machine learningfor identification of species in images are progressing rapidly (reviewed by Waumlldchenamp Maumlder 2018) and are starting to be utilised for scientific assessment of social mediaimages (Di Minin et al 2018) However in our context we continue to be limited by thefact that the observations being reported are inherently difficult to observe thus limitingthe availability of sufficient amounts of training data Nonetheless automation of imageidentification or even social media group administration will be required to scale ourapproach to truly global ecological or evolutionary questions

The relatively low measures of concordance between the data gathered via Facebookand that reported in the literature (Fig 4) or via other digital media platforms (Figs5Bndash5E) raises the important question of which approach more closely reflects reality Someapproaches to studying diet such as fixed videography (Glaudas Kearney amp Alexander2017) andDNAbarcoding of prey remains (reviewed byAlberdi et al 2019) offer promisingfuture prospects for relatively unbiased dietary analysis for many organisms includingsnakes However these approaches are incredibly effort- and cost-intensive limiting their


widespread application Currently it is unclear to what degree our data might bias for oragainst detection of certain interactions However we are encouraged by the detectionin our Facebook dataset of several apparently difficult to detect interactions (eg puffadders (Bitis arietans) consuming amphibians the first reported diet record for Swazi rocksnakes (Inyoka swazicus)) and interactions with incredibly short handling times (eg avine snake (Thelotornis capensis) catching and swallowing a rain frog (Breviceps sp) inunder 20 seconds) It is apparent from our analysis that Google Images may be the leasteffective means for collecting representative diet data at least for our study system This islikely to be the case because not all webpages are indexed by Google (including Facebook)and blogs or media outlets are dominated by eye-catching photos and particularly notableor lengthy encounters On the other hand iNaturalist may provide more representativedata that can be used in corroboration with Facebook data which can be promising forgeographic regions with more involvement (eg United States of America 483000+United Kingdom 28000+ South Africa 7300+ observers)

Our approach has several implications for our understanding of snake biology It iswell-established that diet has played a major role in the evolution of snakes (Greene 1983Colston Costa amp Vitt 2010) and their venoms (Daltry Wuumlster amp Thorpe 1996 Barlow etal 2009 Casewell et al 2013) Additionally snake feeding either through demographiceffects on prey populations risk of predation and lsquolandscape of fearrsquo dynamics or theselective agents for prey anti-predatory adaptations are likely to represent the majorimpacts that snakes have within ecosystems and food webs Understanding these processesis linked inextricably with high-quality natural history data regarding variation in snakediets leading to a recent attempt to centralise and analyse feeding data at a global scale(Grundler 2020) However our understanding of the details of snake diets remainssurprisingly superficial especially in places like Africa where snakebite is a major healthconcern (Harrison et al 2009 Chippaux 2011 Murray Martin amp Iwamura 2020) In thiscontext we think that our novel approach to gathering natural history data can provide apowerful tool to supplement existing datasets and ultimately improve our understandingof snake feeding thereby contextualising studies of snakes their ecological functions andtheir venoms

Our approach has enormous potential beyond our usage of it and we look forward toseeing its application in multiple ecological and evolutionary contexts Even within ourown dataset we have only begun to explore the full potential of our data by addressingspecies-specific questions (Layloo Smith amp Maritz 2017 Maritz Alexander amp Maritz2019Maritz et al 2019 Smith et al 2019) However the dataset lends itself to addressingother questions such as seasonality in feeding and prey preference intraguild predationand the evolution of diet Additionally we see its value in documenting other ephemeraldiscrete event-driven processes similar to predation particularly if they can be captured asphotographs of the types of subjects that are already shared to social media For examplephotographs of pollinators visiting flowers could be crowd-sourced and curated to betterunderstand pollination dynamics images of identifiable individual animals (eg distinctmarkings) could be used to assess seasonal body condition home range size and lifespanand photographs of urban biodiversity could elucidate novel urban ecology interactions


between species or even human-wildlife conflict Importantly images of many of thesetypes of events are being shared on social media platforms already and all that is requiredis for interested researchers to start engaging with those data

CONCLUSIONSEmploying social media as a citizen science platform allowed for the collection of trophicdata across a remarkable diversity of interactions involving African reptiles and amphibiansParticularly the results of the dietary analysis of snakes demonstrate how rapidly andprecisely information can be collected to characterise an ecological process comparedto traditional approaches Additionally the results show a large discordance betweensampling via social media and traditional approaches including the detection of many novelinteractions which emphasises how undersampling can lead to gaps in our understandingFinally the results highlight how social media can outperform traditional citizen scienceand crowdsourcing approaches when observations involve elusive animals or unpredictableevents which is likely due differences in the number of active members and thus overallsampling intensity Beyond herpetological studies the observational power and approachshowcased here has enormous potential for the documentation and investigation of otherrare events that underlie important ecological processes and we emphasise that suchapproaches should no longer be considered ancillary

ACKNOWLEDGEMENTSWe acknowledge Andre Coetzer Tyrone Ping and Luke Verburgt for their roles inconceiving the idea for the Predation Records - Reptiles and Frogs (Sub-Saharan Africa)Facebook group Additionally we thank them Gary Nicolau Nina Perry and Jason Boycefor assisting in the management of the grouprsquos membership and discussions as adminsWe are indebted to the members of the group (and other members of the Facebookcommunity) who have shared observations to and participated in our research projectSpecifically we acknowledge the following members for their extraordinary efforts HelenBadenhorst Toy Bodbijl Norman Barrett Gary Brown Andre Coetzer Nick Evans KyleFinn Daniel Karamitsos Ashley Kemp Luke Kemp Trish McGill Andrea Myburgh GaryNicolau Deon Oosthuizen Tyrone Ping Rian Stander Mike Soroczynski Francois Theartand Ryan van Huyssteen We are grateful for species identifications provided by expertsin the community Particularly we thank Michael Bates Alan Channing Andre CoetzerWerner Conradie Adriaan Engelbrecht Ian Engelbrecht James Harvey Teresa KearneyLuke Kemp Johan Marais Gary Nicolau Dan Parker Tyrone Ping Dominic RollinsonStephen Spawls and Luke Verburgt for their input We recognise the 2017 University ofthe Western Cape Biodiversity and Conservation Biology Herpetology honours studentsfor assisting in the early phase of the literature search


ADDITIONAL INFORMATION AND DECLARATIONS

FundingThis work was supported by the National Research Foundation Thuthuka Grant (UID118090) The funders had no role in study design data collection and analysis decision topublish or preparation of the manuscript

Grant DisclosuresThe following grant information was disclosed by the authorsNational Research Foundation Thuthuka Grant UID 118090

Competing InterestsThe authors declare there are no competing interests

Author Contributionsbull Robin A Maritz conceived and designed the experiments performed the experimentsanalyzed the data prepared figures andor tables authored or reviewed drafts of thepaper and approved the final draftbull Bryan Maritz conceived and designed the experiments authored or reviewed drafts ofthe paper and approved the final draft

Data AvailabilityThe following information was supplied regarding data availability

Datasets and R code used in this study are available on FigshareMaritz Robin Maritz Bryan (2020) Data for lsquolsquoSharing for science High-resolution

trophic interactions revealed rapidly by social mediarsquorsquo figshare Dataset httpsdoiorg106084m9figshare11920128v2

Maritz Robin Maritz Bryan (2020) R code for lsquolsquoSharing for science High-resolution trophic interactions revealed rapidly by social mediarsquorsquo figshare Softwarehttpsdoiorg106084m9figshare12287714v2

Supplemental InformationSupplemental information for this article can be found online at httpdxdoiorg107717peerj9485supplemental-information

REFERENCESAlberdi A Aizpurua O Bohmann K Gopalakrishnan S Lynggaard C NielsenM

Gilbert MT 2019 Promises and pitfalls of using high-throughput sequencing fordiet analysisMolecular Ecology Resources 19327ndash348 DOI 1011111755-099812960

Austen GE BindemannM Griffiths RA Roberts DL 2018 Species identification byconservation practitioners using online images accuracy and agreement betweenexperts PeerJ 6e4157 DOI 107717peerj4157

Bahlai CA Landis DA 2016 Predicting plant attractiveness to pollinators with passivecrowdsourcing Royal Society Open Science 3Article 150677 DOI 101098rsos150677


Barlow A Pook CE Harrison RAWuumlsterW 2009 Coevolution of diet and prey-specific venom activity supports the role of selection in snake venom evo-lution Proceedings of the Royal Society B Biological Sciences 2762443ndash2449DOI 101098rspb20090048

Casewell NRWuumlsterW Vonk FJ Harrison RA Fry BG 2013 Complex cocktailsthe evolutionary novelty of venoms Trends in Ecology amp Evolution 28219ndash229DOI 101016jtree201210020

Chacoff NP Vaacutezquez DP Lomaacutescolo SB Stevani EL Dorado J Padroacuten B 2011Evaluating sampling completeness in a desert plantndashpollinator network Journal ofAnimal Ecology 81190ndash200 DOI 101111j1365-2656201101883x

Chippaux J 2011 Estimate of the burden of snakebites in sub-Saharan Africa a meta-analytic approach Toxicon 57586ndash599 DOI 101016jtoxicon201012022

Cohen J 1960 A coefficient of agreement for nominal scales Educational and Psychologi-cal Measurement 2037ndash46 DOI 101177001316446002000104

Colston TJ Costa GC Vitt LJ 2010 Snake diets and the deep history hypothesis Biologi-cal Journal of the Linnean Society 101476ndash486 DOI 101111j1095-8312201001502x

Daltry JCWuumlsterW Thorpe RS 1996 Diet and snake venom evolution Nature379537ndash554 DOI 101038379537a0

DiMinin E Fink C Tenkanen H Hiippala T 2018Machine learning for trackingillegal wildlife trade on social media Nature Ecology amp Evolution 2406ndash407DOI 101038s41559-018-0466-x

DiMinin E Tenkanen H Toivonen T 2015 Prospects and challenges for social mediadata in conservation science Frontiers in Environmental Science 3Article 63DOI 103389fenvs201500063

Durso AM Seigel RA 2015 A snake in the hand is worth 10000 in the bush Journal ofHerpetology 49503ndash506 DOI 10167015-49-041

Durso AMWillson JDWinne CT 2011 Needles in haystacks Estimating detectionprobability and occupancy of rare and cryptic snakes Biological Conservation1441508ndash1515 DOI 101016jbiocon201101020

Facebook 2019October 30 Facebook reports third quarter 2019 results [Press Release]Available at https investorfbcom investor-newspress-release-details 2019Facebook-Reports-Third-Quarter-2019-Resultsdefaultaspx (accessed on 20 January2020)

FitzSimons VFM 1962 Snakes of Southern Africa Cape Town Purnell and SonsFreelon D 2018 Computational research in the post-API age Political Communication

35665ndash668 DOI 1010801058460920181477506Garvey JE Whiles M 2017 Trophic ecology London CRC PressGlaudas X Kearney TC Alexander GJ 2017Museum specimens bias measures of snake

diet a case study using the ambush-foraging puff adder (Bitis arietans) Herpetologica73121ndash128 DOI 101655herpetologica-d-16-00055

Greene HW 1983 Dietary correlates of the origin and radiation of snakes AmericanZoologist 23431ndash441 DOI 101093icb232431


Greene HW 1997 Snakes The evolution of mystery in nature Berkeley University ofCalifornia Press

Greene HW Jaksić FM 1983 Food-niche relationships among sympatric predatorseffects of level of prey identification Oikos 40151ndash154 DOI 1023073544212

Grundler MC 2020 SquamataBase a natural history database and R package forcomparative biology of snake feeding habits Biodiversity Data Journal 8e49943DOI 103897BDJ8e49943

Harrison RA Hargreaves AWagstaff SC Faragher B Lalloo DG 2009 Snake enven-oming a disease of poverty PLOS Neglected Tropical Diseases 3e569DOI 101371journalpntd0000569

Hegland SJ Dunne J Nielsen A Memmott J 2010How to monitor ecological com-munities cost-efficiently the example of plantndashpollinator networks BiologicalConservation 1432092ndash2101 DOI 101016jbiocon201005018

HernandezMMasonick PWeirauch C 2019 Crowdsourced online images provideinsights into predatorndashprey interactions of putative natural enemies Food Webs21e00126 DOI 101016jfooweb2019e00126

Iverson JB 1982 Biomass in turtle populations a neglected subject Oecologia 5569ndash76DOI 101007bf00386720

Jacobsen NHG 1982 The ecology of the reptiles and amphibians in the Burkea africana-Eragrostis pallens savanna of the Nylsvley Nature Reserve M Sci Thesis Universityof Pretoria South Africa

Jarić I Correia RA Brook BW Buettel JC Courchamp F Di Minin E Firth JA GastonKJ Jepson P Kalinkat G Ladle R Soriano-Redondo A Souza AT Roll U 2020iEcology harnessing large online resources to generate ecological insights Trends inEcology and Evolution 35630ndash639 DOI 101016jtree202003003

Jimeacutenez-Valverde A Pentildea Aguilera P Barve V Burguillo-Madrid L 2019 Photo-sharing platforms key for characterising niche and distribution in poorly studiedtaxa Insect Conservation and Diversity 12389ndash403 DOI 101111icad12351

Jordano P 2016 Sampling networks of ecological interactions Functional Ecology301883ndash1893 DOI 101101025734

Kemp S 2020 January 30 Digital 2020 38 Billion People Use Social Media Availableat httpswearesocialcomblog202001digital-2020-3-8-billion-people-use-social-media (accessed on 10 May 2020)

Keogh J BranchWR Shine R 2000 Feeding ecology reproduction and sexual dimor-phism in the colubrid snake Crotaphopeltis hotamboeia in southern Africa AfricanJournal of Herpetology 49129ndash137 DOI 1010802156457420009635439

Lardner B Rodda GH Adams AA Savidge JA Reed RN 2015 Detection rates of geckosin visual surveys turning confounding variables into useful knowledge Journal ofHerpetology 49522ndash532 DOI 10167014-048

Layloo I Smith C Maritz B 2017 Diet and feeding in the cape cobra Naja nivea AfricanJournal of Herpetology 66147ndash153 DOI 1010802156457420171388297


Leighton GR Hugo PS Roulin A Amar A 2016 Just Google it assessing the use ofGoogle Images to describe geographical variation in visible traits of organismsMethods in Ecology and Evolution 71060ndash1070 DOI 1011112041-210x12562

Maritz B Alexander GJ Maritz RA 2019 The underappreciated extent of cannibalismand ophiophagy in African cobras Ecology 100e02522 DOI 101002ecy2522

Maritz R ConradieW Sardinha CI Peto A Chechene AHC Maritz B 2019 Ophio-phagy and cannibalism in African vine snakes (Colubridae Thelotornis) AfricanJournal of Ecology [early view] DOI 101111aje12702

Marshall BM Strine CT 2019 Exploring snake occurrence records spatial biases andmarginal gains from accessible social media PeerJ 7e8059 DOI 107717peerj8059

McCann K 2007 Protecting biostructure Nature 44629 DOI 101038446029aMikula P Hadrava J Albrecht T Tryjanowski P 2018 Large-scale assessment of

commensalisticndashmutualistic associations between African birds and herbivorousmammals using internet photos PeerJ 6e4520 DOI 107717peerj4520

Mikula P Morelli F Lučan RK Jones DN Tryjanowski P 2016 Bats as prey of diurnalbirds a global perspectiveMammal Review 46160ndash174 DOI 101111mam12060

MirandaM Parrini F Dalerum F 2013 A categorization of recent network approachesto analyse trophic interactionsMethods in Ecology and Evolution 4897ndash905DOI 1011112041-210x12092

Mori E Bari PD Coraglia M 2018 Interference between roe deer and Northern chamoisin the Italian Alps are Facebook groups effective data sources Ethology Ecology ampEvolution 30277ndash284 DOI 1010800394937020171354922

Murray KA Martin G Iwamura T 2020 Focus on snake ecology to fight snakebiteLancet 395e14 DOI 101016S0140-6736(19)32510-3

Naude VN Smyth LKWeideman EA Krochuk BA Amar A 2019 Using web-sourcedphotography to explore the diet of a declining African raptor the Martial Eagle(Polemaetus bellicosus) The Condor 1211ndash9 DOI 101093condorduy015

Paine RT 1988 Road maps of interactions or grist for theoretical development Ecology691648ndash1654 DOI 1023071941141

Pesendorfer MB Dickerson S Dragoo JW 2018 Observation of tool use in stripedskunks how community science and social media help document rare naturalphenomena Ecosphere 9e02484 DOI 101002ecs22484

Petranka JWMurray SS 2001 Effectiveness of removal sampling for determining sala-mander density and biomass a case study in an Appalachian streamside communityJournal of Herpetology 3536ndash44 DOI 1023071566020

Pincheira-Donoso D Bauer AMMeiri S Uetz P 2013 Global taxonomic diversity ofliving reptiles PLOS ONE 8e59741 DOI 101371journalpone0059741

Polis GA 1991 Complex trophic interactions in deserts an empirical critique of food-web theory American Naturalist 138123ndash155 DOI 101086285208

Rodda GH Dean-Bradley K Campbell EW Fritts TH Lardner B Adams AA Reed RN2015 Stability of detectability over 17 years at a single site and other lizard detectioncomparisons from Guam Journal of Herpetology 49513ndash521 DOI 10167014-085


Sauermann H Franzoni C 2015 Crowd science user contribution patterns and theirimplications Proceedings of the National Academy of Sciences of the United States ofAmerica 112679ndash684 DOI 101073pnas1408907112

Shine R BranchWR Harlow PSWebb JK 1998 Reproductive biology and foodhabits of horned adders Bitis caudalis (Viperidae) from southern Africa Copeia1998(2)391ndash401 DOI 1023071447433

Shine R BranchWR Harlow PSWebb JK Shine T 2006a Biology of burrow-ing asps (Atractaspididae) from southern Africa Copeia 2006(1)103ndash115DOI 1016430045-8511(2006)006[0103bobaaf]20co2

Shine R BranchWRWebb JK Harlow PS Shine T 2006b Sexual dimorphismreproductive biology and dietary habits of psammophiine snakes (Colubridae) fromsouthern Africa Copeia 2006(4)650ndash664DOI 1016430045-8511(2006)6[650sdrbad]20co2

Shine R BranchWRWebb JK Harlow PS Shine T Keogh JS 2007 Ecology of cobrasfrom southern Africa Journal of Zoology 272183ndash193DOI 101111j1469-7998200600252x

Smith CCD Layloo I Maritz RA Maritz B 2019 Sexual dichromatism does nottranslate into sex-based difference in morphology or diet for the African boomslangJournal of Zoology 308253ndash258 DOI 101111jzo12670

Steen DA 2010 Snakes in the grass secretive natural histories defy both conventionaland progressive statistics Herpetological Conservation and Biology 5183ndash188

Thompson RM Townsend CR 2000 Is resolution the solution The effect of taxonomicresolution on the calculated properties of three stream food webs Freshwater Biology44413ndash422 DOI 101046j1365-2427200000579x

Toivonen T Heikinheimo V Fink C Hausmann A Hiippala T Jaumlrv O Tenkanen HDi Minin E 2019 Social media data for conservation science a methodologicaloverview Biological Conservation 233298ndash315 DOI 101016jbiocon201901023

Tylianakis JM Didham RK Bascompte J Wardle DA 2008 Global change andspecies interactions in terrestrial ecosystems Ecology Letters 111351ndash1363DOI 101111j1461-0248200801250x

Valiente-Banuet A AizenMA Alcaacutentara JM Arroyo J Cocucci A Galetti M GarciacuteaMB Garciacutea D Goacutemez JM Jordano P Medel R Navarro L Obeso JR Oviedo RRamiacuterez N Rey PJ Traveset A VerduacuteM Zamora R 2014 Beyond species lossthe extinction of ecological interactions in a changing world Functional Ecology29299ndash307 DOI 1011111365-243512356

Vitousek PMMooney HA Lubchenco J Melillo JM 1997Human domination ofEarthrsquos ecosystems Science 277494ndash499 DOI 101126science2775325494

Waumlldchen J Maumlder P 2018Machine learning for image based species identificationMethods in Ecology and Evolution 92216ndash2225 DOI 1011112041-210x13075

Webb JK BranchWR Shine R 2001 Dietary habits and reproductive biology oftyphlopid snakes from southern Africa Journal of Herpetology 35558ndash567DOI 1023071565893


Webb JK Shine R BranchWR Harlow PS 2000 Life-history strategies in basalsnakes reproduction and dietary habits of the African thread snake Leptoty-phlops scutifrons (Serpentes Leptotyphlopidae) Journal of Zoology 250321ndash327DOI 101111j1469-79982000tb00776x

Western D 1974 The distribution density and biomass density of lizards in a semindasharid environment of northern Kenya African Journal of Ecology 1249ndash62DOI 101111j1365-20281974tb00106x


datasets characterising species interactions in ecosystems (McCann 2007 Hegland et al2010 Jordano 2016)

Together the origin and development of social media and the concurrent advances inaccess to mobile cameras represent a disruptive innovation that has changed the mannerand rate at which modern events are recorded and communicated With over 380 billionpeople using social media worldwide (Kemp 2020) the synergy of social media and readilyaccessible mobile cameras has increased the observational effort of researchers by ordersof magnitude Facebook alone has more than 245 billion monthly active users (Facebook2019) making it the digital platform with the largest social networking potential (Kemp2020) Harnessing this power has far-reaching implications for understanding ecologicaland evolutionary processes characterised by difficult to detect discrete transient eventsthrough the resultant increase in observation coverage and depth

Trophic interactions defined as interspecific interactions in which one organismconsumes another form the basis for understanding processes and system characteristicsas diverse as energy flow population dynamics food web dynamics and the evolution ofbehavioural morphological and physiological adaptations by predators and prey (Garveyamp Whiles 2017) Moreover with a world experiencing climatic changes and worseningenvironmental conditions (Vitousek et al 1997) attention to species interactions willbe crucial for understanding ecosystem function and integrity (Tylianakis et al 2008Valiente-Banuet et al 2014) Despite their central position in ecological and evolutionarytheory the characterisation of trophic interactions between species and within food websparticularly those in which such interactions are difficult to study are often incompleteor absent (Paine 1988 Chacoff et al 2011 Miranda Parrini amp Dalerum 2013 Jordano2016) Moreover because certain organismal traits can reduce the detection likelihood of agiven trophic interaction trophic interactions involving terrestrial vertebrates (particularlynon-herbivorous interactions) are underrepresented in the literature (Miranda Parrini ampDalerum 2013) In cases where such datasets exist endotherms tend to be better representedthan ectotherms in trophic studies (Miranda Parrini amp Dalerum 2013)mdashpossibly becauseof ease of sampling or because endothermy often demands higher food intake ratesFinally of the interactions reported organisms involved in lower-trophic-level interactionsoften suffer from taxonomic aggregation (Polis 1991) which can mask complex speciesinteractions and influence metrics associated with food webs community assemblagesand interspecific competition (Greene amp Jaksić 1983 Paine 1988 Thompson amp Townsend2000) Thus a method for improving the quantity and quality of collected trophic data iswarranted and essential

Together reptiles and amphibians (hereafter herpetofauna) include more than 18000ectothermic vertebrate species and account for more than half of all global tetrapoddiversity (Pincheira-Donoso et al 2013) In many terrestrial ecosystems these animalscan make up a large proportion of the total abundance of vertebrates and contributesignificantly to the total biomass of a region (Western 1974 Iverson 1982 Jacobsen 1982Petranka amp Murray 2001) Moreover herpetofauna (mostly amphibians and squamates)often occupy intermediate trophic levels providing important trophic links between small-bodied invertebrate primary consumers and higher trophic levels occupied primarily by























Gastropoda

Other


J




G

D

C

L

I

F

B

K

H

E





lt10 records0

500

1000

1500

2000

2500

3000

3500

400018

8318

8718

9118

9518

9919

0319

0719

1119

1519

1919

2319

2719

3119

3519

3919

4319

4719

5119

5519

5919

6319

6719

7119

7519

7919

8319

8719

9119

9519

9920

0320

0720

1120

1520

19

Year

Cum

mul

ativ

e nu

mbe

r of

rec

ords Facebook

Literature








000

025

050

075

100


Taxonomic level

Pro

port

ion

of r

ecor

ds

FacebookLiterature






Species

Genus

Family

Order

000 025 050 075 100


Taxo

nom

ic le

vel








0

20

40

60

80

100

120


Snake species

Num

ber

of o

bser

vatio

ns


000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Pro

port

ion

of o

bser

vatio

ns

Boaedon capensisB

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Python natalensisC

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Pro

port

ion

of o

bser

vatio

ns

Dispholidus typusD

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Naja niveaE





























































































































Gastropoda

Other


J




G

D

C

L

I

F

B

K

H

E





lt10 records0

500

1000

1500

2000

2500

3000

3500

400018

8318

8718

9118

9518

9919

0319

0719

1119

1519

1919

2319

2719

3119

3519

3919

4319

4719

5119

5519

5919

6319

6719

7119

7519

7919

8319

8719

9119

9519

9920

0320

0720

1120

1520

19

Year

Cum

mul

ativ

e nu

mbe

r of

rec

ords Facebook

Literature








000

025

050

075

100


Taxonomic level

Pro

port

ion

of r

ecor

ds

FacebookLiterature






Species

Genus

Family

Order

000 025 050 075 100


Taxo

nom

ic le

vel








0

20

40

60

80

100

120


Snake species

Num

ber

of o

bser

vatio

ns


000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Pro

port

ion

of o

bser

vatio

ns

Boaedon capensisB

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Python natalensisC

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Pro

port

ion

of o

bser

vatio

ns

Dispholidus typusD

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Naja niveaE









































































































lt10 records0

500

1000

1500

2000

2500

3000

3500

400018

8318

8718

9118

9518

9919

0319

0719

1119

1519

1919

2319

2719

3119

3519

3919

4319

4719

5119

5519

5919

6319

6719

7119

7519

7919

8319

8719

9119

9519

9920

0320

0720

1120

1520

19

Year

Cum

mul

ativ

e nu

mbe

r of

rec

ords Facebook

Literature








000

025

050

075

100


Taxonomic level

Pro

port

ion

of r

ecor

ds

FacebookLiterature






Species

Genus

Family

Order

000 025 050 075 100


Taxo

nom

ic le

vel








0

20

40

60

80

100

120


Snake species

Num

ber

of o

bser

vatio

ns


000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Pro

port

ion

of o

bser

vatio

ns

Boaedon capensisB

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Python natalensisC

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Pro

port

ion

of o

bser

vatio

ns

Dispholidus typusD

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Naja niveaE








































































































000

025

050

075

100


Taxonomic level

Pro

port

ion

of r

ecor

ds

FacebookLiterature






Species

Genus

Family

Order

000 025 050 075 100


Taxo

nom

ic le

vel








0

20

40

60

80

100

120


Snake species

Num

ber

of o

bser

vatio

ns


000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Pro

port

ion

of o

bser

vatio

ns

Boaedon capensisB

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Python natalensisC

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Pro

port

ion

of o

bser

vatio

ns

Dispholidus typusD

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Naja niveaE








































































































Species

Genus

Family

Order

000 025 050 075 100


Taxo

nom

ic le

vel








0

20

40

60

80

100

120


Snake species

Num

ber

of o

bser

vatio

ns


000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Pro

port

ion

of o

bser

vatio

ns

Boaedon capensisB

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Python natalensisC

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Pro

port

ion

of o

bser

vatio

ns

Dispholidus typusD

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Naja niveaE








































































































0

20

40

60

80

100

120


Snake species

Num

ber

of o

bser

vatio

ns


000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Pro

port

ion

of o

bser

vatio

ns

Boaedon capensisB

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Python natalensisC

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Pro

port

ion

of o

bser

vatio

ns

Dispholidus typusD

000

025

050

075

100

frogs

lizar

ds

snak

es

birds

small

mam

mals

large

mam

mals

Prey identity

Naja niveaE








































































































sharing for science: high-resolution trophic interactions ......facebook alone has more than 2.45...

Documents